linux/drivers/edac/edac_pci_sysfs.c

766 lines
19 KiB
C
Raw Normal View History

/*
* (C) 2005, 2006 Linux Networx (http://lnxi.com)
* This file may be distributed under the terms of the
* GNU General Public License.
*
* Written Doug Thompson <norsk5@xmission.com>
*
*/
#include <linux/module.h>
#include <linux/edac.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
#include <linux/ctype.h>
#include "edac_core.h"
#include "edac_module.h"
/* Turn off this whole feature if PCI is not configured */
#ifdef CONFIG_PCI
#define EDAC_PCI_SYMLINK "device"
/* data variables exported via sysfs */
static int check_pci_errors; /* default NO check PCI parity */
static int edac_pci_panic_on_pe; /* default NO panic on PCI Parity */
static int edac_pci_log_pe = 1; /* log PCI parity errors */
static int edac_pci_log_npe = 1; /* log PCI non-parity error errors */
static int edac_pci_poll_msec = 1000; /* one second workq period */
static atomic_t pci_parity_count = ATOMIC_INIT(0);
static atomic_t pci_nonparity_count = ATOMIC_INIT(0);
static struct kobject *edac_pci_top_main_kobj;
static atomic_t edac_pci_sysfs_refcount = ATOMIC_INIT(0);
/* getter functions for the data variables */
int edac_pci_get_check_errors(void)
{
return check_pci_errors;
}
static int edac_pci_get_log_pe(void)
{
return edac_pci_log_pe;
}
static int edac_pci_get_log_npe(void)
{
return edac_pci_log_npe;
}
static int edac_pci_get_panic_on_pe(void)
{
return edac_pci_panic_on_pe;
}
int edac_pci_get_poll_msec(void)
{
return edac_pci_poll_msec;
}
/**************************** EDAC PCI sysfs instance *******************/
static ssize_t instance_pe_count_show(struct edac_pci_ctl_info *pci, char *data)
{
return sprintf(data, "%u\n", atomic_read(&pci->counters.pe_count));
}
static ssize_t instance_npe_count_show(struct edac_pci_ctl_info *pci,
char *data)
{
return sprintf(data, "%u\n", atomic_read(&pci->counters.npe_count));
}
#define to_instance(k) container_of(k, struct edac_pci_ctl_info, kobj)
#define to_instance_attr(a) container_of(a, struct instance_attribute, attr)
/* DEVICE instance kobject release() function */
static void edac_pci_instance_release(struct kobject *kobj)
{
struct edac_pci_ctl_info *pci;
edac_dbg(0, "\n");
/* Form pointer to containing struct, the pci control struct */
pci = to_instance(kobj);
/* decrement reference count on top main kobj */
kobject_put(edac_pci_top_main_kobj);
kfree(pci); /* Free the control struct */
}
/* instance specific attribute structure */
struct instance_attribute {
struct attribute attr;
ssize_t(*show) (struct edac_pci_ctl_info *, char *);
ssize_t(*store) (struct edac_pci_ctl_info *, const char *, size_t);
};
/* Function to 'show' fields from the edac_pci 'instance' structure */
static ssize_t edac_pci_instance_show(struct kobject *kobj,
struct attribute *attr, char *buffer)
{
struct edac_pci_ctl_info *pci = to_instance(kobj);
struct instance_attribute *instance_attr = to_instance_attr(attr);
if (instance_attr->show)
return instance_attr->show(pci, buffer);
return -EIO;
}
/* Function to 'store' fields into the edac_pci 'instance' structure */
static ssize_t edac_pci_instance_store(struct kobject *kobj,
struct attribute *attr,
const char *buffer, size_t count)
{
struct edac_pci_ctl_info *pci = to_instance(kobj);
struct instance_attribute *instance_attr = to_instance_attr(attr);
if (instance_attr->store)
return instance_attr->store(pci, buffer, count);
return -EIO;
}
/* fs_ops table */
static const struct sysfs_ops pci_instance_ops = {
.show = edac_pci_instance_show,
.store = edac_pci_instance_store
};
#define INSTANCE_ATTR(_name, _mode, _show, _store) \
static struct instance_attribute attr_instance_##_name = { \
.attr = {.name = __stringify(_name), .mode = _mode }, \
.show = _show, \
.store = _store, \
};
INSTANCE_ATTR(pe_count, S_IRUGO, instance_pe_count_show, NULL);
INSTANCE_ATTR(npe_count, S_IRUGO, instance_npe_count_show, NULL);
/* pci instance attributes */
static struct instance_attribute *pci_instance_attr[] = {
&attr_instance_pe_count,
&attr_instance_npe_count,
NULL
};
/* the ktype for a pci instance */
static struct kobj_type ktype_pci_instance = {
.release = edac_pci_instance_release,
.sysfs_ops = &pci_instance_ops,
.default_attrs = (struct attribute **)pci_instance_attr,
};
/*
* edac_pci_create_instance_kobj
*
* construct one EDAC PCI instance's kobject for use
*/
static int edac_pci_create_instance_kobj(struct edac_pci_ctl_info *pci, int idx)
{
struct kobject *main_kobj;
int err;
edac_dbg(0, "\n");
/* First bump the ref count on the top main kobj, which will
* track the number of PCI instances we have, and thus nest
* properly on keeping the module loaded
*/
main_kobj = kobject_get(edac_pci_top_main_kobj);
if (!main_kobj) {
err = -ENODEV;
goto error_out;
}
/* And now register this new kobject under the main kobj */
err = kobject_init_and_add(&pci->kobj, &ktype_pci_instance,
edac_pci_top_main_kobj, "pci%d", idx);
if (err != 0) {
edac_dbg(2, "failed to register instance pci%d\n", idx);
kobject_put(edac_pci_top_main_kobj);
goto error_out;
}
kobject_uevent(&pci->kobj, KOBJ_ADD);
edac_dbg(1, "Register instance 'pci%d' kobject\n", idx);
return 0;
/* Error unwind statck */
error_out:
return err;
}
/*
* edac_pci_unregister_sysfs_instance_kobj
*
* unregister the kobj for the EDAC PCI instance
*/
static void edac_pci_unregister_sysfs_instance_kobj(
struct edac_pci_ctl_info *pci)
{
edac_dbg(0, "\n");
/* Unregister the instance kobject and allow its release
* function release the main reference count and then
* kfree the memory
*/
kobject_put(&pci->kobj);
}
/***************************** EDAC PCI sysfs root **********************/
#define to_edacpci(k) container_of(k, struct edac_pci_ctl_info, kobj)
#define to_edacpci_attr(a) container_of(a, struct edac_pci_attr, attr)
/* simple show/store functions for attributes */
static ssize_t edac_pci_int_show(void *ptr, char *buffer)
{
int *value = ptr;
return sprintf(buffer, "%d\n", *value);
}
static ssize_t edac_pci_int_store(void *ptr, const char *buffer, size_t count)
{
int *value = ptr;
if (isdigit(*buffer))
*value = simple_strtoul(buffer, NULL, 0);
return count;
}
struct edac_pci_dev_attribute {
struct attribute attr;
void *value;
ssize_t(*show) (void *, char *);
ssize_t(*store) (void *, const char *, size_t);
};
/* Set of show/store abstract level functions for PCI Parity object */
static ssize_t edac_pci_dev_show(struct kobject *kobj, struct attribute *attr,
char *buffer)
{
struct edac_pci_dev_attribute *edac_pci_dev;
edac_pci_dev = (struct edac_pci_dev_attribute *)attr;
if (edac_pci_dev->show)
return edac_pci_dev->show(edac_pci_dev->value, buffer);
return -EIO;
}
static ssize_t edac_pci_dev_store(struct kobject *kobj,
struct attribute *attr, const char *buffer,
size_t count)
{
struct edac_pci_dev_attribute *edac_pci_dev;
edac_pci_dev = (struct edac_pci_dev_attribute *)attr;
if (edac_pci_dev->store)
return edac_pci_dev->store(edac_pci_dev->value, buffer, count);
return -EIO;
}
static const struct sysfs_ops edac_pci_sysfs_ops = {
.show = edac_pci_dev_show,
.store = edac_pci_dev_store
};
#define EDAC_PCI_ATTR(_name,_mode,_show,_store) \
static struct edac_pci_dev_attribute edac_pci_attr_##_name = { \
.attr = {.name = __stringify(_name), .mode = _mode }, \
.value = &_name, \
.show = _show, \
.store = _store, \
};
#define EDAC_PCI_STRING_ATTR(_name,_data,_mode,_show,_store) \
static struct edac_pci_dev_attribute edac_pci_attr_##_name = { \
.attr = {.name = __stringify(_name), .mode = _mode }, \
.value = _data, \
.show = _show, \
.store = _store, \
};
/* PCI Parity control files */
EDAC_PCI_ATTR(check_pci_errors, S_IRUGO | S_IWUSR, edac_pci_int_show,
edac_pci_int_store);
EDAC_PCI_ATTR(edac_pci_log_pe, S_IRUGO | S_IWUSR, edac_pci_int_show,
edac_pci_int_store);
EDAC_PCI_ATTR(edac_pci_log_npe, S_IRUGO | S_IWUSR, edac_pci_int_show,
edac_pci_int_store);
EDAC_PCI_ATTR(edac_pci_panic_on_pe, S_IRUGO | S_IWUSR, edac_pci_int_show,
edac_pci_int_store);
EDAC_PCI_ATTR(pci_parity_count, S_IRUGO, edac_pci_int_show, NULL);
EDAC_PCI_ATTR(pci_nonparity_count, S_IRUGO, edac_pci_int_show, NULL);
/* Base Attributes of the memory ECC object */
static struct edac_pci_dev_attribute *edac_pci_attr[] = {
&edac_pci_attr_check_pci_errors,
&edac_pci_attr_edac_pci_log_pe,
&edac_pci_attr_edac_pci_log_npe,
&edac_pci_attr_edac_pci_panic_on_pe,
&edac_pci_attr_pci_parity_count,
&edac_pci_attr_pci_nonparity_count,
NULL,
};
/*
* edac_pci_release_main_kobj
*
* This release function is called when the reference count to the
* passed kobj goes to zero.
*
* This kobj is the 'main' kobject that EDAC PCI instances
* link to, and thus provide for proper nesting counts
*/
static void edac_pci_release_main_kobj(struct kobject *kobj)
{
edac_dbg(0, "here to module_put(THIS_MODULE)\n");
kfree(kobj);
/* last reference to top EDAC PCI kobject has been removed,
* NOW release our ref count on the core module
*/
module_put(THIS_MODULE);
}
/* ktype struct for the EDAC PCI main kobj */
static struct kobj_type ktype_edac_pci_main_kobj = {
.release = edac_pci_release_main_kobj,
.sysfs_ops = &edac_pci_sysfs_ops,
.default_attrs = (struct attribute **)edac_pci_attr,
};
/**
* edac_pci_main_kobj_setup()
*
* setup the sysfs for EDAC PCI attributes
* assumes edac_subsys has already been initialized
*/
static int edac_pci_main_kobj_setup(void)
{
int err;
struct bus_type *edac_subsys;
edac_dbg(0, "\n");
/* check and count if we have already created the main kobject */
if (atomic_inc_return(&edac_pci_sysfs_refcount) != 1)
return 0;
/* First time, so create the main kobject and its
* controls and attributes
*/
edac_subsys = edac_get_sysfs_subsys();
if (edac_subsys == NULL) {
edac_dbg(1, "no edac_subsys\n");
err = -ENODEV;
goto decrement_count_fail;
}
/* Bump the reference count on this module to ensure the
* modules isn't unloaded until we deconstruct the top
* level main kobj for EDAC PCI
*/
if (!try_module_get(THIS_MODULE)) {
edac_dbg(1, "try_module_get() failed\n");
err = -ENODEV;
goto mod_get_fail;
}
edac_pci_top_main_kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL);
if (!edac_pci_top_main_kobj) {
edac_dbg(1, "Failed to allocate\n");
err = -ENOMEM;
goto kzalloc_fail;
}
/* Instanstiate the pci object */
err = kobject_init_and_add(edac_pci_top_main_kobj,
&ktype_edac_pci_main_kobj,
&edac_subsys->dev_root->kobj, "pci");
if (err) {
edac_dbg(1, "Failed to register '.../edac/pci'\n");
goto kobject_init_and_add_fail;
}
/* At this point, to 'release' the top level kobject
* for EDAC PCI, then edac_pci_main_kobj_teardown()
* must be used, for resources to be cleaned up properly
*/
kobject_uevent(edac_pci_top_main_kobj, KOBJ_ADD);
edac_dbg(1, "Registered '.../edac/pci' kobject\n");
return 0;
/* Error unwind statck */
kobject_init_and_add_fail:
kfree(edac_pci_top_main_kobj);
kzalloc_fail:
module_put(THIS_MODULE);
mod_get_fail:
edac_put_sysfs_subsys();
decrement_count_fail:
/* if are on this error exit, nothing to tear down */
atomic_dec(&edac_pci_sysfs_refcount);
return err;
}
/*
* edac_pci_main_kobj_teardown()
*
* if no longer linked (needed) remove the top level EDAC PCI
* kobject with its controls and attributes
*/
static void edac_pci_main_kobj_teardown(void)
{
edac_dbg(0, "\n");
/* Decrement the count and only if no more controller instances
* are connected perform the unregisteration of the top level
* main kobj
*/
if (atomic_dec_return(&edac_pci_sysfs_refcount) == 0) {
edac_dbg(0, "called kobject_put on main kobj\n");
kobject_put(edac_pci_top_main_kobj);
i7core_edac: fix kernel crash on unloading i7core_edac. It is easy to trigger this crash on 3.7.0: root@intel_westmere_ep-3:~# modprobe -r i7core_edac EDAC PCI: Removed device 0 for i7core_edac EDAC PCI controller: DEV 0000:fe:03.0 EDAC MC: Removed device 1 for i7core_edac.c i7 core #1: DEV 0000:fe:03.0 EDAC PCI: Removed device 1 for i7core_edac EDAC PCI controller: DEV 0000:ff:03.0 EDAC MC: Removed device 0 for i7core_edac.c i7 core #0: DEV 0000:ff:03.0 BUG: unable to handle kernel NULL pointer dereference at 0000000000000110 IP: [<ffffffff82069ee9>] __blocking_notifier_call_chain+0x29/0x80 PGD 1eaae7067 PUD 1e96e4067 PMD 0 Oops: 0000 [#1] PREEMPT SMP Modules linked in: minix acpi_cpufreq freq_table mperf ioatdma processor edac_core(-) iTCO_wdt coretemp evdev hwmon lpc_ich dca mfd_core crc32c_intel ioapic [last unloaded: i7core_edac] CPU 3 Pid: 1268, comm: modprobe Not tainted 3.7.0-WR5.0.1.0_standard+ #30 Intel Corporation S5520HC/S5520HC RIP: 0010:[<ffffffff82069ee9>] [<ffffffff82069ee9>] __blocking_notifier_call_chain+0x29/0x80 RSP: 0018:ffff8801eb12de28 EFLAGS: 00010246 RAX: 0000000000000000 RBX: 00000000000000f0 RCX: 00000000ffffffff RDX: ffff88012b452800 RSI: 0000000000000002 RDI: 00000000000000f0 RBP: ffff8801eb12de68 R08: 0000000000000000 R09: ffffea0004ad1118 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000 R13: ffff8801eb12dee8 R14: ffff88012b452800 R15: 000000000060e518 FS: 00007f9ea95a9700(0000) GS:ffff8801efc20000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000110 CR3: 00000001262f1000 CR4: 00000000000007e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process modprobe (pid: 1268, threadinfo ffff8801eb12c000, task ffff8801e8421690) Stack: ffff88012c802a00 ffff88012b445ec0 ffff88012c802300 ffff88012b452800 0000000000000000 ffff8801eb12dee8 000000000060e080 000000000060e518 ffff8801eb12de78 ffffffff82069f56 ffff8801eb12dea8 ffffffff824ead7c Call Trace: [<ffffffff82069f56>] blocking_notifier_call_chain+0x16/0x20 [<ffffffff824ead7c>] device_del+0x3c/0x1d0 [<ffffffffa00095a8>] edac_mc_sysfs_exit+0x1c/0x2f [edac_core] [<ffffffffa000961c>] edac_exit+0x4f/0x56 [edac_core] [<ffffffff820a3d2a>] sys_delete_module+0x17a/0x240 [<ffffffff8212da7c>] ? vm_munmap+0x5c/0x80 [<ffffffff82877682>] system_call_fastpath+0x16/0x1b Code: 90 90 55 48 89 e5 48 83 ec 40 48 89 5d d8 4c 89 65 e0 4c 89 6d e8 4c 89 75 f0 4c 89 7d f8 66 66 66 66 90 31 c0 49 89 d6 48 89 fb <48> 8b 57 20 49 89 f5 41 89 cf 4c 8d 67 20 48 85 d2 74 2c 4c 89 RIP [<ffffffff82069ee9>] __blocking_notifier_call_chain+0x29/0x80 RSP <ffff8801eb12de28> CR2: 0000000000000110 ---[ end trace b69acf12ccad1c0d ]--- Usually, edac_subsys is grabbed one time by pci at initialization. But edac_subsys may be released several times if multiple pci MCs exist. The fix just makes the operations balanced. Signed-off-by: Lans Zhang <jia.zhang@windriver.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-12-20 10:16:07 +08:00
edac_put_sysfs_subsys();
}
}
/*
*
* edac_pci_create_sysfs
*
* Create the controls/attributes for the specified EDAC PCI device
*/
int edac_pci_create_sysfs(struct edac_pci_ctl_info *pci)
{
int err;
struct kobject *edac_kobj = &pci->kobj;
edac_dbg(0, "idx=%d\n", pci->pci_idx);
/* create the top main EDAC PCI kobject, IF needed */
err = edac_pci_main_kobj_setup();
if (err)
return err;
/* Create this instance's kobject under the MAIN kobject */
err = edac_pci_create_instance_kobj(pci, pci->pci_idx);
if (err)
goto unregister_cleanup;
err = sysfs_create_link(edac_kobj, &pci->dev->kobj, EDAC_PCI_SYMLINK);
if (err) {
edac_dbg(0, "sysfs_create_link() returned err= %d\n", err);
goto symlink_fail;
}
return 0;
/* Error unwind stack */
symlink_fail:
edac_pci_unregister_sysfs_instance_kobj(pci);
unregister_cleanup:
edac_pci_main_kobj_teardown();
return err;
}
/*
* edac_pci_remove_sysfs
*
* remove the controls and attributes for this EDAC PCI device
*/
void edac_pci_remove_sysfs(struct edac_pci_ctl_info *pci)
{
edac_dbg(0, "index=%d\n", pci->pci_idx);
/* Remove the symlink */
sysfs_remove_link(&pci->kobj, EDAC_PCI_SYMLINK);
/* remove this PCI instance's sysfs entries */
edac_pci_unregister_sysfs_instance_kobj(pci);
/* Call the main unregister function, which will determine
* if this 'pci' is the last instance.
* If it is, the main kobject will be unregistered as a result
*/
edac_dbg(0, "calling edac_pci_main_kobj_teardown()\n");
edac_pci_main_kobj_teardown();
}
/************************ PCI error handling *************************/
static u16 get_pci_parity_status(struct pci_dev *dev, int secondary)
{
int where;
u16 status;
where = secondary ? PCI_SEC_STATUS : PCI_STATUS;
pci_read_config_word(dev, where, &status);
/* If we get back 0xFFFF then we must suspect that the card has been
* pulled but the Linux PCI layer has not yet finished cleaning up.
* We don't want to report on such devices
*/
if (status == 0xFFFF) {
u32 sanity;
pci_read_config_dword(dev, 0, &sanity);
if (sanity == 0xFFFFFFFF)
return 0;
}
status &= PCI_STATUS_DETECTED_PARITY | PCI_STATUS_SIG_SYSTEM_ERROR |
PCI_STATUS_PARITY;
if (status)
/* reset only the bits we are interested in */
pci_write_config_word(dev, where, status);
return status;
}
/* Clear any PCI parity errors logged by this device. */
static void edac_pci_dev_parity_clear(struct pci_dev *dev)
{
u8 header_type;
get_pci_parity_status(dev, 0);
/* read the device TYPE, looking for bridges */
pci_read_config_byte(dev, PCI_HEADER_TYPE, &header_type);
if ((header_type & 0x7F) == PCI_HEADER_TYPE_BRIDGE)
get_pci_parity_status(dev, 1);
}
/*
* PCI Parity polling
*
* Function to retrieve the current parity status
* and decode it
*
*/
static void edac_pci_dev_parity_test(struct pci_dev *dev)
{
unsigned long flags;
u16 status;
u8 header_type;
/* stop any interrupts until we can acquire the status */
local_irq_save(flags);
/* read the STATUS register on this device */
status = get_pci_parity_status(dev, 0);
/* read the device TYPE, looking for bridges */
pci_read_config_byte(dev, PCI_HEADER_TYPE, &header_type);
local_irq_restore(flags);
edac_dbg(4, "PCI STATUS= 0x%04x %s\n", status, dev_name(&dev->dev));
drivers/edac: pci: broken parity regression Using the EDAC code in kernel.org kernel version 2.6.23.8 I am seeing the following problem: In the kernel there is a pci device attribute located in sysfs that is checked by the EDAC PCI scanning code. If that attribute is set, PCI parity/error scannining is skipped for that device. The attribute is: broken_parity_status as is located in /sys/devices/pci<XXX>/0000:XX:YY.Z directorys for PCI devices. I don't think this check was actually implemented. I have a misbehaved card that reports a parity error every 1000 ms: Nov 25 07:28:43 beta kernel: EDAC PCI: Master Data Parity Error on 0000:05:01.0 Nov 25 07:28:44 beta kernel: EDAC PCI: Master Data Parity Error on 0000:05:01.0 Nov 25 07:28:45 beta kernel: EDAC PCI: Master Data Parity Error on 0000:05:01.0 Setting that card's broken_parity_status bit did not mask the error: echo "1" > /sys/bus/pci/devices/0000:05:01.0/broken_parity_status I looked through the EDAC code and did not readily see any reference to broken_parity_status at all (which makes sense based on the behavior I am seeing). I applied the following patch as a proof-of-concept and now EDAC's PCI parity error reporting behaves as documented: bryan Good regression find, bryan. It used to work. sigh. I added more logic to your patch, for more coverage of the error. Doug T Signed-off-by: Bryan Boatright <b1@omega71.com> Signed-off-by: Doug Thompson <dougthompson@xmisson.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 16:14:58 +08:00
/* check the status reg for errors on boards NOT marked as broken
* if broken, we cannot trust any of the status bits
*/
if (status && !dev->broken_parity_status) {
if (status & (PCI_STATUS_SIG_SYSTEM_ERROR)) {
edac_printk(KERN_CRIT, EDAC_PCI,
"Signaled System Error on %s\n",
pci_name(dev));
atomic_inc(&pci_nonparity_count);
}
if (status & (PCI_STATUS_PARITY)) {
edac_printk(KERN_CRIT, EDAC_PCI,
"Master Data Parity Error on %s\n",
pci_name(dev));
atomic_inc(&pci_parity_count);
}
if (status & (PCI_STATUS_DETECTED_PARITY)) {
edac_printk(KERN_CRIT, EDAC_PCI,
"Detected Parity Error on %s\n",
pci_name(dev));
atomic_inc(&pci_parity_count);
}
}
edac_dbg(4, "PCI HEADER TYPE= 0x%02x %s\n",
header_type, dev_name(&dev->dev));
if ((header_type & 0x7F) == PCI_HEADER_TYPE_BRIDGE) {
/* On bridges, need to examine secondary status register */
status = get_pci_parity_status(dev, 1);
edac_dbg(4, "PCI SEC_STATUS= 0x%04x %s\n",
status, dev_name(&dev->dev));
drivers/edac: pci: broken parity regression Using the EDAC code in kernel.org kernel version 2.6.23.8 I am seeing the following problem: In the kernel there is a pci device attribute located in sysfs that is checked by the EDAC PCI scanning code. If that attribute is set, PCI parity/error scannining is skipped for that device. The attribute is: broken_parity_status as is located in /sys/devices/pci<XXX>/0000:XX:YY.Z directorys for PCI devices. I don't think this check was actually implemented. I have a misbehaved card that reports a parity error every 1000 ms: Nov 25 07:28:43 beta kernel: EDAC PCI: Master Data Parity Error on 0000:05:01.0 Nov 25 07:28:44 beta kernel: EDAC PCI: Master Data Parity Error on 0000:05:01.0 Nov 25 07:28:45 beta kernel: EDAC PCI: Master Data Parity Error on 0000:05:01.0 Setting that card's broken_parity_status bit did not mask the error: echo "1" > /sys/bus/pci/devices/0000:05:01.0/broken_parity_status I looked through the EDAC code and did not readily see any reference to broken_parity_status at all (which makes sense based on the behavior I am seeing). I applied the following patch as a proof-of-concept and now EDAC's PCI parity error reporting behaves as documented: bryan Good regression find, bryan. It used to work. sigh. I added more logic to your patch, for more coverage of the error. Doug T Signed-off-by: Bryan Boatright <b1@omega71.com> Signed-off-by: Doug Thompson <dougthompson@xmisson.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 16:14:58 +08:00
/* check the secondary status reg for errors,
* on NOT broken boards
*/
if (status && !dev->broken_parity_status) {
if (status & (PCI_STATUS_SIG_SYSTEM_ERROR)) {
edac_printk(KERN_CRIT, EDAC_PCI, "Bridge "
"Signaled System Error on %s\n",
pci_name(dev));
atomic_inc(&pci_nonparity_count);
}
if (status & (PCI_STATUS_PARITY)) {
edac_printk(KERN_CRIT, EDAC_PCI, "Bridge "
"Master Data Parity Error on "
"%s\n", pci_name(dev));
atomic_inc(&pci_parity_count);
}
if (status & (PCI_STATUS_DETECTED_PARITY)) {
edac_printk(KERN_CRIT, EDAC_PCI, "Bridge "
"Detected Parity Error on %s\n",
pci_name(dev));
atomic_inc(&pci_parity_count);
}
}
}
}
/* reduce some complexity in definition of the iterator */
typedef void (*pci_parity_check_fn_t) (struct pci_dev *dev);
/*
* pci_dev parity list iterator
*
* Scan the PCI device list looking for SERRORs, Master Parity ERRORS or
* Parity ERRORs on primary or secondary devices.
*/
static inline void edac_pci_dev_parity_iterator(pci_parity_check_fn_t fn)
{
struct pci_dev *dev = NULL;
for_each_pci_dev(dev)
fn(dev);
}
/*
* edac_pci_do_parity_check
*
* performs the actual PCI parity check operation
*/
void edac_pci_do_parity_check(void)
{
int before_count;
edac_dbg(3, "\n");
/* if policy has PCI check off, leave now */
if (!check_pci_errors)
return;
before_count = atomic_read(&pci_parity_count);
/* scan all PCI devices looking for a Parity Error on devices and
* bridges.
* The iterator calls pci_get_device() which might sleep, thus
* we cannot disable interrupts in this scan.
*/
edac_pci_dev_parity_iterator(edac_pci_dev_parity_test);
/* Only if operator has selected panic on PCI Error */
if (edac_pci_get_panic_on_pe()) {
/* If the count is different 'after' from 'before' */
if (before_count != atomic_read(&pci_parity_count))
panic("EDAC: PCI Parity Error");
}
}
/*
* edac_pci_clear_parity_errors
*
* function to perform an iteration over the PCI devices
* and clearn their current status
*/
void edac_pci_clear_parity_errors(void)
{
/* Clear any PCI bus parity errors that devices initially have logged
* in their registers.
*/
edac_pci_dev_parity_iterator(edac_pci_dev_parity_clear);
}
/*
* edac_pci_handle_pe
*
* Called to handle a PARITY ERROR event
*/
void edac_pci_handle_pe(struct edac_pci_ctl_info *pci, const char *msg)
{
/* global PE counter incremented by edac_pci_do_parity_check() */
atomic_inc(&pci->counters.pe_count);
if (edac_pci_get_log_pe())
edac_pci_printk(pci, KERN_WARNING,
"Parity Error ctl: %s %d: %s\n",
pci->ctl_name, pci->pci_idx, msg);
/*
* poke all PCI devices and see which one is the troublemaker
* panic() is called if set
*/
edac_pci_do_parity_check();
}
EXPORT_SYMBOL_GPL(edac_pci_handle_pe);
/*
* edac_pci_handle_npe
*
* Called to handle a NON-PARITY ERROR event
*/
void edac_pci_handle_npe(struct edac_pci_ctl_info *pci, const char *msg)
{
/* global NPE counter incremented by edac_pci_do_parity_check() */
atomic_inc(&pci->counters.npe_count);
if (edac_pci_get_log_npe())
edac_pci_printk(pci, KERN_WARNING,
"Non-Parity Error ctl: %s %d: %s\n",
pci->ctl_name, pci->pci_idx, msg);
/*
* poke all PCI devices and see which one is the troublemaker
* panic() is called if set
*/
edac_pci_do_parity_check();
}
EXPORT_SYMBOL_GPL(edac_pci_handle_npe);
/*
* Define the PCI parameter to the module
*/
module_param(check_pci_errors, int, 0644);
MODULE_PARM_DESC(check_pci_errors,
"Check for PCI bus parity errors: 0=off 1=on");
module_param(edac_pci_panic_on_pe, int, 0644);
MODULE_PARM_DESC(edac_pci_panic_on_pe,
"Panic on PCI Bus Parity error: 0=off 1=on");
#endif /* CONFIG_PCI */