According to spec, the pre_div for imx6q should be 1, or the biggest
clock rate we can get is a half of host clock rate. This may cause
we can not get the proper clock rate as we want. e.g. if the desired
clock is 200Mhz, however, the host clock is 200Mhz too, then it causes
the actual clock we get is 100Mhz due to pre_div is 2.
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
Without proper pinctrl state, the card may not be able to work
on high speed stablely. e.g. SDR104.
This patch add pinctrl state switch code according to different
uhs mode include 100mhz sate, 200mhz sate and normal state
(50Mhz and below).
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
Freescale i.MX6Q/DL uSDHC clock tuning progress is a little different from
the standard tuning process defined in host controller spec v3.0.
Thus we use platform_execute_tuning instead of standard sdhci tuning.
The main difference are:
1) not only generate Buffer Read Ready interrupt when tuning is performing.
It generates all other DATA interrupts like the normal data command.
2) SDHCI_CTRL_EXEC_TUNING is not automatically cleared by HW,
instead it's controlled by SW.
3) SDHCI_CTRL_TUNED_CLK is not automatically set by HW,
it's controlled by SW.
4) the clock delay for every tuning is set by SW.
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
The signal voltage switch flow requires to shutdown and output
clock in a specific sequence according to standard host controller
v3.0 spec. In that timing, the card must really receive clock or not.
However, for i.MX6Q, the uSDHC will not output clock even the clock
is enabled until there is command or data in transfer on the bus,
which will then cause singal voltage switch always to fail.
For i.MX6Q, we clear ESDHC_VENDOR_SPEC_FRC_SDCLK_ON bit to let
controller to gate off clock automatically and set that bit
to force clock output if clock is on.
This is required by SD3.0 support.
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
We need a lot of imx6 specific things into common esdhc_set_clock
for support SD3.0 and eMMC DDR mode which is not needed for power pc
platforms, so esdhc_set_clock seems not so common anymore.
Instead of keeping add platform specfics things into this common API,
we choose to move that code into platform driver itself to handle.
This can also exclude the dependency between imx and power pc on this
headfile and is easy for maintain in the future.
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
It helps for platform code to use it send tuning commands.
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
The tuning of some platforms may not follow the standard host control
spec v3.0, e.g. Freescale uSDHC on i.MX6Q/DL.
Add a hook here to allow execute platform specific tuning instead of
standard host controller tuning.
The hook only replaces the tuning process, so it's placed after tuning
checking and before the real tuning process.
Some notes for the tuning hook:
1) it needs handle lock itself if it wants to access host controller
according platform specific implementation.
2) do not need to handle runtime pm since it executes with runtime pm
get already.
Signed-off-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
This avoid duplicated implementation.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
This avoid duplicated implementation and also fixes missing iounmap() and
release_mem_region() calls in sdhci_bcm_kona_remove(). sdhci_pltfm_init()
calls request_mem_region() and ioremap(), thus we need to call the
corresponding iounmap() and release_mem_region() calls in
sdhci_bcm_kona_remove().
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
'dw_mci_socfpga_probe' is used only in this file. Make it static.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
'dw_mci_socfpga_match' is always compiled in.
Hence of_match_ptr is not necessary.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
devm_request_and_ioremap() is deprecated. Use devm_ioremap_resource()
instead.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
In some cases, a NULL pointer dereference happens because data is NULL when
STATE_END_REQUEST case is reached in atmci_tasklet_func.
Cc: <stable@vger.kernel.org> # 3.9+
Signed-off-by: Rodolfo Giometti <giometti@enneenne.com>
Acked-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
When a software timeout occurs, the transfer is not stopped. In DMA case,
it causes DMA channel to be stuck because the transfer is still active
causing following transfers to be queued but not computed.
Cc: <stable@vger.kernel.org> # 3.9+
Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Reported-by: Alexander Morozov <etesial@gmail.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
While speed mode is changed, CMD13 cannot be guaranteed.
According to the spec., it is not recommended to use CMD13
to check the busy completion of the timing change.
If CMD13 is used in this case, CRC error must be ignored.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
In IDMAC mode EVENT_XFER_COMPLETE is set when RI/TI of last descriptor
is done. So if errors are happened in the middle of data transfers,
'dw_mci_stop_dma' during error handing can be called and eventually
prevents this flag to be set. This results in permanent wait for
EVENT_XFER_COMPLETE in 'dw_mci_tasklet_func'. Therefore, if dma
running is stopped forcibly, EVENT_XFER_COMPLETE should be set.
Reported-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
First, compiling warning along with previous change is removed.
[drivers/mmc/host/dw_mmc.c:1890:7: warning: unused variable 'ctrl']
And with the recommendation in manual, IDMAC software reset is followed
by dma-reset of the CTRL register in order to terminate the transfer.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
There are three resets in CTRL register. FIFO reset is especially used
in several points with the same routine. It could be replaced with one
function and the others may be applied similarly if needed. So,
mci_wait_reset() is modified to allow various bit field of reset.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Main change corresponds to dw_mci_command_complete(). And EBE is
divided into read and write. Some minor changes for code readability.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
In error cases, DTO interrupt may or may not be generated depending
on remained data. Stop/Abort command ensures DTO generation for that
situation. Currently if 'stop' field of data is empty, there is no
stop/abort command. So, it could hang waiting DTO. This change
reinforces these cases.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Even if response error is detected in case data command, data transfer
is continued. It means that data can live in FIFO. Current handling
just breaks out the request when seeing the command error. This causes
kernel panic in dw_mci_read_data_pio() [host->data = NULL]. And also,
FIFO should be guaranteed to be empty.
Unable to handle kernel NULL pointer dereference at virtual address 00000018
<...>
[<c02af814>] (dw_mci_read_data_pio+0x68/0x198) from [<c02b04b4>] (dw_mci_interrupt+0x374/0x3a0)
[<c02b04b4>] (dw_mci_interrupt+0x374/0x3a0) from [<c006b094>] (handle_irq_event_percpu+0x50/0x194)
[<c006b094>] (handle_irq_event_percpu+0x50/0x194) from [<c006b214>] (handle_irq_event+0x3c/0x5c)
[<c006b214>] (handle_irq_event+0x3c/0x5c) from [<c006de1c>] (handle_fasteoi_irq+0xa4/0x148)
[<c006de1c>] (handle_fasteoi_irq+0xa4/0x148) from [<c006aa88>] (generic_handle_irq+0x20/0x30)
[<c006aa88>] (generic_handle_irq+0x20/0x30) from [<c000f154>] (handle_IRQ+0x38/0x90)
[<c000f154>] (handle_IRQ+0x38/0x90) from [<c00085bc>] (gic_handle_irq+0x34/0x68)
[<c00085bc>] (gic_handle_irq+0x34/0x68) from [<c0011f40>] (__irq_svc+0x40/0x70)
Exception stack(0xef0b1c00 to 0xef0b1c48)
1c00: 000eb0cf ffffffff 00001300 c01a7738 ef295e10 0000000a c04df298 ef0b1dc0
1c20: ef295ec0 00000000 00000000 00000006 00000000 ef0b1c48 c02b1274 c01a7764
1c40: 20000113 ffffffff
[<c0011f40>] (__irq_svc+0x40/0x70) from [<c01a7764>] (__loop_delay+0x0/0xc)
Code: e1a00005 e0891006 e0662004 e12fff33 (e59a3018)
---[ end trace a7043b9ba9aed1db ]---
Kernel panic - not syncing: Fatal exception in interrupt
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Card Read Threshold should be ensured that the card clock does not stop
in the middle of a block of data being transferred from the card to the
Host. Specially, clock stop is allowed in fast transfer such as HS200
or SDR104 mode. And so, it should be enabled.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
This change helps to choose msize, rx_watermark and tx_watermark
depending on block size for IDMAC mode. For SDIO block size can be
variable, so if these values are set incorrectly, card clock may stop.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Both f_max and f_min will be informed for core layer to request
valid clock rate. But current setting from 'host->bus_hz' may
not represent the max/min frequency properly. Even if host can
actually support high speed than bus_hz, core layer will not
request clock rate over bus_hz. Basically, f_max/f_min can be set
with the values according to spec. And then host will make its best
effort to meet the rate.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
'supports-highspeed' is not one of the quirks but is a capability.
So, it's removed from quirks.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
As host controller can support eMMC's HS200 mode at 1.8V or 1.2V,
these capability will be added.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Exynos's host has divider logic before 'cclk_in' to controller core.
It means that actual clock rate of ciu clock comes from this divider
value. So, source clock should be adjusted along with 'ciu_div' which
indicates the host's divider ratio. Setting clock rate basically fits
the required speed. Specially, 'cclk_in' should have double rate of
target speed in case of DDR 8-bit mode.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Implements variable delay tuning. In this change, exynos host can
determine the correct sampling point for the HS200 and SDR104 speed mode.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
For the speed modes HS200 and SDR104, tuning is needed to determine the
correct sampling point. Actual tuning procedure is provided by specific
host controller driver. This patch defines the tuning command and
tuning data. Additionally, 'struct dw_mci_slot' is moved to header
file to consider the extensive usages in driver.
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Exynos5420 Mobile Storage Host controller has Security Management
Unit (SMU) for channel 0 and channel 1 (mainly for eMMC).
This time, SMU configuration is set for non-encryption mode.
Signed-off-by: Yuvaraj Kumar C D <yuvaraj.cd@samsung.com>
Signed-off-by: Alim Akhtar <alim.akhtar@samsung.com>
Tested-by: Jaehoon Chung <jh80.chung@samsung.com>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Current platform specific private data initialization call
dw_mci_exynos_priv_init() can be used to do platform specific
initialization of SMU and others in future. So the drv_data->init
call has moved to dw_mci_probe().
Signed-off-by: Yuvaraj Kumar C D <yuvaraj.cd@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Tested-by: Jaehoon Chung <jh80.chung@samsung.com>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Currently platform specific private data initialization is done by
dw_mci_socfpga_priv_init and dw_mci_socfpga_parse_dt. As we already have
separate platform specific device tree parser dw_mci_socfpga_parse_dt,
move the dw_mci_socfpga_priv_init code to dw_mci_socfpga_parse_dt.
We can use the dw_mci_socfpga_priv_init to do some actual platform
specific initialization.
Signed-off-by: Yuvaraj Kumar C D <yuvaraj.cd@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Tested-by: Jaehoon Chung <jh80.chung@samsung.com>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Currently platform specific private data initialization is done by
dw_mci_exynos_priv_init and dw_mci_exynos_parse_dt. As we already have
separate platform specific device tree parser dw_mci_exynos_parse_dt,
move the dw_mci_exynos_priv_init code to dw_mci_exynos_parse_dt.
We can use the dw_mci_exynos_priv_init to do some actual platform
specific initialization of SMU and etc.
Signed-off-by: Yuvaraj Kumar C D <yuvaraj.cd@samsung.com>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Tested-by: Jaehoon Chung <jh80.chung@samsung.com>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
The TMOUT register is set to 0xffffffff at probe time but isn't
set after suspend/resume. Add an init of this value.
No problems were observed without this (it will also be set in
__dw_mci_start_request if there is data to send), but it makes the
register dump before and after suspend cleaner.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Reviewed-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Previously the dw_mmc driver would ignore any requests to disable the
card's clock. This doesn't seem like a good thing in general, but had
one extra bad side effect in the following situation:
* mmc core would set clk to 400kHz at boot time while scanning
* mmc core would set clk to 0 since no card, but it would be ignored.
* suspend to ram and resume; clocks in the dw_mmc IP block are now 0
but dw_mmc thinks that they're 400kHz (it ignored the set to 0).
* insert card
* mmc core would set clk to 400kHz which would be considered a no-op.
Note that if there is no card in the slot and we do a suspend/resume
cycle, we _do_ still end up with differences in a dw_mmc register
dump, but the differences are clock related and we've got the clock
disabled both before and after, so this should be OK.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
If the WAKEUP_INT is asserted at wakeup and not cleared, we'll end up
looping around forever. This has been seen to happen on exynos5420
silicon despite the fact that we haven't enabled any wakeup events due
to a silicon errata. It is safe to do on all exynos variants.
Signed-off-by: Doug Anderson <dianders@chromium.org>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
The MMC subsystem handles looking for a card at probe time. Queuing up our
own can race with the rest of the MMC subsystem and cause problems if we
get unlucky with timing. Just remove driver own detection triggering. While
progressing the request from 'mmc_rescan', if 'dw_mci_work_routine_card'
routine is activated, it will cancel the current request. The problem case
is that 'mmc_rescan' is prior to 'dw_mci_work_routine_card' from host own.
Specifically, the following message shows the detection problem in driver's
probing. It would get an err -123 (-ENOMEDIUM) during probe.
[ 4.216595] dwmmc_exynos 12210000.dwmmc1: Using internal DMA controller.
[ 4.395935] dwmmc_exynos 12210000.dwmmc1: Version ID is 250a
[ 4.401948] dwmmc_exynos 12210000.dwmmc1: DW MMC controller at irq 108, 64 bit host data width, 64 deep fifo
[ 4.424430] dwmmc_exynos 12210000.dwmmc1: sdr0 mode (irq=108, width=0)
[ 4.453975] dwmmc_exynos 12210000.dwmmc1: sdr0 mode (irq=108, width=0)
[ 4.459592] mmc_host mmc1: Bus speed (slot 0) = 100000000Hz (slot req 400000Hz, actual 400000HZ div = 125)
[ 4.484258] dwmmc_exynos 12210000.dwmmc1: 1 slots initialized
[ 4.485406] dwmmc_exynos 12210000.dwmmc1: sdr0 mode (irq=108, width=0)
[ 4.487606] dwmmc_exynos 12210000.dwmmc1: sdr0 mode (irq=108, width=0)
[ 4.489794] dwmmc_exynos 12210000.dwmmc1: sdr0 mode (irq=108, width=0)
[ 4.509757] mmc1: error -123 whilst initialising SDIO card
Signed-off-by: Doug Anderson <dianders@chromium.org>
Acked-by: Seungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
afa2c9407f ("sh: ecovec24: Use MMC/SDHI CD and RO GPIO") added
.tmio_flags = TMIO_MMC_USE_GPIO_CD on sh_mobile_sdhi_info, but it needs
<linux/mfd/tmio.h> header. This patch adds it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Reviewed-by: Yusuke Goda <yusuke.goda.sx@renesas.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
cscope says there are no callers for mmc_try_claim_host in the kernel.
No reason to keep it.
Signed-off-by: Grant Grundler <grundler@chromium.org>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
- Fix PV spinlocks triggering jump_label code bug
- Remove extraneous code in the tpm front driver
- Fix ballooning out of pages when non-preemptible
- Fix deadlock when using a 32-bit initial domain with large amount of memory.
- Add xen_nopvpsin parameter to the documentation
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
iQEcBAABAgAGBQJSQvzCAAoJEFjIrFwIi8fJyCIIAMENABapdLhrOiRdQ1Y7T5v1
4bogPDLwpVxHzwo/vnHcNpl35/dUZrC6wQa51Bkoqq0V8o1XmjFy3SY/EBGjEAvw
hh4qxGY0p0NNi6hKrWC8mH9u2TcluZGm1uecabkXUhl9mrAB5oBsfJdbBZ5N69gO
QXXt0j7Xwv1APwH86T0e1Lz+lulhdw2ItXP4osYkEbRYNSaaGnuwsd0Jxcb4DeMk
qhKgP7QMn3C7zDDaapJo1axeYQRBNEtv5M8+0wwMleX4yX1+IBRZeQTsRfMr7RB/
8FhssWiH15xU6Gmzgi/VR8xhTEIbQh5GWsVReGf6pqIYSxGSYTvvyhm0bVRH4JI=
=c+7u
-----END PGP SIGNATURE-----
Merge tag 'stable/for-linus-3.12-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull Xen fixes from Konrad Rzeszutek Wilk:
"Bug-fixes and one update to the kernel-paramters.txt documentation.
- Fix PV spinlocks triggering jump_label code bug
- Remove extraneous code in the tpm front driver
- Fix ballooning out of pages when non-preemptible
- Fix deadlock when using a 32-bit initial domain with large amount
of memory
- Add xen_nopvpsin parameter to the documentation"
* tag 'stable/for-linus-3.12-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/spinlock: Document the xen_nopvspin parameter.
xen/p2m: check MFN is in range before using the m2p table
xen/balloon: don't alloc page while non-preemptible
xen: Do not enable spinlocks before jump_label_init() has executed
tpm: xen-tpmfront: Remove the locality sysfs attribute
tpm: xen-tpmfront: Fix default durations
A few fixes for dm-snapshot, a 32 bit fix for dm-stats, a couple error
handling fixes for dm-multipath. A fix for the thin provisioning target
to not expose non-zero discard limits if discards are disabled.
Lastly, add two DM module parameters which allow users to tune the
emergency memory reserves that DM mainatins per device -- this helps fix
a long-standing issue for dm-multipath. The conservative default
reserve for request-based dm-multipath devices (256) has proven
problematic for users with many multipathed SCSI devices but relatively
little memory. To responsibly select a smaller value users should use
the new nr_bios tracepoint info (via commit 75afb352 "block: Add nr_bios
to block_rq_remap tracepoint") to determine the peak number of bios
their workloads create.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
iQEcBAABAgAGBQJSQMVHAAoJEMUj8QotnQNaOXgIAJS6/XJKMoHfiDJ9M+XD34rZ
Uyr9TEnubX3DKCRBiY23MUcCQn3fx6BjCGv5/c8L4jQFIuLyDi2yatqpwXcbGSJh
G/S/y6u0Axek+ew7TS80OFop4nblW6MoKnoh9/4N55Ofa+1WvKM4ERUGjHGbauyS
TxmLQPToCFPLYRIOZ+imd6hQuIZ1+FFdJFvi7kY9O6Llx2sLD6fWi1iruBd/Da2H
ByMX3biGN45mSpcBzRbSC/FkJ9CRIvT9n82BDPS0o3Tllt8NaVlEDaovB7h4ncc0
bFuT2Z3Q38B9uZ8Lj0bqdGzv3kXMLCkLo6WhWjyUt84hmDPAzRpBwt60jUqWyZs=
=bjVp
-----END PGP SIGNATURE-----
Merge tag 'dm-3.12-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm
Pull device-mapper fixes from Mike Snitzer:
"A few fixes for dm-snapshot, a 32 bit fix for dm-stats, a couple error
handling fixes for dm-multipath. A fix for the thin provisioning
target to not expose non-zero discard limits if discards are disabled.
Lastly, add two DM module parameters which allow users to tune the
emergency memory reserves that DM mainatins per device -- this helps
fix a long-standing issue for dm-multipath. The conservative default
reserve for request-based dm-multipath devices (256) has proven
problematic for users with many multipathed SCSI devices but
relatively little memory. To responsibly select a smaller value users
should use the new nr_bios tracepoint info (via commit 75afb352
"block: Add nr_bios to block_rq_remap tracepoint") to determine the
peak number of bios their workloads create"
* tag 'dm-3.12-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
dm: add reserved_bio_based_ios module parameter
dm: add reserved_rq_based_ios module parameter
dm: lower bio-based mempool reservation
dm thin: do not expose non-zero discard limits if discards disabled
dm mpath: disable WRITE SAME if it fails
dm-snapshot: fix performance degradation due to small hash size
dm snapshot: workaround for a false positive lockdep warning
dm stats: fix possible counter corruption on 32-bit systems
dm mpath: do not fail path on -ENOSPC
Pull MIPS fixes from Ralf Baechle:
- Fix a comment
- A small cleanup the main purpose of which is to work around an
internal compiler error bug in certain Codesource toolchains.
* git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
MIPS: mm: Move some checks out of 'for' loop in DMA operations
MIPS: cpu-features.h: s/MIPS53/MIPS64/
Pull powerpc fixes from Ben Herrenschmidt:
"Here are a few things for -rc2, this time it's all written by me so it
can only be perfect .... right ? :)
So we have the fix to call irq_enter/exit on the irq stack we've been
discussing, plus a cleanup on top to remove an unused (and broken)
stack limit tracking feature (well, make it 32-bit only in fact where
it is used and works properly).
Then we have two things that I wrote over the last couple of days and
made the executive decision to include just because I can (and I'm
sure you won't object .... right ?).
They fix a couple of annoying and long standing "issues":
- We had separate zImages for when booting via Open Firmware vs.
booting via a flat device-tree, while it's trivial to make one that
deals with both
- We wasted a ton of cycles spinning secondary CPUs uselessly at boot
instead of starting them when needed on pseries, thus contributing
significantly to global warming"
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
powerpc/pseries: Do not start secondaries in Open Firmware
powerpc/zImage: make the "OF" wrapper support ePAPR boot
powerpc: Remove ksp_limit on ppc64
powerpc/irq: Run softirqs off the top of the irq stack
Pull x86 fixes from Ingo Molnar:
"An EFI fix and two reboot-quirk fixes"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/reboot: Fix apparent cut-n-paste mistake in Dell reboot workaround
x86/reboot: Add quirk to make Dell C6100 use reboot=pci automatically
x86, efi: Don't map Boot Services on i386
Pull scheduler fixes from Ingo Molnar:
"Three small fixes"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/balancing: Fix cfs_rq->task_h_load calculation
sched/balancing: Fix 'local->avg_load > busiest->avg_load' case in fix_small_imbalance()
sched/balancing: Fix 'local->avg_load > sds->avg_load' case in calculate_imbalance()
Pull perf fixes from Ingo Molnar:
"Assorted standalone fixes"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel: Add model number for Avoton Silvermont
perf: Fix capabilities bitfield compatibility in 'struct perf_event_mmap_page'
perf/x86/intel/uncore: Don't use smp_processor_id() in validate_group()
perf: Update ABI comment
tools lib lk: Uninclude linux/magic.h in debugfs.c
perf tools: Fix old GCC build error in trace-event-parse.c:parse_proc_kallsyms()
perf probe: Fix finder to find lines of given function
perf session: Check for SIGINT in more loops
perf tools: Fix compile with libelf without get_phdrnum
perf tools: Fix buildid cache handling of kallsyms with kcore
perf annotate: Fix objdump line parsing offset validation
perf tools: Fill in new definitions for madvise()/mmap() flags
perf tools: Sharpen the libaudit dependencies test
My old @it.uu.se email address is going away, so update relevant
files to point to my @gmail.com address instead. In sata_promise.c
just delete the address, people can get it from MAINTAINERS.
Signed-off-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The check cpu_needs_post_dma_flush() in mips_dma_sync_sg_for_cpu() and
the check !plat_device_is_coherent() in mips_dma_sync_sg_for_device()
can be moved outside the for loop.
As a side effect, this also avoids a GCC bug that caused kernel compile
to fail with the error:
arch/mips/mm/dma-default.c: In function 'mips_dma_sync_sg_for_cpu':
arch/mips/mm/dma-default.c:316:1: internal compiler error: in add_insn_before, at emit-rtl.c:3852
This gcc failure is seen in Code Sourcery toolchains [e.g. gcc version
4.7.2 (Sourcery CodeBench Lite 2012.09-99)] after commit "MIPS: Optimize
current_cpu_type() for better code."
Signed-off-by: Jayachandran C <jchandra@broadcom.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/5907/
Reviewed-by: Markos Chandras <markos.chandras@imgtec.com>
Tested-by: Markos Chandras <markos.chandras@imgtec.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Which disables in the ticketlock slowpath the Xen PV optimization's.
Useful for diagnosing issues and comparing benchmarks in
over-commit CPU scenarios.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Dong Aisheng