openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/drivers/net/pcmcia/nmclan_cs.c

1587 lines
47 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
/* ----------------------------------------------------------------------------
Linux PCMCIA ethernet adapter driver for the New Media Ethernet LAN.
nmclan_cs.c,v 0.16 1995/07/01 06:42:17 rpao Exp rpao
The Ethernet LAN uses the Advanced Micro Devices (AMD) Am79C940 Media
Access Controller for Ethernet (MACE). It is essentially the Am2150
PCMCIA Ethernet card contained in the Am2150 Demo Kit.
Written by Roger C. Pao <rpao@paonet.org>
Copyright 1995 Roger C. Pao
Linux 2.5 cleanups Copyright Red Hat 2003
This software may be used and distributed according to the terms of
the GNU General Public License.
Ported to Linux 1.3.* network driver environment by
Matti Aarnio <mea@utu.fi>
References
Am2150 Technical Reference Manual, Revision 1.0, August 17, 1993
Am79C940 (MACE) Data Sheet, 1994
Am79C90 (C-LANCE) Data Sheet, 1994
Linux PCMCIA Programmer's Guide v1.17
/usr/src/linux/net/inet/dev.c, Linux kernel 1.2.8
Eric Mears, New Media Corporation
Tom Pollard, New Media Corporation
Dean Siasoyco, New Media Corporation
Ken Lesniak, Silicon Graphics, Inc. <lesniak@boston.sgi.com>
Donald Becker <becker@scyld.com>
David Hinds <dahinds@users.sourceforge.net>
The Linux client driver is based on the 3c589_cs.c client driver by
David Hinds.
The Linux network driver outline is based on the 3c589_cs.c driver,
the 8390.c driver, and the example skeleton.c kernel code, which are
by Donald Becker.
The Am2150 network driver hardware interface code is based on the
OS/9000 driver for the New Media Ethernet LAN by Eric Mears.
Special thanks for testing and help in debugging this driver goes
to Ken Lesniak.
-------------------------------------------------------------------------------
Driver Notes and Issues
-------------------------------------------------------------------------------
1. Developed on a Dell 320SLi
PCMCIA Card Services 2.6.2
Linux dell 1.2.10 #1 Thu Jun 29 20:23:41 PDT 1995 i386
2. rc.pcmcia may require loading pcmcia_core with io_speed=300:
'insmod pcmcia_core.o io_speed=300'.
This will avoid problems with fast systems which causes rx_framecnt
to return random values.
3. If hot extraction does not work for you, use 'ifconfig eth0 down'
before extraction.
4. There is a bad slow-down problem in this driver.
5. Future: Multicast processing. In the meantime, do _not_ compile your
kernel with multicast ip enabled.
-------------------------------------------------------------------------------
History
-------------------------------------------------------------------------------
Log: nmclan_cs.c,v
net: Rationalise email address: Network Specific Parts Clean up the various different email addresses of mine listed in the code to a single current and valid address. As Dave says his network merges for 2.6.28 are now done this seems a good point to send them in where they won't risk disrupting real changes. Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-14 10:01:08 +08:00
* 2.5.75-ac1 2003/07/11 Alan Cox <alan@lxorguk.ukuu.org.uk>
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
* Fixed hang on card eject as we probe it
* Cleaned up to use new style locking.
*
* Revision 0.16 1995/07/01 06:42:17 rpao
* Bug fix: nmclan_reset() called CardServices incorrectly.
*
* Revision 0.15 1995/05/24 08:09:47 rpao
* Re-implement MULTI_TX dev->tbusy handling.
*
* Revision 0.14 1995/05/23 03:19:30 rpao
* Added, in nmclan_config(), "tuple.Attributes = 0;".
* Modified MACE ID check to ignore chip revision level.
* Avoid tx_free_frames race condition between _start_xmit and _interrupt.
*
* Revision 0.13 1995/05/18 05:56:34 rpao
* Statistics changes.
* Bug fix: nmclan_reset did not enable TX and RX: call restore_multicast_list.
* Bug fix: mace_interrupt checks ~MACE_IMR_DEFAULT. Fixes driver lockup.
*
* Revision 0.12 1995/05/14 00:12:23 rpao
* Statistics overhaul.
*
95/05/13 rpao V0.10a
Bug fix: MACE statistics counters used wrong I/O ports.
Bug fix: mace_interrupt() needed to allow statistics to be
processed without RX or TX interrupts pending.
95/05/11 rpao V0.10
Multiple transmit request processing.
Modified statistics to use MACE counters where possible.
95/05/10 rpao V0.09 Bug fix: Must use IO_DATA_PATH_WIDTH_AUTO.
*Released
95/05/10 rpao V0.08
Bug fix: Make all non-exported functions private by using
static keyword.
Bug fix: Test IntrCnt _before_ reading MACE_IR.
95/05/10 rpao V0.07 Statistics.
95/05/09 rpao V0.06 Fix rx_framecnt problem by addition of PCIC wait states.
---------------------------------------------------------------------------- */
#define DRV_NAME "nmclan_cs"
#define DRV_VERSION "0.16"
/* ----------------------------------------------------------------------------
Conditional Compilation Options
---------------------------------------------------------------------------- */
#define MULTI_TX 0
#define RESET_ON_TIMEOUT 1
#define TX_INTERRUPTABLE 1
#define RESET_XILINX 0
/* ----------------------------------------------------------------------------
Include Files
---------------------------------------------------------------------------- */
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/ptrace.h>
#include <linux/slab.h>
#include <linux/string.h>
#include <linux/timer.h>
#include <linux/interrupt.h>
#include <linux/in.h>
#include <linux/delay.h>
#include <linux/ethtool.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <linux/if_arp.h>
#include <linux/ioport.h>
#include <linux/bitops.h>
#include <pcmcia/cs_types.h>
#include <pcmcia/cs.h>
#include <pcmcia/cisreg.h>
#include <pcmcia/cistpl.h>
#include <pcmcia/ds.h>
#include <asm/uaccess.h>
#include <asm/io.h>
#include <asm/system.h>
/* ----------------------------------------------------------------------------
Defines
---------------------------------------------------------------------------- */
#define ETHER_ADDR_LEN ETH_ALEN
/* 6 bytes in an Ethernet Address */
#define MACE_LADRF_LEN 8
/* 8 bytes in Logical Address Filter */
/* Loop Control Defines */
#define MACE_MAX_IR_ITERATIONS 10
#define MACE_MAX_RX_ITERATIONS 12
/*
TBD: Dean brought this up, and I assumed the hardware would
handle it:
If MACE_MAX_RX_ITERATIONS is > 1, rx_framecnt may still be
non-zero when the isr exits. We may not get another interrupt
to process the remaining packets for some time.
*/
/*
The Am2150 has a Xilinx XC3042 field programmable gate array (FPGA)
which manages the interface between the MACE and the PCMCIA bus. It
also includes buffer management for the 32K x 8 SRAM to control up to
four transmit and 12 receive frames at a time.
*/
#define AM2150_MAX_TX_FRAMES 4
#define AM2150_MAX_RX_FRAMES 12
/* Am2150 Ethernet Card I/O Mapping */
#define AM2150_RCV 0x00
#define AM2150_XMT 0x04
#define AM2150_XMT_SKIP 0x09
#define AM2150_RCV_NEXT 0x0A
#define AM2150_RCV_FRAME_COUNT 0x0B
#define AM2150_MACE_BANK 0x0C
#define AM2150_MACE_BASE 0x10
/* MACE Registers */
#define MACE_RCVFIFO 0
#define MACE_XMTFIFO 1
#define MACE_XMTFC 2
#define MACE_XMTFS 3
#define MACE_XMTRC 4
#define MACE_RCVFC 5
#define MACE_RCVFS 6
#define MACE_FIFOFC 7
#define MACE_IR 8
#define MACE_IMR 9
#define MACE_PR 10
#define MACE_BIUCC 11
#define MACE_FIFOCC 12
#define MACE_MACCC 13
#define MACE_PLSCC 14
#define MACE_PHYCC 15
#define MACE_CHIPIDL 16
#define MACE_CHIPIDH 17
#define MACE_IAC 18
/* Reserved */
#define MACE_LADRF 20
#define MACE_PADR 21
/* Reserved */
/* Reserved */
#define MACE_MPC 24
/* Reserved */
#define MACE_RNTPC 26
#define MACE_RCVCC 27
/* Reserved */
#define MACE_UTR 29
#define MACE_RTR1 30
#define MACE_RTR2 31
/* MACE Bit Masks */
#define MACE_XMTRC_EXDEF 0x80
#define MACE_XMTRC_XMTRC 0x0F
#define MACE_XMTFS_XMTSV 0x80
#define MACE_XMTFS_UFLO 0x40
#define MACE_XMTFS_LCOL 0x20
#define MACE_XMTFS_MORE 0x10
#define MACE_XMTFS_ONE 0x08
#define MACE_XMTFS_DEFER 0x04
#define MACE_XMTFS_LCAR 0x02
#define MACE_XMTFS_RTRY 0x01
#define MACE_RCVFS_RCVSTS 0xF000
#define MACE_RCVFS_OFLO 0x8000
#define MACE_RCVFS_CLSN 0x4000
#define MACE_RCVFS_FRAM 0x2000
#define MACE_RCVFS_FCS 0x1000
#define MACE_FIFOFC_RCVFC 0xF0
#define MACE_FIFOFC_XMTFC 0x0F
#define MACE_IR_JAB 0x80
#define MACE_IR_BABL 0x40
#define MACE_IR_CERR 0x20
#define MACE_IR_RCVCCO 0x10
#define MACE_IR_RNTPCO 0x08
#define MACE_IR_MPCO 0x04
#define MACE_IR_RCVINT 0x02
#define MACE_IR_XMTINT 0x01
#define MACE_MACCC_PROM 0x80
#define MACE_MACCC_DXMT2PD 0x40
#define MACE_MACCC_EMBA 0x20
#define MACE_MACCC_RESERVED 0x10
#define MACE_MACCC_DRCVPA 0x08
#define MACE_MACCC_DRCVBC 0x04
#define MACE_MACCC_ENXMT 0x02
#define MACE_MACCC_ENRCV 0x01
#define MACE_PHYCC_LNKFL 0x80
#define MACE_PHYCC_DLNKTST 0x40
#define MACE_PHYCC_REVPOL 0x20
#define MACE_PHYCC_DAPC 0x10
#define MACE_PHYCC_LRT 0x08
#define MACE_PHYCC_ASEL 0x04
#define MACE_PHYCC_RWAKE 0x02
#define MACE_PHYCC_AWAKE 0x01
#define MACE_IAC_ADDRCHG 0x80
#define MACE_IAC_PHYADDR 0x04
#define MACE_IAC_LOGADDR 0x02
#define MACE_UTR_RTRE 0x80
#define MACE_UTR_RTRD 0x40
#define MACE_UTR_RPA 0x20
#define MACE_UTR_FCOLL 0x10
#define MACE_UTR_RCVFCSE 0x08
#define MACE_UTR_LOOP_INCL_MENDEC 0x06
#define MACE_UTR_LOOP_NO_MENDEC 0x04
#define MACE_UTR_LOOP_EXTERNAL 0x02
#define MACE_UTR_LOOP_NONE 0x00
#define MACE_UTR_RESERVED 0x01
/* Switch MACE register bank (only 0 and 1 are valid) */
#define MACEBANK(win_num) outb((win_num), ioaddr + AM2150_MACE_BANK)
#define MACE_IMR_DEFAULT \
(0xFF - \
( \
MACE_IR_CERR | \
MACE_IR_RCVCCO | \
MACE_IR_RNTPCO | \
MACE_IR_MPCO | \
MACE_IR_RCVINT | \
MACE_IR_XMTINT \
) \
)
#undef MACE_IMR_DEFAULT
#define MACE_IMR_DEFAULT 0x00 /* New statistics handling: grab everything */
#define TX_TIMEOUT ((400*HZ)/1000)
/* ----------------------------------------------------------------------------
Type Definitions
---------------------------------------------------------------------------- */
typedef struct _mace_statistics {
/* MACE_XMTFS */
int xmtsv;
int uflo;
int lcol;
int more;
int one;
int defer;
int lcar;
int rtry;
/* MACE_XMTRC */
int exdef;
int xmtrc;
/* RFS1--Receive Status (RCVSTS) */
int oflo;
int clsn;
int fram;
int fcs;
/* RFS2--Runt Packet Count (RNTPC) */
int rfs_rntpc;
/* RFS3--Receive Collision Count (RCVCC) */
int rfs_rcvcc;
/* MACE_IR */
int jab;
int babl;
int cerr;
int rcvcco;
int rntpco;
int mpco;
/* MACE_MPC */
int mpc;
/* MACE_RNTPC */
int rntpc;
/* MACE_RCVCC */
int rcvcc;
} mace_statistics;
typedef struct _mace_private {
[PATCH] pcmcia: embed dev_link_t into struct pcmcia_device Embed dev_link_t into struct pcmcia_device(), as they basically address the same entity. The actual contents of dev_link_t will be cleaned up step by step. This patch includes a bugfix from and signed-off-by Andrew Morton. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-05 17:45:09 +08:00
struct pcmcia_device *p_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
dev_node_t node;
struct net_device_stats linux_stats; /* Linux statistics counters */
mace_statistics mace_stats; /* MACE chip statistics counters */
/* restore_multicast_list() state variables */
int multicast_ladrf[MACE_LADRF_LEN]; /* Logical address filter */
int multicast_num_addrs;
char tx_free_frames; /* Number of free transmit frame buffers */
char tx_irq_disabled; /* MACE TX interrupt disabled */
spinlock_t bank_lock; /* Must be held if you step off bank 0 */
} mace_private;
/* ----------------------------------------------------------------------------
Private Global Variables
---------------------------------------------------------------------------- */
Massive net driver const-ification.
2006-03-04 10:33:57 +08:00
static const char *if_names[]={
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
"Auto", "10baseT", "BNC",
};
/* ----------------------------------------------------------------------------
Parameters
These are the parameters that can be set during loading with
'insmod'.
---------------------------------------------------------------------------- */
MODULE_DESCRIPTION("New Media PCMCIA ethernet driver");
MODULE_LICENSE("GPL");
#define INT_MODULE_PARM(n, v) static int n = v; module_param(n, int, 0)
/* 0=auto, 1=10baseT, 2 = 10base2, default=auto */
INT_MODULE_PARM(if_port, 0);
/* ----------------------------------------------------------------------------
Function Prototypes
---------------------------------------------------------------------------- */
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
static int nmclan_config(struct pcmcia_device *link);
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
static void nmclan_release(struct pcmcia_device *link);
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 nmclan_reset(struct net_device *dev);
static int mace_config(struct net_device *dev, struct ifmap *map);
static int mace_open(struct net_device *dev);
static int mace_close(struct net_device *dev);
netdev: convert pcmcia drivers to netdev_tx_t Update all the pcmcia network drivers for netdev_tx_t. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-01 03:50:49 +08:00
static netdev_tx_t mace_start_xmit(struct sk_buff *skb,
struct net_device *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
static void mace_tx_timeout(struct net_device *dev);
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 mace_interrupt(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
static struct net_device_stats *mace_get_stats(struct net_device *dev);
static int mace_rx(struct net_device *dev, unsigned char RxCnt);
static void restore_multicast_list(struct net_device *dev);
static void set_multicast_list(struct net_device *dev);
drivers/net: const-ify ethtool_ops declarations Signed-off-by: Jeff Garzik <jeff@garzik.org>
2006-09-14 02:30:00 +08:00
static const struct ethtool_ops netdev_ethtool_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
[PATCH] pcmcia: unify detach, REMOVAL_EVENT handlers into one remove callback Unify the "detach" and REMOVAL_EVENT handlers to one "remove" function. Old functionality is preserved, for the moment. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:23:14 +08:00
static void nmclan_detach(struct pcmcia_device *p_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
pcmcia: convert nmclan driver to net_device_ops Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-03-21 03:36:05 +08:00
static const struct net_device_ops mace_netdev_ops = {
.ndo_open = mace_open,
.ndo_stop = mace_close,
.ndo_start_xmit = mace_start_xmit,
.ndo_tx_timeout = mace_tx_timeout,
.ndo_set_config = mace_config,
.ndo_get_stats = mace_get_stats,
.ndo_set_multicast_list = set_multicast_list,
.ndo_change_mtu = eth_change_mtu,
.ndo_set_mac_address = eth_mac_addr,
.ndo_validate_addr = eth_validate_addr,
};
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
/* ----------------------------------------------------------------------------
nmclan_attach
Creates an "instance" of the driver, allocating local data
structures for one device. The device is registered with Card
Services.
---------------------------------------------------------------------------- */
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
static int nmclan_probe(struct pcmcia_device *link)
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
{
mace_private *lp;
struct net_device *dev;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "nmclan_attach()\n");
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
/* Create new ethernet device */
dev = alloc_etherdev(sizeof(mace_private));
if (!dev)
[PATCH] pcmcia: unify attach, EVENT_CARD_INSERTION handlers into one probe callback Unify the EVENT_CARD_INSERTION and "attach" callbacks to one unified probe() callback. As all in-kernel drivers are changed to this new callback, there will be no temporary backwards-compatibility. Inside a probe() function, each driver _must_ set struct pcmcia_device *p_dev->instance and instance->handle correctly. With these patches, the basic driver interface for 16-bit PCMCIA drivers now has the classic four callbacks known also from other buses: int (*probe) (struct pcmcia_device *dev); void (*remove) (struct pcmcia_device *dev); int (*suspend) (struct pcmcia_device *dev); int (*resume) (struct pcmcia_device *dev); Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:25:51 +08:00
return -ENOMEM;
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
lp = netdev_priv(dev);
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
lp->p_dev = link;
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
link->priv = dev;
spin_lock_init(&lp->bank_lock);
link->io.NumPorts1 = 32;
link->io.Attributes1 = IO_DATA_PATH_WIDTH_AUTO;
link->io.IOAddrLines = 5;
link->irq.Attributes = IRQ_TYPE_EXCLUSIVE | IRQ_HANDLE_PRESENT;
link->irq.IRQInfo1 = IRQ_LEVEL_ID;
link->irq.Handler = &mace_interrupt;
link->irq.Instance = dev;
link->conf.Attributes = CONF_ENABLE_IRQ;
link->conf.IntType = INT_MEMORY_AND_IO;
link->conf.ConfigIndex = 1;
link->conf.Present = PRESENT_OPTION;
lp->tx_free_frames=AM2150_MAX_TX_FRAMES;
pcmcia: convert nmclan driver to net_device_ops Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-03-21 03:36:05 +08:00
dev->netdev_ops = &mace_netdev_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
SET_ETHTOOL_OPS(dev, &netdev_ethtool_ops);
dev->watchdog_timeo = TX_TIMEOUT;
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
return nmclan_config(link);
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
} /* nmclan_attach */
/* ----------------------------------------------------------------------------
nmclan_detach
This deletes a driver "instance". The device is de-registered
with Card Services. If it has been released, all local data
structures are freed. Otherwise, the structures will be freed
when the device is released.
---------------------------------------------------------------------------- */
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
static void nmclan_detach(struct pcmcia_device *link)
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 net_device *dev = link->priv;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "nmclan_detach\n");
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
[PATCH] pcmcia: embed dev_link_t into struct pcmcia_device Embed dev_link_t into struct pcmcia_device(), as they basically address the same entity. The actual contents of dev_link_t will be cleaned up step by step. This patch includes a bugfix from and signed-off-by Andrew Morton. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-05 17:45:09 +08:00
if (link->dev_node)
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
unregister_netdev(dev);
[PATCH] pcmcia: use bitfield instead of p_state and state Instead of the two status values struct pcmcia_device->p_state and state, use descriptive bitfields. Most value-checking in drivers was invalid, as the core now only calls the ->remove() (a.k.a. detach) function in case the attachement _and_ configuration was successful. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-02 07:09:29 +08:00
nmclan_release(link);
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
free_netdev(dev);
} /* nmclan_detach */
/* ----------------------------------------------------------------------------
mace_read
Reads a MACE register. This is bank independent; however, the
caller must ensure that this call is not interruptable. We are
assuming that during normal operation, the MACE is always in
bank 0.
---------------------------------------------------------------------------- */
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
static int mace_read(mace_private *lp, unsigned int ioaddr, int 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
{
int data = 0xFF;
unsigned long flags;
switch (reg >> 4) {
case 0: /* register 0-15 */
data = inb(ioaddr + AM2150_MACE_BASE + reg);
break;
case 1: /* register 16-31 */
spin_lock_irqsave(&lp->bank_lock, flags);
MACEBANK(1);
data = inb(ioaddr + AM2150_MACE_BASE + (reg & 0x0F));
MACEBANK(0);
spin_unlock_irqrestore(&lp->bank_lock, flags);
break;
}
return (data & 0xFF);
} /* mace_read */
/* ----------------------------------------------------------------------------
mace_write
Writes to a MACE register. This is bank independent; however,
the caller must ensure that this call is not interruptable. We
are assuming that during normal operation, the MACE is always in
bank 0.
---------------------------------------------------------------------------- */
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
static void mace_write(mace_private *lp, unsigned int ioaddr, int reg,
int data)
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;
switch (reg >> 4) {
case 0: /* register 0-15 */
outb(data & 0xFF, ioaddr + AM2150_MACE_BASE + reg);
break;
case 1: /* register 16-31 */
spin_lock_irqsave(&lp->bank_lock, flags);
MACEBANK(1);
outb(data & 0xFF, ioaddr + AM2150_MACE_BASE + (reg & 0x0F));
MACEBANK(0);
spin_unlock_irqrestore(&lp->bank_lock, flags);
break;
}
} /* mace_write */
/* ----------------------------------------------------------------------------
mace_init
Resets the MACE chip.
---------------------------------------------------------------------------- */
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
static int mace_init(mace_private *lp, unsigned int ioaddr, char *enet_addr)
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 i;
int ct = 0;
/* MACE Software reset */
mace_write(lp, ioaddr, MACE_BIUCC, 1);
while (mace_read(lp, ioaddr, MACE_BIUCC) & 0x01) {
/* Wait for reset bit to be cleared automatically after <= 200ns */;
if(++ct > 500)
{
printk(KERN_ERR "mace: reset failed, card removed ?\n");
return -1;
}
udelay(1);
}
mace_write(lp, ioaddr, MACE_BIUCC, 0);
/* The Am2150 requires that the MACE FIFOs operate in burst mode. */
mace_write(lp, ioaddr, MACE_FIFOCC, 0x0F);
mace_write(lp,ioaddr, MACE_RCVFC, 0); /* Disable Auto Strip Receive */
mace_write(lp, ioaddr, MACE_IMR, 0xFF); /* Disable all interrupts until _open */
/*
* Bit 2-1 PORTSEL[1-0] Port Select.
* 00 AUI/10Base-2
* 01 10Base-T
* 10 DAI Port (reserved in Am2150)
* 11 GPSI
* For this card, only the first two are valid.
* So, PLSCC should be set to
* 0x00 for 10Base-2
* 0x02 for 10Base-T
* Or just set ASEL in PHYCC below!
*/
switch (if_port) {
case 1:
mace_write(lp, ioaddr, MACE_PLSCC, 0x02);
break;
case 2:
mace_write(lp, ioaddr, MACE_PLSCC, 0x00);
break;
default:
mace_write(lp, ioaddr, MACE_PHYCC, /* ASEL */ 4);
/* ASEL Auto Select. When set, the PORTSEL[1-0] bits are overridden,
and the MACE device will automatically select the operating media
interface port. */
break;
}
mace_write(lp, ioaddr, MACE_IAC, MACE_IAC_ADDRCHG | MACE_IAC_PHYADDR);
/* Poll ADDRCHG bit */
ct = 0;
while (mace_read(lp, ioaddr, MACE_IAC) & MACE_IAC_ADDRCHG)
{
if(++ ct > 500)
{
printk(KERN_ERR "mace: ADDRCHG timeout, card removed ?\n");
return -1;
}
}
/* Set PADR register */
for (i = 0; i < ETHER_ADDR_LEN; i++)
mace_write(lp, ioaddr, MACE_PADR, enet_addr[i]);
/* MAC Configuration Control Register should be written last */
/* Let set_multicast_list set this. */
/* mace_write(lp, ioaddr, MACE_MACCC, MACE_MACCC_ENXMT | MACE_MACCC_ENRCV); */
mace_write(lp, ioaddr, MACE_MACCC, 0x00);
return 0;
} /* mace_init */
/* ----------------------------------------------------------------------------
nmclan_config
This routine is scheduled to run after a CARD_INSERTION event
is received, to configure the PCMCIA socket, and to make the
ethernet device available to the system.
---------------------------------------------------------------------------- */
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
static int nmclan_config(struct pcmcia_device *link)
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 net_device *dev = link->priv;
mace_private *lp = netdev_priv(dev);
pcmcia: convert net pcmcia drivers to use new CIS helpers Use the new CIS helpers in net pcmcia drivers, which allows for a few code cleanups. This revision does not remove the phys_addr assignment in 3c589_cs.c -- a bug noted by Komuro <komurojun-mbn@nifty.com> CC: David S. Miller <davem@davemloft.net> CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-19 05:54:24 +08:00
u8 *buf;
size_t len;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
int i, ret;
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr;
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
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "nmclan_config\n");
ret = pcmcia_request_io(link, &link->io);
if (ret)
goto failed;
ret = pcmcia_request_irq(link, &link->irq);
if (ret)
goto failed;
ret = pcmcia_request_configuration(link, &link->conf);
if (ret)
goto failed;
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
dev->irq = link->irq.AssignedIRQ;
dev->base_addr = link->io.BasePort1;
ioaddr = dev->base_addr;
/* Read the ethernet address from the CIS. */
pcmcia: convert net pcmcia drivers to use new CIS helpers Use the new CIS helpers in net pcmcia drivers, which allows for a few code cleanups. This revision does not remove the phys_addr assignment in 3c589_cs.c -- a bug noted by Komuro <komurojun-mbn@nifty.com> CC: David S. Miller <davem@davemloft.net> CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-19 05:54:24 +08:00
len = pcmcia_get_tuple(link, 0x80, &buf);
if (!buf || len < ETHER_ADDR_LEN) {
kfree(buf);
goto failed;
}
memcpy(dev->dev_addr, buf, ETHER_ADDR_LEN);
kfree(buf);
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 configuration by reading the MACE ID. */
{
char sig[2];
sig[0] = mace_read(lp, ioaddr, MACE_CHIPIDL);
sig[1] = mace_read(lp, ioaddr, MACE_CHIPIDH);
if ((sig[0] == 0x40) && ((sig[1] & 0x0F) == 0x09)) {
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "nmclan_cs configured: mace id=%x %x\n",
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
sig[0], sig[1]);
} else {
printk(KERN_NOTICE "nmclan_cs: mace id not found: %x %x should"
" be 0x40 0x?9\n", sig[0], sig[1]);
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
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
}
}
if(mace_init(lp, ioaddr, dev->dev_addr) == -1)
goto failed;
/* The if_port symbol can be set when the module is loaded */
if (if_port <= 2)
dev->if_port = if_port;
else
printk(KERN_NOTICE "nmclan_cs: invalid if_port requested\n");
[PATCH] pcmcia: embed dev_link_t into struct pcmcia_device Embed dev_link_t into struct pcmcia_device(), as they basically address the same entity. The actual contents of dev_link_t will be cleaned up step by step. This patch includes a bugfix from and signed-off-by Andrew Morton. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-05 17:45:09 +08:00
link->dev_node = &lp->node;
pcmcia: remove deprecated handle_to_dev() macro Update remaining users and remove deprecated handle_to_dev() macro CC: Harald Welte <laforge@gnumonks.org> CC: netdev@vger.kernel.org CC: linux-wireless@vger.kernel.org CC: linux-serial@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-11-03 17:27:34 +08:00
SET_NETDEV_DEV(dev, &link->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
i = register_netdev(dev);
if (i != 0) {
printk(KERN_NOTICE "nmclan_cs: register_netdev() failed\n");
[PATCH] pcmcia: embed dev_link_t into struct pcmcia_device Embed dev_link_t into struct pcmcia_device(), as they basically address the same entity. The actual contents of dev_link_t will be cleaned up step by step. This patch includes a bugfix from and signed-off-by Andrew Morton. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-05 17:45:09 +08:00
link->dev_node = 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
goto failed;
}
strcpy(lp->node.dev_name, dev->name);
[NET]: Introduce and use print_mac() and DECLARE_MAC_BUF() This is nicer than the MAC_FMT stuff. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-04 08:59:30 +08:00
printk(KERN_INFO "%s: nmclan: port %#3lx, irq %d, %s port,"
net: convert print_mac to %pM This converts pretty much everything to print_mac. There were a few things that had conflicts which I have just dropped for now, no harm done. I've built an allyesconfig with this and looked at the files that weren't built very carefully, but it's a huge patch. Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-28 06:59:26 +08:00
" hw_addr %pM\n",
[NET]: Introduce and use print_mac() and DECLARE_MAC_BUF() This is nicer than the MAC_FMT stuff. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-04 08:59:30 +08:00
dev->name, dev->base_addr, dev->irq, if_names[dev->if_port],
net: convert print_mac to %pM This converts pretty much everything to print_mac. There were a few things that had conflicts which I have just dropped for now, no harm done. I've built an allyesconfig with this and looked at the files that weren't built very carefully, but it's a huge patch. Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-10-28 06:59:26 +08:00
dev->dev_addr);
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
return 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
failed:
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
nmclan_release(link);
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
} /* nmclan_config */
/* ----------------------------------------------------------------------------
nmclan_release
After a card is removed, nmclan_release() will unregister the
net device, and release the PCMCIA configuration. If the device
is still open, this will be postponed until it is closed.
---------------------------------------------------------------------------- */
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
static void nmclan_release(struct pcmcia_device *link)
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
{
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "nmclan_release\n");
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
pcmcia_disable_device(link);
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
}
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
static int nmclan_suspend(struct pcmcia_device *link)
[PATCH] pcmcia: new suspend core Move the suspend and resume methods out of the event handler, and into special functions. Also use these functions for pre- and post-reset, as almost all drivers already do, and the remaining ones can easily be converted. Bugfix to include/pcmcia/ds.c Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:21:18 +08:00
{
struct net_device *dev = link->priv;
[PATCH] pcmcia: use bitfield instead of p_state and state Instead of the two status values struct pcmcia_device->p_state and state, use descriptive bitfields. Most value-checking in drivers was invalid, as the core now only calls the ->remove() (a.k.a. detach) function in case the attachement _and_ configuration was successful. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-02 07:09:29 +08:00
if (link->open)
[PATCH] pcmcia: default suspend and resume handling In all but one case, the suspend and resume functions of PCMCIA drivers contain mostly of calls to pcmcia_release_configuration() and pcmcia_request_configuration(). Therefore, move this code out of the drivers and into the core. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-02 07:02:33 +08:00
netif_device_detach(dev);
[PATCH] pcmcia: new suspend core Move the suspend and resume methods out of the event handler, and into special functions. Also use these functions for pre- and post-reset, as almost all drivers already do, and the remaining ones can easily be converted. Bugfix to include/pcmcia/ds.c Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:21:18 +08:00
return 0;
}
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
static int nmclan_resume(struct pcmcia_device *link)
[PATCH] pcmcia: new suspend core Move the suspend and resume methods out of the event handler, and into special functions. Also use these functions for pre- and post-reset, as almost all drivers already do, and the remaining ones can easily be converted. Bugfix to include/pcmcia/ds.c Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:21:18 +08:00
{
struct net_device *dev = link->priv;
[PATCH] pcmcia: use bitfield instead of p_state and state Instead of the two status values struct pcmcia_device->p_state and state, use descriptive bitfields. Most value-checking in drivers was invalid, as the core now only calls the ->remove() (a.k.a. detach) function in case the attachement _and_ configuration was successful. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-02 07:09:29 +08:00
if (link->open) {
[PATCH] pcmcia: default suspend and resume handling In all but one case, the suspend and resume functions of PCMCIA drivers contain mostly of calls to pcmcia_release_configuration() and pcmcia_request_configuration(). Therefore, move this code out of the drivers and into the core. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-02 07:02:33 +08:00
nmclan_reset(dev);
netif_device_attach(dev);
[PATCH] pcmcia: new suspend core Move the suspend and resume methods out of the event handler, and into special functions. Also use these functions for pre- and post-reset, as almost all drivers already do, and the remaining ones can easily be converted. Bugfix to include/pcmcia/ds.c Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:21:18 +08:00
}
return 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
/* ----------------------------------------------------------------------------
nmclan_reset
Reset and restore all of the Xilinx and MACE registers.
---------------------------------------------------------------------------- */
static void nmclan_reset(struct net_device *dev)
{
mace_private *lp = netdev_priv(dev);
#if RESET_XILINX
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
struct pcmcia_device *link = &lp->link;
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
conf_reg_t reg;
u_long OrigCorValue;
/* Save original COR value */
reg.Function = 0;
reg.Action = CS_READ;
reg.Offset = CISREG_COR;
reg.Value = 0;
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
pcmcia_access_configuration_register(link, &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
OrigCorValue = reg.Value;
/* Reset Xilinx */
reg.Action = CS_WRITE;
reg.Offset = CISREG_COR;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "nmclan_reset: OrigCorValue=0x%lX, resetting...\n",
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
OrigCorValue);
reg.Value = COR_SOFT_RESET;
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
pcmcia_access_configuration_register(link, &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
/* Need to wait for 20 ms for PCMCIA to finish reset. */
/* Restore original COR configuration index */
reg.Value = COR_LEVEL_REQ | (OrigCorValue & COR_CONFIG_MASK);
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
pcmcia_access_configuration_register(link, &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
/* Xilinx is now completely reset along with the MACE chip. */
lp->tx_free_frames=AM2150_MAX_TX_FRAMES;
#endif /* #if RESET_XILINX */
/* Xilinx is now completely reset along with the MACE chip. */
lp->tx_free_frames=AM2150_MAX_TX_FRAMES;
/* Reinitialize the MACE chip for operation. */
mace_init(lp, dev->base_addr, dev->dev_addr);
mace_write(lp, dev->base_addr, MACE_IMR, MACE_IMR_DEFAULT);
/* Restore the multicast list and enable TX and RX. */
restore_multicast_list(dev);
} /* nmclan_reset */
/* ----------------------------------------------------------------------------
mace_config
[Someone tell me what this is supposed to do? Is if_port a defined
standard? If so, there should be defines to indicate 1=10Base-T,
2=10Base-2, etc. including limited automatic detection.]
---------------------------------------------------------------------------- */
static int mace_config(struct net_device *dev, struct ifmap *map)
{
if ((map->port != (u_char)(-1)) && (map->port != dev->if_port)) {
if (map->port <= 2) {
dev->if_port = map->port;
printk(KERN_INFO "%s: switched to %s port\n", dev->name,
if_names[dev->if_port]);
} else
return -EINVAL;
}
return 0;
} /* mace_config */
/* ----------------------------------------------------------------------------
mace_open
Open device driver.
---------------------------------------------------------------------------- */
static int mace_open(struct net_device *dev)
{
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr = dev->base_addr;
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
mace_private *lp = netdev_priv(dev);
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
struct pcmcia_device *link = lp->p_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
[PATCH] pcmcia: convert DEV_OK to pcmcia_dev_present Instead of the DEV_OK macro, drivers should use pcmcia_dev_present(). Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-05 18:04:33 +08:00
if (!pcmcia_dev_present(link))
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 -ENODEV;
link->open++;
MACEBANK(0);
netif_start_queue(dev);
nmclan_reset(dev);
return 0; /* Always succeed */
} /* mace_open */
/* ----------------------------------------------------------------------------
mace_close
Closes device driver.
---------------------------------------------------------------------------- */
static int mace_close(struct net_device *dev)
{
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr = dev->base_addr;
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
mace_private *lp = netdev_priv(dev);
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
struct pcmcia_device *link = lp->p_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
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
dev_dbg(&link->dev, "%s: shutting down ethercard.\n", dev->name);
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
/* Mask off all interrupts from the MACE chip. */
outb(0xFF, ioaddr + AM2150_MACE_BASE + MACE_IMR);
link->open--;
netif_stop_queue(dev);
return 0;
} /* mace_close */
static void netdev_get_drvinfo(struct net_device *dev,
struct ethtool_drvinfo *info)
{
strcpy(info->driver, DRV_NAME);
strcpy(info->version, DRV_VERSION);
sprintf(info->bus_info, "PCMCIA 0x%lx", dev->base_addr);
}
drivers/net: const-ify ethtool_ops declarations Signed-off-by: Jeff Garzik <jeff@garzik.org>
2006-09-14 02:30:00 +08:00
static const struct ethtool_ops netdev_ethtool_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
.get_drvinfo = netdev_get_drvinfo,
};
/* ----------------------------------------------------------------------------
mace_start_xmit
This routine begins the packet transmit function. When completed,
it will generate a transmit interrupt.
According to /usr/src/linux/net/inet/dev.c, if _start_xmit
returns 0, the "packet is now solely the responsibility of the
driver." If _start_xmit returns non-zero, the "transmission
failed, put skb back into a list."
---------------------------------------------------------------------------- */
static void mace_tx_timeout(struct net_device *dev)
{
mace_private *lp = netdev_priv(dev);
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection dev_link_t * and client_handle_t both mean struct pcmcai_device * by now. Therefore, remove all such indirections. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:21:06 +08:00
struct pcmcia_device *link = lp->p_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
printk(KERN_NOTICE "%s: transmit timed out -- ", dev->name);
#if RESET_ON_TIMEOUT
printk("resetting card\n");
pcmcia: card services header cleanup 16-bit PCMCIA device handling function definitions are moved to ds.h, internal definitions to cs_internal.h. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2008-08-31 21:20:26 +08:00
pcmcia_reset_card(link->socket);
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 /* #if RESET_ON_TIMEOUT */
printk("NOT resetting card\n");
#endif /* #if RESET_ON_TIMEOUT */
dev->trans_start = jiffies;
netif_wake_queue(dev);
}
netdev: convert pcmcia drivers to netdev_tx_t Update all the pcmcia network drivers for netdev_tx_t. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-09-01 03:50:49 +08:00
static netdev_tx_t mace_start_xmit(struct sk_buff *skb,
struct net_device *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
{
mace_private *lp = netdev_priv(dev);
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr = dev->base_addr;
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
netif_stop_queue(dev);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: mace_start_xmit(length = %ld) called.\n",
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
dev->name, (long)skb->len);
#if (!TX_INTERRUPTABLE)
/* Disable MACE TX interrupts. */
outb(MACE_IMR_DEFAULT | MACE_IR_XMTINT,
ioaddr + AM2150_MACE_BASE + MACE_IMR);
lp->tx_irq_disabled=1;
#endif /* #if (!TX_INTERRUPTABLE) */
{
/* This block must not be interrupted by another transmit request!
mace_tx_timeout will take care of timer-based retransmissions from
the upper layers. The interrupt handler is guaranteed never to
service a transmit interrupt while we are in here.
*/
lp->linux_stats.tx_bytes += skb->len;
lp->tx_free_frames--;
/* WARNING: Write the _exact_ number of bytes written in the header! */
/* Put out the word header [must be an outw()] . . . */
outw(skb->len, ioaddr + AM2150_XMT);
/* . . . and the packet [may be any combination of outw() and outb()] */
outsw(ioaddr + AM2150_XMT, skb->data, skb->len >> 1);
if (skb->len & 1) {
/* Odd byte transfer */
outb(skb->data[skb->len-1], ioaddr + AM2150_XMT);
}
dev->trans_start = jiffies;
#if MULTI_TX
if (lp->tx_free_frames > 0)
netif_start_queue(dev);
#endif /* #if MULTI_TX */
}
#if (!TX_INTERRUPTABLE)
/* Re-enable MACE TX interrupts. */
lp->tx_irq_disabled=0;
outb(MACE_IMR_DEFAULT, ioaddr + AM2150_MACE_BASE + MACE_IMR);
#endif /* #if (!TX_INTERRUPTABLE) */
dev_kfree_skb(skb);
net: use NETDEV_TX_OK instead of 0 in ndo_start_xmit() functions This patch is the result of an automatic spatch transformation to convert all ndo_start_xmit() return values of 0 to NETDEV_TX_OK. Some occurences are missed by the automatic conversion, those will be handled in a seperate patch. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2009-06-23 14:03:08 +08:00
return NETDEV_TX_OK;
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
} /* mace_start_xmit */
/* ----------------------------------------------------------------------------
mace_interrupt
The interrupt handler.
---------------------------------------------------------------------------- */
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 mace_interrupt(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 net_device *dev = (struct net_device *) dev_id;
mace_private *lp = netdev_priv(dev);
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr;
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 status;
int IntrCnt = MACE_MAX_IR_ITERATIONS;
if (dev == NULL) {
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("mace_interrupt(): irq 0x%X for unknown device.\n",
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);
return IRQ_NONE;
}
Fix a potential NULL pointer dereference in mace_interrupt() in drivers/net/pcmcia/nmclan_cs.c This patch fixes a potential null dereference bug where we dereference DEV before a null check. This patch simply moves the dereferencing after the null check. Signed-off-by: Micah Gruber <micah.gruber@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-07-24 10:44:56 +08:00
ioaddr = dev->base_addr;
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 (lp->tx_irq_disabled) {
printk(
(lp->tx_irq_disabled?
KERN_NOTICE "%s: Interrupt with tx_irq_disabled "
"[isr=%02X, imr=%02X]\n":
KERN_NOTICE "%s: Re-entering the interrupt handler "
"[isr=%02X, imr=%02X]\n"),
dev->name,
inb(ioaddr + AM2150_MACE_BASE + MACE_IR),
inb(ioaddr + AM2150_MACE_BASE + MACE_IMR)
);
/* WARNING: MACE_IR has been read! */
return IRQ_NONE;
}
if (!netif_device_present(dev)) {
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: interrupt from dead card\n", dev->name);
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_NONE;
}
do {
/* WARNING: MACE_IR is a READ/CLEAR port! */
status = inb(ioaddr + AM2150_MACE_BASE + MACE_IR);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("mace_interrupt: irq 0x%X status 0x%X.\n", irq, status);
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 & MACE_IR_RCVINT) {
mace_rx(dev, MACE_MAX_RX_ITERATIONS);
}
if (status & MACE_IR_XMTINT) {
unsigned char fifofc;
unsigned char xmtrc;
unsigned char xmtfs;
fifofc = inb(ioaddr + AM2150_MACE_BASE + MACE_FIFOFC);
if ((fifofc & MACE_FIFOFC_XMTFC)==0) {
lp->linux_stats.tx_errors++;
outb(0xFF, ioaddr + AM2150_XMT_SKIP);
}
/* Transmit Retry Count (XMTRC, reg 4) */
xmtrc = inb(ioaddr + AM2150_MACE_BASE + MACE_XMTRC);
if (xmtrc & MACE_XMTRC_EXDEF) lp->mace_stats.exdef++;
lp->mace_stats.xmtrc += (xmtrc & MACE_XMTRC_XMTRC);
if (
(xmtfs = inb(ioaddr + AM2150_MACE_BASE + MACE_XMTFS)) &
MACE_XMTFS_XMTSV /* Transmit Status Valid */
) {
lp->mace_stats.xmtsv++;
if (xmtfs & ~MACE_XMTFS_XMTSV) {
if (xmtfs & MACE_XMTFS_UFLO) {
/* Underflow. Indicates that the Transmit FIFO emptied before
the end of frame was reached. */
lp->mace_stats.uflo++;
}
if (xmtfs & MACE_XMTFS_LCOL) {
/* Late Collision */
lp->mace_stats.lcol++;
}
if (xmtfs & MACE_XMTFS_MORE) {
/* MORE than one retry was needed */
lp->mace_stats.more++;
}
if (xmtfs & MACE_XMTFS_ONE) {
/* Exactly ONE retry occurred */
lp->mace_stats.one++;
}
if (xmtfs & MACE_XMTFS_DEFER) {
/* Transmission was defered */
lp->mace_stats.defer++;
}
if (xmtfs & MACE_XMTFS_LCAR) {
/* Loss of carrier */
lp->mace_stats.lcar++;
}
if (xmtfs & MACE_XMTFS_RTRY) {
/* Retry error: transmit aborted after 16 attempts */
lp->mace_stats.rtry++;
}
} /* if (xmtfs & ~MACE_XMTFS_XMTSV) */
} /* if (xmtfs & MACE_XMTFS_XMTSV) */
lp->linux_stats.tx_packets++;
lp->tx_free_frames++;
netif_wake_queue(dev);
} /* if (status & MACE_IR_XMTINT) */
if (status & ~MACE_IMR_DEFAULT & ~MACE_IR_RCVINT & ~MACE_IR_XMTINT) {
if (status & MACE_IR_JAB) {
/* Jabber Error. Excessive transmit duration (20-150ms). */
lp->mace_stats.jab++;
}
if (status & MACE_IR_BABL) {
/* Babble Error. >1518 bytes transmitted. */
lp->mace_stats.babl++;
}
if (status & MACE_IR_CERR) {
/* Collision Error. CERR indicates the absence of the
Signal Quality Error Test message after a packet
transmission. */
lp->mace_stats.cerr++;
}
if (status & MACE_IR_RCVCCO) {
/* Receive Collision Count Overflow; */
lp->mace_stats.rcvcco++;
}
if (status & MACE_IR_RNTPCO) {
/* Runt Packet Count Overflow */
lp->mace_stats.rntpco++;
}
if (status & MACE_IR_MPCO) {
/* Missed Packet Count Overflow */
lp->mace_stats.mpco++;
}
} /* if (status & ~MACE_IMR_DEFAULT & ~MACE_IR_RCVINT & ~MACE_IR_XMTINT) */
} while ((status & ~MACE_IMR_DEFAULT) && (--IntrCnt));
return IRQ_HANDLED;
} /* mace_interrupt */
/* ----------------------------------------------------------------------------
mace_rx
Receives packets.
---------------------------------------------------------------------------- */
static int mace_rx(struct net_device *dev, unsigned char RxCnt)
{
mace_private *lp = netdev_priv(dev);
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr = dev->base_addr;
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 char rx_framecnt;
unsigned short rx_status;
while (
((rx_framecnt = inb(ioaddr + AM2150_RCV_FRAME_COUNT)) > 0) &&
(rx_framecnt <= 12) && /* rx_framecnt==0xFF if card is extracted. */
(RxCnt--)
) {
rx_status = inw(ioaddr + AM2150_RCV);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: in mace_rx(), framecnt 0x%X, rx_status"
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
" 0x%X.\n", dev->name, rx_framecnt, rx_status);
if (rx_status & MACE_RCVFS_RCVSTS) { /* Error, update stats. */
lp->linux_stats.rx_errors++;
if (rx_status & MACE_RCVFS_OFLO) {
lp->mace_stats.oflo++;
}
if (rx_status & MACE_RCVFS_CLSN) {
lp->mace_stats.clsn++;
}
if (rx_status & MACE_RCVFS_FRAM) {
lp->mace_stats.fram++;
}
if (rx_status & MACE_RCVFS_FCS) {
lp->mace_stats.fcs++;
}
} else {
short pkt_len = (rx_status & ~MACE_RCVFS_RCVSTS) - 4;
/* Auto Strip is off, always subtract 4 */
struct sk_buff *skb;
lp->mace_stats.rfs_rntpc += inb(ioaddr + AM2150_RCV);
/* runt packet count */
lp->mace_stats.rfs_rcvcc += inb(ioaddr + AM2150_RCV);
/* rcv collision count */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" receiving packet size 0x%X rx_status"
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
" 0x%X.\n", pkt_len, rx_status);
skb = dev_alloc_skb(pkt_len+2);
if (skb != NULL) {
skb_reserve(skb, 2);
insw(ioaddr + AM2150_RCV, skb_put(skb, pkt_len), pkt_len>>1);
if (pkt_len & 1)
[SK_BUFF]: Convert skb->tail to sk_buff_data_t So that it is also an offset from skb->head, reduces its size from 8 to 4 bytes on 64bit architectures, allowing us to combine the 4 bytes hole left by the layer headers conversion, reducing struct sk_buff size to 256 bytes, i.e. 4 64byte cachelines, and since the sk_buff slab cache is SLAB_HWCACHE_ALIGN... :-) Many calculations that previously required that skb->{transport,network, mac}_header be first converted to a pointer now can be done directly, being meaningful as offsets or pointers. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-04-20 11:29:13 +08:00
*(skb_tail_pointer(skb) - 1) = inb(ioaddr + AM2150_RCV);
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
skb->protocol = eth_type_trans(skb, dev);
netif_rx(skb); /* Send the packet to the upper (protocol) layers. */
lp->linux_stats.rx_packets++;
[PATCH] nmclan_cs: dereferencing skb after netif_rx() From: Florin Malita <fmalita@gmail.com> The skb may be gone after netif_rx(), we can't use 'skb->len' to update the stats. 'pkt_len' should work instead. Coverity CID: 911. Signed-off-by: Florin Malita <fmalita@gmail.com> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Acked-by: Jeff Garzik <jeff@garzik.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-04 17:51:26 +08:00
lp->linux_stats.rx_bytes += pkt_len;
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
outb(0xFF, ioaddr + AM2150_RCV_NEXT); /* skip to next frame */
continue;
} else {
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: couldn't allocate a sk_buff of size"
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
" %d.\n", dev->name, pkt_len);
lp->linux_stats.rx_dropped++;
}
}
outb(0xFF, ioaddr + AM2150_RCV_NEXT); /* skip to next frame */
} /* while */
return 0;
} /* mace_rx */
/* ----------------------------------------------------------------------------
pr_linux_stats
---------------------------------------------------------------------------- */
static void pr_linux_stats(struct net_device_stats *pstats)
{
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("pr_linux_stats\n");
pr_debug(" rx_packets=%-7ld tx_packets=%ld\n",
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
(long)pstats->rx_packets, (long)pstats->tx_packets);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rx_errors=%-7ld tx_errors=%ld\n",
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
(long)pstats->rx_errors, (long)pstats->tx_errors);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rx_dropped=%-7ld tx_dropped=%ld\n",
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
(long)pstats->rx_dropped, (long)pstats->tx_dropped);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" multicast=%-7ld collisions=%ld\n",
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
(long)pstats->multicast, (long)pstats->collisions);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rx_length_errors=%-7ld rx_over_errors=%ld\n",
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
(long)pstats->rx_length_errors, (long)pstats->rx_over_errors);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rx_crc_errors=%-7ld rx_frame_errors=%ld\n",
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
(long)pstats->rx_crc_errors, (long)pstats->rx_frame_errors);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rx_fifo_errors=%-7ld rx_missed_errors=%ld\n",
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
(long)pstats->rx_fifo_errors, (long)pstats->rx_missed_errors);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" tx_aborted_errors=%-7ld tx_carrier_errors=%ld\n",
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
(long)pstats->tx_aborted_errors, (long)pstats->tx_carrier_errors);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" tx_fifo_errors=%-7ld tx_heartbeat_errors=%ld\n",
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
(long)pstats->tx_fifo_errors, (long)pstats->tx_heartbeat_errors);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" tx_window_errors=%ld\n",
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
(long)pstats->tx_window_errors);
} /* pr_linux_stats */
/* ----------------------------------------------------------------------------
pr_mace_stats
---------------------------------------------------------------------------- */
static void pr_mace_stats(mace_statistics *pstats)
{
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("pr_mace_stats\n");
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
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" xmtsv=%-7d uflo=%d\n",
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
pstats->xmtsv, pstats->uflo);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" lcol=%-7d more=%d\n",
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
pstats->lcol, pstats->more);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" one=%-7d defer=%d\n",
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
pstats->one, pstats->defer);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" lcar=%-7d rtry=%d\n",
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
pstats->lcar, pstats->rtry);
/* MACE_XMTRC */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" exdef=%-7d xmtrc=%d\n",
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
pstats->exdef, pstats->xmtrc);
/* RFS1--Receive Status (RCVSTS) */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" oflo=%-7d clsn=%d\n",
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
pstats->oflo, pstats->clsn);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" fram=%-7d fcs=%d\n",
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
pstats->fram, pstats->fcs);
/* RFS2--Runt Packet Count (RNTPC) */
/* RFS3--Receive Collision Count (RCVCC) */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rfs_rntpc=%-7d rfs_rcvcc=%d\n",
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
pstats->rfs_rntpc, pstats->rfs_rcvcc);
/* MACE_IR */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" jab=%-7d babl=%d\n",
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
pstats->jab, pstats->babl);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" cerr=%-7d rcvcco=%d\n",
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
pstats->cerr, pstats->rcvcco);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rntpco=%-7d mpco=%d\n",
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
pstats->rntpco, pstats->mpco);
/* MACE_MPC */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" mpc=%d\n", pstats->mpc);
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
/* MACE_RNTPC */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rntpc=%d\n", pstats->rntpc);
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
/* MACE_RCVCC */
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug(" rcvcc=%d\n", pstats->rcvcc);
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
} /* pr_mace_stats */
/* ----------------------------------------------------------------------------
update_stats
Update statistics. We change to register window 1, so this
should be run single-threaded if the device is active. This is
expected to be a rare operation, and it's simpler for the rest
of the driver to assume that window 0 is always valid rather
than use a special window-state variable.
oflo & uflo should _never_ occur since it would mean the Xilinx
was not able to transfer data between the MACE FIFO and the
card's SRAM fast enough. If this happens, something is
seriously wrong with the hardware.
---------------------------------------------------------------------------- */
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
static void update_stats(unsigned int ioaddr, struct net_device *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
{
mace_private *lp = netdev_priv(dev);
lp->mace_stats.rcvcc += mace_read(lp, ioaddr, MACE_RCVCC);
lp->mace_stats.rntpc += mace_read(lp, ioaddr, MACE_RNTPC);
lp->mace_stats.mpc += mace_read(lp, ioaddr, MACE_MPC);
/* At this point, mace_stats is fully updated for this call.
We may now update the linux_stats. */
/* The MACE has no equivalent for linux_stats field which are commented
out. */
/* lp->linux_stats.multicast; */
lp->linux_stats.collisions =
lp->mace_stats.rcvcco * 256 + lp->mace_stats.rcvcc;
/* Collision: The MACE may retry sending a packet 15 times
before giving up. The retry count is in XMTRC.
Does each retry constitute a collision?
If so, why doesn't the RCVCC record these collisions? */
/* detailed rx_errors: */
lp->linux_stats.rx_length_errors =
lp->mace_stats.rntpco * 256 + lp->mace_stats.rntpc;
/* lp->linux_stats.rx_over_errors */
lp->linux_stats.rx_crc_errors = lp->mace_stats.fcs;
lp->linux_stats.rx_frame_errors = lp->mace_stats.fram;
lp->linux_stats.rx_fifo_errors = lp->mace_stats.oflo;
lp->linux_stats.rx_missed_errors =
lp->mace_stats.mpco * 256 + lp->mace_stats.mpc;
/* detailed tx_errors */
lp->linux_stats.tx_aborted_errors = lp->mace_stats.rtry;
lp->linux_stats.tx_carrier_errors = lp->mace_stats.lcar;
/* LCAR usually results from bad cabling. */
lp->linux_stats.tx_fifo_errors = lp->mace_stats.uflo;
lp->linux_stats.tx_heartbeat_errors = lp->mace_stats.cerr;
/* lp->linux_stats.tx_window_errors; */
return;
} /* update_stats */
/* ----------------------------------------------------------------------------
mace_get_stats
Gathers ethernet statistics from the MACE chip.
---------------------------------------------------------------------------- */
static struct net_device_stats *mace_get_stats(struct net_device *dev)
{
mace_private *lp = netdev_priv(dev);
update_stats(dev->base_addr, dev);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: updating the statistics.\n", dev->name);
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
pr_linux_stats(&lp->linux_stats);
pr_mace_stats(&lp->mace_stats);
return &lp->linux_stats;
} /* net_device_stats */
/* ----------------------------------------------------------------------------
updateCRC
Modified from Am79C90 data sheet.
---------------------------------------------------------------------------- */
#ifdef BROKEN_MULTICAST
static void updateCRC(int *CRC, int bit)
{
int poly[]={
1,1,1,0, 1,1,0,1,
1,0,1,1, 1,0,0,0,
1,0,0,0, 0,0,1,1,
0,0,1,0, 0,0,0,0
}; /* CRC polynomial. poly[n] = coefficient of the x**n term of the
CRC generator polynomial. */
int j;
/* shift CRC and control bit (CRC[32]) */
for (j = 32; j > 0; j--)
CRC[j] = CRC[j-1];
CRC[0] = 0;
/* If bit XOR(control bit) = 1, set CRC = CRC XOR polynomial. */
if (bit ^ CRC[32])
for (j = 0; j < 32; j++)
CRC[j] ^= poly[j];
} /* updateCRC */
/* ----------------------------------------------------------------------------
BuildLAF
Build logical address filter.
Modified from Am79C90 data sheet.
Input
ladrf: logical address filter (contents initialized to 0)
adr: ethernet address
---------------------------------------------------------------------------- */
static void BuildLAF(int *ladrf, int *adr)
{
int CRC[33]={1}; /* CRC register, 1 word/bit + extra control bit */
int i, byte; /* temporary array indices */
int hashcode; /* the output object */
CRC[32]=0;
for (byte = 0; byte < 6; byte++)
for (i = 0; i < 8; i++)
updateCRC(CRC, (adr[byte] >> i) & 1);
hashcode = 0;
for (i = 0; i < 6; i++)
hashcode = (hashcode << 1) + CRC[i];
byte = hashcode >> 3;
ladrf[byte] |= (1 << (hashcode & 7));
#ifdef PCMCIA_DEBUG
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
if (0)
Remove multiple KERN_ prefixes from printk formats Commit 5fd29d6ccbc98884569d6f3105aeca70858b3e0f ("printk: clean up handling of log-levels and newlines") changed printk semantics. printk lines with multiple KERN_<level> prefixes are no longer emitted as before the patch. <level> is now included in the output on each additional use. Remove all uses of multiple KERN_<level>s in formats. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-07-07 04:05:40 +08:00
printk(KERN_DEBUG " adr =%pM\n", adr);
printk(KERN_DEBUG " hashcode = %d(decimal), ladrf[0:63] =", hashcode);
for (i = 0; i < 8; i++)
printk(KERN_CONT " %02X", ladrf[i]);
printk(KERN_CONT "\n");
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
}
#endif
} /* BuildLAF */
/* ----------------------------------------------------------------------------
restore_multicast_list
Restores the multicast filter for MACE chip to the last
set_multicast_list() call.
Input
multicast_num_addrs
multicast_ladrf[]
---------------------------------------------------------------------------- */
static void restore_multicast_list(struct net_device *dev)
{
mace_private *lp = netdev_priv(dev);
int num_addrs = lp->multicast_num_addrs;
int *ladrf = lp->multicast_ladrf;
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr = dev->base_addr;
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 i;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: restoring Rx mode to %d addresses.\n",
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
dev->name, num_addrs);
if (num_addrs > 0) {
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("Attempt to restore multicast list detected.\n");
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
mace_write(lp, ioaddr, MACE_IAC, MACE_IAC_ADDRCHG | MACE_IAC_LOGADDR);
/* Poll ADDRCHG bit */
while (mace_read(lp, ioaddr, MACE_IAC) & MACE_IAC_ADDRCHG)
;
/* Set LADRF register */
for (i = 0; i < MACE_LADRF_LEN; i++)
mace_write(lp, ioaddr, MACE_LADRF, ladrf[i]);
mace_write(lp, ioaddr, MACE_UTR, MACE_UTR_RCVFCSE | MACE_UTR_LOOP_EXTERNAL);
mace_write(lp, ioaddr, MACE_MACCC, MACE_MACCC_ENXMT | MACE_MACCC_ENRCV);
} else if (num_addrs < 0) {
/* Promiscuous mode: receive all packets */
mace_write(lp, ioaddr, MACE_UTR, MACE_UTR_LOOP_EXTERNAL);
mace_write(lp, ioaddr, MACE_MACCC,
MACE_MACCC_PROM | MACE_MACCC_ENXMT | MACE_MACCC_ENRCV
);
} else {
/* Normal mode */
mace_write(lp, ioaddr, MACE_UTR, MACE_UTR_LOOP_EXTERNAL);
mace_write(lp, ioaddr, MACE_MACCC, MACE_MACCC_ENXMT | MACE_MACCC_ENRCV);
}
} /* restore_multicast_list */
/* ----------------------------------------------------------------------------
set_multicast_list
Set or clear the multicast filter for this adaptor.
Input
num_addrs == -1 Promiscuous mode, receive all packets
num_addrs == 0 Normal mode, clear multicast list
num_addrs > 0 Multicast mode, receive normal and MC packets, and do
best-effort filtering.
Output
multicast_num_addrs
multicast_ladrf[]
---------------------------------------------------------------------------- */
static void set_multicast_list(struct net_device *dev)
{
mace_private *lp = netdev_priv(dev);
int adr[ETHER_ADDR_LEN] = {0}; /* Ethernet address */
int i;
struct dev_mc_list *dmi = dev->mc_list;
#ifdef PCMCIA_DEBUG
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +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
static int old;
if (dev->mc_count != old) {
old = dev->mc_count;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: setting Rx mode to %d addresses.\n",
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
dev->name, old);
}
}
#endif
/* Set multicast_num_addrs. */
lp->multicast_num_addrs = dev->mc_count;
/* Set multicast_ladrf. */
if (num_addrs > 0) {
/* Calculate multicast logical address filter */
memset(lp->multicast_ladrf, 0, MACE_LADRF_LEN);
for (i = 0; i < dev->mc_count; i++) {
memcpy(adr, dmi->dmi_addr, ETHER_ADDR_LEN);
dmi = dmi->next;
BuildLAF(lp->multicast_ladrf, adr);
}
}
restore_multicast_list(dev);
} /* set_multicast_list */
#endif /* BROKEN_MULTICAST */
static void restore_multicast_list(struct net_device *dev)
{
pcmcia: replace kio_addr_t with unsigned int everywhere Remove kio_addr_t, and replace it with unsigned int. No known architecture needs more than 32 bits for IO addresses and ports and having a separate type for it is just messy. Signed-off-by: Olof Johansson <olof@lixom.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 14:27:35 +08:00
unsigned int ioaddr = dev->base_addr;
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
mace_private *lp = netdev_priv(dev);
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: restoring Rx mode to %d addresses.\n", dev->name,
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
lp->multicast_num_addrs);
if (dev->flags & IFF_PROMISC) {
/* Promiscuous mode: receive all packets */
mace_write(lp,ioaddr, MACE_UTR, MACE_UTR_LOOP_EXTERNAL);
mace_write(lp, ioaddr, MACE_MACCC,
MACE_MACCC_PROM | MACE_MACCC_ENXMT | MACE_MACCC_ENRCV
);
} else {
/* Normal mode */
mace_write(lp, ioaddr, MACE_UTR, MACE_UTR_LOOP_EXTERNAL);
mace_write(lp, ioaddr, MACE_MACCC, MACE_MACCC_ENXMT | MACE_MACCC_ENRCV);
}
} /* restore_multicast_list */
static void set_multicast_list(struct net_device *dev)
{
mace_private *lp = netdev_priv(dev);
#ifdef PCMCIA_DEBUG
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +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
static int old;
if (dev->mc_count != old) {
old = dev->mc_count;
pcmcia: use dynamic debug infrastructure, deprecate CS_CHECK (net) Convert PCMCIA drivers to use the dynamic debug infrastructure, instead of requiring manual settings of PCMCIA_DEBUG. Only some rare debug checks are now hidden behind "#ifdef DEBUG" or "#if 0". Also, remove all usages of the CS_CHECK macro and replace them with proper Linux style calling and return value checking. The extra error reporting may be dropped, as the PCMCIA core already complains about any (non-driver-author) errors. CC: netdev@vger.kernel.org Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2009-10-24 21:51:05 +08:00
pr_debug("%s: setting Rx mode to %d addresses.\n",
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
dev->name, old);
}
}
#endif
lp->multicast_num_addrs = dev->mc_count;
restore_multicast_list(dev);
} /* set_multicast_list */
[PATCH] pcmcia: id_table for nmclan_cs Add pcmcia_device_id table. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-28 07:28:23 +08:00
static struct pcmcia_device_id nmclan_ids[] = {
PCMCIA_DEVICE_PROD_ID12("New Media Corporation", "Ethernet", 0x085a850b, 0x00b2e941),
[PATCH] pcmcia: fix many device IDs If the product-id-string contains the '+' , '&' ,'_', it was not converted properly from the /etc/pcmcia/config(pcmcia-cs config file). Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-07-28 16:07:24 +08:00
PCMCIA_DEVICE_PROD_ID12("Portable Add-ons", "Ethernet+", 0xebf1d60, 0xad673aaf),
[PATCH] pcmcia: id_table for nmclan_cs Add pcmcia_device_id table. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-28 07:28:23 +08:00
PCMCIA_DEVICE_NULL,
};
MODULE_DEVICE_TABLE(pcmcia, nmclan_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 pcmcia_driver nmclan_cs_driver = {
.owner = THIS_MODULE,
.drv = {
.name = "nmclan_cs",
},
[PATCH] pcmcia: add return value to _config() functions Most of the driver initialization isn't done in the .probe function, but in the internal _config() functions. Make them return a value, so that .probe can properly report whether the probing of the device succeeded or not. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2006-03-31 23:26:06 +08:00
.probe = nmclan_probe,
[PATCH] pcmcia: unify detach, REMOVAL_EVENT handlers into one remove callback Unify the "detach" and REMOVAL_EVENT handlers to one "remove" function. Old functionality is preserved, for the moment. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:23:14 +08:00
.remove = nmclan_detach,
[PATCH] pcmcia: id_table for nmclan_cs Add pcmcia_device_id table. Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-28 07:28:23 +08:00
.id_table = nmclan_ids,
[PATCH] pcmcia: new suspend core Move the suspend and resume methods out of the event handler, and into special functions. Also use these functions for pre- and post-reset, as almost all drivers already do, and the remaining ones can easily be converted. Bugfix to include/pcmcia/ds.c Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
2005-11-15 04:21:18 +08:00
.suspend = nmclan_suspend,
.resume = nmclan_resume,
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 init_nmclan_cs(void)
{
return pcmcia_register_driver(&nmclan_cs_driver);
}
static void __exit exit_nmclan_cs(void)
{
pcmcia_unregister_driver(&nmclan_cs_driver);
}
module_init(init_nmclan_cs);
module_exit(exit_nmclan_cs);