openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/drivers/tty/serial/amba-pl011.c

2658 lines
66 KiB
C
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Driver for AMBA serial ports
*
* Based on drivers/char/serial.c, by Linus Torvalds, Theodore Ts'o.
*
* Copyright 1999 ARM Limited
* Copyright (C) 2000 Deep Blue Solutions Ltd.
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
* Copyright (C) 2010 ST-Ericsson SA
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* This is a generic driver for ARM AMBA-type serial ports. They
* have a lot of 16550-like features, but are not register compatible.
* Note that although they do have CTS, DCD and DSR inputs, they do
* not have an RI input, nor do they have DTR or RTS outputs. If
* required, these have to be supplied via some other means (eg, GPIO)
* and hooked into this driver.
*/
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#if defined(CONFIG_SERIAL_AMBA_PL011_CONSOLE) && defined(CONFIG_MAGIC_SYSRQ)
#define SUPPORT_SYSRQ
#endif
#include <linux/module.h>
#include <linux/ioport.h>
#include <linux/init.h>
#include <linux/console.h>
#include <linux/sysrq.h>
#include <linux/device.h>
#include <linux/tty.h>
#include <linux/tty_flip.h>
#include <linux/serial_core.h>
#include <linux/serial.h>
[ARM] Move AMBA include files to include/linux/amba/ Since the ARM AMBA bus is used on MIPS as well as ARM, we need to make the bus available for other architectures to use. Move the AMBA include files from include/asm-arm/hardware/ to include/linux/amba/ Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-07 21:52:45 +08:00
#include <linux/amba/bus.h>
#include <linux/amba/serial.h>
[ARM] Move asm/hardware/clock.h to linux/clk.h This is needs to be visible to other architectures using the AMBA bus and peripherals. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-08 00:15:52 +08:00
#include <linux/clk.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
#include <linux/dmaengine.h>
#include <linux/dma-mapping.h>
#include <linux/scatterlist.h>
amba pl011: workaround for uart registers lockup This workaround aims to break the deadlock situation which raises during continuous transfer of data for long duration over uart with hardware flow control. It is observed that CTS interrupt cannot be cleared in uart interrupt register (ICR). Hence further transfer over uart gets blocked. It is seen that during such deadlock condition ICR don't get cleared even on multiple write. This leads pass_counter to decrease and finally reach zero. This can be taken as trigger point to run this UART_BT_WA. Workaround backups the register configuration, does soft reset of UART using BIT-0 of PRCC_K_SOFTRST_SET/CLEAR registers and restores the registers. This patch also provides support for uart init and exit function calls if present. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2011-06-13 16:11:33 +08:00
#include <linux/delay.h>
dmaengine: Pass dma_slave_config .device_fc = NULL for all existing users .device_fc is added in struct dma_slave_config recently. All user drivers, which want DMA to be the flow controller must pass this field as false. As earlier driver don't look to use this feature, mark it false for now. Signed-off-by: Viresh Kumar <viresh.kumar@st.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>
2012-02-01 18:42:19 +08:00
#include <linux/types.h>
serial: pl011: honour serial aliases in device tree If the order of UART nodes is changed in the device tree, then tty dev devices are attached to different serial ports causing the console to be directed to a different physical serial port. The "serial" aliases in the device tree should prevent this. This patch ensures that the UART driver creates tty devices that honour these aliases if a device tree is present. Acked-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Will Deacon <will.deacon@arm.com> Acked-by: Rob Herring <rob.herring@calxeda.com> Signed-off-by: Matthew Leach <matthew.leach@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-08-28 23:41:28 +08:00
#include <linux/of.h>
#include <linux/of_device.h>
serial: amba-pl011: adopt pinctrl support Cc: linux-serial@vger.kernel.org Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Shawn Guo <shawn.guo@linaro.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
2012-05-06 22:53:35 +08:00
#include <linux/pinctrl/consumer.h>
ARM: 7433/1: serial: use the new linux/sizes.h Signed-off-by: Alessandro Rubini <rubini@gnudd.com> Acked-by: Giancarlo Asnaghi <giancarlo.asnaghi@st.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2012-06-24 19:46:37 +08:00
#include <linux/sizes.h>
serial/amba-pl011: use devm_* managed resources This switches a bunch of allocation and remapping to use the devm_* garbage collected methods and cleans up the error path and remove() paths consequently. devm_ioremap() is only in <linux/io.h> so fix up the erroneous <asm/*> include as well. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Domenico Andreoli <domenico.andreoli@linux.com> Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-10-15 19:36:01 +08:00
#include <linux/io.h>
drivers: PL011: add ACPI probing for SBSA UART Add the necessary driver boilerplate to let the driver be used when the respective ACPI table is discovered by the ACPI subsystem. [Andre: change table name, add MODULE_DEVICE_TABLE entry and improve commit message] Reviewed-by: Hanjun Guo <hanjun.guo@linaro.org> Signed-off-by: Graeme Gregory <graeme.gregory@linaro.org> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:24 +08:00
#include <linux/acpi.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
#include "amba-pl011.h"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#define UART_NR 14
#define SERIAL_AMBA_MAJOR 204
#define SERIAL_AMBA_MINOR 64
#define SERIAL_AMBA_NR UART_NR
#define AMBA_ISR_PASS_LIMIT 256
[SERIAL] Fix status reporting with PL011 serial driver The receiver status register reports latched error conditions, which must be cleared by writing to it. However, the data register reports unlatched conditions which are associated with the current character. Use the data register to interpret error status rather than the RSR. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-11-19 19:10:35 +08:00
#define UART_DR_ERROR (UART011_DR_OE|UART011_DR_BE|UART011_DR_PE|UART011_DR_FE)
#define UART_DUMMY_DR_RX (1 << 16)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
tty: amba-pl011: add register lookup table Add a register lookup table, which allows the register offsets to be adjusted on a per-port basis. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:26 +08:00
static u16 pl011_std_offsets[REG_ARRAY_SIZE] = {
[REG_DR] = UART01x_DR,
[REG_FR] = UART01x_FR,
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
[REG_LCRH_RX] = UART011_LCRH,
[REG_LCRH_TX] = UART011_LCRH,
tty: amba-pl011: add register lookup table Add a register lookup table, which allows the register offsets to be adjusted on a per-port basis. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:26 +08:00
[REG_IBRD] = UART011_IBRD,
[REG_FBRD] = UART011_FBRD,
[REG_CR] = UART011_CR,
[REG_IFLS] = UART011_IFLS,
[REG_IMSC] = UART011_IMSC,
[REG_RIS] = UART011_RIS,
[REG_MIS] = UART011_MIS,
[REG_ICR] = UART011_ICR,
[REG_DMACR] = UART011_DMACR,
};
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
/* There is by now at least one vendor with differing details, so handle it */
struct vendor_data {
tty: amba-pl011: add register offset table to vendor data Add the register offset table to the vendor data, allowing vendor differences to be described in this table. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:31 +08:00
const u16 *reg_offset;
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
unsigned int ifls;
tty: amba-pl011: add support for 32-bit register access Add support for 32-bit register accesses to the AMBA PL011 UART. This is needed for ZTE UARTs, which require 32-bit accesses as opposed to the more normal 16-bit accesses. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:52 +08:00
bool access_32b;
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
bool oversampling;
ARM: PL011: add DMA burst threshold support for ST variants ST Micro variants has some specific dma burst threshold compensation, which allows them to make better use of a DMA controller. Add support to set this up. Based on a patch from Linus Walleij. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:59:16 +08:00
bool dma_threshold;
serial: pl011: implement workaround for CTS clear event issue Problem Observed: - interrupt status is set by rising or falling edge on CTS line - interrupt status is cleared on a .0. to .1. transition of the interrupt-clear register bit 1. - interrupt-clear register is reset by hardware once the interrupt status is .0.. Remark: It seems not possible to read this register back by the CPU though, but internally this register exists. - when simultaneous set and reset event on the interrupt status happens, then the set-event has priority and the status remains .1.. As a result the interrupt-clear register is not reset to .0., and no new .0. to .1. transition can be detected on it when writing a .1. to it. This implies race condition, the clear must be performed at least one UARTCLK the riding edge of CTS RIS interrupt. Fix: Instead of resetting UART as done in commit c16d51a32bbb61ac8fd96f78b5ce2fccfe0fb4c3 "amba pl011: workaround for uart registers lockup" do the following: write .0. and then .1. to the interrupt-clear register to make sure that this transition is detected. According to the datasheet writing a .0. does not have any effect, but actually it allows to reset the internal interrupt-clear register. Take into account: The .0. needs to last at least for one clk_uart clock period (~ 38 MHz, 26.08ns) This way we can do away with the tasklet and keep only a tiny fix triggered by the variant flag introduced in this patch. Signed-off-by: Guillaume Jaunet <guillaume.jaunet@stericsson.com> Signed-off-by: Christophe Arnal <christophe.arnal@stericsson.com> Signed-off-by: Matthias Locher <Matthias.Locher@stericsson.com> Signed-off-by: Rajanikanth H.V <rajanikanth.hv@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-26 17:17:02 +08:00
bool cts_event_workaround;
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
bool always_enabled;
drivers: PL011: allow to supply fixed option string The SBSA UART has a fixed baud rate and flow control setting, which cannot be changed or queried by software. Add a vendor specific property to always return fixed values when trying to read the console options. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:22 +08:00
bool fixed_options;
ARM: PL011: add support for extended FIFO-size of PL011-r1p5 The latest r1p5-revision of the ARM PL011 UART has 32-byte FIFOs, while all earlier ones have 16-byte FIFOs. This patch suggests a way to set the FIFO-size correctly & flexibly by using a member function named get_fifosize, rather than using the fifosize member variable. The function takes the UARTPeriphID, and returns the correct FIFO size. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-15 13:45:25 +08:00
tty: serial: amba-pl011: revise to use amba_rev macro This patch replaces the ugly bit-operations with the convenient amba_rev macro from the get_fifosize_arm function. The type of get_fifosize member function is also slightly changed to take a pointer to the amba_device. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-05-10 17:05:35 +08:00
unsigned int (*get_fifosize)(struct amba_device *dev);
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
};
tty: serial: amba-pl011: revise to use amba_rev macro This patch replaces the ugly bit-operations with the convenient amba_rev macro from the get_fifosize_arm function. The type of get_fifosize member function is also slightly changed to take a pointer to the amba_device. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-05-10 17:05:35 +08:00
static unsigned int get_fifosize_arm(struct amba_device *dev)
ARM: PL011: add support for extended FIFO-size of PL011-r1p5 The latest r1p5-revision of the ARM PL011 UART has 32-byte FIFOs, while all earlier ones have 16-byte FIFOs. This patch suggests a way to set the FIFO-size correctly & flexibly by using a member function named get_fifosize, rather than using the fifosize member variable. The function takes the UARTPeriphID, and returns the correct FIFO size. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-15 13:45:25 +08:00
{
tty: serial: amba-pl011: revise to use amba_rev macro This patch replaces the ugly bit-operations with the convenient amba_rev macro from the get_fifosize_arm function. The type of get_fifosize member function is also slightly changed to take a pointer to the amba_device. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-05-10 17:05:35 +08:00
return amba_rev(dev) < 3 ? 16 : 32;
ARM: PL011: add support for extended FIFO-size of PL011-r1p5 The latest r1p5-revision of the ARM PL011 UART has 32-byte FIFOs, while all earlier ones have 16-byte FIFOs. This patch suggests a way to set the FIFO-size correctly & flexibly by using a member function named get_fifosize, rather than using the fifosize member variable. The function takes the UARTPeriphID, and returns the correct FIFO size. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-15 13:45:25 +08:00
}
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
static struct vendor_data vendor_arm = {
tty: amba-pl011: add register offset table to vendor data Add the register offset table to the vendor data, allowing vendor differences to be described in this table. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:31 +08:00
.reg_offset = pl011_std_offsets,
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
.ifls = UART011_IFLS_RX4_8|UART011_IFLS_TX4_8,
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
.oversampling = false,
ARM: PL011: add DMA burst threshold support for ST variants ST Micro variants has some specific dma burst threshold compensation, which allows them to make better use of a DMA controller. Add support to set this up. Based on a patch from Linus Walleij. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:59:16 +08:00
.dma_threshold = false,
serial: pl011: implement workaround for CTS clear event issue Problem Observed: - interrupt status is set by rising or falling edge on CTS line - interrupt status is cleared on a .0. to .1. transition of the interrupt-clear register bit 1. - interrupt-clear register is reset by hardware once the interrupt status is .0.. Remark: It seems not possible to read this register back by the CPU though, but internally this register exists. - when simultaneous set and reset event on the interrupt status happens, then the set-event has priority and the status remains .1.. As a result the interrupt-clear register is not reset to .0., and no new .0. to .1. transition can be detected on it when writing a .1. to it. This implies race condition, the clear must be performed at least one UARTCLK the riding edge of CTS RIS interrupt. Fix: Instead of resetting UART as done in commit c16d51a32bbb61ac8fd96f78b5ce2fccfe0fb4c3 "amba pl011: workaround for uart registers lockup" do the following: write .0. and then .1. to the interrupt-clear register to make sure that this transition is detected. According to the datasheet writing a .0. does not have any effect, but actually it allows to reset the internal interrupt-clear register. Take into account: The .0. needs to last at least for one clk_uart clock period (~ 38 MHz, 26.08ns) This way we can do away with the tasklet and keep only a tiny fix triggered by the variant flag introduced in this patch. Signed-off-by: Guillaume Jaunet <guillaume.jaunet@stericsson.com> Signed-off-by: Christophe Arnal <christophe.arnal@stericsson.com> Signed-off-by: Matthias Locher <Matthias.Locher@stericsson.com> Signed-off-by: Rajanikanth H.V <rajanikanth.hv@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-26 17:17:02 +08:00
.cts_event_workaround = false,
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
.always_enabled = false,
drivers: PL011: allow to supply fixed option string The SBSA UART has a fixed baud rate and flow control setting, which cannot be changed or queried by software. Add a vendor specific property to always return fixed values when trying to read the console options. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:22 +08:00
.fixed_options = false,
ARM: PL011: add support for extended FIFO-size of PL011-r1p5 The latest r1p5-revision of the ARM PL011 UART has 32-byte FIFOs, while all earlier ones have 16-byte FIFOs. This patch suggests a way to set the FIFO-size correctly & flexibly by using a member function named get_fifosize, rather than using the fifosize member variable. The function takes the UARTPeriphID, and returns the correct FIFO size. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-15 13:45:25 +08:00
.get_fifosize = get_fifosize_arm,
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
};
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static struct vendor_data vendor_sbsa = {
tty: amba-pl011: add register offset table to vendor data Add the register offset table to the vendor data, allowing vendor differences to be described in this table. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:31 +08:00
.reg_offset = pl011_std_offsets,
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
.oversampling = false,
.dma_threshold = false,
.cts_event_workaround = false,
.always_enabled = true,
.fixed_options = true,
};
tty: amba-pl011: add ST register offset table Add the ST variant register offset table to the driver. Currently, this is an identical copy of the standard version, but this will be modified in the following changes. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:36 +08:00
static u16 pl011_st_offsets[REG_ARRAY_SIZE] = {
[REG_DR] = UART01x_DR,
[REG_ST_DMAWM] = ST_UART011_DMAWM,
[REG_ST_TIMEOUT] = ST_UART011_TIMEOUT,
[REG_FR] = UART01x_FR,
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
[REG_LCRH_RX] = ST_UART011_LCRH_RX,
[REG_LCRH_TX] = ST_UART011_LCRH_TX,
tty: amba-pl011: add ST register offset table Add the ST variant register offset table to the driver. Currently, this is an identical copy of the standard version, but this will be modified in the following changes. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:36 +08:00
[REG_IBRD] = UART011_IBRD,
[REG_FBRD] = UART011_FBRD,
[REG_CR] = UART011_CR,
[REG_IFLS] = UART011_IFLS,
[REG_IMSC] = UART011_IMSC,
[REG_RIS] = UART011_RIS,
[REG_MIS] = UART011_MIS,
[REG_ICR] = UART011_ICR,
[REG_DMACR] = UART011_DMACR,
[REG_ST_XFCR] = ST_UART011_XFCR,
[REG_ST_XON1] = ST_UART011_XON1,
[REG_ST_XON2] = ST_UART011_XON2,
[REG_ST_XOFF1] = ST_UART011_XOFF1,
[REG_ST_XOFF2] = ST_UART011_XOFF2,
[REG_ST_ITCR] = ST_UART011_ITCR,
[REG_ST_ITIP] = ST_UART011_ITIP,
[REG_ST_ABCR] = ST_UART011_ABCR,
[REG_ST_ABIMSC] = ST_UART011_ABIMSC,
};
tty: serial: amba-pl011: revise to use amba_rev macro This patch replaces the ugly bit-operations with the convenient amba_rev macro from the get_fifosize_arm function. The type of get_fifosize member function is also slightly changed to take a pointer to the amba_device. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-05-10 17:05:35 +08:00
static unsigned int get_fifosize_st(struct amba_device *dev)
ARM: PL011: add support for extended FIFO-size of PL011-r1p5 The latest r1p5-revision of the ARM PL011 UART has 32-byte FIFOs, while all earlier ones have 16-byte FIFOs. This patch suggests a way to set the FIFO-size correctly & flexibly by using a member function named get_fifosize, rather than using the fifosize member variable. The function takes the UARTPeriphID, and returns the correct FIFO size. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-15 13:45:25 +08:00
{
return 64;
}
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
static struct vendor_data vendor_st = {
tty: amba-pl011: add ST register offset table Add the ST variant register offset table to the driver. Currently, this is an identical copy of the standard version, but this will be modified in the following changes. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:36 +08:00
.reg_offset = pl011_st_offsets,
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
.ifls = UART011_IFLS_RX_HALF|UART011_IFLS_TX_HALF,
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
.oversampling = true,
ARM: PL011: add DMA burst threshold support for ST variants ST Micro variants has some specific dma burst threshold compensation, which allows them to make better use of a DMA controller. Add support to set this up. Based on a patch from Linus Walleij. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:59:16 +08:00
.dma_threshold = true,
serial: pl011: implement workaround for CTS clear event issue Problem Observed: - interrupt status is set by rising or falling edge on CTS line - interrupt status is cleared on a .0. to .1. transition of the interrupt-clear register bit 1. - interrupt-clear register is reset by hardware once the interrupt status is .0.. Remark: It seems not possible to read this register back by the CPU though, but internally this register exists. - when simultaneous set and reset event on the interrupt status happens, then the set-event has priority and the status remains .1.. As a result the interrupt-clear register is not reset to .0., and no new .0. to .1. transition can be detected on it when writing a .1. to it. This implies race condition, the clear must be performed at least one UARTCLK the riding edge of CTS RIS interrupt. Fix: Instead of resetting UART as done in commit c16d51a32bbb61ac8fd96f78b5ce2fccfe0fb4c3 "amba pl011: workaround for uart registers lockup" do the following: write .0. and then .1. to the interrupt-clear register to make sure that this transition is detected. According to the datasheet writing a .0. does not have any effect, but actually it allows to reset the internal interrupt-clear register. Take into account: The .0. needs to last at least for one clk_uart clock period (~ 38 MHz, 26.08ns) This way we can do away with the tasklet and keep only a tiny fix triggered by the variant flag introduced in this patch. Signed-off-by: Guillaume Jaunet <guillaume.jaunet@stericsson.com> Signed-off-by: Christophe Arnal <christophe.arnal@stericsson.com> Signed-off-by: Matthias Locher <Matthias.Locher@stericsson.com> Signed-off-by: Rajanikanth H.V <rajanikanth.hv@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-26 17:17:02 +08:00
.cts_event_workaround = true,
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
.always_enabled = false,
drivers: PL011: allow to supply fixed option string The SBSA UART has a fixed baud rate and flow control setting, which cannot be changed or queried by software. Add a vendor specific property to always return fixed values when trying to read the console options. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:22 +08:00
.fixed_options = false,
ARM: PL011: add support for extended FIFO-size of PL011-r1p5 The latest r1p5-revision of the ARM PL011 UART has 32-byte FIFOs, while all earlier ones have 16-byte FIFOs. This patch suggests a way to set the FIFO-size correctly & flexibly by using a member function named get_fifosize, rather than using the fifosize member variable. The function takes the UARTPeriphID, and returns the correct FIFO size. Signed-off-by: Jongsung Kim <neidhard.kim@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-15 13:45:25 +08:00
.get_fifosize = get_fifosize_st,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
};
tty: amba-pl011: add support for ZTE UART (EXPERIMENTAL) Add (incomplete) support for the ZTE UART to the AMBA PL011 driver. This is similar to the ARM and ST variants, except it has a different register address layout, and requires 32-bit accesses to the registers. Use the newly introduced register tables and access size support to cope with these differences. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:57 +08:00
static const u16 pl011_zte_offsets[REG_ARRAY_SIZE] = {
[REG_DR] = ZX_UART011_DR,
[REG_FR] = ZX_UART011_FR,
[REG_LCRH_RX] = ZX_UART011_LCRH,
[REG_LCRH_TX] = ZX_UART011_LCRH,
[REG_IBRD] = ZX_UART011_IBRD,
[REG_FBRD] = ZX_UART011_FBRD,
[REG_CR] = ZX_UART011_CR,
[REG_IFLS] = ZX_UART011_IFLS,
[REG_IMSC] = ZX_UART011_IMSC,
[REG_RIS] = ZX_UART011_RIS,
[REG_MIS] = ZX_UART011_MIS,
[REG_ICR] = ZX_UART011_ICR,
[REG_DMACR] = ZX_UART011_DMACR,
};
static struct vendor_data vendor_zte = {
.reg_offset = pl011_zte_offsets,
.access_32b = true,
.ifls = UART011_IFLS_RX4_8|UART011_IFLS_TX4_8,
.get_fifosize = get_fifosize_arm,
};
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/* Deals with DMA transactions */
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
struct pl011_sgbuf {
struct scatterlist sg;
char *buf;
};
struct pl011_dmarx_data {
struct dma_chan *chan;
struct completion complete;
bool use_buf_b;
struct pl011_sgbuf sgbuf_a;
struct pl011_sgbuf sgbuf_b;
dma_cookie_t cookie;
bool running;
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
struct timer_list timer;
unsigned int last_residue;
unsigned long last_jiffies;
bool auto_poll_rate;
unsigned int poll_rate;
unsigned int poll_timeout;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
};
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
struct pl011_dmatx_data {
struct dma_chan *chan;
struct scatterlist sg;
char *buf;
bool queued;
};
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
/*
* We wrap our port structure around the generic uart_port.
*/
struct uart_amba_port {
struct uart_port port;
tty: amba-pl011: add register lookup table Add a register lookup table, which allows the register offsets to be adjusted on a per-port basis. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:26 +08:00
const u16 *reg_offset;
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
struct clk *clk;
const struct vendor_data *vendor;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
unsigned int dmacr; /* dma control reg */
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
unsigned int im; /* interrupt mask */
unsigned int old_status;
ARM: PL011: Separate hardware FIFO size from TTY FIFO size With DMA support, we need to tell the TTY subsystem that the DMA buffer is the size of the FIFO, otherwise things like tty_wait_until_sent() will time out too early. Keep (and use) the hardware value separately from the port->fifosize. This was part of a larger patch from Linus Walleij, with a little modification. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:13:05 +08:00
unsigned int fifosize; /* vendor-specific */
amba-pl011: do not disable RTS during shutdown In present driver, shutdown clears RTS and DTR in CR register. But the documentation "Documentation/serial/driver" suggests not to disable RTS and DTR in shutdown(). Also RTS and DTR is preserved between shutdown and startup calls, i.e. these are restored in startup if they were enabled while doing shutdown. So that if RTS and DTR are set using pl011_set_mctrl then it should continue even after shutdown->startup sequence. For throttling/unthrottling user should call pl011_set_mctrl. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-18 18:23:59 +08:00
unsigned int old_cr; /* state during shutdown */
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
bool autorts;
drivers: PL011: allow to supply fixed option string The SBSA UART has a fixed baud rate and flow control setting, which cannot be changed or queried by software. Add a vendor specific property to always return fixed values when trying to read the console options. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:22 +08:00
unsigned int fixed_baud; /* vendor-set fixed baud rate */
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
char type[12];
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
#ifdef CONFIG_DMA_ENGINE
/* DMA stuff */
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
bool using_tx_dma;
bool using_rx_dma;
struct pl011_dmarx_data dmarx;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
struct pl011_dmatx_data dmatx;
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
bool dma_probed;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
#endif
};
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
static unsigned int pl011_reg_to_offset(const struct uart_amba_port *uap,
unsigned int reg)
{
tty: amba-pl011: add register lookup table Add a register lookup table, which allows the register offsets to be adjusted on a per-port basis. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:26 +08:00
return uap->reg_offset[reg];
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
}
tty: amba-pl011: convert accessor functions to take uart_amba_port Convert the new accessor functions to take the uart_amba_port instead of the port base address. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:03 +08:00
static unsigned int pl011_read(const struct uart_amba_port *uap,
unsigned int reg)
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
{
tty: amba-pl011: add support for 32-bit register access Add support for 32-bit register accesses to the AMBA PL011 UART. This is needed for ZTE UARTs, which require 32-bit accesses as opposed to the more normal 16-bit accesses. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:52 +08:00
void __iomem *addr = uap->port.membase + pl011_reg_to_offset(uap, reg);
tty: amba-pl011: use iotype instead of access_32b to track 32-bit I/O Instead of defining a new field in the uart_amba_port structure, use the existing iotype field of the uart_port structure, which is intended for this purpose. If we need to use 32-bit register access, we set iotype to UPIO_MEM32, otherwise we set it to UPIO_MEM. For early console, specify the "mmio32" option on the kernel command-line. Example: earlycon=pl011,mmio32,0x3ced1000 Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:42 +08:00
return (uap->port.iotype == UPIO_MEM32) ?
readl_relaxed(addr) : readw_relaxed(addr);
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
}
tty: amba-pl011: convert accessor functions to take uart_amba_port Convert the new accessor functions to take the uart_amba_port instead of the port base address. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:03 +08:00
static void pl011_write(unsigned int val, const struct uart_amba_port *uap,
unsigned int reg)
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
{
tty: amba-pl011: add support for 32-bit register access Add support for 32-bit register accesses to the AMBA PL011 UART. This is needed for ZTE UARTs, which require 32-bit accesses as opposed to the more normal 16-bit accesses. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:52 +08:00
void __iomem *addr = uap->port.membase + pl011_reg_to_offset(uap, reg);
tty: amba-pl011: use iotype instead of access_32b to track 32-bit I/O Instead of defining a new field in the uart_amba_port structure, use the existing iotype field of the uart_port structure, which is intended for this purpose. If we need to use 32-bit register access, we set iotype to UPIO_MEM32, otherwise we set it to UPIO_MEM. For early console, specify the "mmio32" option on the kernel command-line. Example: earlycon=pl011,mmio32,0x3ced1000 Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:42 +08:00
if (uap->port.iotype == UPIO_MEM32)
tty: amba-pl011: switch to using relaxed IO accessors Using relaxed IO accessors allows GCC to better optimise this code as we eliminate the heavy memory barriers - for example, GCC can now cache the address of a register across a read-modify-write sequence, rather than reloading the base address, offset and access size flag. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:41:02 +08:00
writel_relaxed(val, addr);
tty: amba-pl011: add support for 32-bit register access Add support for 32-bit register accesses to the AMBA PL011 UART. This is needed for ZTE UARTs, which require 32-bit accesses as opposed to the more normal 16-bit accesses. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:52 +08:00
else
tty: amba-pl011: switch to using relaxed IO accessors Using relaxed IO accessors allows GCC to better optimise this code as we eliminate the heavy memory barriers - for example, GCC can now cache the address of a register across a read-modify-write sequence, rather than reloading the base address, offset and access size flag. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:41:02 +08:00
writew_relaxed(val, addr);
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
}
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
/*
* Reads up to 256 characters from the FIFO or until it's empty and
* inserts them into the TTY layer. Returns the number of characters
* read from the FIFO.
*/
static int pl011_fifo_to_tty(struct uart_amba_port *uap)
{
serial: amba-pl011: fix incorrect integer size in pl011_fifo_to_tty() The UART_DUMMY_DR_RX status bit is equal to (1 << 16), so a u16 is too small to hold that value. The result is that UART_DUMMY_DR_RX is never passed to uart_insert_char(). This means that we're always accepting characters, even when CREAD (in termios) is not set. Signed-off-by: Timur Tabi <timur@codeaurora.org> Reviewed-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-10-08 04:27:16 +08:00
u16 status;
unsigned int ch, flag, max_count = 256;
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
int fifotaken = 0;
while (max_count--) {
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
status = pl011_read(uap, REG_FR);
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
if (status & UART01x_FR_RXFE)
break;
/* Take chars from the FIFO and update status */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
ch = pl011_read(uap, REG_DR) | UART_DUMMY_DR_RX;
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
flag = TTY_NORMAL;
uap->port.icount.rx++;
fifotaken++;
if (unlikely(ch & UART_DR_ERROR)) {
if (ch & UART011_DR_BE) {
ch &= ~(UART011_DR_FE | UART011_DR_PE);
uap->port.icount.brk++;
if (uart_handle_break(&uap->port))
continue;
} else if (ch & UART011_DR_PE)
uap->port.icount.parity++;
else if (ch & UART011_DR_FE)
uap->port.icount.frame++;
if (ch & UART011_DR_OE)
uap->port.icount.overrun++;
ch &= uap->port.read_status_mask;
if (ch & UART011_DR_BE)
flag = TTY_BREAK;
else if (ch & UART011_DR_PE)
flag = TTY_PARITY;
else if (ch & UART011_DR_FE)
flag = TTY_FRAME;
}
if (uart_handle_sysrq_char(&uap->port, ch & 255))
continue;
uart_insert_char(&uap->port, ch, UART011_DR_OE, ch, flag);
}
return fifotaken;
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/*
* All the DMA operation mode stuff goes inside this ifdef.
* This assumes that you have a generic DMA device interface,
* no custom DMA interfaces are supported.
*/
#ifdef CONFIG_DMA_ENGINE
#define PL011_DMA_BUFFER_SIZE PAGE_SIZE
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
static int pl011_sgbuf_init(struct dma_chan *chan, struct pl011_sgbuf *sg,
enum dma_data_direction dir)
{
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
dma_addr_t dma_addr;
sg->buf = dma_alloc_coherent(chan->device->dev,
PL011_DMA_BUFFER_SIZE, &dma_addr, GFP_KERNEL);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (!sg->buf)
return -ENOMEM;
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
sg_init_table(&sg->sg, 1);
sg_set_page(&sg->sg, phys_to_page(dma_addr),
PL011_DMA_BUFFER_SIZE, offset_in_page(dma_addr));
sg_dma_address(&sg->sg) = dma_addr;
serial: pl011: Set length of DMA transfer The DMA engines on some systems require that the dma_length is set when using scatter gather lists. Signed-off-by: Andrew Jackson <Andrew.Jackson@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-07 22:14:43 +08:00
sg_dma_len(&sg->sg) = PL011_DMA_BUFFER_SIZE;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
return 0;
}
static void pl011_sgbuf_free(struct dma_chan *chan, struct pl011_sgbuf *sg,
enum dma_data_direction dir)
{
if (sg->buf) {
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
dma_free_coherent(chan->device->dev,
PL011_DMA_BUFFER_SIZE, sg->buf,
sg_dma_address(&sg->sg));
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
}
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
static void pl011_dma_probe(struct uart_amba_port *uap)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
{
/* DMA is the sole user of the platform data right now */
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
struct amba_pl011_data *plat = dev_get_platdata(uap->port.dev);
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
struct device *dev = uap->port.dev;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
struct dma_slave_config tx_conf = {
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
.dst_addr = uap->port.mapbase +
pl011_reg_to_offset(uap, REG_DR),
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
.dst_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE,
spi, serial: move to dma_transfer_direction fixup usage of dma direction by introducing dma_transfer_direction, this patch moves spi, serial drivers to use new enum Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com> Cc: Grant Likely <grant.likely@secretlab.ca> Cc: Alan Cox <alan@linux.intel.com> Acked-by: Mika Westerberg <mika.westerberg@iki.fi>
2011-10-14 13:17:38 +08:00
.direction = DMA_MEM_TO_DEV,
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
.dst_maxburst = uap->fifosize >> 1,
dmaengine: Pass dma_slave_config .device_fc = NULL for all existing users .device_fc is added in struct dma_slave_config recently. All user drivers, which want DMA to be the flow controller must pass this field as false. As earlier driver don't look to use this feature, mark it false for now. Signed-off-by: Viresh Kumar <viresh.kumar@st.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>
2012-02-01 18:42:19 +08:00
.device_fc = false,
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
};
struct dma_chan *chan;
dma_cap_mask_t mask;
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
uap->dma_probed = true;
chan = dma_request_slave_channel_reason(dev, "tx");
if (IS_ERR(chan)) {
if (PTR_ERR(chan) == -EPROBE_DEFER) {
uap->dma_probed = false;
return;
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
serial: pl011: use generic DMA slave configuration if possible With the new OF DMA binding, it is possible to completely avoid the need for platform_data for configuring a DMA channel. In cases where the platform has already been converted, calling dma_request_slave_channel should get all the necessary information from the device tree. This also adds a binding document specific to the pl011 controller, and extends the generic primecell binding to mention "dmas" and other common properties. Like the patch that converts the dw_dma controller, this is completely untested and is looking for someone to try it out. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Grant Likely <grant.likely@secretlab.ca> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Viresh Kumar <viresh.kumar@linaro.org> Cc: devicetree-discuss@lists.ozlabs.org Cc: linux-arm-kernel@lists.infradead.org
2013-01-29 00:24:37 +08:00
/* We need platform data */
if (!plat || !plat->dma_filter) {
dev_info(uap->port.dev, "no DMA platform data\n");
return;
}
/* Try to acquire a generic DMA engine slave TX channel */
dma_cap_zero(mask);
dma_cap_set(DMA_SLAVE, mask);
chan = dma_request_channel(mask, plat->dma_filter,
plat->dma_tx_param);
if (!chan) {
dev_err(uap->port.dev, "no TX DMA channel!\n");
return;
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
dmaengine_slave_config(chan, &tx_conf);
uap->dmatx.chan = chan;
dev_info(uap->port.dev, "DMA channel TX %s\n",
dma_chan_name(uap->dmatx.chan));
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
/* Optionally make use of an RX channel as well */
serial: pl011: use generic DMA slave configuration if possible With the new OF DMA binding, it is possible to completely avoid the need for platform_data for configuring a DMA channel. In cases where the platform has already been converted, calling dma_request_slave_channel should get all the necessary information from the device tree. This also adds a binding document specific to the pl011 controller, and extends the generic primecell binding to mention "dmas" and other common properties. Like the patch that converts the dw_dma controller, this is completely untested and is looking for someone to try it out. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Grant Likely <grant.likely@secretlab.ca> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Viresh Kumar <viresh.kumar@linaro.org> Cc: devicetree-discuss@lists.ozlabs.org Cc: linux-arm-kernel@lists.infradead.org
2013-01-29 00:24:37 +08:00
chan = dma_request_slave_channel(dev, "rx");
tty/serial: pl011: add generic earlycon support Add earlycon support for the pl011 serial port. This allows enabling the pl011 for console when early_params are processed. This is based on the arm64 earlyprintk support and is intended to replace it. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-19 06:19:57 +08:00
serial: pl011: use generic DMA slave configuration if possible With the new OF DMA binding, it is possible to completely avoid the need for platform_data for configuring a DMA channel. In cases where the platform has already been converted, calling dma_request_slave_channel should get all the necessary information from the device tree. This also adds a binding document specific to the pl011 controller, and extends the generic primecell binding to mention "dmas" and other common properties. Like the patch that converts the dw_dma controller, this is completely untested and is looking for someone to try it out. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Grant Likely <grant.likely@secretlab.ca> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Viresh Kumar <viresh.kumar@linaro.org> Cc: devicetree-discuss@lists.ozlabs.org Cc: linux-arm-kernel@lists.infradead.org
2013-01-29 00:24:37 +08:00
if (!chan && plat->dma_rx_param) {
chan = dma_request_channel(mask, plat->dma_filter, plat->dma_rx_param);
if (!chan) {
dev_err(uap->port.dev, "no RX DMA channel!\n");
return;
}
}
if (chan) {
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
struct dma_slave_config rx_conf = {
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
.src_addr = uap->port.mapbase +
pl011_reg_to_offset(uap, REG_DR),
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
.src_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE,
spi, serial: move to dma_transfer_direction fixup usage of dma direction by introducing dma_transfer_direction, this patch moves spi, serial drivers to use new enum Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com> Cc: Grant Likely <grant.likely@secretlab.ca> Cc: Alan Cox <alan@linux.intel.com> Acked-by: Mika Westerberg <mika.westerberg@iki.fi>
2011-10-14 13:17:38 +08:00
.direction = DMA_DEV_TO_MEM,
serial: pl011: change Rx burst size to half of trigger level The amba-pl011.c driver sets DMA burst size equal to FIFO trigger level. If now exactly DMA burst size bytes are received, the DMAC will retrieve them all and no Rx timeout interrupt will be generated. To fix that set the burst size to half the FIFO trigger level. Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-13 01:47:17 +08:00
.src_maxburst = uap->fifosize >> 2,
dmaengine: Pass dma_slave_config .device_fc = NULL for all existing users .device_fc is added in struct dma_slave_config recently. All user drivers, which want DMA to be the flow controller must pass this field as false. As earlier driver don't look to use this feature, mark it false for now. Signed-off-by: Viresh Kumar <viresh.kumar@st.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>
2012-02-01 18:42:19 +08:00
.device_fc = false,
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
};
serial: pl011: Don't enable RX DMA if residue processing not supported If the DMA engine doesn't support residue processing then the RX DMA handling won't work terribly well if polling is enabled. So, disable RX DMA if residue handling isn't available. Signed-off-by: Andrew Jackson <Andrew.Jackson@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-07 22:14:47 +08:00
struct dma_slave_caps caps;
/*
* Some DMA controllers provide information on their capabilities.
* If the controller does, check for suitable residue processing
* otherwise assime all is well.
*/
if (0 == dma_get_slave_caps(chan, &caps)) {
if (caps.residue_granularity ==
DMA_RESIDUE_GRANULARITY_DESCRIPTOR) {
dma_release_channel(chan);
dev_info(uap->port.dev,
"RX DMA disabled - no residue processing\n");
return;
}
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
dmaengine_slave_config(chan, &rx_conf);
uap->dmarx.chan = chan;
serial: pl011: Add device tree support for RX DMA polling Add equivalent attributes to those provided in the platform data for use when RX DMA is enabled. Signed-off-by: Andrew Jackson <Andrew.Jackson@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-07 22:14:23 +08:00
uap->dmarx.auto_poll_rate = false;
Revert "serial: pl011: use DMA RX polling by default" This reverts commit 17438217a6f5e33d920ed3821a4b857311cc2872 on request of Linus Walleij: Greg can you please drop or revert commit 17438217a6f5e33d920ed3821a4b857311cc2872 "serial: pl011: use DMA RX polling by default" from the TTY tree until this has been sorted out? Cc: Linus Walleij <linus.walleij@linaro.org> Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-18 01:33:18 +08:00
if (plat && plat->dma_rx_poll_enable) {
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
/* Set poll rate if specified. */
if (plat->dma_rx_poll_rate) {
uap->dmarx.auto_poll_rate = false;
uap->dmarx.poll_rate = plat->dma_rx_poll_rate;
} else {
/*
* 100 ms defaults to poll rate if not
* specified. This will be adjusted with
* the baud rate at set_termios.
*/
uap->dmarx.auto_poll_rate = true;
uap->dmarx.poll_rate = 100;
}
/* 3 secs defaults poll_timeout if not specified. */
if (plat->dma_rx_poll_timeout)
uap->dmarx.poll_timeout =
plat->dma_rx_poll_timeout;
else
uap->dmarx.poll_timeout = 3000;
serial: pl011: Add device tree support for RX DMA polling Add equivalent attributes to those provided in the platform data for use when RX DMA is enabled. Signed-off-by: Andrew Jackson <Andrew.Jackson@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-07 22:14:23 +08:00
} else if (!plat && dev->of_node) {
uap->dmarx.auto_poll_rate = of_property_read_bool(
dev->of_node, "auto-poll");
if (uap->dmarx.auto_poll_rate) {
u32 x;
if (0 == of_property_read_u32(dev->of_node,
"poll-rate-ms", &x))
uap->dmarx.poll_rate = x;
else
uap->dmarx.poll_rate = 100;
if (0 == of_property_read_u32(dev->of_node,
"poll-timeout-ms", &x))
uap->dmarx.poll_timeout = x;
else
uap->dmarx.poll_timeout = 3000;
}
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
dev_info(uap->port.dev, "DMA channel RX %s\n",
dma_chan_name(uap->dmarx.chan));
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
static void pl011_dma_remove(struct uart_amba_port *uap)
{
if (uap->dmatx.chan)
dma_release_channel(uap->dmatx.chan);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (uap->dmarx.chan)
dma_release_channel(uap->dmarx.chan);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
/* Forward declare these for the refill routine */
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
static int pl011_dma_tx_refill(struct uart_amba_port *uap);
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
static void pl011_start_tx_pio(struct uart_amba_port *uap);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/*
* The current DMA TX buffer has been sent.
* Try to queue up another DMA buffer.
*/
static void pl011_dma_tx_callback(void *data)
{
struct uart_amba_port *uap = data;
struct pl011_dmatx_data *dmatx = &uap->dmatx;
unsigned long flags;
u16 dmacr;
spin_lock_irqsave(&uap->port.lock, flags);
if (uap->dmatx.queued)
dma_unmap_sg(dmatx->chan->device->dev, &dmatx->sg, 1,
DMA_TO_DEVICE);
dmacr = uap->dmacr;
uap->dmacr = dmacr & ~UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/*
* If TX DMA was disabled, it means that we've stopped the DMA for
* some reason (eg, XOFF received, or we want to send an X-char.)
*
* Note: we need to be careful here of a potential race between DMA
* and the rest of the driver - if the driver disables TX DMA while
* a TX buffer completing, we must update the tx queued status to
* get further refills (hence we check dmacr).
*/
if (!(dmacr & UART011_TXDMAE) || uart_tx_stopped(&uap->port) ||
uart_circ_empty(&uap->port.state->xmit)) {
uap->dmatx.queued = false;
spin_unlock_irqrestore(&uap->port.lock, flags);
return;
}
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
if (pl011_dma_tx_refill(uap) <= 0)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/*
* We didn't queue a DMA buffer for some reason, but we
* have data pending to be sent. Re-enable the TX IRQ.
*/
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
pl011_start_tx_pio(uap);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
spin_unlock_irqrestore(&uap->port.lock, flags);
}
/*
* Try to refill the TX DMA buffer.
* Locking: called with port lock held and IRQs disabled.
* Returns:
* 1 if we queued up a TX DMA buffer.
* 0 if we didn't want to handle this by DMA
* <0 on error
*/
static int pl011_dma_tx_refill(struct uart_amba_port *uap)
{
struct pl011_dmatx_data *dmatx = &uap->dmatx;
struct dma_chan *chan = dmatx->chan;
struct dma_device *dma_dev = chan->device;
struct dma_async_tx_descriptor *desc;
struct circ_buf *xmit = &uap->port.state->xmit;
unsigned int count;
/*
* Try to avoid the overhead involved in using DMA if the
* transaction fits in the first half of the FIFO, by using
* the standard interrupt handling. This ensures that we
* issue a uart_write_wakeup() at the appropriate time.
*/
count = uart_circ_chars_pending(xmit);
if (count < (uap->fifosize >> 1)) {
uap->dmatx.queued = false;
return 0;
}
/*
* Bodge: don't send the last character by DMA, as this
* will prevent XON from notifying us to restart DMA.
*/
count -= 1;
/* Else proceed to copy the TX chars to the DMA buffer and fire DMA */
if (count > PL011_DMA_BUFFER_SIZE)
count = PL011_DMA_BUFFER_SIZE;
if (xmit->tail < xmit->head)
memcpy(&dmatx->buf[0], &xmit->buf[xmit->tail], count);
else {
size_t first = UART_XMIT_SIZE - xmit->tail;
serial: pl011: Avoid assumptions about buffer size when refilling TX DMA The existing code assumed that PL011_DMA_BUFFER_SIZE == UART_XMIT_SIZE, which may not always be the case. This allows for these two being different sizes and not copying too much data. Signed-off-by: Andrew Jackson <Andrew.Jackson@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-07 22:14:39 +08:00
size_t second;
if (first > count)
first = count;
second = count - first;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
memcpy(&dmatx->buf[0], &xmit->buf[xmit->tail], first);
if (second)
memcpy(&dmatx->buf[first], &xmit->buf[0], second);
}
dmatx->sg.length = count;
if (dma_map_sg(dma_dev->dev, &dmatx->sg, 1, DMA_TO_DEVICE) != 1) {
uap->dmatx.queued = false;
dev_dbg(uap->port.dev, "unable to map TX DMA\n");
return -EBUSY;
}
dmaengine/dma_slave: introduce inline wrappers Add inline wrappers for device_prep_slave_sg() and device_prep_dma_cyclic() interfaces to hide new parameter from current users of affected interfaces. Convert current users to use new wrappers instead of direct calls. Suggested by Russell King [https://lkml.org/lkml/2012/2/3/269]. Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com> Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>
2012-03-09 05:11:18 +08:00
desc = dmaengine_prep_slave_sg(chan, &dmatx->sg, 1, DMA_MEM_TO_DEV,
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
if (!desc) {
dma_unmap_sg(dma_dev->dev, &dmatx->sg, 1, DMA_TO_DEVICE);
uap->dmatx.queued = false;
/*
* If DMA cannot be used right now, we complete this
* transaction via IRQ and let the TTY layer retry.
*/
dev_dbg(uap->port.dev, "TX DMA busy\n");
return -EBUSY;
}
/* Some data to go along to the callback */
desc->callback = pl011_dma_tx_callback;
desc->callback_param = uap;
/* All errors should happen at prepare time */
dmaengine_submit(desc);
/* Fire the DMA transaction */
dma_dev->device_issue_pending(chan);
uap->dmacr |= UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
uap->dmatx.queued = true;
/*
* Now we know that DMA will fire, so advance the ring buffer
* with the stuff we just dispatched.
*/
xmit->tail = (xmit->tail + count) & (UART_XMIT_SIZE - 1);
uap->port.icount.tx += count;
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(&uap->port);
return 1;
}
/*
* We received a transmit interrupt without a pending X-char but with
* pending characters.
* Locking: called with port lock held and IRQs disabled.
* Returns:
* false if we want to use PIO to transmit
* true if we queued a DMA buffer
*/
static bool pl011_dma_tx_irq(struct uart_amba_port *uap)
{
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (!uap->using_tx_dma)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return false;
/*
* If we already have a TX buffer queued, but received a
* TX interrupt, it will be because we've just sent an X-char.
* Ensure the TX DMA is enabled and the TX IRQ is disabled.
*/
if (uap->dmatx.queued) {
uap->dmacr |= UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
uap->im &= ~UART011_TXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return true;
}
/*
* We don't have a TX buffer queued, so try to queue one.
Fix common misspellings Fixes generated by 'codespell' and manually reviewed. Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
2011-03-31 09:57:33 +08:00
* If we successfully queued a buffer, mask the TX IRQ.
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
*/
if (pl011_dma_tx_refill(uap) > 0) {
uap->im &= ~UART011_TXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return true;
}
return false;
}
/*
* Stop the DMA transmit (eg, due to received XOFF).
* Locking: called with port lock held and IRQs disabled.
*/
static inline void pl011_dma_tx_stop(struct uart_amba_port *uap)
{
if (uap->dmatx.queued) {
uap->dmacr &= ~UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
}
/*
* Try to start a DMA transmit, or in the case of an XON/OFF
* character queued for send, try to get that character out ASAP.
* Locking: called with port lock held and IRQs disabled.
* Returns:
* false if we want the TX IRQ to be enabled
* true if we have a buffer queued
*/
static inline bool pl011_dma_tx_start(struct uart_amba_port *uap)
{
u16 dmacr;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (!uap->using_tx_dma)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return false;
if (!uap->port.x_char) {
/* no X-char, try to push chars out in DMA mode */
bool ret = true;
if (!uap->dmatx.queued) {
if (pl011_dma_tx_refill(uap) > 0) {
uap->im &= ~UART011_TXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
} else
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
ret = false;
} else if (!(uap->dmacr & UART011_TXDMAE)) {
uap->dmacr |= UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
return ret;
}
/*
* We have an X-char to send. Disable DMA to prevent it loading
* the TX fifo, and then see if we can stuff it into the FIFO.
*/
dmacr = uap->dmacr;
uap->dmacr &= ~UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
if (pl011_read(uap, REG_FR) & UART01x_FR_TXFF) {
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/*
* No space in the FIFO, so enable the transmit interrupt
* so we know when there is space. Note that once we've
* loaded the character, we should just re-enable DMA.
*/
return false;
}
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->port.x_char, uap, REG_DR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
uap->port.icount.tx++;
uap->port.x_char = 0;
/* Success - restore the DMA state */
uap->dmacr = dmacr;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return true;
}
/*
* Flush the transmit buffer.
* Locking: called with port lock held and IRQs disabled.
*/
static void pl011_dma_flush_buffer(struct uart_port *port)
serial: amba-pl011: Use __releases/__acquires annotations Fix the following sparse warnings: drivers/tty/serial/amba-pl011.c:687:20: warning: context imbalance in 'pl011_dma_flush_buffer' - unexpected unlock drivers/tty/serial/amba-pl011.c:1200:13: warning: context imbalance in 'pl011_rx_chars' - unexpected unlock Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-08-10 04:58:51 +08:00
__releases(&uap->port.lock)
__acquires(&uap->port.lock)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (!uap->using_tx_dma)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return;
/* Avoid deadlock with the DMA engine callback */
spin_unlock(&uap->port.lock);
dmaengine_terminate_all(uap->dmatx.chan);
spin_lock(&uap->port.lock);
if (uap->dmatx.queued) {
dma_unmap_sg(uap->dmatx.chan->device->dev, &uap->dmatx.sg, 1,
DMA_TO_DEVICE);
uap->dmatx.queued = false;
uap->dmacr &= ~UART011_TXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
static void pl011_dma_rx_callback(void *data);
static int pl011_dma_rx_trigger_dma(struct uart_amba_port *uap)
{
struct dma_chan *rxchan = uap->dmarx.chan;
struct pl011_dmarx_data *dmarx = &uap->dmarx;
struct dma_async_tx_descriptor *desc;
struct pl011_sgbuf *sgbuf;
if (!rxchan)
return -EIO;
/* Start the RX DMA job */
sgbuf = uap->dmarx.use_buf_b ?
&uap->dmarx.sgbuf_b : &uap->dmarx.sgbuf_a;
dmaengine/dma_slave: introduce inline wrappers Add inline wrappers for device_prep_slave_sg() and device_prep_dma_cyclic() interfaces to hide new parameter from current users of affected interfaces. Convert current users to use new wrappers instead of direct calls. Suggested by Russell King [https://lkml.org/lkml/2012/2/3/269]. Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com> Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>
2012-03-09 05:11:18 +08:00
desc = dmaengine_prep_slave_sg(rxchan, &sgbuf->sg, 1,
spi, serial: move to dma_transfer_direction fixup usage of dma direction by introducing dma_transfer_direction, this patch moves spi, serial drivers to use new enum Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com> Cc: Grant Likely <grant.likely@secretlab.ca> Cc: Alan Cox <alan@linux.intel.com> Acked-by: Mika Westerberg <mika.westerberg@iki.fi>
2011-10-14 13:17:38 +08:00
DMA_DEV_TO_MEM,
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
/*
* If the DMA engine is busy and cannot prepare a
* channel, no big deal, the driver will fall back
* to interrupt mode as a result of this error code.
*/
if (!desc) {
uap->dmarx.running = false;
dmaengine_terminate_all(rxchan);
return -EBUSY;
}
/* Some data to go along to the callback */
desc->callback = pl011_dma_rx_callback;
desc->callback_param = uap;
dmarx->cookie = dmaengine_submit(desc);
dma_async_issue_pending(rxchan);
uap->dmacr |= UART011_RXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
uap->dmarx.running = true;
uap->im &= ~UART011_RXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
return 0;
}
/*
* This is called when either the DMA job is complete, or
* the FIFO timeout interrupt occurred. This must be called
* with the port spinlock uap->port.lock held.
*/
static void pl011_dma_rx_chars(struct uart_amba_port *uap,
u32 pending, bool use_buf_b,
bool readfifo)
{
TTY: switch tty_insert_flip_string Now, we start converting tty buffer functions to actually use tty_port. This will allow us to get rid of the need of tty in many call sites. Only tty_port will needed and hence no more tty_port_tty_get in those paths. tty_insert_flip_string this time. Signed-off-by: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-01-03 22:53:04 +08:00
struct tty_port *port = &uap->port.state->port;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
struct pl011_sgbuf *sgbuf = use_buf_b ?
&uap->dmarx.sgbuf_b : &uap->dmarx.sgbuf_a;
int dma_count = 0;
u32 fifotaken = 0; /* only used for vdbg() */
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
struct pl011_dmarx_data *dmarx = &uap->dmarx;
int dmataken = 0;
if (uap->dmarx.poll_rate) {
/* The data can be taken by polling */
dmataken = sgbuf->sg.length - dmarx->last_residue;
/* Recalculate the pending size */
if (pending >= dmataken)
pending -= dmataken;
}
/* Pick the remain data from the DMA */
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (pending) {
/*
* First take all chars in the DMA pipe, then look in the FIFO.
* Note that tty_insert_flip_buf() tries to take as many chars
* as it can.
*/
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
dma_count = tty_insert_flip_string(port, sgbuf->buf + dmataken,
pending);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
uap->port.icount.rx += dma_count;
if (dma_count < pending)
dev_warn(uap->port.dev,
"couldn't insert all characters (TTY is full?)\n");
}
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
/* Reset the last_residue for Rx DMA poll */
if (uap->dmarx.poll_rate)
dmarx->last_residue = sgbuf->sg.length;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
/*
* Only continue with trying to read the FIFO if all DMA chars have
* been taken first.
*/
if (dma_count == pending && readfifo) {
/* Clear any error flags */
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
pl011_write(UART011_OEIS | UART011_BEIS | UART011_PEIS |
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
UART011_FEIS, uap, REG_ICR);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
/*
* If we read all the DMA'd characters, and we had an
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
* incomplete buffer, that could be due to an rx error, or
* maybe we just timed out. Read any pending chars and check
* the error status.
*
* Error conditions will only occur in the FIFO, these will
* trigger an immediate interrupt and stop the DMA job, so we
* will always find the error in the FIFO, never in the DMA
* buffer.
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
*/
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
fifotaken = pl011_fifo_to_tty(uap);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
spin_unlock(&uap->port.lock);
dev_vdbg(uap->port.dev,
"Took %d chars from DMA buffer and %d chars from the FIFO\n",
dma_count, fifotaken);
TTY: switch tty_flip_buffer_push Now, we start converting tty buffer functions to actually use tty_port. This will allow us to get rid of the need of tty in many call sites. Only tty_port will needed and hence no more tty_port_tty_get in those paths. Now, the one where most of tty_port_tty_get gets removed: tty_flip_buffer_push. IOW we also closed all the races in drivers not using tty_port_tty_get at all yet. Also we move tty_flip_buffer_push declaration from include/linux/tty.h to include/linux/tty_flip.h to all others while we are changing it anyway. Signed-off-by: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-01-03 22:53:06 +08:00
tty_flip_buffer_push(port);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
spin_lock(&uap->port.lock);
}
static void pl011_dma_rx_irq(struct uart_amba_port *uap)
{
struct pl011_dmarx_data *dmarx = &uap->dmarx;
struct dma_chan *rxchan = dmarx->chan;
struct pl011_sgbuf *sgbuf = dmarx->use_buf_b ?
&dmarx->sgbuf_b : &dmarx->sgbuf_a;
size_t pending;
struct dma_tx_state state;
enum dma_status dmastat;
/*
* Pause the transfer so we can trust the current counter,
* do this before we pause the PL011 block, else we may
* overflow the FIFO.
*/
if (dmaengine_pause(rxchan))
dev_err(uap->port.dev, "unable to pause DMA transfer\n");
dmastat = rxchan->device->device_tx_status(rxchan,
dmarx->cookie, &state);
if (dmastat != DMA_PAUSED)
dev_err(uap->port.dev, "unable to pause DMA transfer\n");
/* Disable RX DMA - incoming data will wait in the FIFO */
uap->dmacr &= ~UART011_RXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
uap->dmarx.running = false;
pending = sgbuf->sg.length - state.residue;
BUG_ON(pending > PL011_DMA_BUFFER_SIZE);
/* Then we terminate the transfer - we now know our residue */
dmaengine_terminate_all(rxchan);
/*
* This will take the chars we have so far and insert
* into the framework.
*/
pl011_dma_rx_chars(uap, pending, dmarx->use_buf_b, true);
/* Switch buffer & re-trigger DMA job */
dmarx->use_buf_b = !dmarx->use_buf_b;
if (pl011_dma_rx_trigger_dma(uap)) {
dev_dbg(uap->port.dev, "could not retrigger RX DMA job "
"fall back to interrupt mode\n");
uap->im |= UART011_RXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
}
static void pl011_dma_rx_callback(void *data)
{
struct uart_amba_port *uap = data;
struct pl011_dmarx_data *dmarx = &uap->dmarx;
amba-pl011​/dma: Add check for the residue in DMA callback In DMA-operated uart, I found that rx data can be taken by the UART interrupts during the DMA irq handler. pl011_int is occurred just before it goes inside spin_lock_irq. When it returns to the callback, DMA buffer already has been flushed. Then, pl011_dma_rx_chars gets invalid data. So I add check for the residue as the patch bellow. Signed-off-by: Chanho Min <chanho.min@lge.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-02-20 09:24:40 +08:00
struct dma_chan *rxchan = dmarx->chan;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
bool lastbuf = dmarx->use_buf_b;
amba-pl011​/dma: Add check for the residue in DMA callback In DMA-operated uart, I found that rx data can be taken by the UART interrupts during the DMA irq handler. pl011_int is occurred just before it goes inside spin_lock_irq. When it returns to the callback, DMA buffer already has been flushed. Then, pl011_dma_rx_chars gets invalid data. So I add check for the residue as the patch bellow. Signed-off-by: Chanho Min <chanho.min@lge.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-02-20 09:24:40 +08:00
struct pl011_sgbuf *sgbuf = dmarx->use_buf_b ?
&dmarx->sgbuf_b : &dmarx->sgbuf_a;
size_t pending;
struct dma_tx_state state;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
int ret;
/*
* This completion interrupt occurs typically when the
* RX buffer is totally stuffed but no timeout has yet
* occurred. When that happens, we just want the RX
* routine to flush out the secondary DMA buffer while
* we immediately trigger the next DMA job.
*/
spin_lock_irq(&uap->port.lock);
amba-pl011​/dma: Add check for the residue in DMA callback In DMA-operated uart, I found that rx data can be taken by the UART interrupts during the DMA irq handler. pl011_int is occurred just before it goes inside spin_lock_irq. When it returns to the callback, DMA buffer already has been flushed. Then, pl011_dma_rx_chars gets invalid data. So I add check for the residue as the patch bellow. Signed-off-by: Chanho Min <chanho.min@lge.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-02-20 09:24:40 +08:00
/*
* Rx data can be taken by the UART interrupts during
* the DMA irq handler. So we check the residue here.
*/
rxchan->device->device_tx_status(rxchan, dmarx->cookie, &state);
pending = sgbuf->sg.length - state.residue;
BUG_ON(pending > PL011_DMA_BUFFER_SIZE);
/* Then we terminate the transfer - we now know our residue */
dmaengine_terminate_all(rxchan);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
uap->dmarx.running = false;
dmarx->use_buf_b = !lastbuf;
ret = pl011_dma_rx_trigger_dma(uap);
amba-pl011​/dma: Add check for the residue in DMA callback In DMA-operated uart, I found that rx data can be taken by the UART interrupts during the DMA irq handler. pl011_int is occurred just before it goes inside spin_lock_irq. When it returns to the callback, DMA buffer already has been flushed. Then, pl011_dma_rx_chars gets invalid data. So I add check for the residue as the patch bellow. Signed-off-by: Chanho Min <chanho.min@lge.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-02-20 09:24:40 +08:00
pl011_dma_rx_chars(uap, pending, lastbuf, false);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
spin_unlock_irq(&uap->port.lock);
/*
* Do this check after we picked the DMA chars so we don't
* get some IRQ immediately from RX.
*/
if (ret) {
dev_dbg(uap->port.dev, "could not retrigger RX DMA job "
"fall back to interrupt mode\n");
uap->im |= UART011_RXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
}
/*
* Stop accepting received characters, when we're shutting down or
* suspending this port.
* Locking: called with port lock held and IRQs disabled.
*/
static inline void pl011_dma_rx_stop(struct uart_amba_port *uap)
{
/* FIXME. Just disable the DMA enable */
uap->dmacr &= ~UART011_RXDMAE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
/*
* Timer handler for Rx DMA polling.
* Every polling, It checks the residue in the dma buffer and transfer
* data to the tty. Also, last_residue is updated for the next polling.
*/
static void pl011_dma_rx_poll(unsigned long args)
{
struct uart_amba_port *uap = (struct uart_amba_port *)args;
struct tty_port *port = &uap->port.state->port;
struct pl011_dmarx_data *dmarx = &uap->dmarx;
struct dma_chan *rxchan = uap->dmarx.chan;
unsigned long flags = 0;
unsigned int dmataken = 0;
unsigned int size = 0;
struct pl011_sgbuf *sgbuf;
int dma_count;
struct dma_tx_state state;
sgbuf = dmarx->use_buf_b ? &uap->dmarx.sgbuf_b : &uap->dmarx.sgbuf_a;
rxchan->device->device_tx_status(rxchan, dmarx->cookie, &state);
if (likely(state.residue < dmarx->last_residue)) {
dmataken = sgbuf->sg.length - dmarx->last_residue;
size = dmarx->last_residue - state.residue;
dma_count = tty_insert_flip_string(port, sgbuf->buf + dmataken,
size);
if (dma_count == size)
dmarx->last_residue = state.residue;
dmarx->last_jiffies = jiffies;
}
tty_flip_buffer_push(port);
/*
* If no data is received in poll_timeout, the driver will fall back
* to interrupt mode. We will retrigger DMA at the first interrupt.
*/
if (jiffies_to_msecs(jiffies - dmarx->last_jiffies)
> uap->dmarx.poll_timeout) {
spin_lock_irqsave(&uap->port.lock, flags);
pl011_dma_rx_stop(uap);
serial: pl011: fix fall back from DMA to interrupt mode When falling back from DMA to interrupt mode the receive interrupt has to be re-enabled to catch new incoming data. Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 21:54:47 +08:00
uap->im |= UART011_RXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
spin_unlock_irqrestore(&uap->port.lock, flags);
uap->dmarx.running = false;
dmaengine_terminate_all(rxchan);
del_timer(&uap->dmarx.timer);
} else {
mod_timer(&uap->dmarx.timer,
jiffies + msecs_to_jiffies(uap->dmarx.poll_rate));
}
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
static void pl011_dma_startup(struct uart_amba_port *uap)
{
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
int ret;
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
if (!uap->dma_probed)
pl011_dma_probe(uap);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
if (!uap->dmatx.chan)
return;
serial: pl011: Allocate TX DMA buffer from DMA capable memory Allocating with __GFP_DMA avoids the need for bounce buffers Signed-off-by: Andrew Jackson <Andrew.Jackson@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-07 22:14:35 +08:00
uap->dmatx.buf = kmalloc(PL011_DMA_BUFFER_SIZE, GFP_KERNEL | __GFP_DMA);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
if (!uap->dmatx.buf) {
dev_err(uap->port.dev, "no memory for DMA TX buffer\n");
uap->port.fifosize = uap->fifosize;
return;
}
sg_init_one(&uap->dmatx.sg, uap->dmatx.buf, PL011_DMA_BUFFER_SIZE);
/* The DMA buffer is now the FIFO the TTY subsystem can use */
uap->port.fifosize = PL011_DMA_BUFFER_SIZE;
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
uap->using_tx_dma = true;
if (!uap->dmarx.chan)
goto skip_rx;
/* Allocate and map DMA RX buffers */
ret = pl011_sgbuf_init(uap->dmarx.chan, &uap->dmarx.sgbuf_a,
DMA_FROM_DEVICE);
if (ret) {
dev_err(uap->port.dev, "failed to init DMA %s: %d\n",
"RX buffer A", ret);
goto skip_rx;
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
ret = pl011_sgbuf_init(uap->dmarx.chan, &uap->dmarx.sgbuf_b,
DMA_FROM_DEVICE);
if (ret) {
dev_err(uap->port.dev, "failed to init DMA %s: %d\n",
"RX buffer B", ret);
pl011_sgbuf_free(uap->dmarx.chan, &uap->dmarx.sgbuf_a,
DMA_FROM_DEVICE);
goto skip_rx;
}
uap->using_rx_dma = true;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
skip_rx:
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/* Turn on DMA error (RX/TX will be enabled on demand) */
uap->dmacr |= UART011_DMAONERR;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: add DMA burst threshold support for ST variants ST Micro variants has some specific dma burst threshold compensation, which allows them to make better use of a DMA controller. Add support to set this up. Based on a patch from Linus Walleij. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:59:16 +08:00
/*
* ST Micro variants has some specific dma burst threshold
* compensation. Set this to 16 bytes, so burst will only
* be issued above/below 16 bytes.
*/
if (uap->vendor->dma_threshold)
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
pl011_write(ST_UART011_DMAWM_RX_16 | ST_UART011_DMAWM_TX_16,
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
uap, REG_ST_DMAWM);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (uap->using_rx_dma) {
if (pl011_dma_rx_trigger_dma(uap))
dev_dbg(uap->port.dev, "could not trigger initial "
"RX DMA job, fall back to interrupt mode\n");
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
if (uap->dmarx.poll_rate) {
init_timer(&(uap->dmarx.timer));
uap->dmarx.timer.function = pl011_dma_rx_poll;
uap->dmarx.timer.data = (unsigned long)uap;
mod_timer(&uap->dmarx.timer,
jiffies +
msecs_to_jiffies(uap->dmarx.poll_rate));
uap->dmarx.last_residue = PL011_DMA_BUFFER_SIZE;
uap->dmarx.last_jiffies = jiffies;
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
static void pl011_dma_shutdown(struct uart_amba_port *uap)
{
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (!(uap->using_tx_dma || uap->using_rx_dma))
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
return;
/* Disable RX and TX DMA */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
while (pl011_read(uap, REG_FR) & UART01x_FR_BUSY)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
barrier();
spin_lock_irq(&uap->port.lock);
uap->dmacr &= ~(UART011_DMAONERR | UART011_RXDMAE | UART011_TXDMAE);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->dmacr, uap, REG_DMACR);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
spin_unlock_irq(&uap->port.lock);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (uap->using_tx_dma) {
/* In theory, this should already be done by pl011_dma_flush_buffer */
dmaengine_terminate_all(uap->dmatx.chan);
if (uap->dmatx.queued) {
dma_unmap_sg(uap->dmatx.chan->device->dev, &uap->dmatx.sg, 1,
DMA_TO_DEVICE);
uap->dmatx.queued = false;
}
kfree(uap->dmatx.buf);
uap->using_tx_dma = false;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (uap->using_rx_dma) {
dmaengine_terminate_all(uap->dmarx.chan);
/* Clean up the RX DMA */
pl011_sgbuf_free(uap->dmarx.chan, &uap->dmarx.sgbuf_a, DMA_FROM_DEVICE);
pl011_sgbuf_free(uap->dmarx.chan, &uap->dmarx.sgbuf_b, DMA_FROM_DEVICE);
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
if (uap->dmarx.poll_rate)
del_timer_sync(&uap->dmarx.timer);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
uap->using_rx_dma = false;
}
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
static inline bool pl011_dma_rx_available(struct uart_amba_port *uap)
{
return uap->using_rx_dma;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
static inline bool pl011_dma_rx_running(struct uart_amba_port *uap)
{
return uap->using_rx_dma && uap->dmarx.running;
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
#else
/* Blank functions if the DMA engine is not available */
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
static inline void pl011_dma_probe(struct uart_amba_port *uap)
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
{
}
static inline void pl011_dma_remove(struct uart_amba_port *uap)
{
}
static inline void pl011_dma_startup(struct uart_amba_port *uap)
{
}
static inline void pl011_dma_shutdown(struct uart_amba_port *uap)
{
}
static inline bool pl011_dma_tx_irq(struct uart_amba_port *uap)
{
return false;
}
static inline void pl011_dma_tx_stop(struct uart_amba_port *uap)
{
}
static inline bool pl011_dma_tx_start(struct uart_amba_port *uap)
{
return false;
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
static inline void pl011_dma_rx_irq(struct uart_amba_port *uap)
{
}
static inline void pl011_dma_rx_stop(struct uart_amba_port *uap)
{
}
static inline int pl011_dma_rx_trigger_dma(struct uart_amba_port *uap)
{
return -EIO;
}
static inline bool pl011_dma_rx_available(struct uart_amba_port *uap)
{
return false;
}
static inline bool pl011_dma_rx_running(struct uart_amba_port *uap)
{
return false;
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
#define pl011_dma_flush_buffer NULL
#endif
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 17:12:14 +08:00
static void pl011_stop_tx(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->im &= ~UART011_TXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
pl011_dma_tx_stop(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
static void pl011_tx_chars(struct uart_amba_port *uap, bool from_irq);
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
/* Start TX with programmed I/O only (no DMA) */
static void pl011_start_tx_pio(struct uart_amba_port *uap)
{
uap->im |= UART011_TXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
pl011_tx_chars(uap, false);
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
}
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 17:12:14 +08:00
static void pl011_start_tx(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
if (!pl011_dma_tx_start(uap))
pl011_start_tx_pio(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static void pl011_stop_rx(struct uart_port *port)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->im &= ~(UART011_RXIM|UART011_RTIM|UART011_FEIM|
UART011_PEIM|UART011_BEIM|UART011_OEIM);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
pl011_dma_rx_stop(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static void pl011_enable_ms(struct uart_port *port)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->im |= UART011_RIMIM|UART011_CTSMIM|UART011_DCDMIM|UART011_DSRMIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 21:55:46 +08:00
static void pl011_rx_chars(struct uart_amba_port *uap)
serial: amba-pl011: Use __releases/__acquires annotations Fix the following sparse warnings: drivers/tty/serial/amba-pl011.c:687:20: warning: context imbalance in 'pl011_dma_flush_buffer' - unexpected unlock drivers/tty/serial/amba-pl011.c:1200:13: warning: context imbalance in 'pl011_rx_chars' - unexpected unlock Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-08-10 04:58:51 +08:00
__releases(&uap->port.lock)
__acquires(&uap->port.lock)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
ARM: 6764/1: pl011: factor out FIFO to TTY code This piece of code was just slightly different between the DMA and IRQ paths, in DMA mode we surely shouldn't read more than 256 character either, so factor this out in its own function and use for both DMA and PIO mode. Tested on Ux500 and U300. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:36 +08:00
pl011_fifo_to_tty(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[ARM] 4417/1: Serial: Fix AMBA drivers locking The -rt patch triggered a lockdep warning in the amba serial drivers, which never shows up on UP kernels. On SMP systems this would trigger as well. Release the port lock before calling tty_flip_buffer_push() and reacquire it after the call. This matches the code in the 8250 serial driver. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-05-30 04:53:50 +08:00
spin_unlock(&uap->port.lock);
TTY: switch tty_flip_buffer_push Now, we start converting tty buffer functions to actually use tty_port. This will allow us to get rid of the need of tty in many call sites. Only tty_port will needed and hence no more tty_port_tty_get in those paths. Now, the one where most of tty_port_tty_get gets removed: tty_flip_buffer_push. IOW we also closed all the races in drivers not using tty_port_tty_get at all yet. Also we move tty_flip_buffer_push declaration from include/linux/tty.h to include/linux/tty_flip.h to all others while we are changing it anyway. Signed-off-by: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-01-03 22:53:06 +08:00
tty_flip_buffer_push(&uap->port.state->port);
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
/*
* If we were temporarily out of DMA mode for a while,
* attempt to switch back to DMA mode again.
*/
if (pl011_dma_rx_available(uap)) {
if (pl011_dma_rx_trigger_dma(uap)) {
dev_dbg(uap->port.dev, "could not trigger RX DMA job "
"fall back to interrupt mode again\n");
uap->im |= UART011_RXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
} else {
amba-pl011: fix build error if CONFIG_DMA_ENGINE is not enabled Following patch will fix build error if CONFIG_DMA_ENGINE is unset. Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-03 10:10:37 +08:00
#ifdef CONFIG_DMA_ENGINE
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
/* Start Rx DMA poll */
if (uap->dmarx.poll_rate) {
uap->dmarx.last_jiffies = jiffies;
uap->dmarx.last_residue = PL011_DMA_BUFFER_SIZE;
mod_timer(&uap->dmarx.timer,
jiffies +
msecs_to_jiffies(uap->dmarx.poll_rate));
}
amba-pl011: fix build error if CONFIG_DMA_ENGINE is not enabled Following patch will fix build error if CONFIG_DMA_ENGINE is unset. Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-03 10:10:37 +08:00
#endif
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
}
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
}
[ARM] 4417/1: Serial: Fix AMBA drivers locking The -rt patch triggered a lockdep warning in the amba serial drivers, which never shows up on UP kernels. On SMP systems this would trigger as well. Release the port lock before calling tty_flip_buffer_push() and reacquire it after the call. This matches the code in the 8250 serial driver. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-05-30 04:53:50 +08:00
spin_lock(&uap->port.lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
static bool pl011_tx_char(struct uart_amba_port *uap, unsigned char c,
bool from_irq)
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
{
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
if (unlikely(!from_irq) &&
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_read(uap, REG_FR) & UART01x_FR_TXFF)
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
return false; /* unable to transmit character */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(c, uap, REG_DR);
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
uap->port.icount.tx++;
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
return true;
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
}
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
static void pl011_tx_chars(struct uart_amba_port *uap, bool from_irq)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: kill off uart_info We moved this into uart_state, now move the fields out of the separate structure and kill it off. Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-09-20 04:13:28 +08:00
struct circ_buf *xmit = &uap->port.state->xmit;
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
int count = uap->fifosize >> 1;
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (uap->port.x_char) {
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
if (!pl011_tx_char(uap, uap->port.x_char, from_irq))
return;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->port.x_char = 0;
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
--count;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
if (uart_circ_empty(xmit) || uart_tx_stopped(&uap->port)) {
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 17:12:14 +08:00
pl011_stop_tx(&uap->port);
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
return;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/* If we are using DMA mode, try to send some characters. */
if (pl011_dma_tx_irq(uap))
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
return;
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
do {
if (likely(from_irq) && count-- == 0)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
break;
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
if (!pl011_tx_char(uap, xmit->buf[xmit->tail], from_irq))
break;
xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
} while (!uart_circ_empty(xmit));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(&uap->port);
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
if (uart_circ_empty(xmit))
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 17:12:14 +08:00
pl011_stop_tx(&uap->port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static void pl011_modem_status(struct uart_amba_port *uap)
{
unsigned int status, delta;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
status = pl011_read(uap, REG_FR) & UART01x_FR_MODEM_ANY;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
delta = status ^ uap->old_status;
uap->old_status = status;
if (!delta)
return;
if (delta & UART01x_FR_DCD)
uart_handle_dcd_change(&uap->port, status & UART01x_FR_DCD);
Revert "uart: pl011: Add support to ZTE ZX296702 uart" This reverts commit 8cd90e50d1408c65c355084b1c7f8f9085f49c6b as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:11:24 +08:00
if (delta & UART01x_FR_DSR)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->port.icount.dsr++;
Revert "uart: pl011: Add support to ZTE ZX296702 uart" This reverts commit 8cd90e50d1408c65c355084b1c7f8f9085f49c6b as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:11:24 +08:00
if (delta & UART01x_FR_CTS)
uart_handle_cts_change(&uap->port, status & UART01x_FR_CTS);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial: move delta_msr_wait into the tty_port This is used by various drivers not just serial and can be extracted as commonality Signed-off-by: Alan Cox <alan@linux.intel.com>
2009-09-20 04:13:31 +08:00
wake_up_interruptible(&uap->port.state->port.delta_msr_wait);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
drivers: PL011: move cts_event workaround into separate function To avoid lines with more than 80 characters and to make the pl011_int() function more readable, move the workaround out into a separate function. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:20 +08:00
static void check_apply_cts_event_workaround(struct uart_amba_port *uap)
{
unsigned int dummy_read;
if (!uap->vendor->cts_event_workaround)
return;
/* workaround to make sure that all bits are unlocked.. */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(0x00, uap, REG_ICR);
drivers: PL011: move cts_event workaround into separate function To avoid lines with more than 80 characters and to make the pl011_int() function more readable, move the workaround out into a separate function. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:20 +08:00
/*
* WA: introduce 26ns(1 uart clk) delay before W1C;
* single apb access will incur 2 pclk(133.12Mhz) delay,
* so add 2 dummy reads
*/
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
dummy_read = pl011_read(uap, REG_ICR);
dummy_read = pl011_read(uap, REG_ICR);
drivers: PL011: move cts_event workaround into separate function To avoid lines with more than 80 characters and to make the pl011_int() function more readable, move the workaround out into a separate function. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:20 +08:00
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 21:55:46 +08:00
static irqreturn_t pl011_int(int irq, void *dev_id)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct uart_amba_port *uap = dev_id;
ARM: PL011: Ensure IRQs are disabled in UART interrupt handler As the DMA support introduces a separate interrupt-time callback, our interrupt handler will not be the only handler which takes the port lock, so we need to ensure that IRQs are disabled. We must use the _irqsave variant so we don't inadvertently enable interrupts. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:16:09 +08:00
unsigned long flags;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned int status, pass_counter = AMBA_ISR_PASS_LIMIT;
drivers: PL011: replace UART_MIS reading with _RIS & _IMSC The PL011 register UART_MIS is actually a bitwise AND of the UART_RIS and the UART_MISC register. Since the SBSA UART does not include the _MIS register, use the two separate registers to get the same behaviour. Since we are inside the spinlock and we read the _IMSC register only once, there should be no race issue. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:19 +08:00
u16 imsc;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
int handled = 0;
ARM: PL011: Ensure IRQs are disabled in UART interrupt handler As the DMA support introduces a separate interrupt-time callback, our interrupt handler will not be the only handler which takes the port lock, so we need to ensure that IRQs are disabled. We must use the _irqsave variant so we don't inadvertently enable interrupts. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:16:09 +08:00
spin_lock_irqsave(&uap->port.lock, flags);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
imsc = pl011_read(uap, REG_IMSC);
status = pl011_read(uap, REG_RIS) & imsc;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (status) {
do {
drivers: PL011: move cts_event workaround into separate function To avoid lines with more than 80 characters and to make the pl011_int() function more readable, move the workaround out into a separate function. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:20 +08:00
check_apply_cts_event_workaround(uap);
Revert "uart: pl011: Introduce register accessor" This reverts commit 7b753f318d1456c8e7740f3bd96d1dbb362d5449 as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:13:39 +08:00
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
pl011_write(status & ~(UART011_TXIS|UART011_RTIS|
UART011_RXIS),
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
uap, REG_ICR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ARM: 6763/1: pl011: add optional RX DMA to PL011 v2 This adds an optional RX DMA codepath for the devices that support this by using the apropriate burst sizes instead of pulling single bytes. Includes portions of code written by Russell King during a PL08x hacking session. This has been tested on U300 and Ux500. Tested-by: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Tested-by: Grzegorz Sygieda <grzegorz.sygieda@tieto.com> Tested-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Per Forlin <per.friden@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-24 20:21:08 +08:00
if (status & (UART011_RTIS|UART011_RXIS)) {
if (pl011_dma_rx_running(uap))
pl011_dma_rx_irq(uap);
else
pl011_rx_chars(uap);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (status & (UART011_DSRMIS|UART011_DCDMIS|
UART011_CTSMIS|UART011_RIMIS))
pl011_modem_status(uap);
serial/amba-pl011: Refactor and simplify TX FIFO handling Commit 734745c serial/amba-pl011: Activate TX IRQ passively adds some complexity and overhead in the form of a softirq mechanism for transmitting in the absence of interrupts. This patch simplifies the code flow to reduce the reliance on subtle behaviour and avoid fragility under future maintenance. To this end, the TX softirq mechanism is removed and instead pl011_start_tx() will now simply stuff the FIFO until full (guaranteeing future TX IRQs), or until there are no more chars to write (in which case we don't care whether an IRQ happens). Signed-off-by: Dave Martin <Dave.Martin@arm.com> Signed-off-by: Jakub Kicinski <kubakici@wp.pl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-27 23:49:05 +08:00
if (status & UART011_TXIS)
pl011_tx_chars(uap, true);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial: pl011: implement workaround for CTS clear event issue Problem Observed: - interrupt status is set by rising or falling edge on CTS line - interrupt status is cleared on a .0. to .1. transition of the interrupt-clear register bit 1. - interrupt-clear register is reset by hardware once the interrupt status is .0.. Remark: It seems not possible to read this register back by the CPU though, but internally this register exists. - when simultaneous set and reset event on the interrupt status happens, then the set-event has priority and the status remains .1.. As a result the interrupt-clear register is not reset to .0., and no new .0. to .1. transition can be detected on it when writing a .1. to it. This implies race condition, the clear must be performed at least one UARTCLK the riding edge of CTS RIS interrupt. Fix: Instead of resetting UART as done in commit c16d51a32bbb61ac8fd96f78b5ce2fccfe0fb4c3 "amba pl011: workaround for uart registers lockup" do the following: write .0. and then .1. to the interrupt-clear register to make sure that this transition is detected. According to the datasheet writing a .0. does not have any effect, but actually it allows to reset the internal interrupt-clear register. Take into account: The .0. needs to last at least for one clk_uart clock period (~ 38 MHz, 26.08ns) This way we can do away with the tasklet and keep only a tiny fix triggered by the variant flag introduced in this patch. Signed-off-by: Guillaume Jaunet <guillaume.jaunet@stericsson.com> Signed-off-by: Christophe Arnal <christophe.arnal@stericsson.com> Signed-off-by: Matthias Locher <Matthias.Locher@stericsson.com> Signed-off-by: Rajanikanth H.V <rajanikanth.hv@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-26 17:17:02 +08:00
if (pass_counter-- == 0)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
break;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
status = pl011_read(uap, REG_RIS) & imsc;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
} while (status != 0);
handled = 1;
}
ARM: PL011: Ensure IRQs are disabled in UART interrupt handler As the DMA support introduces a separate interrupt-time callback, our interrupt handler will not be the only handler which takes the port lock, so we need to ensure that IRQs are disabled. We must use the _irqsave variant so we don't inadvertently enable interrupts. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:16:09 +08:00
spin_unlock_irqrestore(&uap->port.lock, flags);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return IRQ_RETVAL(handled);
}
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static unsigned int pl011_tx_empty(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
unsigned int status = pl011_read(uap, REG_FR);
Revert "uart: pl011: Add support to ZTE ZX296702 uart" This reverts commit 8cd90e50d1408c65c355084b1c7f8f9085f49c6b as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:11:24 +08:00
return status & (UART01x_FR_BUSY|UART01x_FR_TXFF) ? 0 : TIOCSER_TEMT;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static unsigned int pl011_get_mctrl(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned int result = 0;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
unsigned int status = pl011_read(uap, REG_FR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
amba-pl011, rename BIT macro amba-pl011, rename BIT macro Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:31 +08:00
#define TIOCMBIT(uartbit, tiocmbit) \
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (status & uartbit) \
result |= tiocmbit
amba-pl011, rename BIT macro amba-pl011, rename BIT macro Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:31 +08:00
TIOCMBIT(UART01x_FR_DCD, TIOCM_CAR);
Revert "uart: pl011: Add support to ZTE ZX296702 uart" This reverts commit 8cd90e50d1408c65c355084b1c7f8f9085f49c6b as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:11:24 +08:00
TIOCMBIT(UART01x_FR_DSR, TIOCM_DSR);
TIOCMBIT(UART01x_FR_CTS, TIOCM_CTS);
TIOCMBIT(UART011_FR_RI, TIOCM_RNG);
amba-pl011, rename BIT macro amba-pl011, rename BIT macro Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:31 +08:00
#undef TIOCMBIT
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return result;
}
static void pl011_set_mctrl(struct uart_port *port, unsigned int mctrl)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned int cr;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
cr = pl011_read(uap, REG_CR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
amba-pl011, rename BIT macro amba-pl011, rename BIT macro Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:31 +08:00
#define TIOCMBIT(tiocmbit, uartbit) \
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (mctrl & tiocmbit) \
cr |= uartbit; \
else \
cr &= ~uartbit
amba-pl011, rename BIT macro amba-pl011, rename BIT macro Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:31 +08:00
TIOCMBIT(TIOCM_RTS, UART011_CR_RTS);
TIOCMBIT(TIOCM_DTR, UART011_CR_DTR);
TIOCMBIT(TIOCM_OUT1, UART011_CR_OUT1);
TIOCMBIT(TIOCM_OUT2, UART011_CR_OUT2);
TIOCMBIT(TIOCM_LOOP, UART011_CR_LBE);
ARM: 5933/1: amba-pl011: support hardware flow control Enable/disable automatic hardware flow control as requested by the termios. The controller does not allow us to control the RTS line when auto-RTS is enabled, so we enable auto-RTS only if the kernel has not disabled RTS. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-02-12 13:43:11 +08:00
if (uap->autorts) {
/* We need to disable auto-RTS if we want to turn RTS off */
TIOCMBIT(TIOCM_RTS, UART011_CR_RTSEN);
}
amba-pl011, rename BIT macro amba-pl011, rename BIT macro Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-19 14:40:31 +08:00
#undef TIOCMBIT
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(cr, uap, REG_CR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static void pl011_break_ctl(struct uart_port *port, int break_state)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned long flags;
unsigned int lcr_h;
spin_lock_irqsave(&uap->port.lock, flags);
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
lcr_h = pl011_read(uap, REG_LCRH_TX);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (break_state == -1)
lcr_h |= UART01x_LCRH_BRK;
else
lcr_h &= ~UART01x_LCRH_BRK;
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
pl011_write(lcr_h, uap, REG_LCRH_TX);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
spin_unlock_irqrestore(&uap->port.lock, flags);
}
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
#ifdef CONFIG_CONSOLE_POLL
tty/serial/amba-pl011: Quiesce interrupts in poll_get_char We need to quiesce interrupts in the poll_get_char routine, otherwise, if used with KGDB NMI debugger, we'll keep reentering the NMI. Quiescing interrupts is pretty straightforward, except for TXIM interrupt. The interrupt has "ready to transmit" meaning, so it's almost always raised, and the only way to silence it is to mask it. But that's OK, ops->start_tx will unmask it. Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:55 +08:00
static void pl011_quiesce_irqs(struct uart_port *port)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
tty/serial/amba-pl011: Quiesce interrupts in poll_get_char We need to quiesce interrupts in the poll_get_char routine, otherwise, if used with KGDB NMI debugger, we'll keep reentering the NMI. Quiescing interrupts is pretty straightforward, except for TXIM interrupt. The interrupt has "ready to transmit" meaning, so it's almost always raised, and the only way to silence it is to mask it. But that's OK, ops->start_tx will unmask it. Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:55 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(pl011_read(uap, REG_MIS), uap, REG_ICR);
tty/serial/amba-pl011: Quiesce interrupts in poll_get_char We need to quiesce interrupts in the poll_get_char routine, otherwise, if used with KGDB NMI debugger, we'll keep reentering the NMI. Quiescing interrupts is pretty straightforward, except for TXIM interrupt. The interrupt has "ready to transmit" meaning, so it's almost always raised, and the only way to silence it is to mask it. But that's OK, ops->start_tx will unmask it. Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:55 +08:00
/*
* There is no way to clear TXIM as this is "ready to transmit IRQ", so
* we simply mask it. start_tx() will unmask it.
*
* Note we can race with start_tx(), and if the race happens, the
* polling user might get another interrupt just after we clear it.
* But it should be OK and can happen even w/o the race, e.g.
* controller immediately got some new data and raised the IRQ.
*
* And whoever uses polling routines assumes that it manages the device
* (including tx queue), so we're also fine with start_tx()'s caller
* side.
*/
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(pl011_read(uap, REG_IMSC) & ~UART011_TXIM, uap,
REG_IMSC);
tty/serial/amba-pl011: Quiesce interrupts in poll_get_char We need to quiesce interrupts in the poll_get_char routine, otherwise, if used with KGDB NMI debugger, we'll keep reentering the NMI. Quiescing interrupts is pretty straightforward, except for TXIM interrupt. The interrupt has "ready to transmit" meaning, so it's almost always raised, and the only way to silence it is to mask it. But that's OK, ops->start_tx will unmask it. Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:55 +08:00
}
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static int pl011_get_poll_char(struct uart_port *port)
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
unsigned int status;
tty/serial/amba-pl011: Quiesce interrupts in poll_get_char We need to quiesce interrupts in the poll_get_char routine, otherwise, if used with KGDB NMI debugger, we'll keep reentering the NMI. Quiescing interrupts is pretty straightforward, except for TXIM interrupt. The interrupt has "ready to transmit" meaning, so it's almost always raised, and the only way to silence it is to mask it. But that's OK, ops->start_tx will unmask it. Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:55 +08:00
/*
* The caller might need IRQs lowered, e.g. if used with KDB NMI
* debugger.
*/
pl011_quiesce_irqs(port);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
status = pl011_read(uap, REG_FR);
kgdb,8250,pl011: Return immediately from console poll The design of the kdb shell requires that every device that can provide input to kdb have a polling routine that exits immediately if there is no character available. This is required in order to get the page scrolling mechanism working. Changing the kernel debugger I/O API to require all polling character routines to exit immediately if there is no data allows the kernel debugger to process multiple input channels. NO_POLL_CHAR will be the return code to the polling routine when ever there is no character available. CC: linux-serial@vger.kernel.org Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
2010-05-21 10:04:22 +08:00
if (status & UART01x_FR_RXFE)
return NO_POLL_CHAR;
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
return pl011_read(uap, REG_DR);
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
}
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static void pl011_put_poll_char(struct uart_port *port,
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
unsigned char ch)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
while (pl011_read(uap, REG_FR) & UART01x_FR_TXFF)
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
barrier();
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(ch, uap, REG_DR);
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
}
#endif /* CONFIG_CONSOLE_POLL */
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
static int pl011_hwinit(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
int retval;
serial/amba-pl011: move custom pin control to driver We had a boot regression in Ux500 in the merge window because two orthogonal pin control schemes for the PL011 were merged at the same time: - One using the .init() and .exit() hooks into the platform for Ux500 putting the pins into default vs sleep state respectively as the port was started/stopped. commit a09806607fd20bed2f8c41fe22793386790a14aa "ARM: ux500: switch to using pinctrl for uart0" - One hogging the default setting at PL011 probe() commit 258e055111d3cde2607e0d04eb91da2f7a59b591 "serial: amba-pl011: adopt pinctrl support" To get a solution that works for both let's scrap the stuff in the platform callbacks, instead have the driver itself select default and sleep states when the port is started/stopped. Hopefully this works for all clients. Platform callbacks are bad for device tree migration anyway, so this rids us of another problem in Ux500. Cc: Shawn Guo <shawn.guo@linaro.org> Cc: Russell King <rmk+kernel@arm.linux.org.uk> Reported-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-24 03:18:46 +08:00
/* Optionaly enable pins to be muxed in and configured */
tty: serial: modify PL011 driver to use pinctrl PM helpers This augments the PL011 UART driver to utilize the new pinctrl core PM helpers to transition the driver to default and sleep states, cutting away some boilerplate code. Cc: Hebbar Gururaja <gururaja.hebbar@ti.com> Cc: Mark Brown <broonie@kernel.org> Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com> Cc: Kevin Hilman <khilman@linaro.org> Cc: Stephen Warren <swarren@wwwdotorg.org> Acked-by: Wolfram Sang <wsa@the-dreams.de> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-06-05 21:36:42 +08:00
pinctrl_pm_select_default_state(port->dev);
serial/amba-pl011: move custom pin control to driver We had a boot regression in Ux500 in the merge window because two orthogonal pin control schemes for the PL011 were merged at the same time: - One using the .init() and .exit() hooks into the platform for Ux500 putting the pins into default vs sleep state respectively as the port was started/stopped. commit a09806607fd20bed2f8c41fe22793386790a14aa "ARM: ux500: switch to using pinctrl for uart0" - One hogging the default setting at PL011 probe() commit 258e055111d3cde2607e0d04eb91da2f7a59b591 "serial: amba-pl011: adopt pinctrl support" To get a solution that works for both let's scrap the stuff in the platform callbacks, instead have the driver itself select default and sleep states when the port is started/stopped. Hopefully this works for all clients. Platform callbacks are bad for device tree migration anyway, so this rids us of another problem in Ux500. Cc: Shawn Guo <shawn.guo@linaro.org> Cc: Russell King <rmk+kernel@arm.linux.org.uk> Reported-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-24 03:18:46 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Try to enable the clock producer.
*/
drivers/tty/serial/amba-pl0{10,11}.c: use clk_prepare_enable and clk_disable_unprepare Clk_prepare_enable and clk_disable_unprepare combine clk_prepare and clk_enable, and clk_disable and clk_unprepare. The9 make the code more concise, and ensure that clk_unprepare is called when clk_enable fails. A simplified version of the semantic patch that introduces calls to these functions is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression e; @@ - clk_prepare(e); - clk_enable(e); + clk_prepare_enable(e); @@ expression e; @@ - clk_disable(e); - clk_unprepare(e); + clk_disable_unprepare(e); // </smpl> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-08-27 00:01:01 +08:00
retval = clk_prepare_enable(uap->clk);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (retval)
serial: amba-pl011: Remove redundant label The label 'out' is only used to return the error code. We can return the error code directly and remove 'out' label. Signed-off-by: Tushar Behera <tushar.b@samsung.com> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-06-26 18:05:35 +08:00
return retval;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->port.uartclk = clk_get_rate(uap->clk);
serial: PL011: clear pending interrupts Chanho Min reported that when the boot loader transfers control to the kernel, there may be pending interrupts causing the UART to lock up in an eternal loop trying to pick tokens from the FIFO (since the RX interrupt flag indicates there are tokens) while in practice there are no tokens - in fact there is only a pending IRQ flag. This patch address the issue with a combination of two patches suggested by Russell King that clears and mask all interrupts at probe() and clears any pending error and RX interrupts at port startup time. We suspect the spurious interrupts are a side-effect of switching the UART from FIFO to non-FIFO mode. Cc: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Reported-by: Chanho Min <chanho0207@gmail.com> Suggested-by: Russell King <linux@arm.linux.org.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Jong-Sung Kim <neidhard.kim@lge.com> Cc: stable <stable@vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-13 20:27:23 +08:00
/* Clear pending error and receive interrupts */
tty: amba-pl011: add register accessor functions Add register accessor functions to amba-pl011. Much of this transformation was done using the sed expression below, with any left-overs fixed up manually afterwards, and code formatted to remain within coding style. s/readw(\(uap->port.membase\|regs\|port->membase\) +/pl011_read(\1,/g s/writew(\(.*\) +/pl011_write(\1,/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:50:58 +08:00
pl011_write(UART011_OEIS | UART011_BEIS | UART011_PEIS |
UART011_FEIS | UART011_RTIS | UART011_RXIS,
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
uap, REG_ICR);
serial: PL011: clear pending interrupts Chanho Min reported that when the boot loader transfers control to the kernel, there may be pending interrupts causing the UART to lock up in an eternal loop trying to pick tokens from the FIFO (since the RX interrupt flag indicates there are tokens) while in practice there are no tokens - in fact there is only a pending IRQ flag. This patch address the issue with a combination of two patches suggested by Russell King that clears and mask all interrupts at probe() and clears any pending error and RX interrupts at port startup time. We suspect the spurious interrupts are a side-effect of switching the UART from FIFO to non-FIFO mode. Cc: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Reported-by: Chanho Min <chanho0207@gmail.com> Suggested-by: Russell King <linux@arm.linux.org.uk> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Jong-Sung Kim <neidhard.kim@lge.com> Cc: stable <stable@vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-13 20:27:23 +08:00
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
/*
* Save interrupts enable mask, and enable RX interrupts in case if
* the interrupt is used for NMI entry.
*/
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
uap->im = pl011_read(uap, REG_IMSC);
pl011_write(UART011_RTIM | UART011_RXIM, uap, REG_IMSC);
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
if (dev_get_platdata(uap->port.dev)) {
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
struct amba_pl011_data *plat;
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
plat = dev_get_platdata(uap->port.dev);
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
if (plat->init)
plat->init();
}
return 0;
}
tty: amba-pl011: add helper to detect split LCRH register Add a helper to detect the split LCRH register found on ST variants. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:08 +08:00
static bool pl011_split_lcrh(const struct uart_amba_port *uap)
{
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
return pl011_reg_to_offset(uap, REG_LCRH_RX) !=
pl011_reg_to_offset(uap, REG_LCRH_TX);
tty: amba-pl011: add helper to detect split LCRH register Add a helper to detect the split LCRH register found on ST variants. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:08 +08:00
}
serial: amba-pl011: factor out code for writing LCR_H register The code to cope with a split tx/rx LCR_H register is non-trivial so put it into it's own function to avoid duplication. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:59 +08:00
static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
{
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
pl011_write(lcr_h, uap, REG_LCRH_RX);
tty: amba-pl011: add helper to detect split LCRH register Add a helper to detect the split LCRH register found on ST variants. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:08 +08:00
if (pl011_split_lcrh(uap)) {
serial: amba-pl011: factor out code for writing LCR_H register The code to cope with a split tx/rx LCR_H register is non-trivial so put it into it's own function to avoid duplication. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:59 +08:00
int i;
/*
* Wait 10 PCLKs before writing LCRH_TX register,
* to get this delay write read only register 10 times
*/
for (i = 0; i < 10; ++i)
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(0xff, uap, REG_MIS);
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
pl011_write(lcr_h, uap, REG_LCRH_TX);
serial: amba-pl011: factor out code for writing LCR_H register The code to cope with a split tx/rx LCR_H register is non-trivial so put it into it's own function to avoid duplication. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:59 +08:00
}
}
drivers: PL011: refactor pl011_startup() Split the pl011_startup() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:15 +08:00
static int pl011_allocate_irq(struct uart_amba_port *uap)
{
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
drivers: PL011: refactor pl011_startup() Split the pl011_startup() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:15 +08:00
return request_irq(uap->port.irq, pl011_int, 0, "uart-pl011", uap);
}
/*
* Enable interrupts, only timeouts when using DMA
* if initial RX DMA job failed, start in interrupt mode
* as well.
*/
static void pl011_enable_interrupts(struct uart_amba_port *uap)
{
spin_lock_irq(&uap->port.lock);
/* Clear out any spuriously appearing RX interrupts */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(UART011_RTIS | UART011_RXIS, uap, REG_ICR);
drivers: PL011: refactor pl011_startup() Split the pl011_startup() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:15 +08:00
uap->im = UART011_RTIM;
if (!pl011_dma_rx_running(uap))
uap->im |= UART011_RXIM;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
drivers: PL011: refactor pl011_startup() Split the pl011_startup() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:15 +08:00
spin_unlock_irq(&uap->port.lock);
}
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
static int pl011_startup(struct uart_port *port)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
unsigned int cr;
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
int retval;
retval = pl011_hwinit(port);
if (retval)
goto clk_dis;
drivers: PL011: refactor pl011_startup() Split the pl011_startup() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:15 +08:00
retval = pl011_allocate_irq(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (retval)
goto clk_dis;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->vendor->ifls, uap, REG_IFLS);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial/amba-pl011: Activate TX IRQ passively The current PL011 driver transmits a dummy character when the UART is opened, to assert the TX IRQ for the first time (see pl011_startup()). The UART is put in loopback mode temporarily, so the receiver presumably shouldn't see anything. However... At least some platforms containing a PL011 send characters down the wire even when loopback mode is enabled. This means that a spurious NUL character may be seen at the receiver when the PL011 is opened through the TTY layer. The current code also temporarily sets the baud rate to maximum and the character width to the minimum, to that the dummy TX completes as quickly as possible. If this is seen by the receiver it will result in a framing error and can knock the receiver out of sync -- turning subsequent output into garbage until synchronisation is reestablished. (Particularly problematic during boot with systemd.) To avoid spurious transmissions, this patch removes assumptions about whether the TX IRQ will fire until at least one TX IRQ has been seen. Instead, the UART will unmask the TX IRQ and then slow-start via polling and timer-based soft IRQs initially. If the TTY layer writes enough data to fill the FIFO to the interrupt threshold in one go, the TX IRQ should assert, at which point the driver changes to fully interrupt-driven TX. In this way, the TX IRQ is activated as a side-effect instead of being done deliberately. This should also mean that the driver works on the SBSA Generic UART[1] (a cut-down PL011) without invasive changes. The Generic UART lacks some features needed for the dummy TX approach to work (FIFO disabling and loopback). [1] Server Base System Architecture (ARM-DEN-0029-v2.3) http://infocenter.arm.com/ (click-thru required :/) Signed-off-by: Dave Martin <Dave.Martin@arm.com> Tested-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-04 20:27:33 +08:00
spin_lock_irq(&uap->port.lock);
serial: amba-pl011: preseserve hardware settings during initialisation During initialisation, a UART may already be in use for a console, so take care to preserve things like baud rate and data format to avoid corrupting console output. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:19:00 +08:00
amba-pl011: do not disable RTS during shutdown In present driver, shutdown clears RTS and DTR in CR register. But the documentation "Documentation/serial/driver" suggests not to disable RTS and DTR in shutdown(). Also RTS and DTR is preserved between shutdown and startup calls, i.e. these are restored in startup if they were enabled while doing shutdown. So that if RTS and DTR are set using pl011_set_mctrl then it should continue even after shutdown->startup sequence. For throttling/unthrottling user should call pl011_set_mctrl. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-18 18:23:59 +08:00
/* restore RTS and DTR */
cr = uap->old_cr & (UART011_CR_RTS | UART011_CR_DTR);
cr |= UART01x_CR_UARTEN | UART011_CR_RXE | UART011_CR_TXE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(cr, uap, REG_CR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial: amba-pl011: use port lock to guard control register access When the pl011 is being used for a console, pl011_console_write forces the control register (CR) to enable the UART for transmission and then restores this to the original value afterwards. It does this while holding the port lock. Unfortunately, when the uart is started or shutdown - say in response to userland using the serial device for a terminal - then this updates the control register without any locking. This means we can have pl011_console_write Save CR pl011_startup Initialise CR, e.g. enable receive pl011_console_write Restore old CR with receive not enabled this result is a serial port which doesn't respond to any input. A similar race in reverse could happen when the device is shutdown. We can fix these problems by taking the port lock when updating CR. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:58 +08:00
spin_unlock_irq(&uap->port.lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* initialise the old status of the modem signals
*/
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
uap->old_status = pl011_read(uap, REG_FR) & UART01x_FR_MODEM_ANY;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
/* Startup DMA */
pl011_dma_startup(uap);
drivers: PL011: refactor pl011_startup() Split the pl011_startup() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:15 +08:00
pl011_enable_interrupts(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return 0;
clk_dis:
drivers/tty/serial/amba-pl0{10,11}.c: use clk_prepare_enable and clk_disable_unprepare Clk_prepare_enable and clk_disable_unprepare combine clk_prepare and clk_enable, and clk_disable and clk_unprepare. The9 make the code more concise, and ensure that clk_unprepare is called when clk_enable fails. A simplified version of the semantic patch that introduces calls to these functions is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression e; @@ - clk_prepare(e); - clk_enable(e); + clk_prepare_enable(e); @@ expression e; @@ - clk_disable(e); - clk_unprepare(e); + clk_disable_unprepare(e); // </smpl> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-08-27 00:01:01 +08:00
clk_disable_unprepare(uap->clk);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return retval;
}
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static int sbsa_uart_startup(struct uart_port *port)
{
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
int retval;
retval = pl011_hwinit(port);
if (retval)
return retval;
retval = pl011_allocate_irq(uap);
if (retval)
return retval;
/* The SBSA UART does not support any modem status lines. */
uap->old_status = 0;
pl011_enable_interrupts(uap);
return 0;
}
ARM: 6157/2: PL011 TX/RX split of LCR for ST-Ericssons derivative In the ST-Ericsson version of the PL011 the TX and RX have different control registers. Cc: Alessandro Rubini <rubini@unipv.it> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-02 15:13:52 +08:00
static void pl011_shutdown_channel(struct uart_amba_port *uap,
unsigned int lcrh)
{
Revert "uart: pl011: Introduce register accessor" This reverts commit 7b753f318d1456c8e7740f3bd96d1dbb362d5449 as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:13:39 +08:00
unsigned long val;
ARM: 6157/2: PL011 TX/RX split of LCR for ST-Ericssons derivative In the ST-Ericsson version of the PL011 the TX and RX have different control registers. Cc: Alessandro Rubini <rubini@unipv.it> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-02 15:13:52 +08:00
tty: amba-pl011: convert accessor functions to take uart_amba_port Convert the new accessor functions to take the uart_amba_port instead of the port base address. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:03 +08:00
val = pl011_read(uap, lcrh);
Revert "uart: pl011: Introduce register accessor" This reverts commit 7b753f318d1456c8e7740f3bd96d1dbb362d5449 as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:13:39 +08:00
val &= ~(UART01x_LCRH_BRK | UART01x_LCRH_FEN);
tty: amba-pl011: convert accessor functions to take uart_amba_port Convert the new accessor functions to take the uart_amba_port instead of the port base address. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:03 +08:00
pl011_write(val, uap, lcrh);
ARM: 6157/2: PL011 TX/RX split of LCR for ST-Ericssons derivative In the ST-Ericsson version of the PL011 the TX and RX have different control registers. Cc: Alessandro Rubini <rubini@unipv.it> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-02 15:13:52 +08:00
}
drivers: PL011: refactor pl011_shutdown() Split the pl011_shutdown() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:16 +08:00
/*
* disable the port. It should not disable RTS and DTR.
* Also RTS and DTR state should be preserved to restore
* it during startup().
*/
static void pl011_disable_uart(struct uart_amba_port *uap)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
amba-pl011: do not disable RTS during shutdown In present driver, shutdown clears RTS and DTR in CR register. But the documentation "Documentation/serial/driver" suggests not to disable RTS and DTR in shutdown(). Also RTS and DTR is preserved between shutdown and startup calls, i.e. these are restored in startup if they were enabled while doing shutdown. So that if RTS and DTR are set using pl011_set_mctrl then it should continue even after shutdown->startup sequence. For throttling/unthrottling user should call pl011_set_mctrl. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-18 18:23:59 +08:00
unsigned int cr;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ARM: 5933/1: amba-pl011: support hardware flow control Enable/disable automatic hardware flow control as requested by the termios. The controller does not allow us to control the RTS line when auto-RTS is enabled, so we enable auto-RTS only if the kernel has not disabled RTS. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-02-12 13:43:11 +08:00
uap->autorts = false;
serial: amba-pl011: use port lock to guard control register access When the pl011 is being used for a console, pl011_console_write forces the control register (CR) to enable the UART for transmission and then restores this to the original value afterwards. It does this while holding the port lock. Unfortunately, when the uart is started or shutdown - say in response to userland using the serial device for a terminal - then this updates the control register without any locking. This means we can have pl011_console_write Save CR pl011_startup Initialise CR, e.g. enable receive pl011_console_write Restore old CR with receive not enabled this result is a serial port which doesn't respond to any input. A similar race in reverse could happen when the device is shutdown. We can fix these problems by taking the port lock when updating CR. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:58 +08:00
spin_lock_irq(&uap->port.lock);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
cr = pl011_read(uap, REG_CR);
amba-pl011: do not disable RTS during shutdown In present driver, shutdown clears RTS and DTR in CR register. But the documentation "Documentation/serial/driver" suggests not to disable RTS and DTR in shutdown(). Also RTS and DTR is preserved between shutdown and startup calls, i.e. these are restored in startup if they were enabled while doing shutdown. So that if RTS and DTR are set using pl011_set_mctrl then it should continue even after shutdown->startup sequence. For throttling/unthrottling user should call pl011_set_mctrl. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-18 18:23:59 +08:00
uap->old_cr = cr;
cr &= UART011_CR_RTS | UART011_CR_DTR;
cr |= UART01x_CR_UARTEN | UART011_CR_TXE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(cr, uap, REG_CR);
serial: amba-pl011: use port lock to guard control register access When the pl011 is being used for a console, pl011_console_write forces the control register (CR) to enable the UART for transmission and then restores this to the original value afterwards. It does this while holding the port lock. Unfortunately, when the uart is started or shutdown - say in response to userland using the serial device for a terminal - then this updates the control register without any locking. This means we can have pl011_console_write Save CR pl011_startup Initialise CR, e.g. enable receive pl011_console_write Restore old CR with receive not enabled this result is a serial port which doesn't respond to any input. A similar race in reverse could happen when the device is shutdown. We can fix these problems by taking the port lock when updating CR. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:58 +08:00
spin_unlock_irq(&uap->port.lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* disable break condition and fifos
*/
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
pl011_shutdown_channel(uap, REG_LCRH_RX);
tty: amba-pl011: add helper to detect split LCRH register Add a helper to detect the split LCRH register found on ST variants. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:08 +08:00
if (pl011_split_lcrh(uap))
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
pl011_shutdown_channel(uap, REG_LCRH_TX);
drivers: PL011: refactor pl011_shutdown() Split the pl011_shutdown() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:16 +08:00
}
static void pl011_disable_interrupts(struct uart_amba_port *uap)
{
spin_lock_irq(&uap->port.lock);
/* mask all interrupts and clear all pending ones */
uap->im = 0;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(uap->im, uap, REG_IMSC);
pl011_write(0xffff, uap, REG_ICR);
drivers: PL011: refactor pl011_shutdown() Split the pl011_shutdown() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:16 +08:00
spin_unlock_irq(&uap->port.lock);
}
static void pl011_shutdown(struct uart_port *port)
{
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
pl011_disable_interrupts(uap);
pl011_dma_shutdown(uap);
free_irq(uap->port.irq, uap);
pl011_disable_uart(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Shut down the clock producer
*/
drivers/tty/serial/amba-pl0{10,11}.c: use clk_prepare_enable and clk_disable_unprepare Clk_prepare_enable and clk_disable_unprepare combine clk_prepare and clk_enable, and clk_disable and clk_unprepare. The9 make the code more concise, and ensure that clk_unprepare is called when clk_enable fails. A simplified version of the semantic patch that introduces calls to these functions is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression e; @@ - clk_prepare(e); - clk_enable(e); + clk_prepare_enable(e); @@ expression e; @@ - clk_disable(e); - clk_unprepare(e); + clk_disable_unprepare(e); // </smpl> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-08-27 00:01:01 +08:00
clk_disable_unprepare(uap->clk);
serial/amba-pl011: move custom pin control to driver We had a boot regression in Ux500 in the merge window because two orthogonal pin control schemes for the PL011 were merged at the same time: - One using the .init() and .exit() hooks into the platform for Ux500 putting the pins into default vs sleep state respectively as the port was started/stopped. commit a09806607fd20bed2f8c41fe22793386790a14aa "ARM: ux500: switch to using pinctrl for uart0" - One hogging the default setting at PL011 probe() commit 258e055111d3cde2607e0d04eb91da2f7a59b591 "serial: amba-pl011: adopt pinctrl support" To get a solution that works for both let's scrap the stuff in the platform callbacks, instead have the driver itself select default and sleep states when the port is started/stopped. Hopefully this works for all clients. Platform callbacks are bad for device tree migration anyway, so this rids us of another problem in Ux500. Cc: Shawn Guo <shawn.guo@linaro.org> Cc: Russell King <rmk+kernel@arm.linux.org.uk> Reported-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-24 03:18:46 +08:00
/* Optionally let pins go into sleep states */
tty: serial: modify PL011 driver to use pinctrl PM helpers This augments the PL011 UART driver to utilize the new pinctrl core PM helpers to transition the driver to default and sleep states, cutting away some boilerplate code. Cc: Hebbar Gururaja <gururaja.hebbar@ti.com> Cc: Mark Brown <broonie@kernel.org> Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com> Cc: Kevin Hilman <khilman@linaro.org> Cc: Stephen Warren <swarren@wwwdotorg.org> Acked-by: Wolfram Sang <wsa@the-dreams.de> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-06-05 21:36:42 +08:00
pinctrl_pm_select_sleep_state(port->dev);
amba pl011: workaround for uart registers lockup This workaround aims to break the deadlock situation which raises during continuous transfer of data for long duration over uart with hardware flow control. It is observed that CTS interrupt cannot be cleared in uart interrupt register (ICR). Hence further transfer over uart gets blocked. It is seen that during such deadlock condition ICR don't get cleared even on multiple write. This leads pass_counter to decrease and finally reach zero. This can be taken as trigger point to run this UART_BT_WA. Workaround backups the register configuration, does soft reset of UART using BIT-0 of PRCC_K_SOFTRST_SET/CLEAR registers and restores the registers. This patch also provides support for uart init and exit function calls if present. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2011-06-13 16:11:33 +08:00
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
if (dev_get_platdata(uap->port.dev)) {
amba pl011: workaround for uart registers lockup This workaround aims to break the deadlock situation which raises during continuous transfer of data for long duration over uart with hardware flow control. It is observed that CTS interrupt cannot be cleared in uart interrupt register (ICR). Hence further transfer over uart gets blocked. It is seen that during such deadlock condition ICR don't get cleared even on multiple write. This leads pass_counter to decrease and finally reach zero. This can be taken as trigger point to run this UART_BT_WA. Workaround backups the register configuration, does soft reset of UART using BIT-0 of PRCC_K_SOFTRST_SET/CLEAR registers and restores the registers. This patch also provides support for uart init and exit function calls if present. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2011-06-13 16:11:33 +08:00
struct amba_pl011_data *plat;
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
plat = dev_get_platdata(uap->port.dev);
amba pl011: workaround for uart registers lockup This workaround aims to break the deadlock situation which raises during continuous transfer of data for long duration over uart with hardware flow control. It is observed that CTS interrupt cannot be cleared in uart interrupt register (ICR). Hence further transfer over uart gets blocked. It is seen that during such deadlock condition ICR don't get cleared even on multiple write. This leads pass_counter to decrease and finally reach zero. This can be taken as trigger point to run this UART_BT_WA. Workaround backups the register configuration, does soft reset of UART using BIT-0 of PRCC_K_SOFTRST_SET/CLEAR registers and restores the registers. This patch also provides support for uart init and exit function calls if present. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2011-06-13 16:11:33 +08:00
if (plat->exit)
plat->exit();
}
serial: pl011: Fix build breakage with !CONFIG_DMA_ENGINE Commit 479e9b94fdce7bc4 ("serial: Refactor uart_flush_buffer() from uart_close") refactored the uart_flush_buffer() in uart_close() into those drivers that define a flush_buffer() method. However, in the amba-pl011 driver configured without CONFIG_DMA_ENGINE, flush_buffer() is a NULL method, so the direct call fails to compile. Check and call the flush_buffer() method through the ops table instead. Reported-by: Fengguang Wu <fengguang.wu@intel.com> Cc: Russell King <linux@arm.linux.org.uk> Signed-off-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-06 22:06:12 +08:00
if (uap->port.ops->flush_buffer)
uap->port.ops->flush_buffer(port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static void sbsa_uart_shutdown(struct uart_port *port)
{
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
pl011_disable_interrupts(uap);
free_irq(uap->port.irq, uap);
if (uap->port.ops->flush_buffer)
uap->port.ops->flush_buffer(port);
}
drivers: PL011: refactor pl011_set_termios() Split the pl011_set_termios() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:17 +08:00
static void
pl011_setup_status_masks(struct uart_port *port, struct ktermios *termios)
{
port->read_status_mask = UART011_DR_OE | 255;
if (termios->c_iflag & INPCK)
port->read_status_mask |= UART011_DR_FE | UART011_DR_PE;
if (termios->c_iflag & (IGNBRK | BRKINT | PARMRK))
port->read_status_mask |= UART011_DR_BE;
/*
* Characters to ignore
*/
port->ignore_status_mask = 0;
if (termios->c_iflag & IGNPAR)
port->ignore_status_mask |= UART011_DR_FE | UART011_DR_PE;
if (termios->c_iflag & IGNBRK) {
port->ignore_status_mask |= UART011_DR_BE;
/*
* If we're ignoring parity and break indicators,
* ignore overruns too (for real raw support).
*/
if (termios->c_iflag & IGNPAR)
port->ignore_status_mask |= UART011_DR_OE;
}
/*
* Ignore all characters if CREAD is not set.
*/
if ((termios->c_cflag & CREAD) == 0)
port->ignore_status_mask |= UART_DUMMY_DR_RX;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static void
[PATCH] tty: switch to ktermios This is the grungy swap all the occurrences in the right places patch that goes with the updates. At this point we have the same functionality as before (except that sgttyb() returns speeds not zero) and are ready to begin turning new stuff on providing nobody reports lots of bugs If you are a tty driver author converting an out of tree driver the only impact should be termios->ktermios name changes for the speed/property setting functions from your upper layers. If you are implementing your own TCGETS function before then your driver was broken already and its about to get a whole lot more painful for you so please fix it 8) Also fill in c_ispeed/ospeed on init for most devices, although the current code will do this for you anyway but I'd like eventually to lose that extra paranoia [akpm@osdl.org: bluetooth fix] [mp3@de.ibm.com: sclp fix] [mp3@de.ibm.com: warning fix for tty3270] [hugh@veritas.com: fix tty_ioctl powerpc build] [jdike@addtoit.com: uml: fix ->set_termios declaration] Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Martin Peschke <mp3@de.ibm.com> Acked-by: Peter Oberparleiter <oberpar@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 18:38:45 +08:00
pl011_set_termios(struct uart_port *port, struct ktermios *termios,
struct ktermios *old)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned int lcr_h, old_cr;
unsigned long flags;
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
unsigned int baud, quot, clkdiv;
if (uap->vendor->oversampling)
clkdiv = 8;
else
clkdiv = 16;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Ask the core to calculate the divisor for us.
*/
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
baud = uart_get_baud_rate(port, termios, old, 0,
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
port->uartclk / clkdiv);
amba-pl011: fix build error if CONFIG_DMA_ENGINE is not enabled Following patch will fix build error if CONFIG_DMA_ENGINE is unset. Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-03 10:10:37 +08:00
#ifdef CONFIG_DMA_ENGINE
ARM: PL011: Add support for Rx DMA buffer polling. In DMA support, The received data is not pushed to tty until the DMA buffer is filled. But some megabyte rate chips such as BT expect fast response and data should be pushed immediately. In order to fix this issue, We suggest the use of the timer for polling DMA buffer. In our test, no data loss occurred at high-baudrate as compared with interrupt- driven (We tested with 3Mbps). We changes: - We add timer for polling. If we set poll_timer to 10, every 10ms, timer handler checks the residue in the dma buffer and transfer data to the tty. Also, last_residue is updated for the next polling. - poll_timeout is used to prevent the timer's system cost. If poll_timeout is set to 3000 and no data is received in 3 seconds, we inactivate poll timer and driver falls back to interrupt-driven. When data is received again in FIFO and UART irq is occurred, we switch back to DMA mode and start polling. - We use consistent DMA mappings to avoid from the frequent cache operation of the timer function for default. - pl011_dma_rx_chars is modified. the pending size is recalculated because data can be taken by polling. - the polling time is adjusted if dma rx poll is enabled but no rate is specified. Ideal polling interval to push 1 character at every interval is the reciprocal of 'baud rate / 10 line bits per character / 1000 ms per sec'. But It is very aggressive to system. Experimentally, '10000000 / baud' is suitable to receive dozens of characters. the poll rate can be specified statically by dma_rx_poll_rate of the platform data as well. Changes compared to v1: - Use of consistent DMA mappings. - Added dma_rx_poll_rate in platform data to specify the polling interval. - Added dma_rx_poll_timeout in platform data to specify the polling timeout. Changes compared to v2: - Use of consistent DMA mappings for default. - Added dma_rx_poll_enable in platform data to adjust the polling time according to the baud rate. - remove unnecessary lock from the polling function. Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-27 17:38:11 +08:00
/*
* Adjust RX DMA polling rate with baud rate if not specified.
*/
if (uap->dmarx.auto_poll_rate)
uap->dmarx.poll_rate = DIV_ROUND_UP(10000000, baud);
amba-pl011: fix build error if CONFIG_DMA_ENGINE is not enabled Following patch will fix build error if CONFIG_DMA_ENGINE is unset. Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Chanho Min <chanho.min@lge.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-04-03 10:10:37 +08:00
#endif
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
if (baud > port->uartclk/16)
quot = DIV_ROUND_CLOSEST(port->uartclk * 8, baud);
else
quot = DIV_ROUND_CLOSEST(port->uartclk * 4, baud);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
switch (termios->c_cflag & CSIZE) {
case CS5:
lcr_h = UART01x_LCRH_WLEN_5;
break;
case CS6:
lcr_h = UART01x_LCRH_WLEN_6;
break;
case CS7:
lcr_h = UART01x_LCRH_WLEN_7;
break;
default: // CS8
lcr_h = UART01x_LCRH_WLEN_8;
break;
}
if (termios->c_cflag & CSTOPB)
lcr_h |= UART01x_LCRH_STP2;
if (termios->c_cflag & PARENB) {
lcr_h |= UART01x_LCRH_PEN;
if (!(termios->c_cflag & PARODD))
lcr_h |= UART01x_LCRH_EPS;
}
ARM: PL011: Separate hardware FIFO size from TTY FIFO size With DMA support, we need to tell the TTY subsystem that the DMA buffer is the size of the FIFO, otherwise things like tty_wait_until_sent() will time out too early. Keep (and use) the hardware value separately from the port->fifosize. This was part of a larger patch from Linus Walleij, with a little modification. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:13:05 +08:00
if (uap->fifosize > 1)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
lcr_h |= UART01x_LCRH_FEN;
spin_lock_irqsave(&port->lock, flags);
/*
* Update the per-port timeout.
*/
uart_update_timeout(port, termios->c_cflag, baud);
drivers: PL011: refactor pl011_set_termios() Split the pl011_set_termios() function into smaller chunks to allow easier reuse later when adding SBSA support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:17 +08:00
pl011_setup_status_masks(port, termios);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (UART_ENABLE_MS(port, termios->c_cflag))
pl011_enable_ms(port);
/* first, disable everything */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
old_cr = pl011_read(uap, REG_CR);
pl011_write(0, uap, REG_CR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ARM: 5933/1: amba-pl011: support hardware flow control Enable/disable automatic hardware flow control as requested by the termios. The controller does not allow us to control the RTS line when auto-RTS is enabled, so we enable auto-RTS only if the kernel has not disabled RTS. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-02-12 13:43:11 +08:00
if (termios->c_cflag & CRTSCTS) {
if (old_cr & UART011_CR_RTS)
old_cr |= UART011_CR_RTSEN;
old_cr |= UART011_CR_CTSEN;
uap->autorts = true;
} else {
old_cr &= ~(UART011_CR_CTSEN | UART011_CR_RTSEN);
uap->autorts = false;
}
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
if (uap->vendor->oversampling) {
if (baud > port->uartclk / 16)
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
old_cr |= ST_UART011_CR_OVSFACT;
else
old_cr &= ~ST_UART011_CR_OVSFACT;
}
serial: pl011: handle corruption at high clock speeds This works around a few glitches in the ST version of the PL011 serial driver when using very high baud rates, as we do in the Ux500: 3, 3.25, 4 and 4.05 Mbps. Problem Observed/rootcause: When using high baud-rates, and the baudrate*8 is getting close to the provided clock frequency (so a division factor close to 1), when using bursts of characters (so they are abutted), then it seems as if there is not enough time to detect the beginning of the start-bit which is a timing reference for the entire character, and thus the sampling moment of character bits is moving towards the end of each bit, instead of the middle. Fix: Increase slightly the RX baud rate of the UART above the theoretical baudrate by 5%. This will definitely give more margin time to the UART_RX to correctly sample the data at the middle of the bit period. Also fix the ages old copy-paste error in the very stressed comment, it's referencing the registers used in the PL010 driver rather than the PL011 ones. Signed-off-by: Guillaume Jaunet <guillaume.jaunet@stericsson.com> Signed-off-by: Christophe Arnal <christophe.arnal@stericsson.com> Signed-off-by: Matthias Locher <matthias.locher@stericsson.com> Signed-off-by: Rajanikanth HV <rajanikanth.hv@stericsson.com> Cc: stable <stable@vger.kernel.org> Cc: Bibek Basu <bibek.basu@stericsson.com> Cc: Par-Gunnar Hjalmdahl <par-gunnar.hjalmdahl@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-26 23:21:36 +08:00
/*
* Workaround for the ST Micro oversampling variants to
* increase the bitrate slightly, by lowering the divisor,
* to avoid delayed sampling of start bit at high speeds,
* else we see data corruption.
*/
if (uap->vendor->oversampling) {
if ((baud >= 3000000) && (baud < 3250000) && (quot > 1))
quot -= 1;
else if ((baud > 3250000) && (quot > 2))
quot -= 2;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* Set baud rate */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(quot & 0x3f, uap, REG_FBRD);
pl011_write(quot >> 6, uap, REG_IBRD);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* ----------v----------v----------v----------v-----
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
* NOTE: REG_LCRH_TX and REG_LCRH_RX MUST BE WRITTEN AFTER
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
* REG_FBRD & REG_IBRD.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
* ----------^----------^----------^----------^-----
*/
serial: amba-pl011: factor out code for writing LCR_H register The code to cope with a split tx/rx LCR_H register is non-trivial so put it into it's own function to avoid duplication. Signed-off-by: Jon Medhurst <tixy@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-10 18:18:59 +08:00
pl011_write_lcr_h(uap, lcr_h);
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(old_cr, uap, REG_CR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
spin_unlock_irqrestore(&port->lock, flags);
}
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static void
sbsa_uart_set_termios(struct uart_port *port, struct ktermios *termios,
struct ktermios *old)
{
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
unsigned long flags;
tty_termios_encode_baud_rate(termios, uap->fixed_baud, uap->fixed_baud);
/* The SBSA UART only supports 8n1 without hardware flow control. */
termios->c_cflag &= ~(CSIZE | CSTOPB | PARENB | PARODD);
termios->c_cflag &= ~(CMSPAR | CRTSCTS);
termios->c_cflag |= CS8 | CLOCAL;
spin_lock_irqsave(&port->lock, flags);
uart_update_timeout(port, CS8, uap->fixed_baud);
pl011_setup_status_masks(port, termios);
spin_unlock_irqrestore(&port->lock, flags);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static const char *pl011_type(struct uart_port *port)
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
ARM: PL011: include revision number in boot-time port printk Include the revision number of the PL011 primecell in the boot-time port printk to allow proper identification of the peripheral. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-28 17:16:54 +08:00
return uap->port.type == PORT_AMBA ? uap->type : NULL;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
/*
* Release the memory region(s) being used by 'port'
*/
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static void pl011_release_port(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
release_mem_region(port->mapbase, SZ_4K);
}
/*
* Request the memory region(s) being used by 'port'
*/
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static int pl011_request_port(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
return request_mem_region(port->mapbase, SZ_4K, "uart-pl011")
!= NULL ? 0 : -EBUSY;
}
/*
* Configure/autoconfigure the port.
*/
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static void pl011_config_port(struct uart_port *port, int flags)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
if (flags & UART_CONFIG_TYPE) {
port->type = PORT_AMBA;
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
pl011_request_port(port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
/*
* verify the new serial_struct (for TIOCSSERIAL).
*/
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
static int pl011_verify_port(struct uart_port *port, struct serial_struct *ser)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
int ret = 0;
if (ser->type != PORT_UNKNOWN && ser->type != PORT_AMBA)
ret = -EINVAL;
drivers/serial: use nr_irqs Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-08-20 11:49:55 +08:00
if (ser->irq < 0 || ser->irq >= nr_irqs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ret = -EINVAL;
if (ser->baud_base < 9600)
ret = -EINVAL;
return ret;
}
static struct uart_ops amba_pl011_pops = {
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
.tx_empty = pl011_tx_empty,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
.set_mctrl = pl011_set_mctrl,
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
.get_mctrl = pl011_get_mctrl,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
.stop_tx = pl011_stop_tx,
.start_tx = pl011_start_tx,
.stop_rx = pl011_stop_rx,
.enable_ms = pl011_enable_ms,
.break_ctl = pl011_break_ctl,
.startup = pl011_startup,
.shutdown = pl011_shutdown,
ARM: PL011: Add support for transmit DMA Add DMA engine support for transmit to the PL011 driver. Based on a patch from Linus Walliej, with the following changes: - remove RX DMA support. As PL011 doesn't give us receive timeout interrupts, we only get notified of received data when the RX DMA has completed. This rather sucks for interactive use of the TTY. - remove abuse of completions. Completions are supposed to be for events, not to tell what condition buffers are in. Replace it with a simple 'queued' bool. - fix locking - it is only safe to access the circular buffer with the port lock held. - only map the DMA buffer when required - if we're ever behind an IOMMU this helps keep IOMMU usage down, and also ensures that we're legal when we change the scatterlist entry length. - fix XON/XOFF sending - we must send XON/XOFF characters out as soon as possible - waiting for up to 4095 characters in the DMA buffer to be sent first is not acceptable. - fix XON/XOFF receive handling - we need to stop DMA when instructed to by the TTY layer, and restart it again when instructed to. There is a subtle problem here: we must not completely empty the circular buffer with DMA, otherwise we will not be notified of XON. - change the 'enable_dma' flag into a 'using DMA' flag, and track whether we can use TX DMA by whether the channel pointer is non-NULL. This gives us more control over whether we use DMA in the driver. - we don't need to have the TX DMA buffer continually allocated for each port - instead, allocate it when the port starts up, and free it when it's shut down. Update the 'using DMA' flag if we get the buffer, and adjust the TTY FIFO size appropriately. - if we're going to use PIO to send characters, use the existing IRQ based functionality rather than reimplementing it. This also ensures we call uart_write_wakeup() at the appropriate time, otherwise we'll stall. - use DMA engine helper functions for type safety. - fix init when built as a module - we can't have to initcall functions, so we must settle on one. This means we can eliminate the deferred DMA initialization. - there is no need to terminate transfers on a failed prep_slave_sg() call - nothing has been setup, so nothing needs to be terminated. This avoids a potential deadlock in the DMA engine code (tasklet->callback->failed prepare->terminate->tasklet_disable which then ends up waiting for the tasklet to finish running.) - Dan says that the submission callback should not return an error: | dma_submit_error() is something I should have removed after commit | a0587bcf "ioat1: move descriptor allocation from submit to prep" all | errors should be notified by prep failing to return a descriptor | handle. Negative dma_cookie_t values are only returned by the | dma_async_memcpy* calls which translate a prep failure into -ENOMEM. So remove the error handling at that point. This also solves the potential deadlock mentioned in the previous comment. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:24:39 +08:00
.flush_buffer = pl011_dma_flush_buffer,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
.set_termios = pl011_set_termios,
.type = pl011_type,
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
.release_port = pl011_release_port,
.request_port = pl011_request_port,
.config_port = pl011_config_port,
.verify_port = pl011_verify_port,
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
#ifdef CONFIG_CONSOLE_POLL
tty/serial/amba-pl011: Implement poll_init callback The callback is used to initialize the hardware, nothing else should be done, i.e. we should not request interrupts (but we can and do unmask some of them, as they might be useful for NMI entry). As a side-effect, the patch also fixes a division by zero[1] when booting with kgdboc options specified (e.g. kgdboc=ttyAMA0,115200n8). The issue happens because serial core calls set_termios callback, but the driver doesn't know clock frequency, and thus cannot calculate proper baud rate values. [1] WARNING: at drivers/tty/serial/serial_core.c:400 uart_get_baud_rate+0xe8/0x14c() Modules linked in: [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) [<c0020ae8>] (warn_slowpath_common+0x4c/0x64) from [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) [<c0020b1c>] (warn_slowpath_null+0x1c/0x24) from [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) [<c0185ed8>] (uart_get_baud_rate+0xe8/0x14c) from [<c0187078>] (pl011_set_termios+0x48/0x278) [<c0187078>] (pl011_set_termios+0x48/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) ---[ end trace 7d41c9186f342c40 ]--- Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0187098>] (pl011_set_termios+0x68/0x278) [<c0187098>] (pl011_set_termios+0x68/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Division by zero in kernel. [<c0018e50>] (unwind_backtrace+0x0/0xf0) from [<c014546c>] (Ldiv0+0x8/0x10) [<c014546c>] (Ldiv0+0x8/0x10) from [<c0183a98>] (uart_update_timeout+0x4c/0x5c) [<c0183a98>] (uart_update_timeout+0x4c/0x5c) from [<c01870f8>] (pl011_set_termios+0xc8/0x278) [<c01870f8>] (pl011_set_termios+0xc8/0x278) from [<c01850b0>] (uart_set_options+0xe8/0x114) [<c01850b0>] (uart_set_options+0xe8/0x114) from [<c0185de4>] (uart_poll_init+0xd4/0xe0) [<c0185de4>] (uart_poll_init+0xd4/0xe0) from [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) [<c016da8c>] (tty_find_polling_driver+0x100/0x17c) from [<c0188538>] (configure_kgdboc+0xc8/0x1b8) [<c0188538>] (configure_kgdboc+0xc8/0x1b8) from [<c00088a4>] (do_one_initcall+0x30/0x168) [<c00088a4>] (do_one_initcall+0x30/0x168) from [<c033784c>] (do_basic_setup+0x94/0xc8) [<c033784c>] (do_basic_setup+0x94/0xc8) from [<c03378e0>] (kernel_init+0x60/0xf4) [<c03378e0>] (kernel_init+0x60/0xf4) from [<c00144a0>] (kernel_thread_exit+0x0/0x8) Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-09-25 05:27:54 +08:00
.poll_init = pl011_hwinit,
serial/amba-pl011: fix ages old copy-paste errors The PL011 driver has a number of symbols referring to "pl01x" (probably once shared with the pl010 driver) and some even named "pl010" (probably a pure copy-paste artifact). Lets name all local static functions with the prefix pl011_* for clarity. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-06-17 21:44:19 +08:00
.poll_get_char = pl011_get_poll_char,
.poll_put_char = pl011_put_poll_char,
kgdb: kgdboc pl011 I/O module Implement the serial polling hooks for the pl011 uart for use with kgdboc. This patch was specifically tested on the ARM Versatile AB reference platform. [ mingo@elte.hu: minor cleanups. ] Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-02-21 03:33:39 +08:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
};
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static void sbsa_uart_set_mctrl(struct uart_port *port, unsigned int mctrl)
{
}
static unsigned int sbsa_uart_get_mctrl(struct uart_port *port)
{
return 0;
}
static const struct uart_ops sbsa_uart_pops = {
.tx_empty = pl011_tx_empty,
.set_mctrl = sbsa_uart_set_mctrl,
.get_mctrl = sbsa_uart_get_mctrl,
.stop_tx = pl011_stop_tx,
.start_tx = pl011_start_tx,
.stop_rx = pl011_stop_rx,
.startup = sbsa_uart_startup,
.shutdown = sbsa_uart_shutdown,
.set_termios = sbsa_uart_set_termios,
.type = pl011_type,
.release_port = pl011_release_port,
.request_port = pl011_request_port,
.config_port = pl011_config_port,
.verify_port = pl011_verify_port,
#ifdef CONFIG_CONSOLE_POLL
.poll_init = pl011_hwinit,
.poll_get_char = pl011_get_poll_char,
.poll_put_char = pl011_put_poll_char,
#endif
};
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static struct uart_amba_port *amba_ports[UART_NR];
#ifdef CONFIG_SERIAL_AMBA_PL011_CONSOLE
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-21 04:00:09 +08:00
static void pl011_console_putchar(struct uart_port *port, int ch)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
serial: amba-pl011: Use container_of() to get uart_amba_port Universally adopt container_of() for all pointer conversion from uart_port to uart_amba_port. Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org> Cc: Peter Hurley <peter@hurleysoftware.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Cc: linux-serial@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-09-03 19:51:55 +08:00
struct uart_amba_port *uap =
container_of(port, struct uart_amba_port, port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
while (pl011_read(uap, REG_FR) & UART01x_FR_TXFF)
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-21 04:00:09 +08:00
barrier();
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(ch, uap, REG_DR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static void
pl011_console_write(struct console *co, const char *s, unsigned int count)
{
struct uart_amba_port *uap = amba_ports[co->index];
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
unsigned int status, old_cr = 0, new_cr;
serial: amba-pl011: lock console writes against interrupts Protect against pl011_console_write() and the interrupt for the console UART running concurrently on different CPUs. Otherwise the console_write could spin for a long time waiting for the UART to become not busy, while the other CPU continuously services UART interrupts and keeps the UART busy. The checks for sysrq and oops_in_progress are taken from 8250.c. Cc: stable <stable@vger.kernel.org> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Reviewed-by: Bibek Basu <bibek.basu@stericsson.com> Reviewed-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-17 18:52:28 +08:00
unsigned long flags;
int locked = 1;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
clk_enable(uap->clk);
serial: amba-pl011: lock console writes against interrupts Protect against pl011_console_write() and the interrupt for the console UART running concurrently on different CPUs. Otherwise the console_write could spin for a long time waiting for the UART to become not busy, while the other CPU continuously services UART interrupts and keeps the UART busy. The checks for sysrq and oops_in_progress are taken from 8250.c. Cc: stable <stable@vger.kernel.org> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Reviewed-by: Bibek Basu <bibek.basu@stericsson.com> Reviewed-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-17 18:52:28 +08:00
local_irq_save(flags);
if (uap->port.sysrq)
locked = 0;
else if (oops_in_progress)
locked = spin_trylock(&uap->port.lock);
else
spin_lock(&uap->port.lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* First save the CR then disable the interrupts
*/
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
if (!uap->vendor->always_enabled) {
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
old_cr = pl011_read(uap, REG_CR);
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
new_cr = old_cr & ~UART011_CR_CTSEN;
new_cr |= UART01x_CR_UARTEN | UART011_CR_TXE;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(new_cr, uap, REG_CR);
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-21 04:00:09 +08:00
uart_console_write(&uap->port, s, count, pl011_console_putchar);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Finally, wait for transmitter to become empty
* and restore the TCR
*/
do {
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
status = pl011_read(uap, REG_FR);
Revert "uart: pl011: Add support to ZTE ZX296702 uart" This reverts commit 8cd90e50d1408c65c355084b1c7f8f9085f49c6b as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:11:24 +08:00
} while (status & UART01x_FR_BUSY);
drivers: PL011: allow avoiding UART enabling/disabling The SBSA UART should not be enabled or disabled (it is always on), and consequently the spec lacks the UART_CR register. Add a vendor specific property to skip disabling or enabling of the UART. This will be used later by the SBSA UART support. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:21 +08:00
if (!uap->vendor->always_enabled)
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(old_cr, uap, REG_CR);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial: amba-pl011: lock console writes against interrupts Protect against pl011_console_write() and the interrupt for the console UART running concurrently on different CPUs. Otherwise the console_write could spin for a long time waiting for the UART to become not busy, while the other CPU continuously services UART interrupts and keeps the UART busy. The checks for sysrq and oops_in_progress are taken from 8250.c. Cc: stable <stable@vger.kernel.org> Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com> Reviewed-by: Bibek Basu <bibek.basu@stericsson.com> Reviewed-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-17 18:52:28 +08:00
if (locked)
spin_unlock(&uap->port.lock);
local_irq_restore(flags);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
clk_disable(uap->clk);
}
static void __init
pl011_console_get_options(struct uart_amba_port *uap, int *baud,
int *parity, int *bits)
{
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
if (pl011_read(uap, REG_CR) & UART01x_CR_UARTEN) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned int lcr_h, ibrd, fbrd;
tty: amba-pl011: clean up LCR register offsets As we can detect when the LCR register is split between TX and RX, we don't need three entries in the table to deal with this. Reduce this down to two entries by converting the REG_ST_LCRH_* entries to standard REG_LCRH_* and remove REG_LCRH. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:41 +08:00
lcr_h = pl011_read(uap, REG_LCRH_TX);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*parity = 'n';
if (lcr_h & UART01x_LCRH_PEN) {
if (lcr_h & UART01x_LCRH_EPS)
*parity = 'e';
else
*parity = 'o';
}
if ((lcr_h & 0x60) == UART01x_LCRH_WLEN_7)
*bits = 7;
else
*bits = 8;
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
ibrd = pl011_read(uap, REG_IBRD);
fbrd = pl011_read(uap, REG_FBRD);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*baud = uap->port.uartclk * 4 / (64 * ibrd + fbrd);
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
ARM: PL011: Allow better handling of vendor data Rather than copying all vendor data into the port structure, copy just that which is frequently used, and keep a pointer to the remaining vendor data structure. This makes it easier to add vendor quirks in the future. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:48:26 +08:00
if (uap->vendor->oversampling) {
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
if (pl011_read(uap, REG_CR)
ARM: 6158/2: PL011 baudrate extension for ST-Ericssons derivative Implementation of the ST-Ericsson baudrate extension in the PL011 block. In this modified variant it is possible to change the sampling factor from 16 to 8, and thanks to this we can get higher baudrates while still using the same peripheral clock. Also replace the simple division to determine the baud divisor with DIV_ROUND_CLOSEST() rather than a simple integer division. Cc: Alessandro Rubini <rubini@unipv.it> Cc: Jerzy Kasenberg <jerzy.kasenberg@tieto.com> Signed-off-by: Marcin Mielczarczyk <marcin.mielczarczyk@tieto.com> Signed-off-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-06-03 03:40:22 +08:00
& ST_UART011_CR_OVSFACT)
*baud *= 2;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
static int __init pl011_console_setup(struct console *co, char *options)
{
struct uart_amba_port *uap;
int baud = 38400;
int bits = 8;
int parity = 'n';
int flow = 'n';
clk: amba-pl011: convert to clk_prepare()/clk_unprepare() Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-09-22 18:35:30 +08:00
int ret;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Check whether an invalid uart number has been specified, and
* if so, search for the first available port that does have
* console support.
*/
if (co->index >= UART_NR)
co->index = 0;
uap = amba_ports[co->index];
[ARM] Fix AMBA serial drivers for non-first serial ports Using console=ttyAM1 or console=ttyAMA1 resulted in an oops during boot due to trying to drive the console before that port had been registered. Fix this by checking whether the port is present before allowing console setup to proceed. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-23 02:59:42 +08:00
if (!uap)
return -ENODEV;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
serial/amba-pl011: move custom pin control to driver We had a boot regression in Ux500 in the merge window because two orthogonal pin control schemes for the PL011 were merged at the same time: - One using the .init() and .exit() hooks into the platform for Ux500 putting the pins into default vs sleep state respectively as the port was started/stopped. commit a09806607fd20bed2f8c41fe22793386790a14aa "ARM: ux500: switch to using pinctrl for uart0" - One hogging the default setting at PL011 probe() commit 258e055111d3cde2607e0d04eb91da2f7a59b591 "serial: amba-pl011: adopt pinctrl support" To get a solution that works for both let's scrap the stuff in the platform callbacks, instead have the driver itself select default and sleep states when the port is started/stopped. Hopefully this works for all clients. Platform callbacks are bad for device tree migration anyway, so this rids us of another problem in Ux500. Cc: Shawn Guo <shawn.guo@linaro.org> Cc: Russell King <rmk+kernel@arm.linux.org.uk> Reported-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-24 03:18:46 +08:00
/* Allow pins to be muxed in and configured */
tty: serial: modify PL011 driver to use pinctrl PM helpers This augments the PL011 UART driver to utilize the new pinctrl core PM helpers to transition the driver to default and sleep states, cutting away some boilerplate code. Cc: Hebbar Gururaja <gururaja.hebbar@ti.com> Cc: Mark Brown <broonie@kernel.org> Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com> Cc: Kevin Hilman <khilman@linaro.org> Cc: Stephen Warren <swarren@wwwdotorg.org> Acked-by: Wolfram Sang <wsa@the-dreams.de> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-06-05 21:36:42 +08:00
pinctrl_pm_select_default_state(uap->port.dev);
serial/amba-pl011: move custom pin control to driver We had a boot regression in Ux500 in the merge window because two orthogonal pin control schemes for the PL011 were merged at the same time: - One using the .init() and .exit() hooks into the platform for Ux500 putting the pins into default vs sleep state respectively as the port was started/stopped. commit a09806607fd20bed2f8c41fe22793386790a14aa "ARM: ux500: switch to using pinctrl for uart0" - One hogging the default setting at PL011 probe() commit 258e055111d3cde2607e0d04eb91da2f7a59b591 "serial: amba-pl011: adopt pinctrl support" To get a solution that works for both let's scrap the stuff in the platform callbacks, instead have the driver itself select default and sleep states when the port is started/stopped. Hopefully this works for all clients. Platform callbacks are bad for device tree migration anyway, so this rids us of another problem in Ux500. Cc: Shawn Guo <shawn.guo@linaro.org> Cc: Russell King <rmk+kernel@arm.linux.org.uk> Reported-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-24 03:18:46 +08:00
clk: amba-pl011: convert to clk_prepare()/clk_unprepare() Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-09-22 18:35:30 +08:00
ret = clk_prepare(uap->clk);
if (ret)
return ret;
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
if (dev_get_platdata(uap->port.dev)) {
amba pl011: workaround for uart registers lockup This workaround aims to break the deadlock situation which raises during continuous transfer of data for long duration over uart with hardware flow control. It is observed that CTS interrupt cannot be cleared in uart interrupt register (ICR). Hence further transfer over uart gets blocked. It is seen that during such deadlock condition ICR don't get cleared even on multiple write. This leads pass_counter to decrease and finally reach zero. This can be taken as trigger point to run this UART_BT_WA. Workaround backups the register configuration, does soft reset of UART using BIT-0 of PRCC_K_SOFTRST_SET/CLEAR registers and restores the registers. This patch also provides support for uart init and exit function calls if present. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2011-06-13 16:11:33 +08:00
struct amba_pl011_data *plat;
serial: use dev_get_platdata() Use the wrapper function for retrieving the platform data instead of accessing dev->platform_data directly. Signed-off-by: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-07-30 16:06:57 +08:00
plat = dev_get_platdata(uap->port.dev);
amba pl011: workaround for uart registers lockup This workaround aims to break the deadlock situation which raises during continuous transfer of data for long duration over uart with hardware flow control. It is observed that CTS interrupt cannot be cleared in uart interrupt register (ICR). Hence further transfer over uart gets blocked. It is seen that during such deadlock condition ICR don't get cleared even on multiple write. This leads pass_counter to decrease and finally reach zero. This can be taken as trigger point to run this UART_BT_WA. Workaround backups the register configuration, does soft reset of UART using BIT-0 of PRCC_K_SOFTRST_SET/CLEAR registers and restores the registers. This patch also provides support for uart init and exit function calls if present. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2011-06-13 16:11:33 +08:00
if (plat->init)
plat->init();
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->port.uartclk = clk_get_rate(uap->clk);
drivers: PL011: allow to supply fixed option string The SBSA UART has a fixed baud rate and flow control setting, which cannot be changed or queried by software. Add a vendor specific property to always return fixed values when trying to read the console options. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:22 +08:00
if (uap->vendor->fixed_options) {
baud = uap->fixed_baud;
} else {
if (options)
uart_parse_options(options,
&baud, &parity, &bits, &flow);
else
pl011_console_get_options(uap, &baud, &parity, &bits);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return uart_set_options(&uap->port, co, baud, parity, bits, flow);
}
[ARM] 2907/1: GCC 4 serial driver compile fixes Patch from Vincent Sanders When building the ARM platforms several serial drivers fail to compile with GCC 4.01 due to extern/static ambiguity. Signed-off-by: Vincent Sanders <vince@arm.linux.org.uk> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-09-15 05:36:03 +08:00
static struct uart_driver amba_reg;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static struct console amba_console = {
.name = "ttyAMA",
.write = pl011_console_write,
.device = uart_console_device,
.setup = pl011_console_setup,
.flags = CON_PRINTBUFFER,
.index = -1,
.data = &amba_reg,
};
#define AMBA_CONSOLE (&amba_console)
tty/serial: pl011: add generic earlycon support Add earlycon support for the pl011 serial port. This allows enabling the pl011 for console when early_params are processed. This is based on the arm64 earlyprintk support and is intended to replace it. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-19 06:19:57 +08:00
static void pl011_putc(struct uart_port *port, int c)
{
tty: amba-pl011: fix earlycon register offsets The REG_x macros are indices into a table, not register offsets. Since earlycon does not have access to the vendor data, we can currently only support standard ARM PL011 devices. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Tested-by: Huang Shijie <shijie.huang@arm.com> Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:41 +08:00
while (readl(port->membase + UART01x_FR) & UART01x_FR_TXFF)
tty/serial: pl011: add generic earlycon support Add earlycon support for the pl011 serial port. This allows enabling the pl011 for console when early_params are processed. This is based on the arm64 earlyprintk support and is intended to replace it. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-19 06:19:57 +08:00
;
tty: amba-pl011: use iotype instead of access_32b to track 32-bit I/O Instead of defining a new field in the uart_amba_port structure, use the existing iotype field of the uart_port structure, which is intended for this purpose. If we need to use 32-bit register access, we set iotype to UPIO_MEM32, otherwise we set it to UPIO_MEM. For early console, specify the "mmio32" option on the kernel command-line. Example: earlycon=pl011,mmio32,0x3ced1000 Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:42 +08:00
if (port->iotype == UPIO_MEM32)
writel(c, port->membase + UART01x_DR);
else
writeb(c, port->membase + UART01x_DR);
tty: amba-pl011: fix earlycon register offsets The REG_x macros are indices into a table, not register offsets. Since earlycon does not have access to the vendor data, we can currently only support standard ARM PL011 devices. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Tested-by: Huang Shijie <shijie.huang@arm.com> Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:41 +08:00
while (readl(port->membase + UART01x_FR) & UART01x_FR_BUSY)
tty/serial: pl011: add generic earlycon support Add earlycon support for the pl011 serial port. This allows enabling the pl011 for console when early_params are processed. This is based on the arm64 earlyprintk support and is intended to replace it. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-19 06:19:57 +08:00
;
}
static void pl011_early_write(struct console *con, const char *s, unsigned n)
{
struct earlycon_device *dev = con->data;
uart_console_write(&dev->port, s, n, pl011_putc);
}
static int __init pl011_early_console_setup(struct earlycon_device *device,
const char *opt)
{
if (!device->port.membase)
return -ENODEV;
device->con->write = pl011_early_write;
return 0;
}
EARLYCON_DECLARE(pl011, pl011_early_console_setup);
tty/serial: pl011: add DT based earlycon support Enable DT based earlycon initialization for the pl011 uart. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Jiri Slaby <jslaby@suse.cz> Acked-by: Grant Likely <grant.likely@linaro.org>
2014-03-27 21:08:03 +08:00
OF_EARLYCON_DECLARE(pl011, "arm,pl011", pl011_early_console_setup);
tty/serial: pl011: add generic earlycon support Add earlycon support for the pl011 serial port. This allows enabling the pl011 for console when early_params are processed. This is based on the arm64 earlyprintk support and is intended to replace it. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Russell King <linux@arm.linux.org.uk> Cc: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-19 06:19:57 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#else
#define AMBA_CONSOLE NULL
#endif
static struct uart_driver amba_reg = {
.owner = THIS_MODULE,
.driver_name = "ttyAMA",
.dev_name = "ttyAMA",
.major = SERIAL_AMBA_MAJOR,
.minor = SERIAL_AMBA_MINOR,
.nr = UART_NR,
.cons = AMBA_CONSOLE,
};
serial: pl011: honour serial aliases in device tree If the order of UART nodes is changed in the device tree, then tty dev devices are attached to different serial ports causing the console to be directed to a different physical serial port. The "serial" aliases in the device tree should prevent this. This patch ensures that the UART driver creates tty devices that honour these aliases if a device tree is present. Acked-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Will Deacon <will.deacon@arm.com> Acked-by: Rob Herring <rob.herring@calxeda.com> Signed-off-by: Matthew Leach <matthew.leach@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-08-28 23:41:28 +08:00
static int pl011_probe_dt_alias(int index, struct device *dev)
{
struct device_node *np;
static bool seen_dev_with_alias = false;
static bool seen_dev_without_alias = false;
int ret = index;
if (!IS_ENABLED(CONFIG_OF))
return ret;
np = dev->of_node;
if (!np)
return ret;
ret = of_alias_get_id(np, "serial");
if (IS_ERR_VALUE(ret)) {
seen_dev_without_alias = true;
ret = index;
} else {
seen_dev_with_alias = true;
if (ret >= ARRAY_SIZE(amba_ports) || amba_ports[ret] != NULL) {
dev_warn(dev, "requested serial port %d not available.\n", ret);
ret = index;
}
}
if (seen_dev_with_alias && seen_dev_without_alias)
dev_warn(dev, "aliased and non-aliased serial devices found in device tree. Serial port enumeration may be unpredictable.\n");
return ret;
}
drivers: PL011: avoid potential unregister_driver call Although we care about not unregistering the driver if there are still ports connected during the .remove callback, we do miss this check in the pl011_probe function. So if the current port allocation fails, but there are other ports already registered, we will kill those. So factor out the port removal into a separate function and use that in the probe function, too. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:14 +08:00
/* unregisters the driver also if no more ports are left */
static void pl011_unregister_port(struct uart_amba_port *uap)
{
int i;
bool busy = false;
for (i = 0; i < ARRAY_SIZE(amba_ports); i++) {
if (amba_ports[i] == uap)
amba_ports[i] = NULL;
else if (amba_ports[i])
busy = true;
}
pl011_dma_remove(uap);
if (!busy)
uart_unregister_driver(&amba_reg);
}
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
static int pl011_find_free_port(void)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
int i;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
for (i = 0; i < ARRAY_SIZE(amba_ports); i++)
if (amba_ports[i] == NULL)
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
return i;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
return -EBUSY;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
static int pl011_setup_port(struct device *dev, struct uart_amba_port *uap,
struct resource *mmiobase, int index)
{
void __iomem *base;
serial: pl011: honour serial aliases in device tree If the order of UART nodes is changed in the device tree, then tty dev devices are attached to different serial ports causing the console to be directed to a different physical serial port. The "serial" aliases in the device tree should prevent this. This patch ensures that the UART driver creates tty devices that honour these aliases if a device tree is present. Acked-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Will Deacon <will.deacon@arm.com> Acked-by: Rob Herring <rob.herring@calxeda.com> Signed-off-by: Matthew Leach <matthew.leach@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-08-28 23:41:28 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
base = devm_ioremap_resource(dev, mmiobase);
serial: amba-pl011: Fix devm_ioremap_resource return value check Value returned by devm_ioremap_resource() was checked for non-NULL but devm_ioremap_resource() returns IOMEM_ERR_PTR, not NULL. In case of error this could lead to dereference of ERR_PTR. Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com> Cc: <stable@vger.kernel.org> Fixes: 3873e2d7f63a ("drivers: PL011: refactor pl011_probe()") Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-07-09 21:21:41 +08:00
if (IS_ERR(base))
return PTR_ERR(base);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
index = pl011_probe_dt_alias(index, dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
amba-pl011: do not disable RTS during shutdown In present driver, shutdown clears RTS and DTR in CR register. But the documentation "Documentation/serial/driver" suggests not to disable RTS and DTR in shutdown(). Also RTS and DTR is preserved between shutdown and startup calls, i.e. these are restored in startup if they were enabled while doing shutdown. So that if RTS and DTR are set using pl011_set_mctrl then it should continue even after shutdown->startup sequence. For throttling/unthrottling user should call pl011_set_mctrl. Signed-off-by: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2012-01-18 18:23:59 +08:00
uap->old_cr = 0;
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
uap->port.dev = dev;
uap->port.mapbase = mmiobase->start;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->port.membase = base;
ARM: PL011: Separate hardware FIFO size from TTY FIFO size With DMA support, we need to tell the TTY subsystem that the DMA buffer is the size of the FIFO, otherwise things like tty_wait_until_sent() will time out too early. Keep (and use) the hardware value separately from the port->fifosize. This was part of a larger patch from Linus Walleij, with a little modification. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-23 01:13:05 +08:00
uap->port.fifosize = uap->fifosize;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
uap->port.flags = UPF_BOOT_AUTOCONF;
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
uap->port.line = index;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
amba_ports[index] = uap;
serial: PL011: move interrupt clearing Commit 360f748b204275229f8398cb2f9f53955db1503b "serial: PL011: clear pending interrupts" attempts to clear interrupts by writing to a yet-unassigned memory address. This fixes the issue. The breaking patch is marked for stable so should be carried along with the other patch. Cc: Shreshtha Kumar Sahu <shreshthakumar.sahu@stericsson.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: stable <stable@vger.kernel.org> Cc: Nicolas Pitre <nico@fluxnic.net> Reported-by: Viresh Kumar <viresh.kumar@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-03-22 03:15:18 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
return 0;
}
ARM: PL011: include revision number in boot-time port printk Include the revision number of the PL011 primecell in the boot-time port printk to allow proper identification of the peripheral. Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-28 17:16:54 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
static int pl011_register_port(struct uart_amba_port *uap)
{
int ret;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
/* Ensure interrupts from this UART are masked and cleared */
tty: amba-pl011: prepare REG_* register indexes Prepare for REG_* register accessors. This change involves introducing pl011_reg_to_offset() to convert REG_* to the hardware register offset, and converting all call sites to use REG_* names. We need to fix up locations where we check for equivalence of register offsets as well. Much of this change was made via these sed expressions: s/ST_UART01[1x]\(_[^_]*\|_LCRH_[TR]X\)\>/REG_ST\1/ s/UART01[1x]_\(DR\|RSR\|ECR\|FR\|ILPR\|[IF]BRD\|LCRH\|CR\|IFLS\|IMSC\|RIS\|MIS\|ICR\|DMACR\)\>/REG_\1/g Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-03 22:51:13 +08:00
pl011_write(0, uap, REG_IMSC);
pl011_write(0xffff, uap, REG_ICR);
serial: pl011: Move uart_register_driver call to device probe uart_register_driver call binds the driver to a specific device node through tty_register_driver call. This should typically happen during device probe call. In a multiplatform scenario, it is possible that multiple serial drivers are part of the kernel. Currently the driver registration fails if multiple serial drivers with same default major/minor numbers are included in the kernel. A typical case is observed with amba-pl011 and samsung-uart drivers. Signed-off-by: Tushar Behera <tushar.behera@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-01-20 17:02:35 +08:00
if (!amba_reg.state) {
ret = uart_register_driver(&amba_reg);
if (ret < 0) {
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
dev_err(uap->port.dev,
drivers/tty: amba: defer DMA probe until the DMA channel is required. Fix a race condition that happens when device_initcall(pl011_dma_initicall) is executed before all the devices have been probed - this issue was observed on a hisi_6220 SoC (HiKey board from Linaro). The deferred driver probing framework relies on late_initcall to trigger deferred probes so it is just possible that, even with a valid DMA driver ready to be loaded, we fail to synchronize with it. The proposed implementation delays probing the DMA until dma_startup. As this is invoked on port startup and port resume - but DMA probing is only required once - we avoid calling multiple times using a new field in uart_amba_port to track this scenario. This commit allows for subsequent attempts to associate an external DMA if the DMA driver itself is not available (but present in the deferred probe pending list). Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-03-07 02:05:40 +08:00
"Failed to register AMBA-PL011 driver\n");
serial: pl011: Move uart_register_driver call to device probe uart_register_driver call binds the driver to a specific device node through tty_register_driver call. This should typically happen during device probe call. In a multiplatform scenario, it is possible that multiple serial drivers are part of the kernel. Currently the driver registration fails if multiple serial drivers with same default major/minor numbers are included in the kernel. A typical case is observed with amba-pl011 and samsung-uart drivers. Signed-off-by: Tushar Behera <tushar.behera@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-01-20 17:02:35 +08:00
return ret;
}
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ret = uart_add_one_port(&amba_reg, &uap->port);
drivers: PL011: avoid potential unregister_driver call Although we care about not unregistering the driver if there are still ports connected during the .remove callback, we do miss this check in the pl011_probe function. So if the current port allocation fails, but there are other ports already registered, we will kill those. So factor out the port removal into a separate function and use that in the probe function, too. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:14 +08:00
if (ret)
pl011_unregister_port(uap);
serial: amba-pl011: Remove redundant label The label 'out' is only used to return the error code. We can return the error code directly and remove 'out' label. Signed-off-by: Tushar Behera <tushar.b@samsung.com> Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-06-26 18:05:35 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return ret;
}
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
static int pl011_probe(struct amba_device *dev, const struct amba_id *id)
{
struct uart_amba_port *uap;
struct vendor_data *vendor = id->data;
int portnr, ret;
portnr = pl011_find_free_port();
if (portnr < 0)
return portnr;
uap = devm_kzalloc(&dev->dev, sizeof(struct uart_amba_port),
GFP_KERNEL);
if (!uap)
return -ENOMEM;
uap->clk = devm_clk_get(&dev->dev, NULL);
if (IS_ERR(uap->clk))
return PTR_ERR(uap->clk);
tty: amba-pl011: add register offset table to vendor data Add the register offset table to the vendor data, allowing vendor differences to be described in this table. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:31 +08:00
uap->reg_offset = vendor->reg_offset;
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
uap->vendor = vendor;
uap->fifosize = vendor->get_fifosize(dev);
tty: amba-pl011: use iotype instead of access_32b to track 32-bit I/O Instead of defining a new field in the uart_amba_port structure, use the existing iotype field of the uart_port structure, which is intended for this purpose. If we need to use 32-bit register access, we set iotype to UPIO_MEM32, otherwise we set it to UPIO_MEM. For early console, specify the "mmio32" option on the kernel command-line. Example: earlycon=pl011,mmio32,0x3ced1000 Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:42 +08:00
uap->port.iotype = vendor->access_32b ? UPIO_MEM32 : UPIO_MEM;
drivers: PL011: refactor pl011_probe() Currently the pl011_probe() function is relying on some AMBA IDs and a device tree node to initialize the driver and a port. Both features are not necessarily required for the driver: - we lack AMBA IDs in the ARM SBSA generic UART and - we lack a DT node in ACPI systems. So lets refactor the function to ease later reuse. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:18 +08:00
uap->port.irq = dev->irq[0];
uap->port.ops = &amba_pl011_pops;
snprintf(uap->type, sizeof(uap->type), "PL011 rev%u", amba_rev(dev));
ret = pl011_setup_port(&dev->dev, uap, &dev->res, portnr);
if (ret)
return ret;
amba_set_drvdata(dev, uap);
return pl011_register_port(uap);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static int pl011_remove(struct amba_device *dev)
{
struct uart_amba_port *uap = amba_get_drvdata(dev);
uart_remove_one_port(&amba_reg, &uap->port);
drivers: PL011: avoid potential unregister_driver call Although we care about not unregistering the driver if there are still ports connected during the .remove callback, we do miss this check in the pl011_probe function. So if the current port allocation fails, but there are other ports already registered, we will kill those. So factor out the port removal into a separate function and use that in the probe function, too. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:14 +08:00
pl011_unregister_port(uap);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return 0;
}
serial: pl011: Convert to modern PM ops Convert to modern PM ops and use the SIMPLE_DEV_PM_OPS macro to set up the PM callbacks. Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-03 18:04:28 +08:00
#ifdef CONFIG_PM_SLEEP
static int pl011_suspend(struct device *dev)
ARM: 5626/1: add suspend/resume functions to amba-pl011 serial driver Add suspend/resume functions to the AMBA pl011 serial driver. Signed-off-by: Leo Chen <leochen@broadcom.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-07-29 06:43:33 +08:00
{
serial: pl011: Convert to modern PM ops Convert to modern PM ops and use the SIMPLE_DEV_PM_OPS macro to set up the PM callbacks. Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-03 18:04:28 +08:00
struct uart_amba_port *uap = dev_get_drvdata(dev);
ARM: 5626/1: add suspend/resume functions to amba-pl011 serial driver Add suspend/resume functions to the AMBA pl011 serial driver. Signed-off-by: Leo Chen <leochen@broadcom.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-07-29 06:43:33 +08:00
if (!uap)
return -EINVAL;
return uart_suspend_port(&amba_reg, &uap->port);
}
serial: pl011: Convert to modern PM ops Convert to modern PM ops and use the SIMPLE_DEV_PM_OPS macro to set up the PM callbacks. Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-03 18:04:28 +08:00
static int pl011_resume(struct device *dev)
ARM: 5626/1: add suspend/resume functions to amba-pl011 serial driver Add suspend/resume functions to the AMBA pl011 serial driver. Signed-off-by: Leo Chen <leochen@broadcom.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-07-29 06:43:33 +08:00
{
serial: pl011: Convert to modern PM ops Convert to modern PM ops and use the SIMPLE_DEV_PM_OPS macro to set up the PM callbacks. Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-03 18:04:28 +08:00
struct uart_amba_port *uap = dev_get_drvdata(dev);
ARM: 5626/1: add suspend/resume functions to amba-pl011 serial driver Add suspend/resume functions to the AMBA pl011 serial driver. Signed-off-by: Leo Chen <leochen@broadcom.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-07-29 06:43:33 +08:00
if (!uap)
return -EINVAL;
return uart_resume_port(&amba_reg, &uap->port);
}
#endif
serial: pl011: Convert to modern PM ops Convert to modern PM ops and use the SIMPLE_DEV_PM_OPS macro to set up the PM callbacks. Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-03 18:04:28 +08:00
static SIMPLE_DEV_PM_OPS(pl011_dev_pm_ops, pl011_suspend, pl011_resume);
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static int sbsa_uart_probe(struct platform_device *pdev)
{
struct uart_amba_port *uap;
struct resource *r;
int portnr, ret;
int baudrate;
/*
* Check the mandatory baud rate parameter in the DT node early
* so that we can easily exit with the error.
*/
if (pdev->dev.of_node) {
struct device_node *np = pdev->dev.of_node;
ret = of_property_read_u32(np, "current-speed", &baudrate);
if (ret)
return ret;
} else {
baudrate = 115200;
}
portnr = pl011_find_free_port();
if (portnr < 0)
return portnr;
uap = devm_kzalloc(&pdev->dev, sizeof(struct uart_amba_port),
GFP_KERNEL);
if (!uap)
return -ENOMEM;
tty: amba-pl011: add register offset table to vendor data Add the register offset table to the vendor data, allowing vendor differences to be described in this table. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Reviewed-by: Peter Hurley <peter@hurleysoftware.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-11-17 01:40:31 +08:00
uap->reg_offset = vendor_sbsa.reg_offset;
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
uap->vendor = &vendor_sbsa;
uap->fifosize = 32;
tty: amba-pl011: use iotype instead of access_32b to track 32-bit I/O Instead of defining a new field in the uart_amba_port structure, use the existing iotype field of the uart_port structure, which is intended for this purpose. If we need to use 32-bit register access, we set iotype to UPIO_MEM32, otherwise we set it to UPIO_MEM. For early console, specify the "mmio32" option on the kernel command-line. Example: earlycon=pl011,mmio32,0x3ced1000 Signed-off-by: Timur Tabi <timur@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-01-05 05:37:42 +08:00
uap->port.iotype = vendor_sbsa.access_32b ? UPIO_MEM32 : UPIO_MEM;
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
uap->port.irq = platform_get_irq(pdev, 0);
uap->port.ops = &sbsa_uart_pops;
uap->fixed_baud = baudrate;
snprintf(uap->type, sizeof(uap->type), "SBSA");
r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
ret = pl011_setup_port(&pdev->dev, uap, r, portnr);
if (ret)
return ret;
platform_set_drvdata(pdev, uap);
return pl011_register_port(uap);
}
static int sbsa_uart_remove(struct platform_device *pdev)
{
struct uart_amba_port *uap = platform_get_drvdata(pdev);
uart_remove_one_port(&amba_reg, &uap->port);
pl011_unregister_port(uap);
return 0;
}
static const struct of_device_id sbsa_uart_of_match[] = {
{ .compatible = "arm,sbsa-uart", },
{},
};
MODULE_DEVICE_TABLE(of, sbsa_uart_of_match);
drivers: PL011: add ACPI probing for SBSA UART Add the necessary driver boilerplate to let the driver be used when the respective ACPI table is discovered by the ACPI subsystem. [Andre: change table name, add MODULE_DEVICE_TABLE entry and improve commit message] Reviewed-by: Hanjun Guo <hanjun.guo@linaro.org> Signed-off-by: Graeme Gregory <graeme.gregory@linaro.org> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:24 +08:00
static const struct acpi_device_id sbsa_uart_acpi_match[] = {
{ "ARMH0011", 0 },
{},
};
MODULE_DEVICE_TABLE(acpi, sbsa_uart_acpi_match);
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
static struct platform_driver arm_sbsa_uart_platform_driver = {
.probe = sbsa_uart_probe,
.remove = sbsa_uart_remove,
.driver = {
.name = "sbsa-uart",
.of_match_table = of_match_ptr(sbsa_uart_of_match),
drivers: PL011: add ACPI probing for SBSA UART Add the necessary driver boilerplate to let the driver be used when the respective ACPI table is discovered by the ACPI subsystem. [Andre: change table name, add MODULE_DEVICE_TABLE entry and improve commit message] Reviewed-by: Hanjun Guo <hanjun.guo@linaro.org> Signed-off-by: Graeme Gregory <graeme.gregory@linaro.org> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:24 +08:00
.acpi_match_table = ACPI_PTR(sbsa_uart_acpi_match),
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
},
};
ARM: Fix section build warnings for AMBA drivers Found in the Versatile build: WARNING: drivers/built-in.o(.data+0x14c): Section mismatch in reference from the variable pl061_gpio_driver to the (unknown reference) .init.data:(unknown) The variable pl061_gpio_driver references the (unknown reference) __initdata (unknown) WARNING: drivers/built-in.o(.data+0x40f8): Section mismatch in reference from the variable pl011_driver to the (unknown reference) .init.data:(unknown) The variable pl011_driver references the (unknown reference) __initdata (unknown) WARNING: drivers/built-in.o(.data+0x5ab4): Section mismatch in reference from the variable pl031_driver to the (unknown reference) .init.data:(unknown) The variable pl031_driver references the (unknown reference) __initdata (unknown) Basically, amba_id structures must not be __initdata. Also fix: WARNING: drivers/built-in.o(.data+0x138): Section mismatch in reference from the variable pl061_gpio_driver to the function .init.text:pl061_probe() The variable pl061_gpio_driver references the function __init pl061_probe() which is an incorrectly annotated probe function. Fix it to reflect the other AMBA bus probe functions by removing the __init attributation. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-07-27 15:50:16 +08:00
static struct amba_id pl011_ids[] = {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
.id = 0x00041011,
.mask = 0x000fffff,
[ARM] 5541/1: serial/amba-pl011.c: add support for the modified port found in Nomadik The Nomadik 8815 SoC has a slightly modified version of the PL011 block. The patch uses the different ID value as a key to select a vendor structure that is used to keep track of the differences, as suggested by Russell King. Signed-off-by: Alessandro Rubini <rubini@unipv.it> Acked-by: Andrea Gallo <andrea.gallo@stericsson.com> Acked-by: Linus Walleij <linus.walleij@stericsson.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-05 00:43:04 +08:00
.data = &vendor_arm,
},
{
.id = 0x00380802,
.mask = 0x00ffffff,
.data = &vendor_st,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
},
{ 0, 0 },
};
serial: pl011: Enable module alias autogeneration for AMBA drivers Signed-off-by: Dave Martin <dave.martin@linaro.org>
2011-10-05 22:15:22 +08:00
MODULE_DEVICE_TABLE(amba, pl011_ids);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static struct amba_driver pl011_driver = {
.drv = {
.name = "uart-pl011",
serial: pl011: Convert to modern PM ops Convert to modern PM ops and use the SIMPLE_DEV_PM_OPS macro to set up the PM callbacks. Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-12-03 18:04:28 +08:00
.pm = &pl011_dev_pm_ops,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
},
.id_table = pl011_ids,
.probe = pl011_probe,
.remove = pl011_remove,
};
static int __init pl011_init(void)
{
printk(KERN_INFO "Serial: AMBA PL011 UART driver\n");
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
if (platform_driver_register(&arm_sbsa_uart_platform_driver))
pr_warn("could not register SBSA UART platform driver\n");
Revert "uart: pl011: Add support to ZTE ZX296702 uart" This reverts commit 8cd90e50d1408c65c355084b1c7f8f9085f49c6b as with this patch the serial console is broken on lots of platforms. Reported-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jun Nie <jun.nie@linaro.org> Acked-by: Will Deacon <will.deacon@arm.com> Tested-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-09-05 00:11:24 +08:00
return amba_driver_register(&pl011_driver);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
static void __exit pl011_exit(void)
{
drivers: PL011: add support for the ARM SBSA generic UART The ARM Server Base System Architecture[1] document describes a generic UART which is a subset of the PL011 UART. It lacks DMA support, baud rate control and modem status line control, among other things. The idea is to move the UART initialization and setup into the firmware (which does this job today already) and let the kernel just use the UART for sending and receiving characters. We use the recent refactoring to build a new struct uart_ops variable which points to some new functions avoiding access to the missing registers. We reuse as much existing PL011 code as possible. In contrast to the PL011 the SBSA UART does not define any AMBA or PrimeCell relations, so we go with a pretty generic probe function which only uses platform device functions. A DT binding is provided with this patch, ACPI support is added in a separate one. Signed-off-by: Andre Przywara <andre.przywara@arm.com> Tested-by: Mark Langsdorf <mlangsdo@redhat.com> Tested-by: Naresh Bhat <nbhat@cavium.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-05-22 00:26:23 +08:00
platform_driver_unregister(&arm_sbsa_uart_platform_driver);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
amba_driver_unregister(&pl011_driver);
}
[ARM] 5505/1: serial amba-pl011: move to arch_initcall for earlier console Signed-off-by: Alessandro Rubini <rubini@unipv.it>" Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-05 12:54:13 +08:00
/*
* While this can be a module, if builtin it's most likely the console
* So let's leave module_exit but move module_init to an earlier place
*/
arch_initcall(pl011_init);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
module_exit(pl011_exit);
MODULE_AUTHOR("ARM Ltd/Deep Blue Solutions Ltd");
MODULE_DESCRIPTION("ARM AMBA serial port driver");
MODULE_LICENSE("GPL");