2018-01-26 14:12:23 -06:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
2006-07-31 15:21:33 +08:00
|
|
|
/*
|
2018-03-09 16:36:33 -06:00
|
|
|
* Implement the AER root port service driver. The driver registers an IRQ
|
|
|
|
* handler. When a root port triggers an AER interrupt, the IRQ handler
|
2025-03-07 17:17:15 -06:00
|
|
|
* collects Root Port status and schedules work.
|
2006-07-31 15:21:33 +08:00
|
|
|
*
|
|
|
|
* Copyright (C) 2006 Intel Corp.
|
|
|
|
* Tom Long Nguyen (tom.l.nguyen@intel.com)
|
|
|
|
* Zhang Yanmin (yanmin.zhang@intel.com)
|
2018-06-08 08:40:00 -05:00
|
|
|
*
|
|
|
|
* (C) Copyright 2009 Hewlett-Packard Development Company, L.P.
|
|
|
|
* Andrew Patterson <andrew.patterson@hp.com>
|
2006-07-31 15:21:33 +08:00
|
|
|
*/
|
|
|
|
|
2019-05-07 18:24:50 -05:00
|
|
|
#define pr_fmt(fmt) "AER: " fmt
|
|
|
|
#define dev_fmt pr_fmt
|
|
|
|
|
2019-08-27 18:18:22 +03:00
|
|
|
#include <linux/bitops.h>
|
2018-06-08 08:39:38 -05:00
|
|
|
#include <linux/cper.h>
|
2024-12-16 18:10:12 +02:00
|
|
|
#include <linux/dev_printk.h>
|
2006-07-31 15:21:33 +08:00
|
|
|
#include <linux/pci.h>
|
2011-01-07 00:55:09 +01:00
|
|
|
#include <linux/pci-acpi.h>
|
2009-10-07 17:09:06 +04:00
|
|
|
#include <linux/sched.h>
|
2006-07-31 15:21:33 +08:00
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/pm.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/interrupt.h>
|
|
|
|
#include <linux/delay.h>
|
2018-06-08 08:33:39 -05:00
|
|
|
#include <linux/kfifo.h>
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
#include <linux/ratelimit.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 17:04:11 +09:00
|
|
|
#include <linux/slab.h>
|
2018-06-08 08:39:45 -05:00
|
|
|
#include <acpi/apei.h>
|
2023-09-21 02:03:36 +08:00
|
|
|
#include <acpi/ghes.h>
|
2018-06-08 08:39:38 -05:00
|
|
|
#include <ras/ras_event.h>
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2018-06-08 08:48:47 -05:00
|
|
|
#include "../pci.h"
|
|
|
|
#include "portdrv.h"
|
2018-06-08 08:41:28 -05:00
|
|
|
|
2024-12-16 18:10:12 +02:00
|
|
|
#define aer_printk(level, pdev, fmt, arg...) \
|
|
|
|
dev_printk(level, &(pdev)->dev, fmt, ##arg)
|
|
|
|
|
2018-09-18 17:58:43 -06:00
|
|
|
#define AER_ERROR_SOURCES_MAX 128
|
2018-06-08 08:41:28 -05:00
|
|
|
|
2018-06-30 15:07:17 -05:00
|
|
|
#define AER_MAX_TYPEOF_COR_ERRS 16 /* as per PCI_ERR_COR_STATUS */
|
2019-08-27 15:21:44 -07:00
|
|
|
#define AER_MAX_TYPEOF_UNCOR_ERRS 27 /* as per PCI_ERR_UNCOR_STATUS*/
|
2018-06-30 15:07:17 -05:00
|
|
|
|
2018-06-08 08:41:28 -05:00
|
|
|
struct aer_err_source {
|
2023-12-06 16:42:31 -06:00
|
|
|
u32 status; /* PCI_ERR_ROOT_STATUS */
|
|
|
|
u32 id; /* PCI_ERR_ROOT_ERR_SRC */
|
2018-06-08 08:41:28 -05:00
|
|
|
};
|
|
|
|
|
|
|
|
struct aer_rpc {
|
|
|
|
struct pci_dev *rpd; /* Root Port device */
|
2018-09-18 17:58:43 -06:00
|
|
|
DECLARE_KFIFO(aer_fifo, struct aer_err_source, AER_ERROR_SOURCES_MAX);
|
2018-06-08 08:41:28 -05:00
|
|
|
};
|
|
|
|
|
2025-05-22 18:21:21 -05:00
|
|
|
/* AER info for the device */
|
|
|
|
struct aer_info {
|
2018-06-30 15:07:17 -05:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Fields for all AER capable devices. They indicate the errors
|
|
|
|
* "as seen by this device". Note that this may mean that if an
|
2025-03-07 17:17:15 -06:00
|
|
|
* Endpoint is causing problems, the AER counters may increment
|
|
|
|
* at its link partner (e.g. Root Port) because the errors will be
|
|
|
|
* "seen" by the link partner and not the problematic Endpoint
|
2018-06-30 15:07:17 -05:00
|
|
|
* itself (which may report all counters as 0 as it never saw any
|
|
|
|
* problems).
|
|
|
|
*/
|
|
|
|
/* Counters for different type of correctable errors */
|
|
|
|
u64 dev_cor_errs[AER_MAX_TYPEOF_COR_ERRS];
|
|
|
|
/* Counters for different type of fatal uncorrectable errors */
|
|
|
|
u64 dev_fatal_errs[AER_MAX_TYPEOF_UNCOR_ERRS];
|
|
|
|
/* Counters for different type of nonfatal uncorrectable errors */
|
|
|
|
u64 dev_nonfatal_errs[AER_MAX_TYPEOF_UNCOR_ERRS];
|
|
|
|
/* Total number of ERR_COR sent by this device */
|
|
|
|
u64 dev_total_cor_errs;
|
|
|
|
/* Total number of ERR_FATAL sent by this device */
|
|
|
|
u64 dev_total_fatal_errs;
|
|
|
|
/* Total number of ERR_NONFATAL sent by this device */
|
|
|
|
u64 dev_total_nonfatal_errs;
|
|
|
|
|
|
|
|
/*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Fields for Root Ports & Root Complex Event Collectors only; these
|
2018-06-30 15:07:17 -05:00
|
|
|
* indicate the total number of ERR_COR, ERR_FATAL, and ERR_NONFATAL
|
2025-03-07 17:17:15 -06:00
|
|
|
* messages received by the Root Port / Event Collector, INCLUDING the
|
|
|
|
* ones that are generated internally (by the Root Port itself)
|
2018-06-30 15:07:17 -05:00
|
|
|
*/
|
|
|
|
u64 rootport_total_cor_errs;
|
|
|
|
u64 rootport_total_fatal_errs;
|
|
|
|
u64 rootport_total_nonfatal_errs;
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
|
|
|
|
/* Ratelimits for errors */
|
|
|
|
struct ratelimit_state correctable_ratelimit;
|
|
|
|
struct ratelimit_state nonfatal_ratelimit;
|
2018-06-30 15:07:17 -05:00
|
|
|
};
|
|
|
|
|
2018-06-08 08:41:28 -05:00
|
|
|
#define AER_LOG_TLP_MASKS (PCI_ERR_UNC_POISON_TLP| \
|
|
|
|
PCI_ERR_UNC_ECRC| \
|
|
|
|
PCI_ERR_UNC_UNSUP| \
|
|
|
|
PCI_ERR_UNC_COMP_ABORT| \
|
|
|
|
PCI_ERR_UNC_UNX_COMP| \
|
|
|
|
PCI_ERR_UNC_MALF_TLP)
|
|
|
|
|
|
|
|
#define SYSTEM_ERROR_INTR_ON_MESG_MASK (PCI_EXP_RTCTL_SECEE| \
|
|
|
|
PCI_EXP_RTCTL_SENFEE| \
|
|
|
|
PCI_EXP_RTCTL_SEFEE)
|
|
|
|
#define ROOT_PORT_INTR_ON_MESG_MASK (PCI_ERR_ROOT_CMD_COR_EN| \
|
|
|
|
PCI_ERR_ROOT_CMD_NONFATAL_EN| \
|
|
|
|
PCI_ERR_ROOT_CMD_FATAL_EN)
|
|
|
|
#define ERR_COR_ID(d) (d & 0xffff)
|
|
|
|
#define ERR_UNCOR_ID(d) (d >> 16)
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2022-04-18 15:02:37 +00:00
|
|
|
#define AER_ERR_STATUS_MASK (PCI_ERR_ROOT_UNCOR_RCV | \
|
|
|
|
PCI_ERR_ROOT_COR_RCV | \
|
|
|
|
PCI_ERR_ROOT_MULTI_COR_RCV | \
|
|
|
|
PCI_ERR_ROOT_MULTI_UNCOR_RCV)
|
|
|
|
|
2025-05-17 00:52:23 +08:00
|
|
|
static bool pcie_aer_disable;
|
2020-03-23 17:26:02 -07:00
|
|
|
static pci_ers_result_t aer_root_reset(struct pci_dev *dev);
|
2007-10-05 13:17:58 -07:00
|
|
|
|
|
|
|
void pci_no_aer(void)
|
|
|
|
{
|
2025-05-17 00:52:23 +08:00
|
|
|
pcie_aer_disable = true;
|
2007-10-05 13:17:58 -07:00
|
|
|
}
|
|
|
|
|
2010-08-21 01:50:52 +02:00
|
|
|
bool pci_aer_available(void)
|
|
|
|
{
|
|
|
|
return !pcie_aer_disable && pci_msi_enabled();
|
|
|
|
}
|
|
|
|
|
2018-06-08 08:40:00 -05:00
|
|
|
#ifdef CONFIG_PCIE_ECRC
|
|
|
|
|
|
|
|
#define ECRC_POLICY_DEFAULT 0 /* ECRC set by BIOS */
|
|
|
|
#define ECRC_POLICY_OFF 1 /* ECRC off for performance */
|
|
|
|
#define ECRC_POLICY_ON 2 /* ECRC on for data integrity */
|
|
|
|
|
|
|
|
static int ecrc_policy = ECRC_POLICY_DEFAULT;
|
|
|
|
|
2019-01-28 13:57:28 +02:00
|
|
|
static const char * const ecrc_policy_str[] = {
|
2018-06-08 08:40:00 -05:00
|
|
|
[ECRC_POLICY_DEFAULT] = "bios",
|
|
|
|
[ECRC_POLICY_OFF] = "off",
|
|
|
|
[ECRC_POLICY_ON] = "on"
|
|
|
|
};
|
|
|
|
|
|
|
|
/**
|
2021-03-11 00:17:17 +00:00
|
|
|
* enable_ecrc_checking - enable PCIe ECRC checking for a device
|
2018-06-08 08:40:00 -05:00
|
|
|
* @dev: the PCI device
|
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Return: 0 on success, or negative on failure.
|
2018-06-08 08:40:00 -05:00
|
|
|
*/
|
|
|
|
static int enable_ecrc_checking(struct pci_dev *dev)
|
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-06-08 08:40:00 -05:00
|
|
|
u32 reg32;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2018-06-08 08:40:00 -05:00
|
|
|
return -ENODEV;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_CAP, ®32);
|
2018-06-08 08:40:00 -05:00
|
|
|
if (reg32 & PCI_ERR_CAP_ECRC_GENC)
|
|
|
|
reg32 |= PCI_ERR_CAP_ECRC_GENE;
|
|
|
|
if (reg32 & PCI_ERR_CAP_ECRC_CHKC)
|
|
|
|
reg32 |= PCI_ERR_CAP_ECRC_CHKE;
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_CAP, reg32);
|
2018-06-08 08:40:00 -05:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2025-03-07 17:17:15 -06:00
|
|
|
* disable_ecrc_checking - disable PCIe ECRC checking for a device
|
2018-06-08 08:40:00 -05:00
|
|
|
* @dev: the PCI device
|
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Return: 0 on success, or negative on failure.
|
2018-06-08 08:40:00 -05:00
|
|
|
*/
|
|
|
|
static int disable_ecrc_checking(struct pci_dev *dev)
|
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-06-08 08:40:00 -05:00
|
|
|
u32 reg32;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2018-06-08 08:40:00 -05:00
|
|
|
return -ENODEV;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_CAP, ®32);
|
2018-06-08 08:40:00 -05:00
|
|
|
reg32 &= ~(PCI_ERR_CAP_ECRC_GENE | PCI_ERR_CAP_ECRC_CHKE);
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_CAP, reg32);
|
2018-06-08 08:40:00 -05:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2024-11-02 12:45:37 -05:00
|
|
|
* pcie_set_ecrc_checking - set/unset PCIe ECRC checking for a device based
|
|
|
|
* on global policy
|
2018-06-08 08:40:00 -05:00
|
|
|
* @dev: the PCI device
|
|
|
|
*/
|
|
|
|
void pcie_set_ecrc_checking(struct pci_dev *dev)
|
|
|
|
{
|
2023-01-12 12:51:11 +05:30
|
|
|
if (!pcie_aer_is_native(dev))
|
|
|
|
return;
|
|
|
|
|
2018-06-08 08:40:00 -05:00
|
|
|
switch (ecrc_policy) {
|
|
|
|
case ECRC_POLICY_DEFAULT:
|
|
|
|
return;
|
|
|
|
case ECRC_POLICY_OFF:
|
|
|
|
disable_ecrc_checking(dev);
|
|
|
|
break;
|
|
|
|
case ECRC_POLICY_ON:
|
|
|
|
enable_ecrc_checking(dev);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* pcie_ecrc_get_policy - parse kernel command-line ecrc option
|
2019-08-27 18:18:23 +03:00
|
|
|
* @str: ECRC policy from kernel command line to use
|
2018-06-08 08:40:00 -05:00
|
|
|
*/
|
|
|
|
void pcie_ecrc_get_policy(char *str)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2019-01-28 13:57:28 +02:00
|
|
|
i = match_string(ecrc_policy_str, ARRAY_SIZE(ecrc_policy_str), str);
|
|
|
|
if (i < 0)
|
2018-06-08 08:40:00 -05:00
|
|
|
return;
|
|
|
|
|
|
|
|
ecrc_policy = i;
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_PCIE_ECRC */
|
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
#define PCI_EXP_AER_FLAGS (PCI_EXP_DEVCTL_CERE | PCI_EXP_DEVCTL_NFERE | \
|
|
|
|
PCI_EXP_DEVCTL_FERE | PCI_EXP_DEVCTL_URRE)
|
2018-06-08 08:39:45 -05:00
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
int pcie_aer_is_native(struct pci_dev *dev)
|
2018-06-08 08:39:45 -05:00
|
|
|
{
|
2020-05-26 16:18:29 -07:00
|
|
|
struct pci_host_bridge *host = pci_find_host_bridge(dev->bus);
|
2018-06-08 08:39:45 -05:00
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
if (!dev->aer_cap)
|
2018-07-03 18:27:43 -05:00
|
|
|
return 0;
|
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
return pcie_ports_native || host->native_aer;
|
2018-06-08 08:39:45 -05:00
|
|
|
}
|
module: Convert symbol namespace to string literal
Clean up the existing export namespace code along the same lines of
commit 33def8498fdd ("treewide: Convert macro and uses of __section(foo)
to __section("foo")") and for the same reason, it is not desired for the
namespace argument to be a macro expansion itself.
Scripted using
git grep -l -e MODULE_IMPORT_NS -e EXPORT_SYMBOL_NS | while read file;
do
awk -i inplace '
/^#define EXPORT_SYMBOL_NS/ {
gsub(/__stringify\(ns\)/, "ns");
print;
next;
}
/^#define MODULE_IMPORT_NS/ {
gsub(/__stringify\(ns\)/, "ns");
print;
next;
}
/MODULE_IMPORT_NS/ {
$0 = gensub(/MODULE_IMPORT_NS\(([^)]*)\)/, "MODULE_IMPORT_NS(\"\\1\")", "g");
}
/EXPORT_SYMBOL_NS/ {
if ($0 ~ /(EXPORT_SYMBOL_NS[^(]*)\(([^,]+),/) {
if ($0 !~ /(EXPORT_SYMBOL_NS[^(]*)\(([^,]+), ([^)]+)\)/ &&
$0 !~ /(EXPORT_SYMBOL_NS[^(]*)\(\)/ &&
$0 !~ /^my/) {
getline line;
gsub(/[[:space:]]*\\$/, "");
gsub(/[[:space:]]/, "", line);
$0 = $0 " " line;
}
$0 = gensub(/(EXPORT_SYMBOL_NS[^(]*)\(([^,]+), ([^)]+)\)/,
"\\1(\\2, \"\\3\")", "g");
}
}
{ print }' $file;
done
Requested-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://mail.google.com/mail/u/2/#inbox/FMfcgzQXKWgMmjdFwwdsfgxzKpVHWPlc
Acked-by: Greg KH <gregkh@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2024-12-02 15:59:47 +01:00
|
|
|
EXPORT_SYMBOL_NS_GPL(pcie_aer_is_native, "CXL");
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2023-07-10 18:21:36 -05:00
|
|
|
static int pci_enable_pcie_error_reporting(struct pci_dev *dev)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2020-06-15 09:32:18 +02:00
|
|
|
int rc;
|
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
if (!pcie_aer_is_native(dev))
|
2018-06-08 08:33:39 -05:00
|
|
|
return -EIO;
|
|
|
|
|
2020-06-15 09:32:18 +02:00
|
|
|
rc = pcie_capability_set_word(dev, PCI_EXP_DEVCTL, PCI_EXP_AER_FLAGS);
|
|
|
|
return pcibios_err_to_errno(rc);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
2020-03-23 17:26:08 -07:00
|
|
|
int pci_aer_clear_nonfatal_status(struct pci_dev *dev)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-07-19 17:58:05 -05:00
|
|
|
u32 status, sev;
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
if (!pcie_aer_is_native(dev))
|
2018-07-17 10:31:23 -05:00
|
|
|
return -EIO;
|
|
|
|
|
2018-07-19 17:58:05 -05:00
|
|
|
/* Clear status bits for ERR_NONFATAL errors only */
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, &status);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_SEVER, &sev);
|
2018-07-19 17:58:05 -05:00
|
|
|
status &= ~sev;
|
2018-06-08 08:33:39 -05:00
|
|
|
if (status)
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, status);
|
2018-06-08 08:33:39 -05:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2020-03-23 17:26:08 -07:00
|
|
|
EXPORT_SYMBOL_GPL(pci_aer_clear_nonfatal_status);
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2018-07-19 17:55:58 -05:00
|
|
|
void pci_aer_clear_fatal_status(struct pci_dev *dev)
|
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-07-19 17:55:58 -05:00
|
|
|
u32 status, sev;
|
|
|
|
|
2020-05-26 16:18:29 -07:00
|
|
|
if (!pcie_aer_is_native(dev))
|
2018-07-17 10:31:23 -05:00
|
|
|
return;
|
|
|
|
|
2018-07-19 17:55:58 -05:00
|
|
|
/* Clear status bits for ERR_FATAL errors only */
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, &status);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_SEVER, &sev);
|
2018-07-19 17:55:58 -05:00
|
|
|
status &= sev;
|
|
|
|
if (status)
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, status);
|
2018-07-19 17:55:58 -05:00
|
|
|
}
|
|
|
|
|
PCI/AER: Add pci_aer_raw_clear_status() to unconditionally clear Error Status
Per the SFI _OSC and DPC Updates ECN [1] implementation note flowchart, the
OS seems to be expected to clear AER status even if it doesn't have
ownership of the AER capability. Unlike the DPC capability, where a DPC
ECN [2] specifies a window when the OS is allowed to access DPC registers
even if it doesn't have ownership, there is no clear model for AER.
Add pci_aer_raw_clear_status() to clear the AER error status registers
unconditionally. This is intended for use only by the EDR path (see [2]).
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] Downstream Port Containment Related Enhancements ECN, Jan 28, 2019,
affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/12888
[bhelgaas: changelog]
Link: https://lore.kernel.org/r/c19ad28f3633cce67448609e89a75635da0da07d.1585000084.git.sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-03-23 17:26:05 -07:00
|
|
|
/**
|
|
|
|
* pci_aer_raw_clear_status - Clear AER error registers.
|
|
|
|
* @dev: the PCI device
|
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Clear AER error status registers unconditionally, regardless of
|
PCI/AER: Add pci_aer_raw_clear_status() to unconditionally clear Error Status
Per the SFI _OSC and DPC Updates ECN [1] implementation note flowchart, the
OS seems to be expected to clear AER status even if it doesn't have
ownership of the AER capability. Unlike the DPC capability, where a DPC
ECN [2] specifies a window when the OS is allowed to access DPC registers
even if it doesn't have ownership, there is no clear model for AER.
Add pci_aer_raw_clear_status() to clear the AER error status registers
unconditionally. This is intended for use only by the EDR path (see [2]).
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] Downstream Port Containment Related Enhancements ECN, Jan 28, 2019,
affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/12888
[bhelgaas: changelog]
Link: https://lore.kernel.org/r/c19ad28f3633cce67448609e89a75635da0da07d.1585000084.git.sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-03-23 17:26:05 -07:00
|
|
|
* whether they're owned by firmware or the OS.
|
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Return: 0 on success, or negative on failure.
|
PCI/AER: Add pci_aer_raw_clear_status() to unconditionally clear Error Status
Per the SFI _OSC and DPC Updates ECN [1] implementation note flowchart, the
OS seems to be expected to clear AER status even if it doesn't have
ownership of the AER capability. Unlike the DPC capability, where a DPC
ECN [2] specifies a window when the OS is allowed to access DPC registers
even if it doesn't have ownership, there is no clear model for AER.
Add pci_aer_raw_clear_status() to clear the AER error status registers
unconditionally. This is intended for use only by the EDR path (see [2]).
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] Downstream Port Containment Related Enhancements ECN, Jan 28, 2019,
affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/12888
[bhelgaas: changelog]
Link: https://lore.kernel.org/r/c19ad28f3633cce67448609e89a75635da0da07d.1585000084.git.sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-03-23 17:26:05 -07:00
|
|
|
*/
|
|
|
|
int pci_aer_raw_clear_status(struct pci_dev *dev)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-06-08 08:33:39 -05:00
|
|
|
u32 status;
|
|
|
|
int port_type;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2018-06-08 08:33:39 -05:00
|
|
|
return -EIO;
|
|
|
|
|
|
|
|
port_type = pci_pcie_type(dev);
|
2020-12-02 11:26:29 -06:00
|
|
|
if (port_type == PCI_EXP_TYPE_ROOT_PORT ||
|
|
|
|
port_type == PCI_EXP_TYPE_RC_EC) {
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_ROOT_STATUS, &status);
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_ROOT_STATUS, status);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_STATUS, &status);
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_COR_STATUS, status);
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, &status);
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, status);
|
2018-06-08 08:33:39 -05:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2020-03-23 17:26:08 -07:00
|
|
|
int pci_aer_clear_status(struct pci_dev *dev)
|
PCI/AER: Add pci_aer_raw_clear_status() to unconditionally clear Error Status
Per the SFI _OSC and DPC Updates ECN [1] implementation note flowchart, the
OS seems to be expected to clear AER status even if it doesn't have
ownership of the AER capability. Unlike the DPC capability, where a DPC
ECN [2] specifies a window when the OS is allowed to access DPC registers
even if it doesn't have ownership, there is no clear model for AER.
Add pci_aer_raw_clear_status() to clear the AER error status registers
unconditionally. This is intended for use only by the EDR path (see [2]).
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] Downstream Port Containment Related Enhancements ECN, Jan 28, 2019,
affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/12888
[bhelgaas: changelog]
Link: https://lore.kernel.org/r/c19ad28f3633cce67448609e89a75635da0da07d.1585000084.git.sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-03-23 17:26:05 -07:00
|
|
|
{
|
2020-05-26 16:18:29 -07:00
|
|
|
if (!pcie_aer_is_native(dev))
|
PCI/AER: Add pci_aer_raw_clear_status() to unconditionally clear Error Status
Per the SFI _OSC and DPC Updates ECN [1] implementation note flowchart, the
OS seems to be expected to clear AER status even if it doesn't have
ownership of the AER capability. Unlike the DPC capability, where a DPC
ECN [2] specifies a window when the OS is allowed to access DPC registers
even if it doesn't have ownership, there is no clear model for AER.
Add pci_aer_raw_clear_status() to clear the AER error status registers
unconditionally. This is intended for use only by the EDR path (see [2]).
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] Downstream Port Containment Related Enhancements ECN, Jan 28, 2019,
affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/12888
[bhelgaas: changelog]
Link: https://lore.kernel.org/r/c19ad28f3633cce67448609e89a75635da0da07d.1585000084.git.sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-03-23 17:26:05 -07:00
|
|
|
return -EIO;
|
|
|
|
|
|
|
|
return pci_aer_raw_clear_status(dev);
|
|
|
|
}
|
|
|
|
|
2019-10-18 16:52:21 +00:00
|
|
|
void pci_save_aer_state(struct pci_dev *dev)
|
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2019-10-18 16:52:21 +00:00
|
|
|
struct pci_cap_saved_state *save_state;
|
|
|
|
u32 *cap;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2019-10-18 16:52:21 +00:00
|
|
|
return;
|
|
|
|
|
|
|
|
save_state = pci_find_saved_ext_cap(dev, PCI_EXT_CAP_ID_ERR);
|
|
|
|
if (!save_state)
|
|
|
|
return;
|
|
|
|
|
|
|
|
cap = &save_state->cap.data[0];
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_MASK, cap++);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_SEVER, cap++);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_MASK, cap++);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_CAP, cap++);
|
2019-10-18 16:52:21 +00:00
|
|
|
if (pcie_cap_has_rtctl(dev))
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_ROOT_COMMAND, cap++);
|
2019-10-18 16:52:21 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void pci_restore_aer_state(struct pci_dev *dev)
|
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2019-10-18 16:52:21 +00:00
|
|
|
struct pci_cap_saved_state *save_state;
|
|
|
|
u32 *cap;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2019-10-18 16:52:21 +00:00
|
|
|
return;
|
|
|
|
|
|
|
|
save_state = pci_find_saved_ext_cap(dev, PCI_EXT_CAP_ID_ERR);
|
|
|
|
if (!save_state)
|
|
|
|
return;
|
|
|
|
|
|
|
|
cap = &save_state->cap.data[0];
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_UNCOR_MASK, *cap++);
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_UNCOR_SEVER, *cap++);
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_COR_MASK, *cap++);
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_CAP, *cap++);
|
2019-10-18 16:52:21 +00:00
|
|
|
if (pcie_cap_has_rtctl(dev))
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_ROOT_COMMAND, *cap++);
|
2019-10-18 16:52:21 +00:00
|
|
|
}
|
|
|
|
|
2018-06-21 16:48:26 -07:00
|
|
|
void pci_aer_init(struct pci_dev *dev)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2019-10-18 16:52:21 +00:00
|
|
|
int n;
|
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
dev->aer_cap = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR);
|
2019-10-18 16:52:21 +00:00
|
|
|
if (!dev->aer_cap)
|
|
|
|
return;
|
2018-06-30 15:07:17 -05:00
|
|
|
|
2025-05-22 18:21:21 -05:00
|
|
|
dev->aer_info = kzalloc(sizeof(*dev->aer_info), GFP_KERNEL);
|
2019-10-18 16:52:21 +00:00
|
|
|
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
ratelimit_state_init(&dev->aer_info->correctable_ratelimit,
|
|
|
|
DEFAULT_RATELIMIT_INTERVAL, DEFAULT_RATELIMIT_BURST);
|
|
|
|
ratelimit_state_init(&dev->aer_info->nonfatal_ratelimit,
|
|
|
|
DEFAULT_RATELIMIT_INTERVAL, DEFAULT_RATELIMIT_BURST);
|
|
|
|
|
2019-10-18 16:52:21 +00:00
|
|
|
/*
|
|
|
|
* We save/restore PCI_ERR_UNCOR_MASK, PCI_ERR_UNCOR_SEVER,
|
|
|
|
* PCI_ERR_COR_MASK, and PCI_ERR_CAP. Root and Root Complex Event
|
2025-03-07 17:17:15 -06:00
|
|
|
* Collectors also implement PCI_ERR_ROOT_COMMAND (PCIe r6.0, sec
|
|
|
|
* 7.8.4.9).
|
2019-10-18 16:52:21 +00:00
|
|
|
*/
|
|
|
|
n = pcie_cap_has_rtctl(dev) ? 5 : 4;
|
|
|
|
pci_add_ext_cap_save_buffer(dev, PCI_EXT_CAP_ID_ERR, sizeof(u32) * n);
|
2018-06-30 15:07:17 -05:00
|
|
|
|
2020-03-23 17:26:08 -07:00
|
|
|
pci_aer_clear_status(dev);
|
2022-01-25 08:18:18 +01:00
|
|
|
|
2022-01-25 08:18:20 +01:00
|
|
|
if (pci_aer_available())
|
|
|
|
pci_enable_pcie_error_reporting(dev);
|
|
|
|
|
2022-01-25 08:18:18 +01:00
|
|
|
pcie_set_ecrc_checking(dev);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
2018-06-30 15:07:17 -05:00
|
|
|
void pci_aer_exit(struct pci_dev *dev)
|
|
|
|
{
|
2025-05-22 18:21:21 -05:00
|
|
|
kfree(dev->aer_info);
|
|
|
|
dev->aer_info = NULL;
|
2018-06-30 15:07:17 -05:00
|
|
|
}
|
|
|
|
|
2018-06-08 08:39:38 -05:00
|
|
|
#define AER_AGENT_RECEIVER 0
|
|
|
|
#define AER_AGENT_REQUESTER 1
|
|
|
|
#define AER_AGENT_COMPLETER 2
|
|
|
|
#define AER_AGENT_TRANSMITTER 3
|
|
|
|
|
|
|
|
#define AER_AGENT_REQUESTER_MASK(t) ((t == AER_CORRECTABLE) ? \
|
|
|
|
0 : (PCI_ERR_UNC_COMP_TIME|PCI_ERR_UNC_UNSUP))
|
|
|
|
#define AER_AGENT_COMPLETER_MASK(t) ((t == AER_CORRECTABLE) ? \
|
|
|
|
0 : PCI_ERR_UNC_COMP_ABORT)
|
|
|
|
#define AER_AGENT_TRANSMITTER_MASK(t) ((t == AER_CORRECTABLE) ? \
|
|
|
|
(PCI_ERR_COR_REP_ROLL|PCI_ERR_COR_REP_TIMER) : 0)
|
|
|
|
|
|
|
|
#define AER_GET_AGENT(t, e) \
|
|
|
|
((e & AER_AGENT_COMPLETER_MASK(t)) ? AER_AGENT_COMPLETER : \
|
|
|
|
(e & AER_AGENT_REQUESTER_MASK(t)) ? AER_AGENT_REQUESTER : \
|
|
|
|
(e & AER_AGENT_TRANSMITTER_MASK(t)) ? AER_AGENT_TRANSMITTER : \
|
|
|
|
AER_AGENT_RECEIVER)
|
|
|
|
|
|
|
|
#define AER_PHYSICAL_LAYER_ERROR 0
|
|
|
|
#define AER_DATA_LINK_LAYER_ERROR 1
|
|
|
|
#define AER_TRANSACTION_LAYER_ERROR 2
|
|
|
|
|
|
|
|
#define AER_PHYSICAL_LAYER_ERROR_MASK(t) ((t == AER_CORRECTABLE) ? \
|
|
|
|
PCI_ERR_COR_RCVR : 0)
|
|
|
|
#define AER_DATA_LINK_LAYER_ERROR_MASK(t) ((t == AER_CORRECTABLE) ? \
|
|
|
|
(PCI_ERR_COR_BAD_TLP| \
|
|
|
|
PCI_ERR_COR_BAD_DLLP| \
|
|
|
|
PCI_ERR_COR_REP_ROLL| \
|
|
|
|
PCI_ERR_COR_REP_TIMER) : PCI_ERR_UNC_DLP)
|
|
|
|
|
|
|
|
#define AER_GET_LAYER_ERROR(t, e) \
|
|
|
|
((e & AER_PHYSICAL_LAYER_ERROR_MASK(t)) ? AER_PHYSICAL_LAYER_ERROR : \
|
|
|
|
(e & AER_DATA_LINK_LAYER_ERROR_MASK(t)) ? AER_DATA_LINK_LAYER_ERROR : \
|
|
|
|
AER_TRANSACTION_LAYER_ERROR)
|
|
|
|
|
|
|
|
/*
|
|
|
|
* AER error strings
|
|
|
|
*/
|
2023-12-06 16:42:29 -06:00
|
|
|
static const char * const aer_error_severity_string[] = {
|
|
|
|
"Uncorrectable (Non-Fatal)",
|
|
|
|
"Uncorrectable (Fatal)",
|
|
|
|
"Correctable"
|
2018-06-08 08:39:38 -05:00
|
|
|
};
|
|
|
|
|
|
|
|
static const char *aer_error_layer[] = {
|
|
|
|
"Physical Layer",
|
|
|
|
"Data Link Layer",
|
|
|
|
"Transaction Layer"
|
|
|
|
};
|
|
|
|
|
2020-07-07 18:48:17 -05:00
|
|
|
static const char *aer_correctable_error_string[] = {
|
2018-06-26 11:44:15 -04:00
|
|
|
"RxErr", /* Bit Position 0 */
|
2018-06-08 08:39:38 -05:00
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
2018-06-26 11:44:15 -04:00
|
|
|
"BadTLP", /* Bit Position 6 */
|
|
|
|
"BadDLLP", /* Bit Position 7 */
|
|
|
|
"Rollover", /* Bit Position 8 */
|
2018-06-08 08:39:38 -05:00
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
2018-06-26 11:44:15 -04:00
|
|
|
"Timeout", /* Bit Position 12 */
|
|
|
|
"NonFatalErr", /* Bit Position 13 */
|
|
|
|
"CorrIntErr", /* Bit Position 14 */
|
|
|
|
"HeaderOF", /* Bit Position 15 */
|
2020-07-07 18:48:17 -05:00
|
|
|
NULL, /* Bit Position 16 */
|
|
|
|
NULL, /* Bit Position 17 */
|
|
|
|
NULL, /* Bit Position 18 */
|
|
|
|
NULL, /* Bit Position 19 */
|
|
|
|
NULL, /* Bit Position 20 */
|
|
|
|
NULL, /* Bit Position 21 */
|
|
|
|
NULL, /* Bit Position 22 */
|
|
|
|
NULL, /* Bit Position 23 */
|
|
|
|
NULL, /* Bit Position 24 */
|
|
|
|
NULL, /* Bit Position 25 */
|
|
|
|
NULL, /* Bit Position 26 */
|
|
|
|
NULL, /* Bit Position 27 */
|
|
|
|
NULL, /* Bit Position 28 */
|
|
|
|
NULL, /* Bit Position 29 */
|
|
|
|
NULL, /* Bit Position 30 */
|
|
|
|
NULL, /* Bit Position 31 */
|
2018-06-08 08:39:38 -05:00
|
|
|
};
|
|
|
|
|
2020-07-07 18:48:17 -05:00
|
|
|
static const char *aer_uncorrectable_error_string[] = {
|
2018-06-08 08:39:38 -05:00
|
|
|
"Undefined", /* Bit Position 0 */
|
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
2018-06-26 11:44:15 -04:00
|
|
|
"DLP", /* Bit Position 4 */
|
|
|
|
"SDES", /* Bit Position 5 */
|
2018-06-08 08:39:38 -05:00
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
|
|
|
NULL,
|
2018-06-26 11:44:15 -04:00
|
|
|
"TLP", /* Bit Position 12 */
|
|
|
|
"FCP", /* Bit Position 13 */
|
|
|
|
"CmpltTO", /* Bit Position 14 */
|
|
|
|
"CmpltAbrt", /* Bit Position 15 */
|
|
|
|
"UnxCmplt", /* Bit Position 16 */
|
|
|
|
"RxOF", /* Bit Position 17 */
|
|
|
|
"MalfTLP", /* Bit Position 18 */
|
2018-06-08 08:39:38 -05:00
|
|
|
"ECRC", /* Bit Position 19 */
|
2018-06-26 11:44:15 -04:00
|
|
|
"UnsupReq", /* Bit Position 20 */
|
|
|
|
"ACSViol", /* Bit Position 21 */
|
|
|
|
"UncorrIntErr", /* Bit Position 22 */
|
|
|
|
"BlockedTLP", /* Bit Position 23 */
|
|
|
|
"AtomicOpBlocked", /* Bit Position 24 */
|
|
|
|
"TLPBlockedErr", /* Bit Position 25 */
|
2019-08-27 15:21:44 -07:00
|
|
|
"PoisonTLPBlocked", /* Bit Position 26 */
|
2020-07-07 18:48:17 -05:00
|
|
|
NULL, /* Bit Position 27 */
|
|
|
|
NULL, /* Bit Position 28 */
|
|
|
|
NULL, /* Bit Position 29 */
|
|
|
|
NULL, /* Bit Position 30 */
|
|
|
|
NULL, /* Bit Position 31 */
|
2018-06-08 08:39:38 -05:00
|
|
|
};
|
|
|
|
|
|
|
|
static const char *aer_agent_string[] = {
|
|
|
|
"Receiver ID",
|
|
|
|
"Requester ID",
|
|
|
|
"Completer ID",
|
|
|
|
"Transmitter ID"
|
|
|
|
};
|
|
|
|
|
2018-06-21 16:48:28 -07:00
|
|
|
#define aer_stats_dev_attr(name, stats_array, strings_array, \
|
|
|
|
total_string, total_field) \
|
|
|
|
static ssize_t \
|
|
|
|
name##_show(struct device *dev, struct device_attribute *attr, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
unsigned int i; \
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev); \
|
2025-05-22 18:21:21 -05:00
|
|
|
u64 *stats = pdev->aer_info->stats_array; \
|
2021-06-03 00:01:07 +00:00
|
|
|
size_t len = 0; \
|
2018-06-21 16:48:28 -07:00
|
|
|
\
|
2025-05-22 18:21:21 -05:00
|
|
|
for (i = 0; i < ARRAY_SIZE(pdev->aer_info->stats_array); i++) { \
|
2018-06-21 16:48:28 -07:00
|
|
|
if (strings_array[i]) \
|
2021-06-03 00:01:07 +00:00
|
|
|
len += sysfs_emit_at(buf, len, "%s %llu\n", \
|
|
|
|
strings_array[i], \
|
|
|
|
stats[i]); \
|
2018-06-21 16:48:28 -07:00
|
|
|
else if (stats[i]) \
|
2021-06-03 00:01:07 +00:00
|
|
|
len += sysfs_emit_at(buf, len, \
|
|
|
|
#stats_array "_bit[%d] %llu\n",\
|
|
|
|
i, stats[i]); \
|
2018-06-21 16:48:28 -07:00
|
|
|
} \
|
2021-06-03 00:01:07 +00:00
|
|
|
len += sysfs_emit_at(buf, len, "TOTAL_%s %llu\n", total_string, \
|
2025-05-22 18:21:21 -05:00
|
|
|
pdev->aer_info->total_field); \
|
2021-06-03 00:01:07 +00:00
|
|
|
return len; \
|
2018-06-21 16:48:28 -07:00
|
|
|
} \
|
|
|
|
static DEVICE_ATTR_RO(name)
|
|
|
|
|
|
|
|
aer_stats_dev_attr(aer_dev_correctable, dev_cor_errs,
|
|
|
|
aer_correctable_error_string, "ERR_COR",
|
|
|
|
dev_total_cor_errs);
|
|
|
|
aer_stats_dev_attr(aer_dev_fatal, dev_fatal_errs,
|
|
|
|
aer_uncorrectable_error_string, "ERR_FATAL",
|
|
|
|
dev_total_fatal_errs);
|
|
|
|
aer_stats_dev_attr(aer_dev_nonfatal, dev_nonfatal_errs,
|
|
|
|
aer_uncorrectable_error_string, "ERR_NONFATAL",
|
|
|
|
dev_total_nonfatal_errs);
|
|
|
|
|
2018-06-21 16:48:29 -07:00
|
|
|
#define aer_stats_rootport_attr(name, field) \
|
|
|
|
static ssize_t \
|
|
|
|
name##_show(struct device *dev, struct device_attribute *attr, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev); \
|
2025-05-22 18:21:21 -05:00
|
|
|
return sysfs_emit(buf, "%llu\n", pdev->aer_info->field); \
|
2018-06-21 16:48:29 -07:00
|
|
|
} \
|
|
|
|
static DEVICE_ATTR_RO(name)
|
|
|
|
|
|
|
|
aer_stats_rootport_attr(aer_rootport_total_err_cor,
|
|
|
|
rootport_total_cor_errs);
|
|
|
|
aer_stats_rootport_attr(aer_rootport_total_err_fatal,
|
|
|
|
rootport_total_fatal_errs);
|
|
|
|
aer_stats_rootport_attr(aer_rootport_total_err_nonfatal,
|
|
|
|
rootport_total_nonfatal_errs);
|
|
|
|
|
2018-06-21 16:48:28 -07:00
|
|
|
static struct attribute *aer_stats_attrs[] __ro_after_init = {
|
|
|
|
&dev_attr_aer_dev_correctable.attr,
|
|
|
|
&dev_attr_aer_dev_fatal.attr,
|
|
|
|
&dev_attr_aer_dev_nonfatal.attr,
|
2018-06-21 16:48:29 -07:00
|
|
|
&dev_attr_aer_rootport_total_err_cor.attr,
|
|
|
|
&dev_attr_aer_rootport_total_err_fatal.attr,
|
|
|
|
&dev_attr_aer_rootport_total_err_nonfatal.attr,
|
2018-06-21 16:48:28 -07:00
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
|
|
|
static umode_t aer_stats_attrs_are_visible(struct kobject *kobj,
|
|
|
|
struct attribute *a, int n)
|
|
|
|
{
|
|
|
|
struct device *dev = kobj_to_dev(kobj);
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev);
|
|
|
|
|
2025-05-22 18:21:21 -05:00
|
|
|
if (!pdev->aer_info)
|
2018-06-21 16:48:28 -07:00
|
|
|
return 0;
|
|
|
|
|
2018-06-21 16:48:29 -07:00
|
|
|
if ((a == &dev_attr_aer_rootport_total_err_cor.attr ||
|
|
|
|
a == &dev_attr_aer_rootport_total_err_fatal.attr ||
|
|
|
|
a == &dev_attr_aer_rootport_total_err_nonfatal.attr) &&
|
2020-12-02 11:26:29 -06:00
|
|
|
((pci_pcie_type(pdev) != PCI_EXP_TYPE_ROOT_PORT) &&
|
|
|
|
(pci_pcie_type(pdev) != PCI_EXP_TYPE_RC_EC)))
|
2018-06-21 16:48:29 -07:00
|
|
|
return 0;
|
|
|
|
|
2018-06-21 16:48:28 -07:00
|
|
|
return a->mode;
|
|
|
|
}
|
|
|
|
|
|
|
|
const struct attribute_group aer_stats_attr_group = {
|
|
|
|
.attrs = aer_stats_attrs,
|
|
|
|
.is_visible = aer_stats_attrs_are_visible,
|
|
|
|
};
|
|
|
|
|
2025-05-22 18:21:26 -05:00
|
|
|
/*
|
|
|
|
* Ratelimit interval
|
|
|
|
* <=0: disabled with ratelimit.interval = 0
|
|
|
|
* >0: enabled with ratelimit.interval in ms
|
|
|
|
*/
|
|
|
|
#define aer_ratelimit_interval_attr(name, ratelimit) \
|
|
|
|
static ssize_t \
|
|
|
|
name##_show(struct device *dev, struct device_attribute *attr, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev); \
|
|
|
|
\
|
|
|
|
return sysfs_emit(buf, "%d\n", \
|
|
|
|
pdev->aer_info->ratelimit.interval); \
|
|
|
|
} \
|
|
|
|
\
|
|
|
|
static ssize_t \
|
|
|
|
name##_store(struct device *dev, struct device_attribute *attr, \
|
|
|
|
const char *buf, size_t count) \
|
|
|
|
{ \
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev); \
|
|
|
|
int interval; \
|
|
|
|
\
|
|
|
|
if (!capable(CAP_SYS_ADMIN)) \
|
|
|
|
return -EPERM; \
|
|
|
|
\
|
|
|
|
if (kstrtoint(buf, 0, &interval) < 0) \
|
|
|
|
return -EINVAL; \
|
|
|
|
\
|
|
|
|
if (interval <= 0) \
|
|
|
|
interval = 0; \
|
|
|
|
else \
|
|
|
|
interval = msecs_to_jiffies(interval); \
|
|
|
|
\
|
|
|
|
pdev->aer_info->ratelimit.interval = interval; \
|
|
|
|
\
|
|
|
|
return count; \
|
|
|
|
} \
|
|
|
|
static DEVICE_ATTR_RW(name);
|
|
|
|
|
|
|
|
#define aer_ratelimit_burst_attr(name, ratelimit) \
|
|
|
|
static ssize_t \
|
|
|
|
name##_show(struct device *dev, struct device_attribute *attr, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev); \
|
|
|
|
\
|
|
|
|
return sysfs_emit(buf, "%d\n", \
|
|
|
|
pdev->aer_info->ratelimit.burst); \
|
|
|
|
} \
|
|
|
|
\
|
|
|
|
static ssize_t \
|
|
|
|
name##_store(struct device *dev, struct device_attribute *attr, \
|
|
|
|
const char *buf, size_t count) \
|
|
|
|
{ \
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev); \
|
|
|
|
int burst; \
|
|
|
|
\
|
|
|
|
if (!capable(CAP_SYS_ADMIN)) \
|
|
|
|
return -EPERM; \
|
|
|
|
\
|
|
|
|
if (kstrtoint(buf, 0, &burst) < 0) \
|
|
|
|
return -EINVAL; \
|
|
|
|
\
|
|
|
|
pdev->aer_info->ratelimit.burst = burst; \
|
|
|
|
\
|
|
|
|
return count; \
|
|
|
|
} \
|
|
|
|
static DEVICE_ATTR_RW(name);
|
|
|
|
|
|
|
|
#define aer_ratelimit_attrs(name) \
|
|
|
|
aer_ratelimit_interval_attr(name##_ratelimit_interval_ms, \
|
|
|
|
name##_ratelimit) \
|
|
|
|
aer_ratelimit_burst_attr(name##_ratelimit_burst, \
|
|
|
|
name##_ratelimit)
|
|
|
|
|
|
|
|
aer_ratelimit_attrs(correctable)
|
|
|
|
aer_ratelimit_attrs(nonfatal)
|
|
|
|
|
|
|
|
static struct attribute *aer_attrs[] = {
|
|
|
|
&dev_attr_correctable_ratelimit_interval_ms.attr,
|
|
|
|
&dev_attr_correctable_ratelimit_burst.attr,
|
|
|
|
&dev_attr_nonfatal_ratelimit_interval_ms.attr,
|
|
|
|
&dev_attr_nonfatal_ratelimit_burst.attr,
|
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
|
|
|
static umode_t aer_attrs_are_visible(struct kobject *kobj,
|
|
|
|
struct attribute *a, int n)
|
|
|
|
{
|
|
|
|
struct device *dev = kobj_to_dev(kobj);
|
|
|
|
struct pci_dev *pdev = to_pci_dev(dev);
|
|
|
|
|
|
|
|
if (!pdev->aer_info)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
return a->mode;
|
|
|
|
}
|
|
|
|
|
|
|
|
const struct attribute_group aer_attr_group = {
|
|
|
|
.name = "aer",
|
|
|
|
.attrs = aer_attrs,
|
|
|
|
.is_visible = aer_attrs_are_visible,
|
|
|
|
};
|
|
|
|
|
2018-06-21 16:48:28 -07:00
|
|
|
static void pci_dev_aer_stats_incr(struct pci_dev *pdev,
|
|
|
|
struct aer_err_info *info)
|
|
|
|
{
|
2019-08-27 18:18:22 +03:00
|
|
|
unsigned long status = info->status & ~info->mask;
|
|
|
|
int i, max = -1;
|
2018-06-21 16:48:28 -07:00
|
|
|
u64 *counter = NULL;
|
2025-05-22 18:21:21 -05:00
|
|
|
struct aer_info *aer_info = pdev->aer_info;
|
2018-06-21 16:48:28 -07:00
|
|
|
|
2025-05-22 18:21:21 -05:00
|
|
|
if (!aer_info)
|
2018-06-21 16:48:28 -07:00
|
|
|
return;
|
|
|
|
|
|
|
|
switch (info->severity) {
|
|
|
|
case AER_CORRECTABLE:
|
2025-05-22 18:21:21 -05:00
|
|
|
aer_info->dev_total_cor_errs++;
|
|
|
|
counter = &aer_info->dev_cor_errs[0];
|
2018-06-21 16:48:28 -07:00
|
|
|
max = AER_MAX_TYPEOF_COR_ERRS;
|
|
|
|
break;
|
|
|
|
case AER_NONFATAL:
|
2025-05-22 18:21:21 -05:00
|
|
|
aer_info->dev_total_nonfatal_errs++;
|
|
|
|
counter = &aer_info->dev_nonfatal_errs[0];
|
2018-06-21 16:48:28 -07:00
|
|
|
max = AER_MAX_TYPEOF_UNCOR_ERRS;
|
|
|
|
break;
|
|
|
|
case AER_FATAL:
|
2025-05-22 18:21:21 -05:00
|
|
|
aer_info->dev_total_fatal_errs++;
|
|
|
|
counter = &aer_info->dev_fatal_errs[0];
|
2018-06-21 16:48:28 -07:00
|
|
|
max = AER_MAX_TYPEOF_UNCOR_ERRS;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2019-08-27 18:18:22 +03:00
|
|
|
for_each_set_bit(i, &status, max)
|
|
|
|
counter[i]++;
|
2018-06-21 16:48:28 -07:00
|
|
|
}
|
|
|
|
|
2018-06-21 16:48:29 -07:00
|
|
|
static void pci_rootport_aer_stats_incr(struct pci_dev *pdev,
|
|
|
|
struct aer_err_source *e_src)
|
|
|
|
{
|
2025-05-22 18:21:21 -05:00
|
|
|
struct aer_info *aer_info = pdev->aer_info;
|
2018-06-21 16:48:29 -07:00
|
|
|
|
2025-05-22 18:21:21 -05:00
|
|
|
if (!aer_info)
|
2018-06-21 16:48:29 -07:00
|
|
|
return;
|
|
|
|
|
|
|
|
if (e_src->status & PCI_ERR_ROOT_COR_RCV)
|
2025-05-22 18:21:21 -05:00
|
|
|
aer_info->rootport_total_cor_errs++;
|
2018-06-21 16:48:29 -07:00
|
|
|
|
|
|
|
if (e_src->status & PCI_ERR_ROOT_UNCOR_RCV) {
|
|
|
|
if (e_src->status & PCI_ERR_ROOT_FATAL_RCV)
|
2025-05-22 18:21:21 -05:00
|
|
|
aer_info->rootport_total_fatal_errs++;
|
2018-06-21 16:48:29 -07:00
|
|
|
else
|
2025-05-22 18:21:21 -05:00
|
|
|
aer_info->rootport_total_nonfatal_errs++;
|
2018-06-21 16:48:29 -07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
static int aer_ratelimit(struct pci_dev *dev, unsigned int severity)
|
|
|
|
{
|
|
|
|
switch (severity) {
|
|
|
|
case AER_NONFATAL:
|
|
|
|
return __ratelimit(&dev->aer_info->nonfatal_ratelimit);
|
|
|
|
case AER_CORRECTABLE:
|
|
|
|
return __ratelimit(&dev->aer_info->correctable_ratelimit);
|
|
|
|
default:
|
|
|
|
return 1; /* Don't ratelimit fatal errors */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2025-05-22 18:21:18 -05:00
|
|
|
static void __aer_print_error(struct pci_dev *dev, struct aer_err_info *info)
|
2018-06-08 08:39:38 -05:00
|
|
|
{
|
2020-07-07 18:48:17 -05:00
|
|
|
const char **strings;
|
2019-08-27 18:18:22 +03:00
|
|
|
unsigned long status = info->status & ~info->mask;
|
2025-05-22 18:21:18 -05:00
|
|
|
const char *level = info->level;
|
|
|
|
const char *errmsg;
|
2019-08-27 18:18:22 +03:00
|
|
|
int i;
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2025-05-22 18:21:18 -05:00
|
|
|
if (info->severity == AER_CORRECTABLE)
|
2020-07-07 18:48:17 -05:00
|
|
|
strings = aer_correctable_error_string;
|
2025-05-22 18:21:18 -05:00
|
|
|
else
|
2020-07-07 18:48:17 -05:00
|
|
|
strings = aer_uncorrectable_error_string;
|
|
|
|
|
2019-08-27 18:18:22 +03:00
|
|
|
for_each_set_bit(i, &status, 32) {
|
2020-07-07 18:48:17 -05:00
|
|
|
errmsg = strings[i];
|
|
|
|
if (!errmsg)
|
|
|
|
errmsg = "Unknown Error Bit";
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2024-12-16 18:10:12 +02:00
|
|
|
aer_printk(level, dev, " [%2d] %-22s%s\n", i, errmsg,
|
2018-06-08 08:39:38 -05:00
|
|
|
info->first_error == i ? " (First)" : "");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2025-05-22 18:21:13 -05:00
|
|
|
static void aer_print_source(struct pci_dev *dev, struct aer_err_info *info,
|
|
|
|
bool found)
|
|
|
|
{
|
|
|
|
u16 source = info->id;
|
|
|
|
|
|
|
|
pci_info(dev, "%s%s error message received from %04x:%02x:%02x.%d%s\n",
|
|
|
|
info->multi_error_valid ? "Multiple " : "",
|
|
|
|
aer_error_severity_string[info->severity],
|
|
|
|
pci_domain_nr(dev->bus), PCI_BUS_NUM(source),
|
|
|
|
PCI_SLOT(source), PCI_FUNC(source),
|
|
|
|
found ? "" : " (no details found");
|
|
|
|
}
|
|
|
|
|
2025-05-22 18:21:22 -05:00
|
|
|
void aer_print_error(struct aer_err_info *info, int i)
|
2018-06-08 08:39:38 -05:00
|
|
|
{
|
2025-05-22 18:21:22 -05:00
|
|
|
struct pci_dev *dev;
|
|
|
|
int layer, agent, id;
|
2025-05-22 18:21:18 -05:00
|
|
|
const char *level = info->level;
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2025-05-22 18:21:22 -05:00
|
|
|
if (WARN_ON_ONCE(i >= AER_MAX_MULTI_ERR_DEVICES))
|
|
|
|
return;
|
|
|
|
|
|
|
|
dev = info->dev[i];
|
|
|
|
id = pci_dev_id(dev);
|
|
|
|
|
2025-05-22 18:21:16 -05:00
|
|
|
pci_dev_aer_stats_incr(dev, info);
|
2025-05-22 18:21:17 -05:00
|
|
|
trace_aer_event(pci_name(dev), (info->status & ~info->mask),
|
|
|
|
info->severity, info->tlp_header_valid, &info->tlp);
|
2025-05-22 18:21:16 -05:00
|
|
|
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
if (!info->ratelimit_print[i])
|
|
|
|
return;
|
|
|
|
|
2018-06-08 08:39:38 -05:00
|
|
|
if (!info->status) {
|
|
|
|
pci_err(dev, "PCIe Bus Error: severity=%s, type=Inaccessible, (Unregistered Agent ID)\n",
|
|
|
|
aer_error_severity_string[info->severity]);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
layer = AER_GET_LAYER_ERROR(info->severity, info->status);
|
|
|
|
agent = AER_GET_AGENT(info->severity, info->status);
|
|
|
|
|
2024-12-16 18:10:12 +02:00
|
|
|
aer_printk(level, dev, "PCIe Bus Error: severity=%s, type=%s, (%s)\n",
|
2020-06-19 01:55:11 +10:00
|
|
|
aer_error_severity_string[info->severity],
|
|
|
|
aer_error_layer[layer], aer_agent_string[agent]);
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2024-12-16 18:10:12 +02:00
|
|
|
aer_printk(level, dev, " device [%04x:%04x] error status/mask=%08x/%08x\n",
|
2020-06-19 01:55:11 +10:00
|
|
|
dev->vendor, dev->device, info->status, info->mask);
|
2018-06-08 08:39:38 -05:00
|
|
|
|
|
|
|
__aer_print_error(dev, info);
|
|
|
|
|
|
|
|
if (info->tlp_header_valid)
|
2025-05-22 18:21:19 -05:00
|
|
|
pcie_print_tlp_log(dev, &info->tlp, level, dev_fmt(" "));
|
2018-06-08 08:39:38 -05:00
|
|
|
|
|
|
|
out:
|
|
|
|
if (info->id && info->error_dev_num > 1 && info->id == id)
|
|
|
|
pci_err(dev, " Error of this Agent is reported first\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_ACPI_APEI_PCIEAER
|
|
|
|
int cper_severity_to_aer(int cper_severity)
|
|
|
|
{
|
|
|
|
switch (cper_severity) {
|
|
|
|
case CPER_SEV_RECOVERABLE:
|
|
|
|
return AER_NONFATAL;
|
|
|
|
case CPER_SEV_FATAL:
|
|
|
|
return AER_FATAL;
|
|
|
|
default:
|
|
|
|
return AER_CORRECTABLE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(cper_severity_to_aer);
|
2023-10-18 19:17:05 +02:00
|
|
|
#endif
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2023-10-18 19:17:05 +02:00
|
|
|
void pci_print_aer(struct pci_dev *dev, int aer_severity,
|
|
|
|
struct aer_capability_regs *aer)
|
2018-06-08 08:39:38 -05:00
|
|
|
{
|
|
|
|
int layer, agent, tlp_header_valid = 0;
|
|
|
|
u32 status, mask;
|
2025-05-22 18:21:15 -05:00
|
|
|
struct aer_err_info info = {
|
|
|
|
.severity = aer_severity,
|
|
|
|
.first_error = PCI_ERR_CAP_FEP(aer->cap_control),
|
|
|
|
};
|
2018-06-08 08:39:38 -05:00
|
|
|
|
|
|
|
if (aer_severity == AER_CORRECTABLE) {
|
|
|
|
status = aer->cor_status;
|
|
|
|
mask = aer->cor_mask;
|
2025-05-22 18:21:18 -05:00
|
|
|
info.level = KERN_WARNING;
|
2018-06-08 08:39:38 -05:00
|
|
|
} else {
|
|
|
|
status = aer->uncor_status;
|
|
|
|
mask = aer->uncor_mask;
|
2025-05-22 18:21:18 -05:00
|
|
|
info.level = KERN_ERR;
|
2018-06-08 08:39:38 -05:00
|
|
|
tlp_header_valid = status & AER_LOG_TLP_MASKS;
|
|
|
|
}
|
|
|
|
|
|
|
|
info.status = status;
|
|
|
|
info.mask = mask;
|
2025-05-22 18:21:15 -05:00
|
|
|
|
2025-05-22 18:21:16 -05:00
|
|
|
pci_dev_aer_stats_incr(dev, &info);
|
2025-05-22 18:21:17 -05:00
|
|
|
trace_aer_event(pci_name(dev), (status & ~mask),
|
|
|
|
aer_severity, tlp_header_valid, &aer->header_log);
|
2025-05-22 18:21:16 -05:00
|
|
|
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
if (!aer_ratelimit(dev, info.severity))
|
|
|
|
return;
|
|
|
|
|
2025-05-22 18:21:15 -05:00
|
|
|
layer = AER_GET_LAYER_ERROR(aer_severity, status);
|
|
|
|
agent = AER_GET_AGENT(aer_severity, status);
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2025-05-22 18:21:20 -05:00
|
|
|
aer_printk(info.level, dev, "aer_status: 0x%08x, aer_mask: 0x%08x\n",
|
|
|
|
status, mask);
|
2018-06-08 08:39:38 -05:00
|
|
|
__aer_print_error(dev, &info);
|
2025-05-22 18:21:20 -05:00
|
|
|
aer_printk(info.level, dev, "aer_layer=%s, aer_agent=%s\n",
|
|
|
|
aer_error_layer[layer], aer_agent_string[agent]);
|
2018-06-08 08:39:38 -05:00
|
|
|
|
|
|
|
if (aer_severity != AER_CORRECTABLE)
|
2025-05-22 18:21:20 -05:00
|
|
|
aer_printk(info.level, dev, "aer_uncor_severity: 0x%08x\n",
|
|
|
|
aer->uncor_severity);
|
2018-06-08 08:39:38 -05:00
|
|
|
|
|
|
|
if (tlp_header_valid)
|
2025-05-22 18:21:19 -05:00
|
|
|
pcie_print_tlp_log(dev, &aer->header_log, info.level,
|
|
|
|
dev_fmt(" "));
|
2018-06-08 08:39:38 -05:00
|
|
|
}
|
module: Convert symbol namespace to string literal
Clean up the existing export namespace code along the same lines of
commit 33def8498fdd ("treewide: Convert macro and uses of __section(foo)
to __section("foo")") and for the same reason, it is not desired for the
namespace argument to be a macro expansion itself.
Scripted using
git grep -l -e MODULE_IMPORT_NS -e EXPORT_SYMBOL_NS | while read file;
do
awk -i inplace '
/^#define EXPORT_SYMBOL_NS/ {
gsub(/__stringify\(ns\)/, "ns");
print;
next;
}
/^#define MODULE_IMPORT_NS/ {
gsub(/__stringify\(ns\)/, "ns");
print;
next;
}
/MODULE_IMPORT_NS/ {
$0 = gensub(/MODULE_IMPORT_NS\(([^)]*)\)/, "MODULE_IMPORT_NS(\"\\1\")", "g");
}
/EXPORT_SYMBOL_NS/ {
if ($0 ~ /(EXPORT_SYMBOL_NS[^(]*)\(([^,]+),/) {
if ($0 !~ /(EXPORT_SYMBOL_NS[^(]*)\(([^,]+), ([^)]+)\)/ &&
$0 !~ /(EXPORT_SYMBOL_NS[^(]*)\(\)/ &&
$0 !~ /^my/) {
getline line;
gsub(/[[:space:]]*\\$/, "");
gsub(/[[:space:]]/, "", line);
$0 = $0 " " line;
}
$0 = gensub(/(EXPORT_SYMBOL_NS[^(]*)\(([^,]+), ([^)]+)\)/,
"\\1(\\2, \"\\3\")", "g");
}
}
{ print }' $file;
done
Requested-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://mail.google.com/mail/u/2/#inbox/FMfcgzQXKWgMmjdFwwdsfgxzKpVHWPlc
Acked-by: Greg KH <gregkh@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2024-12-02 15:59:47 +01:00
|
|
|
EXPORT_SYMBOL_NS_GPL(pci_print_aer, "CXL");
|
2018-06-08 08:39:38 -05:00
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
/**
|
|
|
|
* add_error_device - list device to be handled
|
|
|
|
* @e_info: pointer to error info
|
|
|
|
* @dev: pointer to pci_dev to be added
|
|
|
|
*/
|
|
|
|
static int add_error_device(struct aer_err_info *e_info, struct pci_dev *dev)
|
|
|
|
{
|
2025-05-22 18:21:23 -05:00
|
|
|
int i = e_info->error_dev_num;
|
|
|
|
|
|
|
|
if (i >= AER_MAX_MULTI_ERR_DEVICES)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
e_info->dev[i] = pci_dev_get(dev);
|
|
|
|
e_info->error_dev_num++;
|
|
|
|
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
/*
|
|
|
|
* Ratelimit AER log messages. "dev" is either the source
|
|
|
|
* identified by the root's Error Source ID or it has an unmasked
|
|
|
|
* error logged in its own AER Capability. Messages are emitted
|
|
|
|
* when "ratelimit_print[i]" is non-zero. If we will print detail
|
|
|
|
* for a downstream device, make sure we print the Error Source ID
|
|
|
|
* from the root as well.
|
|
|
|
*/
|
|
|
|
if (aer_ratelimit(dev, e_info->severity)) {
|
|
|
|
e_info->ratelimit_print[i] = 1;
|
|
|
|
e_info->root_ratelimit_print = 1;
|
|
|
|
}
|
2025-05-22 18:21:23 -05:00
|
|
|
return 0;
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* is_error_source - check whether the device is source of reported error
|
|
|
|
* @dev: pointer to pci_dev to be checked
|
|
|
|
* @e_info: pointer to reported error info
|
|
|
|
*/
|
|
|
|
static bool is_error_source(struct pci_dev *dev, struct aer_err_info *e_info)
|
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-06-08 08:33:39 -05:00
|
|
|
u32 status, mask;
|
|
|
|
u16 reg16;
|
|
|
|
|
|
|
|
/*
|
2025-03-07 17:17:15 -06:00
|
|
|
* When bus ID is equal to 0, it might be a bad ID
|
|
|
|
* reported by Root Port.
|
2018-06-08 08:33:39 -05:00
|
|
|
*/
|
|
|
|
if ((PCI_BUS_NUM(e_info->id) != 0) &&
|
|
|
|
!(dev->bus->bus_flags & PCI_BUS_FLAGS_NO_AERSID)) {
|
|
|
|
/* Device ID match? */
|
2023-08-07 21:48:57 +08:00
|
|
|
if (e_info->id == pci_dev_id(dev))
|
2018-06-08 08:33:39 -05:00
|
|
|
return true;
|
|
|
|
|
2025-03-07 17:17:15 -06:00
|
|
|
/* Continue ID comparing if there is no multiple error */
|
2018-06-08 08:33:39 -05:00
|
|
|
if (!e_info->multi_error_valid)
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When either
|
2025-03-07 17:17:15 -06:00
|
|
|
* 1) bus ID is equal to 0. Some ports might lose the bus
|
|
|
|
* ID of error source id;
|
2018-06-08 08:33:39 -05:00
|
|
|
* 2) bus flag PCI_BUS_FLAGS_NO_AERSID is set
|
|
|
|
* 3) There are multiple errors and prior ID comparing fails;
|
|
|
|
* We check AER status registers to find possible reporter.
|
|
|
|
*/
|
|
|
|
if (atomic_read(&dev->enable_cnt) == 0)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/* Check if AER is enabled */
|
|
|
|
pcie_capability_read_word(dev, PCI_EXP_DEVCTL, ®16);
|
|
|
|
if (!(reg16 & PCI_EXP_AER_FLAGS))
|
|
|
|
return false;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2018-06-08 08:33:39 -05:00
|
|
|
return false;
|
|
|
|
|
|
|
|
/* Check if error is recorded */
|
|
|
|
if (e_info->severity == AER_CORRECTABLE) {
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_STATUS, &status);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_MASK, &mask);
|
2018-06-08 08:33:39 -05:00
|
|
|
} else {
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS, &status);
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_MASK, &mask);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
if (status & ~mask)
|
|
|
|
return true;
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int find_device_iter(struct pci_dev *dev, void *data)
|
|
|
|
{
|
|
|
|
struct aer_err_info *e_info = (struct aer_err_info *)data;
|
|
|
|
|
|
|
|
if (is_error_source(dev, e_info)) {
|
|
|
|
/* List this device */
|
|
|
|
if (add_error_device(e_info, dev)) {
|
|
|
|
/* We cannot handle more... Stop iteration */
|
2025-06-20 00:20:30 +05:30
|
|
|
pci_err(dev, "Exceeded max supported (%d) devices with errors logged\n",
|
|
|
|
AER_MAX_MULTI_ERR_DEVICES);
|
2018-06-08 08:33:39 -05:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If there is only a single error, stop iteration */
|
|
|
|
if (!e_info->multi_error_valid)
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* find_source_device - search through device hierarchy for source device
|
|
|
|
* @parent: pointer to Root Port pci_dev data structure
|
2025-03-07 17:17:15 -06:00
|
|
|
* @e_info: including detailed error information such as ID
|
2018-06-08 08:33:39 -05:00
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Return: true if found.
|
2018-06-08 08:33:39 -05:00
|
|
|
*
|
|
|
|
* Invoked by DPC when error is detected at the Root Port.
|
|
|
|
* Caller of this function must set id, severity, and multi_error_valid of
|
|
|
|
* struct aer_err_info pointed by @e_info properly. This function must fill
|
|
|
|
* e_info->error_dev_num and e_info->dev[], based on the given information.
|
|
|
|
*/
|
|
|
|
static bool find_source_device(struct pci_dev *parent,
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
struct aer_err_info *e_info)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
|
|
|
struct pci_dev *dev = parent;
|
|
|
|
int result;
|
|
|
|
|
|
|
|
/* Must reset in this function */
|
|
|
|
e_info->error_dev_num = 0;
|
|
|
|
|
|
|
|
/* Is Root Port an agent that sends error message? */
|
|
|
|
result = find_device_iter(dev, e_info);
|
|
|
|
if (result)
|
|
|
|
return true;
|
|
|
|
|
2020-11-20 16:10:34 -08:00
|
|
|
if (pci_pcie_type(parent) == PCI_EXP_TYPE_RC_EC)
|
|
|
|
pcie_walk_rcec(parent, find_device_iter, e_info);
|
|
|
|
else
|
|
|
|
pci_walk_bus(parent->subordinate, find_device_iter, e_info);
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2025-05-22 18:21:10 -05:00
|
|
|
if (!e_info->error_dev_num)
|
2018-06-08 08:33:39 -05:00
|
|
|
return false;
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
#ifdef CONFIG_PCIEAER_CXL
|
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
/**
|
2023-10-18 19:17:11 +02:00
|
|
|
* pci_aer_unmask_internal_errors - unmask internal errors
|
2025-03-07 17:17:15 -06:00
|
|
|
* @dev: pointer to the pci_dev data structure
|
2023-10-18 19:17:11 +02:00
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Unmask internal errors in the Uncorrectable and Correctable Error
|
2023-10-18 19:17:11 +02:00
|
|
|
* Mask registers.
|
|
|
|
*
|
|
|
|
* Note: AER must be enabled and supported by the device which must be
|
|
|
|
* checked in advance, e.g. with pcie_aer_is_native().
|
|
|
|
*/
|
|
|
|
static void pci_aer_unmask_internal_errors(struct pci_dev *dev)
|
|
|
|
{
|
|
|
|
int aer = dev->aer_cap;
|
|
|
|
u32 mask;
|
|
|
|
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_MASK, &mask);
|
|
|
|
mask &= ~PCI_ERR_UNC_INTN;
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_UNCOR_MASK, mask);
|
|
|
|
|
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_MASK, &mask);
|
|
|
|
mask &= ~PCI_ERR_COR_INTERNAL;
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_COR_MASK, mask);
|
|
|
|
}
|
|
|
|
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
static bool is_cxl_mem_dev(struct pci_dev *dev)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* The capability, status, and control fields in Device 0,
|
|
|
|
* Function 0 DVSEC control the CXL functionality of the
|
|
|
|
* entire device (CXL 3.0, 8.1.3).
|
|
|
|
*/
|
|
|
|
if (dev->devfn != PCI_DEVFN(0, 0))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* CXL Memory Devices must have the 502h class code set (CXL
|
|
|
|
* 3.0, 8.1.12.1).
|
|
|
|
*/
|
|
|
|
if ((dev->class >> 8) != PCI_CLASS_MEMORY_CXL)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool cxl_error_is_native(struct pci_dev *dev)
|
|
|
|
{
|
|
|
|
struct pci_host_bridge *host = pci_find_host_bridge(dev->bus);
|
|
|
|
|
|
|
|
return (pcie_ports_native || host->native_aer);
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool is_internal_error(struct aer_err_info *info)
|
|
|
|
{
|
|
|
|
if (info->severity == AER_CORRECTABLE)
|
|
|
|
return info->status & PCI_ERR_COR_INTERNAL;
|
|
|
|
|
|
|
|
return info->status & PCI_ERR_UNC_INTN;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int cxl_rch_handle_error_iter(struct pci_dev *dev, void *data)
|
|
|
|
{
|
|
|
|
struct aer_err_info *info = (struct aer_err_info *)data;
|
|
|
|
const struct pci_error_handlers *err_handler;
|
|
|
|
|
|
|
|
if (!is_cxl_mem_dev(dev) || !cxl_error_is_native(dev))
|
|
|
|
return 0;
|
|
|
|
|
2025-03-07 17:17:15 -06:00
|
|
|
/* Protect dev->driver */
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
device_lock(&dev->dev);
|
|
|
|
|
|
|
|
err_handler = dev->driver ? dev->driver->err_handler : NULL;
|
|
|
|
if (!err_handler)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
if (info->severity == AER_CORRECTABLE) {
|
|
|
|
if (err_handler->cor_error_detected)
|
|
|
|
err_handler->cor_error_detected(dev);
|
|
|
|
} else if (err_handler->error_detected) {
|
|
|
|
if (info->severity == AER_NONFATAL)
|
|
|
|
err_handler->error_detected(dev, pci_channel_io_normal);
|
|
|
|
else if (info->severity == AER_FATAL)
|
|
|
|
err_handler->error_detected(dev, pci_channel_io_frozen);
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
device_unlock(&dev->dev);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void cxl_rch_handle_error(struct pci_dev *dev, struct aer_err_info *info)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Internal errors of an RCEC indicate an AER error in an
|
|
|
|
* RCH's downstream port. Check and handle them in the CXL.mem
|
|
|
|
* device driver.
|
|
|
|
*/
|
|
|
|
if (pci_pcie_type(dev) == PCI_EXP_TYPE_RC_EC &&
|
|
|
|
is_internal_error(info))
|
|
|
|
pcie_walk_rcec(dev, cxl_rch_handle_error_iter, info);
|
|
|
|
}
|
|
|
|
|
2023-10-18 19:17:11 +02:00
|
|
|
static int handles_cxl_error_iter(struct pci_dev *dev, void *data)
|
|
|
|
{
|
|
|
|
bool *handles_cxl = data;
|
|
|
|
|
|
|
|
if (!*handles_cxl)
|
|
|
|
*handles_cxl = is_cxl_mem_dev(dev) && cxl_error_is_native(dev);
|
|
|
|
|
|
|
|
/* Non-zero terminates iteration */
|
|
|
|
return *handles_cxl;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool handles_cxl_errors(struct pci_dev *rcec)
|
|
|
|
{
|
|
|
|
bool handles_cxl = false;
|
|
|
|
|
|
|
|
if (pci_pcie_type(rcec) == PCI_EXP_TYPE_RC_EC &&
|
|
|
|
pcie_aer_is_native(rcec))
|
|
|
|
pcie_walk_rcec(rcec, handles_cxl_error_iter, &handles_cxl);
|
|
|
|
|
|
|
|
return handles_cxl;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void cxl_rch_enable_rcec(struct pci_dev *rcec)
|
|
|
|
{
|
|
|
|
if (!handles_cxl_errors(rcec))
|
|
|
|
return;
|
|
|
|
|
|
|
|
pci_aer_unmask_internal_errors(rcec);
|
|
|
|
pci_info(rcec, "CXL: Internal errors unmasked");
|
|
|
|
}
|
|
|
|
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
#else
|
2023-10-18 19:17:11 +02:00
|
|
|
static inline void cxl_rch_enable_rcec(struct pci_dev *dev) { }
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
static inline void cxl_rch_handle_error(struct pci_dev *dev,
|
|
|
|
struct aer_err_info *info) { }
|
|
|
|
#endif
|
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
/**
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
* pci_aer_handle_error - handle logging error into an event log
|
2018-06-08 08:33:39 -05:00
|
|
|
* @dev: pointer to pci_dev data structure of error source device
|
|
|
|
* @info: comprehensive error information
|
|
|
|
*
|
|
|
|
* Invoked when an error being detected by Root Port.
|
|
|
|
*/
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
static void pci_aer_handle_error(struct pci_dev *dev, struct aer_err_info *info)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = dev->aer_cap;
|
2018-06-08 08:33:39 -05:00
|
|
|
|
|
|
|
if (info->severity == AER_CORRECTABLE) {
|
|
|
|
/*
|
|
|
|
* Correctable error does not need software intervention.
|
|
|
|
* No need to go through error recovery process.
|
|
|
|
*/
|
2020-05-29 17:56:09 -05:00
|
|
|
if (aer)
|
|
|
|
pci_write_config_dword(dev, aer + PCI_ERR_COR_STATUS,
|
2018-06-08 08:33:39 -05:00
|
|
|
info->status);
|
2022-11-30 15:11:21 -07:00
|
|
|
if (pcie_aer_is_native(dev)) {
|
|
|
|
struct pci_driver *pdrv = dev->driver;
|
|
|
|
|
|
|
|
if (pdrv && pdrv->err_handler &&
|
|
|
|
pdrv->err_handler->cor_error_detected)
|
|
|
|
pdrv->err_handler->cor_error_detected(dev);
|
PCI/ERR: Clear PCIe Device Status errors only if OS owns AER
pcie_clear_device_status() resets the error bits in the PCIe Device Status
Register (PCI_EXP_DEVSTA).
Previously we did this unconditionally, but on ACPI systems, the _OSC AER
bit negotiates control of the AER capability. Per sec 4.5.1 of the System
Firmware Intermediary _OSC and DPC Updates ECN [1], this bit also covers
other error enable/status bits including the following:
Correctable Error Reporting Enable
Non-Fatal Error Reporting Enable
Fatal Error Reporting Enable
Unsupported Request Reporting Enable
These bits are all in the PCIe Device Control register (the ECN omitted
"Reporting", but I think that's a typo), so by implication the _OSC AER bit
also applies to the error status bits in the PCIe Device Status register:
Correctable Error Detected
Non-Fatal Error Detected
Fatal Error Detected
Unsupported Request Detected
Clear the PCIe Device Status error bits only when the OS controls the AER
capability and related error enable/status bits. If platform firmware
controls the AER capability, firmware is responsible for clearing these
bits.
One call path leading here is:
ghes_do_proc
ghes_handle_aer
aer_recover_queue
schedule_work(&aer_recover_work)
...
aer_recover_work_func
pcie_do_recovery
pcie_clear_device_status
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[bhelgaas: commit log, move test from pcie_clear_device_status() to callers]
Link: https://lore.kernel.org/r/20200622113523.891666-1-Jonathan.Cameron@huawei.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-06-22 19:35:23 +08:00
|
|
|
pcie_clear_device_status(dev);
|
2022-11-30 15:11:21 -07:00
|
|
|
}
|
2018-06-08 08:33:39 -05:00
|
|
|
} else if (info->severity == AER_NONFATAL)
|
2020-03-23 17:26:02 -07:00
|
|
|
pcie_do_recovery(dev, pci_channel_io_normal, aer_root_reset);
|
2018-06-08 08:33:39 -05:00
|
|
|
else if (info->severity == AER_FATAL)
|
2020-03-23 17:26:02 -07:00
|
|
|
pcie_do_recovery(dev, pci_channel_io_frozen, aer_root_reset);
|
PCI/AER: Forward RCH downstream port-detected errors to the CXL.mem dev handler
In Restricted CXL Device (RCD) mode a CXL device is exposed as an
RCiEP, but CXL downstream and upstream ports are not enumerated and
not visible in the PCIe hierarchy. [1] Protocol and link errors from
these non-enumerated ports are signaled as internal AER errors, either
Uncorrectable Internal Error (UIE) or Corrected Internal Errors (CIE)
via an RCEC.
Restricted CXL host (RCH) downstream port-detected errors have the
Requester ID of the RCEC set in the RCEC's AER Error Source ID
register. A CXL handler must then inspect the error status in various
CXL registers residing in the dport's component register space (CXL
RAS capability) or the dport's RCRB (PCIe AER extended
capability). [2]
Errors showing up in the RCEC's error handler must be handled and
connected to the CXL subsystem. Implement this by forwarding the error
to all CXL devices below the RCEC. Since the entire CXL device is
controlled only using PCIe Configuration Space of device 0, function
0, only pass it there [3]. The error handling is limited to currently
supported devices with the Memory Device class code set (CXL Type 3
Device, PCI_CLASS_MEMORY_CXL, 502h), handle downstream port errors in
the device's cxl_pci driver. Support for other CXL Device Types
(e.g. a CXL.cache Device) can be added later.
To handle downstream port errors in addition to errors directed to the
CXL endpoint device, a handler must also inspect the CXL RAS and PCIe
AER capabilities of the CXL downstream port the device is connected
to.
Since CXL downstream port errors are signaled using internal errors,
the handler requires those errors to be unmasked. This is subject of a
follow-on patch.
The reason for choosing this implementation is that the AER service
driver claims the RCEC device, but does not allow it to register a
custom specific handler to support CXL. Connecting the RCEC hard-wired
with a CXL handler does not work, as the CXL subsystem might not be
present all the time. The alternative to add an implementation to the
portdrv to allow the registration of a custom RCEC error handler isn't
worth doing it as CXL would be its only user. Instead, just check for
an CXL RCEC and pass it down to the connected CXL device's error
handler. With this approach the code can entirely be implemented in
the PCIe AER driver and is independent of the CXL subsystem. The CXL
driver only provides the handler.
[1] CXL 3.0 spec: 9.11.8 CXL Devices Attached to an RCH
[2] CXL 3.0 spec, 12.2.1.1 RCH Downstream Port-detected Errors
[3] CXL 3.0 spec, 8.1.3 PCIe DVSEC for CXL Devices
Co-developed-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Terry Bowman <terry.bowman@amd.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
Cc: Oliver O'Halloran <oohall@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-pci@vger.kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20231018171713.1883517-18-rrichter@amd.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2023-10-18 19:17:10 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static void handle_error_source(struct pci_dev *dev, struct aer_err_info *info)
|
|
|
|
{
|
|
|
|
cxl_rch_handle_error(dev, info);
|
|
|
|
pci_aer_handle_error(dev, info);
|
2018-09-20 10:27:09 -06:00
|
|
|
pci_dev_put(dev);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_ACPI_APEI_PCIEAER
|
|
|
|
|
2020-05-27 16:01:48 -05:00
|
|
|
#define AER_RECOVER_RING_SIZE 16
|
2018-06-08 08:33:39 -05:00
|
|
|
|
|
|
|
struct aer_recover_entry {
|
|
|
|
u8 bus;
|
|
|
|
u8 devfn;
|
|
|
|
u16 domain;
|
|
|
|
int severity;
|
|
|
|
struct aer_capability_regs *regs;
|
|
|
|
};
|
|
|
|
|
|
|
|
static DEFINE_KFIFO(aer_recover_ring, struct aer_recover_entry,
|
|
|
|
AER_RECOVER_RING_SIZE);
|
|
|
|
|
|
|
|
static void aer_recover_work_func(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct aer_recover_entry entry;
|
|
|
|
struct pci_dev *pdev;
|
|
|
|
|
|
|
|
while (kfifo_get(&aer_recover_ring, &entry)) {
|
|
|
|
pdev = pci_get_domain_bus_and_slot(entry.domain, entry.bus,
|
|
|
|
entry.devfn);
|
|
|
|
if (!pdev) {
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
pr_err_ratelimited("%04x:%02x:%02x.%x: no pci_dev found\n",
|
|
|
|
entry.domain, entry.bus,
|
|
|
|
PCI_SLOT(entry.devfn),
|
|
|
|
PCI_FUNC(entry.devfn));
|
2018-06-08 08:33:39 -05:00
|
|
|
continue;
|
|
|
|
}
|
2023-10-18 19:17:05 +02:00
|
|
|
pci_print_aer(pdev, entry.severity, entry.regs);
|
2024-11-02 12:45:37 -05:00
|
|
|
|
2023-09-21 02:03:36 +08:00
|
|
|
/*
|
2024-11-02 12:45:37 -05:00
|
|
|
* Memory for aer_capability_regs(entry.regs) is being
|
|
|
|
* allocated from the ghes_estatus_pool to protect it from
|
|
|
|
* overwriting when multiple sections are present in the
|
|
|
|
* error status. Thus free the same after processing the
|
|
|
|
* data.
|
2023-09-21 02:03:36 +08:00
|
|
|
*/
|
|
|
|
ghes_estatus_pool_region_free((unsigned long)entry.regs,
|
2024-11-02 12:45:37 -05:00
|
|
|
sizeof(struct aer_capability_regs));
|
2023-09-21 02:03:36 +08:00
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
if (entry.severity == AER_NONFATAL)
|
2018-09-20 10:27:12 -06:00
|
|
|
pcie_do_recovery(pdev, pci_channel_io_normal,
|
2020-03-23 17:26:02 -07:00
|
|
|
aer_root_reset);
|
2018-06-08 08:33:39 -05:00
|
|
|
else if (entry.severity == AER_FATAL)
|
2018-09-20 10:27:12 -06:00
|
|
|
pcie_do_recovery(pdev, pci_channel_io_frozen,
|
2020-03-23 17:26:02 -07:00
|
|
|
aer_root_reset);
|
2018-06-08 08:33:39 -05:00
|
|
|
pci_dev_put(pdev);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Mutual exclusion for writers of aer_recover_ring, reader side don't
|
|
|
|
* need lock, because there is only one reader and lock is not needed
|
|
|
|
* between reader and writer.
|
|
|
|
*/
|
|
|
|
static DEFINE_SPINLOCK(aer_recover_ring_lock);
|
|
|
|
static DECLARE_WORK(aer_recover_work, aer_recover_work_func);
|
|
|
|
|
|
|
|
void aer_recover_queue(int domain, unsigned int bus, unsigned int devfn,
|
|
|
|
int severity, struct aer_capability_regs *aer_regs)
|
|
|
|
{
|
|
|
|
struct aer_recover_entry entry = {
|
|
|
|
.bus = bus,
|
|
|
|
.devfn = devfn,
|
|
|
|
.domain = domain,
|
|
|
|
.severity = severity,
|
|
|
|
.regs = aer_regs,
|
|
|
|
};
|
|
|
|
|
2018-12-14 11:29:37 -06:00
|
|
|
if (kfifo_in_spinlocked(&aer_recover_ring, &entry, 1,
|
2018-09-18 17:58:44 -06:00
|
|
|
&aer_recover_ring_lock))
|
2018-06-08 08:33:39 -05:00
|
|
|
schedule_work(&aer_recover_work);
|
|
|
|
else
|
2021-03-29 19:02:01 +08:00
|
|
|
pr_err("buffer overflow in recovery for %04x:%02x:%02x.%x\n",
|
2018-06-08 08:33:39 -05:00
|
|
|
domain, bus, PCI_SLOT(devfn), PCI_FUNC(devfn));
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(aer_recover_queue);
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/**
|
2018-07-19 16:16:55 -05:00
|
|
|
* aer_get_device_error_info - read error status from dev and store it to info
|
2018-06-08 08:33:39 -05:00
|
|
|
* @info: pointer to structure to store the error record
|
2025-05-22 18:21:22 -05:00
|
|
|
* @i: index into info->dev[]
|
2018-06-08 08:33:39 -05:00
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Return: 1 on success, 0 on error.
|
2018-06-08 08:33:39 -05:00
|
|
|
*
|
|
|
|
* Note that @info is reused among all error devices. Clear fields properly.
|
|
|
|
*/
|
2025-05-22 18:21:22 -05:00
|
|
|
int aer_get_device_error_info(struct aer_err_info *info, int i)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2025-05-22 18:21:22 -05:00
|
|
|
struct pci_dev *dev;
|
|
|
|
int type, aer;
|
2024-02-06 15:57:14 +02:00
|
|
|
u32 aercc;
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2025-05-22 18:21:22 -05:00
|
|
|
if (i >= AER_MAX_MULTI_ERR_DEVICES)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
dev = info->dev[i];
|
|
|
|
aer = dev->aer_cap;
|
|
|
|
type = pci_pcie_type(dev);
|
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
/* Must reset in this function */
|
|
|
|
info->status = 0;
|
|
|
|
info->tlp_header_valid = 0;
|
|
|
|
|
|
|
|
/* The device might not support AER */
|
2020-05-29 17:56:09 -05:00
|
|
|
if (!aer)
|
2018-06-08 08:33:39 -05:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (info->severity == AER_CORRECTABLE) {
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_STATUS,
|
2018-06-08 08:33:39 -05:00
|
|
|
&info->status);
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_COR_MASK,
|
2018-06-08 08:33:39 -05:00
|
|
|
&info->mask);
|
|
|
|
if (!(info->status & ~info->mask))
|
|
|
|
return 0;
|
2020-11-20 16:10:27 -08:00
|
|
|
} else if (type == PCI_EXP_TYPE_ROOT_PORT ||
|
2020-11-20 16:10:34 -08:00
|
|
|
type == PCI_EXP_TYPE_RC_EC ||
|
2020-11-20 16:10:27 -08:00
|
|
|
type == PCI_EXP_TYPE_DOWNSTREAM ||
|
2018-09-20 10:27:10 -06:00
|
|
|
info->severity == AER_NONFATAL) {
|
2018-06-08 08:33:39 -05:00
|
|
|
|
|
|
|
/* Link is still healthy for IO reads */
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_STATUS,
|
2018-06-08 08:33:39 -05:00
|
|
|
&info->status);
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_UNCOR_MASK,
|
2018-06-08 08:33:39 -05:00
|
|
|
&info->mask);
|
|
|
|
if (!(info->status & ~info->mask))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* Get First Error Pointer */
|
2024-02-06 15:57:14 +02:00
|
|
|
pci_read_config_dword(dev, aer + PCI_ERR_CAP, &aercc);
|
|
|
|
info->first_error = PCI_ERR_CAP_FEP(aercc);
|
2018-06-08 08:33:39 -05:00
|
|
|
|
|
|
|
if (info->status & AER_LOG_TLP_MASKS) {
|
|
|
|
info->tlp_header_valid = 1;
|
2025-01-14 19:08:39 +02:00
|
|
|
pcie_read_tlp_log(dev, aer + PCI_ERR_HEADER_LOG,
|
|
|
|
aer + PCI_ERR_PREFIX_LOG,
|
|
|
|
aer_tlp_log_len(dev, aercc),
|
2025-02-07 18:18:36 +02:00
|
|
|
aercc & PCI_ERR_CAP_TLP_LOG_FLIT,
|
2025-01-14 19:08:39 +02:00
|
|
|
&info->tlp);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void aer_process_err_devices(struct aer_err_info *e_info)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2025-03-07 17:17:15 -06:00
|
|
|
/* Report all before handling them, to not lose records by reset etc. */
|
2018-06-08 08:33:39 -05:00
|
|
|
for (i = 0; i < e_info->error_dev_num && e_info->dev[i]; i++) {
|
2025-05-22 18:21:22 -05:00
|
|
|
if (aer_get_device_error_info(e_info, i))
|
|
|
|
aer_print_error(e_info, i);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
for (i = 0; i < e_info->error_dev_num && e_info->dev[i]; i++) {
|
2025-05-22 18:21:22 -05:00
|
|
|
if (aer_get_device_error_info(e_info, i))
|
2018-06-08 08:33:39 -05:00
|
|
|
handle_error_source(e_info->dev[i], e_info);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2025-05-22 18:21:09 -05:00
|
|
|
* aer_isr_one_error_type - consume a Correctable or Uncorrectable Error
|
|
|
|
* detected by Root Port or RCEC
|
|
|
|
* @root: pointer to Root Port or RCEC that signaled AER interrupt
|
|
|
|
* @info: pointer to AER error info
|
|
|
|
*/
|
|
|
|
static void aer_isr_one_error_type(struct pci_dev *root,
|
|
|
|
struct aer_err_info *info)
|
|
|
|
{
|
2025-05-22 18:21:10 -05:00
|
|
|
bool found;
|
2025-05-22 18:21:09 -05:00
|
|
|
|
2025-05-22 18:21:10 -05:00
|
|
|
found = find_source_device(root, info);
|
PCI/AER: Ratelimit correctable and non-fatal error logging
Spammy devices can flood kernel logs with AER errors and slow/stall
execution. Add per-device ratelimits for AER correctable and non-fatal
uncorrectable errors that use the kernel defaults (10 per 5s). Logging of
fatal errors is not ratelimited.
There are two AER logging entry points:
- aer_print_error() is used by DPC and native AER
- pci_print_aer() is used by GHES and CXL
The native AER aer_print_error() case includes a loop that may log details
from multiple devices, which are ratelimited individually. If we log
details for any device, we also log the Error Source ID from the Root Port
or RCEC.
If no such device details are found, we still log the Error Source from the
ERR_* Message, ratelimited by the Root Port or RCEC that received it.
The DPC aer_print_error() case is not ratelimited, since this only happens
for fatal errors.
The CXL pci_print_aer() case is ratelimited by the Error Source device.
The GHES pci_print_aer() case is via aer_recover_work_func(), which
searches for the Error Source device. If the device is not found, there's
no per-device ratelimit, so we use a system-wide ratelimit that covers all
error types (correctable, non-fatal, and fatal).
Sargun at Meta reported internally that a flood of AER errors causes RCU
CPU stall warnings and CSD-lock warnings.
Tested using aer-inject[1]. Sent 11 AER errors. Observed 10 errors logged
while AER stats (cat /sys/bus/pci/devices/<dev>/aer_dev_correctable) show
true count of 11.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/gong.chen/aer-inject.git
[bhelgaas: commit log, factor out trace_aer_event() and aer_print_rp_info()
changes to previous patches, enable Error Source logging if any downstream
detail will be printed, don't ratelimit fatal errors, "aer_report" ->
"aer_info", "cor_log_ratelimit" -> "correctable_ratelimit",
"uncor_log_ratelimit" -> "nonfatal_ratelimit"]
Reported-by: Sargun Dhillon <sargun@meta.com>
Signed-off-by: Jon Pan-Doh <pandoh@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Link: https://patch.msgid.link/20250522232339.1525671-19-helgaas@kernel.org
2025-05-22 18:21:24 -05:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If we're going to log error messages, we've already set
|
|
|
|
* "info->root_ratelimit_print" and "info->ratelimit_print[i]" to
|
|
|
|
* non-zero (which enables printing) because this is either an
|
|
|
|
* ERR_FATAL or we found a device with an error logged in its AER
|
|
|
|
* Capability.
|
|
|
|
*
|
|
|
|
* If we didn't find the Error Source device, at least log the
|
|
|
|
* Requester ID from the ERR_* Message received by the Root Port or
|
|
|
|
* RCEC, ratelimited by the RP or RCEC.
|
|
|
|
*/
|
|
|
|
if (info->root_ratelimit_print ||
|
|
|
|
(!found && aer_ratelimit(root, info->severity)))
|
|
|
|
aer_print_source(root, info, found);
|
|
|
|
|
2025-05-22 18:21:10 -05:00
|
|
|
if (found)
|
2025-05-22 18:21:09 -05:00
|
|
|
aer_process_err_devices(info);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* aer_isr_one_error - consume error(s) signaled by an AER interrupt from
|
|
|
|
* Root Port or RCEC
|
|
|
|
* @root: pointer to Root Port or RCEC that signaled AER interrupt
|
2018-06-08 08:33:39 -05:00
|
|
|
* @e_src: pointer to an error source
|
|
|
|
*/
|
2025-05-22 18:21:09 -05:00
|
|
|
static void aer_isr_one_error(struct pci_dev *root,
|
2025-05-22 18:21:14 -05:00
|
|
|
struct aer_err_source *e_src)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2025-05-22 18:21:14 -05:00
|
|
|
u32 status = e_src->status;
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2025-05-22 18:21:09 -05:00
|
|
|
pci_rootport_aer_stats_incr(root, e_src);
|
2018-06-21 16:48:29 -07:00
|
|
|
|
2018-06-08 08:33:39 -05:00
|
|
|
/*
|
|
|
|
* There is a possibility that both correctable error and
|
|
|
|
* uncorrectable error being logged. Report correctable error first.
|
|
|
|
*/
|
2025-05-22 18:21:14 -05:00
|
|
|
if (status & PCI_ERR_ROOT_COR_RCV) {
|
|
|
|
int multi = status & PCI_ERR_ROOT_MULTI_COR_RCV;
|
|
|
|
struct aer_err_info e_info = {
|
|
|
|
.id = ERR_COR_ID(e_src->id),
|
|
|
|
.severity = AER_CORRECTABLE,
|
2025-05-22 18:21:18 -05:00
|
|
|
.level = KERN_WARNING,
|
2025-05-22 18:21:14 -05:00
|
|
|
.multi_error_valid = multi ? 1 : 0,
|
|
|
|
};
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2025-05-22 18:21:09 -05:00
|
|
|
aer_isr_one_error_type(root, &e_info);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
2025-05-22 18:21:14 -05:00
|
|
|
if (status & PCI_ERR_ROOT_UNCOR_RCV) {
|
|
|
|
int fatal = status & PCI_ERR_ROOT_FATAL_RCV;
|
|
|
|
int multi = status & PCI_ERR_ROOT_MULTI_UNCOR_RCV;
|
|
|
|
struct aer_err_info e_info = {
|
|
|
|
.id = ERR_UNCOR_ID(e_src->id),
|
|
|
|
.severity = fatal ? AER_FATAL : AER_NONFATAL,
|
2025-05-22 18:21:18 -05:00
|
|
|
.level = KERN_ERR,
|
2025-05-22 18:21:14 -05:00
|
|
|
.multi_error_valid = multi ? 1 : 0,
|
|
|
|
};
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2025-05-22 18:21:09 -05:00
|
|
|
aer_isr_one_error_type(root, &e_info);
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2025-03-07 17:17:15 -06:00
|
|
|
* aer_isr - consume errors detected by Root Port
|
2019-08-27 18:18:23 +03:00
|
|
|
* @irq: IRQ assigned to Root Port
|
|
|
|
* @context: pointer to Root Port data structure
|
2018-06-08 08:33:39 -05:00
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Invoked, as DPC, when Root Port records new detected error
|
2018-06-08 08:33:39 -05:00
|
|
|
*/
|
2018-09-18 17:58:46 -06:00
|
|
|
static irqreturn_t aer_isr(int irq, void *context)
|
2018-06-08 08:33:39 -05:00
|
|
|
{
|
2018-09-18 17:58:46 -06:00
|
|
|
struct pcie_device *dev = (struct pcie_device *)context;
|
|
|
|
struct aer_rpc *rpc = get_service_data(dev);
|
treewide: Remove uninitialized_var() usage
Using uninitialized_var() is dangerous as it papers over real bugs[1]
(or can in the future), and suppresses unrelated compiler warnings
(e.g. "unused variable"). If the compiler thinks it is uninitialized,
either simply initialize the variable or make compiler changes.
In preparation for removing[2] the[3] macro[4], remove all remaining
needless uses with the following script:
git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \
xargs perl -pi -e \
's/\buninitialized_var\(([^\)]+)\)/\1/g;
s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;'
drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid
pathological white-space.
No outstanding warnings were found building allmodconfig with GCC 9.3.0
for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64,
alpha, and m68k.
[1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/
[2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/
[3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/
[4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/
Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5
Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB
Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers
Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs
Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-03 13:09:38 -07:00
|
|
|
struct aer_err_source e_src;
|
2018-06-08 08:33:39 -05:00
|
|
|
|
2018-09-18 17:58:46 -06:00
|
|
|
if (kfifo_is_empty(&rpc->aer_fifo))
|
|
|
|
return IRQ_NONE;
|
|
|
|
|
2018-09-18 17:58:43 -06:00
|
|
|
while (kfifo_get(&rpc->aer_fifo, &e_src))
|
2025-05-22 18:21:09 -05:00
|
|
|
aer_isr_one_error(rpc->rpd, &e_src);
|
2018-09-18 17:58:46 -06:00
|
|
|
return IRQ_HANDLED;
|
2018-06-08 08:33:39 -05:00
|
|
|
}
|
|
|
|
|
2018-06-08 08:31:57 -05:00
|
|
|
/**
|
|
|
|
* aer_irq - Root Port's ISR
|
|
|
|
* @irq: IRQ assigned to Root Port
|
|
|
|
* @context: pointer to Root Port data structure
|
|
|
|
*
|
|
|
|
* Invoked when Root Port detects AER messages.
|
|
|
|
*/
|
2018-10-11 12:34:11 -06:00
|
|
|
static irqreturn_t aer_irq(int irq, void *context)
|
2018-06-08 08:31:57 -05:00
|
|
|
{
|
|
|
|
struct pcie_device *pdev = (struct pcie_device *)context;
|
|
|
|
struct aer_rpc *rpc = get_service_data(pdev);
|
2018-09-18 17:58:43 -06:00
|
|
|
struct pci_dev *rp = rpc->rpd;
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = rp->aer_cap;
|
2018-09-18 17:58:43 -06:00
|
|
|
struct aer_err_source e_src = {};
|
2018-06-08 08:31:57 -05:00
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(rp, aer + PCI_ERR_ROOT_STATUS, &e_src.status);
|
2022-04-18 15:02:37 +00:00
|
|
|
if (!(e_src.status & AER_ERR_STATUS_MASK))
|
2018-06-08 08:31:57 -05:00
|
|
|
return IRQ_NONE;
|
|
|
|
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(rp, aer + PCI_ERR_ROOT_ERR_SRC, &e_src.id);
|
|
|
|
pci_write_config_dword(rp, aer + PCI_ERR_ROOT_STATUS, e_src.status);
|
2018-06-08 08:31:57 -05:00
|
|
|
|
2018-09-18 17:58:43 -06:00
|
|
|
if (!kfifo_put(&rpc->aer_fifo, e_src))
|
2018-06-08 08:31:57 -05:00
|
|
|
return IRQ_HANDLED;
|
|
|
|
|
2018-09-18 17:58:46 -06:00
|
|
|
return IRQ_WAKE_THREAD;
|
2018-06-08 08:31:57 -05:00
|
|
|
}
|
|
|
|
|
2023-05-12 08:00:12 +08:00
|
|
|
static void aer_enable_irq(struct pci_dev *pdev)
|
|
|
|
{
|
|
|
|
int aer = pdev->aer_cap;
|
|
|
|
u32 reg32;
|
|
|
|
|
|
|
|
/* Enable Root Port's interrupt in response to error messages */
|
|
|
|
pci_read_config_dword(pdev, aer + PCI_ERR_ROOT_COMMAND, ®32);
|
|
|
|
reg32 |= ROOT_PORT_INTR_ON_MESG_MASK;
|
|
|
|
pci_write_config_dword(pdev, aer + PCI_ERR_ROOT_COMMAND, reg32);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void aer_disable_irq(struct pci_dev *pdev)
|
|
|
|
{
|
|
|
|
int aer = pdev->aer_cap;
|
|
|
|
u32 reg32;
|
|
|
|
|
2025-03-07 17:17:15 -06:00
|
|
|
/* Disable Root Port's interrupt in response to error messages */
|
2023-05-12 08:00:12 +08:00
|
|
|
pci_read_config_dword(pdev, aer + PCI_ERR_ROOT_COMMAND, ®32);
|
|
|
|
reg32 &= ~ROOT_PORT_INTR_ON_MESG_MASK;
|
|
|
|
pci_write_config_dword(pdev, aer + PCI_ERR_ROOT_COMMAND, reg32);
|
|
|
|
}
|
|
|
|
|
2010-04-15 13:10:53 +09:00
|
|
|
/**
|
|
|
|
* aer_enable_rootport - enable Root Port's interrupts when receiving messages
|
|
|
|
* @rpc: pointer to a Root Port data structure
|
|
|
|
*
|
|
|
|
* Invoked when PCIe bus loads AER service driver.
|
|
|
|
*/
|
|
|
|
static void aer_enable_rootport(struct aer_rpc *rpc)
|
|
|
|
{
|
2018-04-09 16:04:42 -06:00
|
|
|
struct pci_dev *pdev = rpc->rpd;
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = pdev->aer_cap;
|
2010-04-15 13:10:53 +09:00
|
|
|
u16 reg16;
|
|
|
|
u32 reg32;
|
|
|
|
|
|
|
|
/* Clear PCIe Capability's Device Status */
|
2012-07-24 17:20:11 +08:00
|
|
|
pcie_capability_read_word(pdev, PCI_EXP_DEVSTA, ®16);
|
|
|
|
pcie_capability_write_word(pdev, PCI_EXP_DEVSTA, reg16);
|
2010-04-15 13:10:53 +09:00
|
|
|
|
|
|
|
/* Disable system error generation in response to error messages */
|
2012-07-24 17:20:11 +08:00
|
|
|
pcie_capability_clear_word(pdev, PCI_EXP_RTCTL,
|
|
|
|
SYSTEM_ERROR_INTR_ON_MESG_MASK);
|
2010-04-15 13:10:53 +09:00
|
|
|
|
|
|
|
/* Clear error status */
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(pdev, aer + PCI_ERR_ROOT_STATUS, ®32);
|
|
|
|
pci_write_config_dword(pdev, aer + PCI_ERR_ROOT_STATUS, reg32);
|
|
|
|
pci_read_config_dword(pdev, aer + PCI_ERR_COR_STATUS, ®32);
|
|
|
|
pci_write_config_dword(pdev, aer + PCI_ERR_COR_STATUS, reg32);
|
|
|
|
pci_read_config_dword(pdev, aer + PCI_ERR_UNCOR_STATUS, ®32);
|
|
|
|
pci_write_config_dword(pdev, aer + PCI_ERR_UNCOR_STATUS, reg32);
|
2010-04-15 13:10:53 +09:00
|
|
|
|
2023-05-12 08:00:12 +08:00
|
|
|
aer_enable_irq(pdev);
|
2010-04-15 13:10:53 +09:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* aer_disable_rootport - disable Root Port's interrupts when receiving messages
|
|
|
|
* @rpc: pointer to a Root Port data structure
|
|
|
|
*
|
|
|
|
* Invoked when PCIe bus unloads AER service driver.
|
|
|
|
*/
|
|
|
|
static void aer_disable_rootport(struct aer_rpc *rpc)
|
|
|
|
{
|
2018-04-09 16:04:42 -06:00
|
|
|
struct pci_dev *pdev = rpc->rpd;
|
2020-05-29 17:56:09 -05:00
|
|
|
int aer = pdev->aer_cap;
|
2010-04-15 13:10:53 +09:00
|
|
|
u32 reg32;
|
|
|
|
|
2023-05-12 08:00:12 +08:00
|
|
|
aer_disable_irq(pdev);
|
2010-04-15 13:10:53 +09:00
|
|
|
|
|
|
|
/* Clear Root's error status reg */
|
2020-05-29 17:56:09 -05:00
|
|
|
pci_read_config_dword(pdev, aer + PCI_ERR_ROOT_STATUS, ®32);
|
|
|
|
pci_write_config_dword(pdev, aer + PCI_ERR_ROOT_STATUS, reg32);
|
2010-04-15 13:10:53 +09:00
|
|
|
}
|
|
|
|
|
2006-07-31 15:21:33 +08:00
|
|
|
/**
|
|
|
|
* aer_remove - clean up resources
|
|
|
|
* @dev: pointer to the pcie_dev data structure
|
|
|
|
*
|
|
|
|
* Invoked when PCI Express bus unloads or AER probe fails.
|
2010-04-15 13:22:11 +09:00
|
|
|
*/
|
2006-07-31 15:21:33 +08:00
|
|
|
static void aer_remove(struct pcie_device *dev)
|
|
|
|
{
|
|
|
|
struct aer_rpc *rpc = get_service_data(dev);
|
|
|
|
|
2018-09-18 17:58:47 -06:00
|
|
|
aer_disable_rootport(rpc);
|
2006-07-31 15:21:33 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* aer_probe - initialize resources
|
|
|
|
* @dev: pointer to the pcie_dev data structure
|
|
|
|
*
|
|
|
|
* Invoked when PCI Express bus loads AER service driver.
|
2010-04-15 13:22:11 +09:00
|
|
|
*/
|
2012-11-21 15:35:00 -05:00
|
|
|
static int aer_probe(struct pcie_device *dev)
|
2006-07-31 15:21:33 +08:00
|
|
|
{
|
|
|
|
int status;
|
|
|
|
struct aer_rpc *rpc;
|
2018-09-18 17:58:47 -06:00
|
|
|
struct device *device = &dev->device;
|
2019-05-07 18:24:50 -05:00
|
|
|
struct pci_dev *port = dev->port;
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2022-05-09 18:14:41 +00:00
|
|
|
BUILD_BUG_ON(ARRAY_SIZE(aer_correctable_error_string) <
|
|
|
|
AER_MAX_TYPEOF_COR_ERRS);
|
|
|
|
BUILD_BUG_ON(ARRAY_SIZE(aer_uncorrectable_error_string) <
|
|
|
|
AER_MAX_TYPEOF_UNCOR_ERRS);
|
|
|
|
|
2020-12-02 11:26:29 -06:00
|
|
|
/* Limit to Root Ports or Root Complex Event Collectors */
|
|
|
|
if ((pci_pcie_type(port) != PCI_EXP_TYPE_RC_EC) &&
|
|
|
|
(pci_pcie_type(port) != PCI_EXP_TYPE_ROOT_PORT))
|
|
|
|
return -ENODEV;
|
|
|
|
|
2018-09-18 17:58:47 -06:00
|
|
|
rpc = devm_kzalloc(device, sizeof(struct aer_rpc), GFP_KERNEL);
|
2019-05-07 18:24:46 -05:00
|
|
|
if (!rpc)
|
2006-07-31 15:21:33 +08:00
|
|
|
return -ENOMEM;
|
2019-05-07 18:24:46 -05:00
|
|
|
|
2019-05-07 18:24:50 -05:00
|
|
|
rpc->rpd = port;
|
2020-01-23 16:26:31 +08:00
|
|
|
INIT_KFIFO(rpc->aer_fifo);
|
2018-09-18 17:58:46 -06:00
|
|
|
set_service_data(dev, rpc);
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2018-09-18 17:58:47 -06:00
|
|
|
status = devm_request_threaded_irq(device, dev->irq, aer_irq, aer_isr,
|
|
|
|
IRQF_SHARED, "aerdrv", dev);
|
PCI: pcie, aer: checkpatch style cleanup in pcie/aer/*
Before:
drivers/pci/pcie/aer/aer_inject.c
total: 4 errors, 4 warnings, 473 lines checked
drivers/pci/pcie/aer/aerdrv.c
total: 5 errors, 2 warnings, 333 lines checked
drivers/pci/pcie/aer/aerdrv.h
total: 1 errors, 0 warnings, 139 lines checked
drivers/pci/pcie/aer/aerdrv_core.c
total: 4 errors, 3 warnings, 872 lines checked
drivers/pci/pcie/aer/aerdrv_errprint.c
total: 12 errors, 11 warnings, 248 lines checked
After:
drivers/pci/pcie/aer/aer_inject.c
total: 0 errors, 0 warnings, 466 lines checked
drivers/pci/pcie/aer/aerdrv.c
total: 0 errors, 0 warnings, 335 lines checked
drivers/pci/pcie/aer/aerdrv.h
total: 0 errors, 0 warnings, 139 lines checked
drivers/pci/pcie/aer/aerdrv_core.c
total: 0 errors, 0 warnings, 869 lines checked
drivers/pci/pcie/aer/aerdrv_errprint.c
total: 0 errors, 10 warnings, 247 lines checked
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Reviewed-by: Andrew Patterson <andrew.patterson@hp.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
2009-09-07 17:07:29 +09:00
|
|
|
if (status) {
|
2019-05-07 18:24:50 -05:00
|
|
|
pci_err(port, "request AER IRQ %d failed\n", dev->irq);
|
2006-07-31 15:21:33 +08:00
|
|
|
return status;
|
|
|
|
}
|
|
|
|
|
2023-10-18 19:17:11 +02:00
|
|
|
cxl_rch_enable_rcec(port);
|
2006-07-31 15:21:33 +08:00
|
|
|
aer_enable_rootport(rpc);
|
2019-05-07 18:24:50 -05:00
|
|
|
pci_info(port, "enabled with IRQ %d\n", dev->irq);
|
2016-11-21 15:34:02 -06:00
|
|
|
return 0;
|
2006-07-31 15:21:33 +08:00
|
|
|
}
|
|
|
|
|
2024-04-16 12:32:24 +08:00
|
|
|
static int aer_suspend(struct pcie_device *dev)
|
|
|
|
{
|
|
|
|
struct aer_rpc *rpc = get_service_data(dev);
|
|
|
|
|
|
|
|
aer_disable_rootport(rpc);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int aer_resume(struct pcie_device *dev)
|
|
|
|
{
|
|
|
|
struct aer_rpc *rpc = get_service_data(dev);
|
|
|
|
|
|
|
|
aer_enable_rootport(rpc);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-07-31 15:21:33 +08:00
|
|
|
/**
|
2020-11-20 16:10:33 -08:00
|
|
|
* aer_root_reset - reset Root Port hierarchy, RCEC, or RCiEP
|
|
|
|
* @dev: pointer to Root Port, RCEC, or RCiEP
|
2006-07-31 15:21:33 +08:00
|
|
|
*
|
2020-12-02 11:26:29 -06:00
|
|
|
* Invoked by Port Bus driver when performing reset.
|
2010-04-15 13:22:11 +09:00
|
|
|
*/
|
2006-07-31 15:21:33 +08:00
|
|
|
static pci_ers_result_t aer_root_reset(struct pci_dev *dev)
|
|
|
|
{
|
2020-12-02 11:26:29 -06:00
|
|
|
int type = pci_pcie_type(dev);
|
|
|
|
struct pci_dev *root;
|
|
|
|
int aer;
|
|
|
|
struct pci_host_bridge *host = pci_find_host_bridge(dev->bus);
|
2010-04-15 13:09:13 +09:00
|
|
|
u32 reg32;
|
2018-07-19 18:04:09 -05:00
|
|
|
int rc;
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2020-11-20 16:10:33 -08:00
|
|
|
/*
|
|
|
|
* Only Root Ports and RCECs have AER Root Command and Root Status
|
|
|
|
* registers. If "dev" is an RCiEP, the relevant registers are in
|
|
|
|
* the RCEC.
|
|
|
|
*/
|
|
|
|
if (type == PCI_EXP_TYPE_RC_END)
|
|
|
|
root = dev->rcec;
|
|
|
|
else
|
2021-01-04 15:02:57 -08:00
|
|
|
root = pcie_find_root_port(dev);
|
2020-11-20 16:10:33 -08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If the platform retained control of AER, an RCiEP may not have
|
|
|
|
* an RCEC visible to us, so dev->rcec ("root") may be NULL. In
|
|
|
|
* that case, firmware is responsible for these registers.
|
|
|
|
*/
|
|
|
|
aer = root ? root->aer_cap : 0;
|
2020-12-02 11:26:29 -06:00
|
|
|
|
2023-05-12 08:00:12 +08:00
|
|
|
if ((host->native_aer || pcie_ports_native) && aer)
|
|
|
|
aer_disable_irq(root);
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2020-11-20 16:10:33 -08:00
|
|
|
if (type == PCI_EXP_TYPE_RC_EC || type == PCI_EXP_TYPE_RC_END) {
|
2021-08-17 23:35:00 +05:30
|
|
|
rc = pcie_reset_flr(dev, PCI_RESET_DO_RESET);
|
2021-08-17 23:34:53 +05:30
|
|
|
if (!rc)
|
|
|
|
pci_info(dev, "has been reset\n");
|
|
|
|
else
|
|
|
|
pci_info(dev, "not reset (no FLR support: %d)\n", rc);
|
2020-12-02 11:26:29 -06:00
|
|
|
} else {
|
|
|
|
rc = pci_bus_error_reset(dev);
|
2021-01-04 15:02:59 -08:00
|
|
|
pci_info(dev, "%s Port link has been reset (%d)\n",
|
|
|
|
pci_is_root_bus(dev->bus) ? "Root" : "Downstream", rc);
|
2020-12-02 11:26:29 -06:00
|
|
|
}
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2020-12-02 11:26:29 -06:00
|
|
|
if ((host->native_aer || pcie_ports_native) && aer) {
|
PCI/AER: Write AER Capability only when we control it
If an OS has not been granted AER control via _OSC, it should not make
changes to PCI_ERR_ROOT_COMMAND and PCI_ERR_ROOT_STATUS related registers.
Per section 4.5.1 of the System Firmware Intermediary (SFI) _OSC and DPC
Updates ECN [1], this bit also covers these aspects of the PCI Express
Advanced Error Reporting. Based on the above and earlier discussion [2],
make the following changes:
Add a check for the native case (i.e., AER control via _OSC)
Note that the previous "clear, reset, enable" order suggests that the reset
might cause errors that we should ignore. After this commit, those errors
(if any) will remain logged in the PCI_ERR_ROOT_STATUS register.
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] https://lore.kernel.org/linux-pci/20201020162820.GA370938@bjorn-Precision-5520/
Link: https://lore.kernel.org/r/20201121001036.8560-2-sean.v.kelley@intel.com
Tested-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> # non-native/no RCEC
Signed-off-by: Sean V Kelley <sean.v.kelley@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-11-20 16:10:22 -08:00
|
|
|
/* Clear Root Error Status */
|
2020-12-02 11:26:29 -06:00
|
|
|
pci_read_config_dword(root, aer + PCI_ERR_ROOT_STATUS, ®32);
|
|
|
|
pci_write_config_dword(root, aer + PCI_ERR_ROOT_STATUS, reg32);
|
2010-04-15 13:09:13 +09:00
|
|
|
|
2023-05-12 08:00:12 +08:00
|
|
|
aer_enable_irq(root);
|
PCI/AER: Write AER Capability only when we control it
If an OS has not been granted AER control via _OSC, it should not make
changes to PCI_ERR_ROOT_COMMAND and PCI_ERR_ROOT_STATUS related registers.
Per section 4.5.1 of the System Firmware Intermediary (SFI) _OSC and DPC
Updates ECN [1], this bit also covers these aspects of the PCI Express
Advanced Error Reporting. Based on the above and earlier discussion [2],
make the following changes:
Add a check for the native case (i.e., AER control via _OSC)
Note that the previous "clear, reset, enable" order suggests that the reset
might cause errors that we should ignore. After this commit, those errors
(if any) will remain logged in the PCI_ERR_ROOT_STATUS register.
[1] System Firmware Intermediary (SFI) _OSC and DPC Updates ECN, Feb 24,
2020, affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/14076
[2] https://lore.kernel.org/linux-pci/20201020162820.GA370938@bjorn-Precision-5520/
Link: https://lore.kernel.org/r/20201121001036.8560-2-sean.v.kelley@intel.com
Tested-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> # non-native/no RCEC
Signed-off-by: Sean V Kelley <sean.v.kelley@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2020-11-20 16:10:22 -08:00
|
|
|
}
|
2006-07-31 15:21:33 +08:00
|
|
|
|
2018-07-19 18:04:09 -05:00
|
|
|
return rc ? PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_RECOVERED;
|
2006-07-31 15:21:33 +08:00
|
|
|
}
|
|
|
|
|
2018-06-08 08:31:42 -05:00
|
|
|
static struct pcie_port_service_driver aerdriver = {
|
|
|
|
.name = "aer",
|
2020-12-02 11:26:29 -06:00
|
|
|
.port_type = PCIE_ANY_PORT,
|
2018-06-08 08:31:42 -05:00
|
|
|
.service = PCIE_PORT_SERVICE_AER,
|
|
|
|
|
|
|
|
.probe = aer_probe,
|
2024-04-16 12:32:24 +08:00
|
|
|
.suspend = aer_suspend,
|
|
|
|
.resume = aer_resume,
|
2018-06-08 08:31:42 -05:00
|
|
|
.remove = aer_remove,
|
|
|
|
};
|
|
|
|
|
2006-07-31 15:21:33 +08:00
|
|
|
/**
|
2025-03-07 17:17:15 -06:00
|
|
|
* pcie_aer_init - register AER service driver
|
2006-07-31 15:21:33 +08:00
|
|
|
*
|
2025-03-07 17:17:15 -06:00
|
|
|
* Invoked when AER service driver is loaded.
|
2010-04-15 13:22:11 +09:00
|
|
|
*/
|
2018-09-20 10:27:06 -06:00
|
|
|
int __init pcie_aer_init(void)
|
2006-07-31 15:21:33 +08:00
|
|
|
{
|
PCI/AER: Use only _OSC to determine AER ownership
Per the PCI Firmware spec, r3.2, sec 4.5.1, the OS can request control of
AER via bit 3 of the _OSC Control Field. In the returned value of the
Control Field:
The firmware sets [bit 3] to 1 to grant control over PCI Express Advanced
Error Reporting. ... after control is transferred to the operating
system, firmware must not modify the Advanced Error Reporting Capability.
If control of this feature was requested and denied or was not requested,
firmware returns this bit set to 0.
Previously the pci_root driver looked at the HEST FIRMWARE_FIRST bit to
determine whether to request ownership of the AER Capability. This was
based on ACPI spec v6.3, sec 18.3.2.4, and similar sections, which say
things like:
Bit [0] - FIRMWARE_FIRST: If set, indicates that system firmware will
handle errors from this source first.
Bit [1] - GLOBAL: If set, indicates that the settings contained in this
structure apply globally to all PCI Express Devices.
These ACPI references don't say anything about ownership of the AER
Capability.
Remove use of the FIRMWARE_FIRST bit and rely only on the _OSC bit to
determine whether we have control of the AER Capability.
Link: https://lore.kernel.org/r/20181115231605.24352-1-mr.nuke.me@gmail.com/ v1
Link: https://lore.kernel.org/r/20190326172343.28946-1-mr.nuke.me@gmail.com/ v2
Link: https://lore.kernel.org/r/67af2931705bed9a588b5a39d369cb70b9942190.1587925636.git.sathyanarayanan.kuppuswamy@linux.intel.com
[bhelgaas: commit log, note: Alex posted this identical patch 18 months
ago, and I failed to apply it then, so I made him the author, added links
to his postings, and added his Signed-off-by]
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jon Derrick <jonathan.derrick@intel.com>
2020-04-27 18:25:13 -05:00
|
|
|
if (!pci_aer_available())
|
2009-09-16 22:40:22 +02:00
|
|
|
return -ENXIO;
|
2007-02-27 10:22:00 +01:00
|
|
|
return pcie_port_service_register(&aerdriver);
|
2006-07-31 15:21:33 +08:00
|
|
|
}
|