arch/x86/kernel/cpu/mce/internal.h | 48 ++++- arch/x86/kernel/cpu/mce/core.c | 45 ++--- arch/x86/kernel/cpu/mce/intel.c | 303 ++++++++++++---------------- arch/x86/kernel/cpu/mce/threshold.c | 115 +++++++++++ 4 files changed, 304 insertions(+), 207 deletions(-)
Linux CMCI storm mitigation is a big hammer that just disables the CMCI interrupt globally and switches to polling all banks. There are two problems with this: 1) It really is a big hammer. It means that errors reported in other banks from different functional units are all subject to the same polling delay before being processed. 2) Intel systems signal some uncorrected errors using CMCI (e.g. memory controller patrol scrub on Icelake Xeon and newer). Delaying processing these error reports negates some of the benefit of the patrol scrubber providing early notice of errors before they are consumed and cause a machine check. This series throws away the old storm implementation and replaces it with one that keeps track of the weather on each separate machine check bank. When a storm is detected from a bank. On Intel the storm is mitigated by setting a very high threshold for corrected errors to signal CMCI. This threshold does not affect signaling CMCI for uncorrected errors. Signed-off-by: Tony Luck <tony.luck@intel.com> --- Changes since v8: Fixed issue reported by lkp with randconfig build with neither CONFIG_X86_MCE_INTEL not CONFIG_X86_MCE_AMD set by making a cleaner division between the storm tracking code in threshold.c with the restof the code using more function accessors that can be stubbed out. Tony Luck (3): x86/mce: Remove old CMCI storm mitigation code x86/mce: Add per-bank CMCI storm mitigation x86/mce: Handle Intel threshold interrupt storms arch/x86/kernel/cpu/mce/internal.h | 48 ++++- arch/x86/kernel/cpu/mce/core.c | 45 ++--- arch/x86/kernel/cpu/mce/intel.c | 303 ++++++++++++---------------- arch/x86/kernel/cpu/mce/threshold.c | 115 +++++++++++ 4 files changed, 304 insertions(+), 207 deletions(-) base-commit: 6465e260f48790807eef06b583b38ca9789b6072 -- 2.41.0
Linux CMCI storm mitigation is a big hammer that just disables the CMCI interrupt globally and switches to polling all banks. There are two problems with this: 1) It really is a big hammer. It means that errors reported in other banks from different functional units are all subject to the same polling delay before being processed. 2) Intel systems signal some uncorrected errors using CMCI (e.g. memory controller patrol scrub on Icelake Xeon and newer). Delaying processing these error reports negates some of the benefit of the patrol scrubber providing early notice of errors before they are consumed and cause a machine check. This series throws away the old storm implementation and replaces it with one that keeps track of the weather on each separate machine check bank. When a storm is detected from a bank. On Intel the storm is mitigated by setting a very high threshold for corrected errors to signal CMCI. This threshold does not affect signaling CMCI for uncorrected errors. Signed-off-by: Tony Luck <tony.luck@intel.com> --- Changes since v9 (based on Boris reviews) #1 Better commit comment on flow. Added detail that both timer poll and CMCI feed results of scanning each bank into the history calculation. Also added comment in code where mce_trac_storm() is called. #2 Set a flag for banks that don't support CMCI so they can be excluded from history processing #3 Skip history processing if CMCI globally disabled with boot argument mce=cmci_disable #4 Move struct mca_storm_desc definition to internal.h (I had argued against the need for this, but the new "poll_mode" flag added in change #2 needs to be set in intel.c). #5 Add #define NUM_HISTORY_BITS instead of hard-coded "64". #6 Rebase to v6.7-rc1 Tony Luck (3): x86/mce: Remove old CMCI storm mitigation code x86/mce: Add per-bank CMCI storm mitigation x86/mce: Handle Intel threshold interrupt storms arch/x86/kernel/cpu/mce/internal.h | 66 +++++- arch/x86/kernel/cpu/mce/core.c | 53 +++-- arch/x86/kernel/cpu/mce/intel.c | 304 ++++++++++++---------------- arch/x86/kernel/cpu/mce/threshold.c | 115 +++++++++++ 4 files changed, 332 insertions(+), 206 deletions(-) base-commit: b85ea95d086471afb4ad062012a4d73cd328fa86 -- 2.41.0
When a "storm" of CMCI is detected this code mitigates by
disabling CMCI interrupt signalling from all of the banks
owned by the CPU that saw the storm.
There are problems with this approach:
1) It is very coarse grained. In all likelihood only one of the
banks was generating the interrupts, but CMCI is disabled for all.
This means Linux may delay seeing and processing errors logged
from other banks.
2) Although CMCI stands for Corrected Machine Check Interrupt, it
is also used to signal when an uncorrected error is logged. This
is a problem because these errors should be handled in a timely
manner.
Delete all this code in preparation for a finer grained solution.
Reviewed-by: Yazen Ghannam <yazen.ghannam@amd.com>
Tested-by: Yazen Ghannam <yazen.ghannam@amd.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
arch/x86/kernel/cpu/mce/internal.h | 6 --
arch/x86/kernel/cpu/mce/core.c | 20 +---
arch/x86/kernel/cpu/mce/intel.c | 145 -----------------------------
3 files changed, 1 insertion(+), 170 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index e13a26c9c0ac..b18e99016ce5 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -41,9 +41,6 @@ struct dentry *mce_get_debugfs_dir(void);
extern mce_banks_t mce_banks_ce_disabled;
#ifdef CONFIG_X86_MCE_INTEL
-unsigned long cmci_intel_adjust_timer(unsigned long interval);
-bool mce_intel_cmci_poll(void);
-void mce_intel_hcpu_update(unsigned long cpu);
void cmci_disable_bank(int bank);
void intel_init_cmci(void);
void intel_init_lmce(void);
@@ -51,9 +48,6 @@ void intel_clear_lmce(void);
bool intel_filter_mce(struct mce *m);
bool intel_mce_usable_address(struct mce *m);
#else
-# define cmci_intel_adjust_timer mce_adjust_timer_default
-static inline bool mce_intel_cmci_poll(void) { return false; }
-static inline void mce_intel_hcpu_update(unsigned long cpu) { }
static inline void cmci_disable_bank(int bank) { }
static inline void intel_init_cmci(void) { }
static inline void intel_init_lmce(void) { }
diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 7b397370b4d6..117848a63aff 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -1601,13 +1601,6 @@ static unsigned long check_interval = INITIAL_CHECK_INTERVAL;
static DEFINE_PER_CPU(unsigned long, mce_next_interval); /* in jiffies */
static DEFINE_PER_CPU(struct timer_list, mce_timer);
-static unsigned long mce_adjust_timer_default(unsigned long interval)
-{
- return interval;
-}
-
-static unsigned long (*mce_adjust_timer)(unsigned long interval) = mce_adjust_timer_default;
-
static void __start_timer(struct timer_list *t, unsigned long interval)
{
unsigned long when = jiffies + interval;
@@ -1637,15 +1630,9 @@ static void mce_timer_fn(struct timer_list *t)
iv = __this_cpu_read(mce_next_interval);
- if (mce_available(this_cpu_ptr(&cpu_info))) {
+ if (mce_available(this_cpu_ptr(&cpu_info)))
mc_poll_banks();
- if (mce_intel_cmci_poll()) {
- iv = mce_adjust_timer(iv);
- goto done;
- }
- }
-
/*
* Alert userspace if needed. If we logged an MCE, reduce the polling
* interval, otherwise increase the polling interval.
@@ -1655,7 +1642,6 @@ static void mce_timer_fn(struct timer_list *t)
else
iv = min(iv * 2, round_jiffies_relative(check_interval * HZ));
-done:
__this_cpu_write(mce_next_interval, iv);
__start_timer(t, iv);
}
@@ -1995,7 +1981,6 @@ static void mce_zhaoxin_feature_init(struct cpuinfo_x86 *c)
intel_init_cmci();
intel_init_lmce();
- mce_adjust_timer = cmci_intel_adjust_timer;
}
static void mce_zhaoxin_feature_clear(struct cpuinfo_x86 *c)
@@ -2008,7 +1993,6 @@ static void __mcheck_cpu_init_vendor(struct cpuinfo_x86 *c)
switch (c->x86_vendor) {
case X86_VENDOR_INTEL:
mce_intel_feature_init(c);
- mce_adjust_timer = cmci_intel_adjust_timer;
break;
case X86_VENDOR_AMD: {
@@ -2665,8 +2649,6 @@ static void mce_reenable_cpu(void)
static int mce_cpu_dead(unsigned int cpu)
{
- mce_intel_hcpu_update(cpu);
-
/* intentionally ignoring frozen here */
if (!cpuhp_tasks_frozen)
cmci_rediscover();
diff --git a/arch/x86/kernel/cpu/mce/intel.c b/arch/x86/kernel/cpu/mce/intel.c
index 52bce533ddcc..fc4ffc434023 100644
--- a/arch/x86/kernel/cpu/mce/intel.c
+++ b/arch/x86/kernel/cpu/mce/intel.c
@@ -41,15 +41,6 @@
*/
static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned);
-/*
- * CMCI storm detection backoff counter
- *
- * During storm, we reset this counter to INITIAL_CHECK_INTERVAL in case we've
- * encountered an error. If not, we decrement it by one. We signal the end of
- * the CMCI storm when it reaches 0.
- */
-static DEFINE_PER_CPU(int, cmci_backoff_cnt);
-
/*
* cmci_discover_lock protects against parallel discovery attempts
* which could race against each other.
@@ -64,21 +55,6 @@ static DEFINE_RAW_SPINLOCK(cmci_discover_lock);
static DEFINE_SPINLOCK(cmci_poll_lock);
#define CMCI_THRESHOLD 1
-#define CMCI_POLL_INTERVAL (30 * HZ)
-#define CMCI_STORM_INTERVAL (HZ)
-#define CMCI_STORM_THRESHOLD 15
-
-static DEFINE_PER_CPU(unsigned long, cmci_time_stamp);
-static DEFINE_PER_CPU(unsigned int, cmci_storm_cnt);
-static DEFINE_PER_CPU(unsigned int, cmci_storm_state);
-
-enum {
- CMCI_STORM_NONE,
- CMCI_STORM_ACTIVE,
- CMCI_STORM_SUBSIDED,
-};
-
-static atomic_t cmci_storm_on_cpus;
static int cmci_supported(int *banks)
{
@@ -134,124 +110,6 @@ static bool lmce_supported(void)
return tmp & FEAT_CTL_LMCE_ENABLED;
}
-bool mce_intel_cmci_poll(void)
-{
- if (__this_cpu_read(cmci_storm_state) == CMCI_STORM_NONE)
- return false;
-
- /*
- * Reset the counter if we've logged an error in the last poll
- * during the storm.
- */
- if (machine_check_poll(0, this_cpu_ptr(&mce_banks_owned)))
- this_cpu_write(cmci_backoff_cnt, INITIAL_CHECK_INTERVAL);
- else
- this_cpu_dec(cmci_backoff_cnt);
-
- return true;
-}
-
-void mce_intel_hcpu_update(unsigned long cpu)
-{
- if (per_cpu(cmci_storm_state, cpu) == CMCI_STORM_ACTIVE)
- atomic_dec(&cmci_storm_on_cpus);
-
- per_cpu(cmci_storm_state, cpu) = CMCI_STORM_NONE;
-}
-
-static void cmci_toggle_interrupt_mode(bool on)
-{
- unsigned long flags, *owned;
- int bank;
- u64 val;
-
- raw_spin_lock_irqsave(&cmci_discover_lock, flags);
- owned = this_cpu_ptr(mce_banks_owned);
- for_each_set_bit(bank, owned, MAX_NR_BANKS) {
- rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
-
- if (on)
- val |= MCI_CTL2_CMCI_EN;
- else
- val &= ~MCI_CTL2_CMCI_EN;
-
- wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
- }
- raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
-}
-
-unsigned long cmci_intel_adjust_timer(unsigned long interval)
-{
- if ((this_cpu_read(cmci_backoff_cnt) > 0) &&
- (__this_cpu_read(cmci_storm_state) == CMCI_STORM_ACTIVE)) {
- mce_notify_irq();
- return CMCI_STORM_INTERVAL;
- }
-
- switch (__this_cpu_read(cmci_storm_state)) {
- case CMCI_STORM_ACTIVE:
-
- /*
- * We switch back to interrupt mode once the poll timer has
- * silenced itself. That means no events recorded and the timer
- * interval is back to our poll interval.
- */
- __this_cpu_write(cmci_storm_state, CMCI_STORM_SUBSIDED);
- if (!atomic_sub_return(1, &cmci_storm_on_cpus))
- pr_notice("CMCI storm subsided: switching to interrupt mode\n");
-
- fallthrough;
-
- case CMCI_STORM_SUBSIDED:
- /*
- * We wait for all CPUs to go back to SUBSIDED state. When that
- * happens we switch back to interrupt mode.
- */
- if (!atomic_read(&cmci_storm_on_cpus)) {
- __this_cpu_write(cmci_storm_state, CMCI_STORM_NONE);
- cmci_toggle_interrupt_mode(true);
- cmci_recheck();
- }
- return CMCI_POLL_INTERVAL;
- default:
-
- /* We have shiny weather. Let the poll do whatever it thinks. */
- return interval;
- }
-}
-
-static bool cmci_storm_detect(void)
-{
- unsigned int cnt = __this_cpu_read(cmci_storm_cnt);
- unsigned long ts = __this_cpu_read(cmci_time_stamp);
- unsigned long now = jiffies;
- int r;
-
- if (__this_cpu_read(cmci_storm_state) != CMCI_STORM_NONE)
- return true;
-
- if (time_before_eq(now, ts + CMCI_STORM_INTERVAL)) {
- cnt++;
- } else {
- cnt = 1;
- __this_cpu_write(cmci_time_stamp, now);
- }
- __this_cpu_write(cmci_storm_cnt, cnt);
-
- if (cnt <= CMCI_STORM_THRESHOLD)
- return false;
-
- cmci_toggle_interrupt_mode(false);
- __this_cpu_write(cmci_storm_state, CMCI_STORM_ACTIVE);
- r = atomic_add_return(1, &cmci_storm_on_cpus);
- mce_timer_kick(CMCI_STORM_INTERVAL);
- this_cpu_write(cmci_backoff_cnt, INITIAL_CHECK_INTERVAL);
-
- if (r == 1)
- pr_notice("CMCI storm detected: switching to poll mode\n");
- return true;
-}
-
/*
* The interrupt handler. This is called on every event.
* Just call the poller directly to log any events.
@@ -260,9 +118,6 @@ static bool cmci_storm_detect(void)
*/
static void intel_threshold_interrupt(void)
{
- if (cmci_storm_detect())
- return;
-
machine_check_poll(MCP_TIMESTAMP, this_cpu_ptr(&mce_banks_owned));
}
--
2.41.0The following commit has been merged into the ras/core branch of tip:
Commit-ID: 3ed57b41a4125609e9fd03e32228aec61d95fe1f
Gitweb: https://git.kernel.org/tip/3ed57b41a4125609e9fd03e32228aec61d95fe1f
Author: Tony Luck <tony.luck@intel.com>
AuthorDate: Wed, 15 Nov 2023 11:54:48 -08:00
Committer: Borislav Petkov (AMD) <bp@alien8.de>
CommitterDate: Fri, 15 Dec 2023 13:44:12 +01:00
x86/mce: Remove old CMCI storm mitigation code
When a "storm" of corrected machine check interrupts (CMCI) is detected
this code mitigates by disabling CMCI interrupt signalling from all of
the banks owned by the CPU that saw the storm.
There are problems with this approach:
1) It is very coarse grained. In all likelihood only one of the banks
was generating the interrupts, but CMCI is disabled for all. This
means Linux may delay seeing and processing errors logged from other
banks.
2) Although CMCI stands for Corrected Machine Check Interrupt, it is
also used to signal when an uncorrected error is logged. This is
a problem because these errors should be handled in a timely manner.
Delete all this code in preparation for a finer grained solution.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Yazen Ghannam <yazen.ghannam@amd.com>
Tested-by: Yazen Ghannam <yazen.ghannam@amd.com>
Link: https://lore.kernel.org/r/20231115195450.12963-2-tony.luck@intel.com
---
arch/x86/kernel/cpu/mce/core.c | 20 +----
arch/x86/kernel/cpu/mce/intel.c | 145 +----------------------------
arch/x86/kernel/cpu/mce/internal.h | 6 +-
3 files changed, 1 insertion(+), 170 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 1642018..b2ef487 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -1617,13 +1617,6 @@ static unsigned long check_interval = INITIAL_CHECK_INTERVAL;
static DEFINE_PER_CPU(unsigned long, mce_next_interval); /* in jiffies */
static DEFINE_PER_CPU(struct timer_list, mce_timer);
-static unsigned long mce_adjust_timer_default(unsigned long interval)
-{
- return interval;
-}
-
-static unsigned long (*mce_adjust_timer)(unsigned long interval) = mce_adjust_timer_default;
-
static void __start_timer(struct timer_list *t, unsigned long interval)
{
unsigned long when = jiffies + interval;
@@ -1653,15 +1646,9 @@ static void mce_timer_fn(struct timer_list *t)
iv = __this_cpu_read(mce_next_interval);
- if (mce_available(this_cpu_ptr(&cpu_info))) {
+ if (mce_available(this_cpu_ptr(&cpu_info)))
mc_poll_banks();
- if (mce_intel_cmci_poll()) {
- iv = mce_adjust_timer(iv);
- goto done;
- }
- }
-
/*
* Alert userspace if needed. If we logged an MCE, reduce the polling
* interval, otherwise increase the polling interval.
@@ -1671,7 +1658,6 @@ static void mce_timer_fn(struct timer_list *t)
else
iv = min(iv * 2, round_jiffies_relative(check_interval * HZ));
-done:
__this_cpu_write(mce_next_interval, iv);
__start_timer(t, iv);
}
@@ -2011,7 +1997,6 @@ static void mce_zhaoxin_feature_init(struct cpuinfo_x86 *c)
intel_init_cmci();
intel_init_lmce();
- mce_adjust_timer = cmci_intel_adjust_timer;
}
static void mce_zhaoxin_feature_clear(struct cpuinfo_x86 *c)
@@ -2024,7 +2009,6 @@ static void __mcheck_cpu_init_vendor(struct cpuinfo_x86 *c)
switch (c->x86_vendor) {
case X86_VENDOR_INTEL:
mce_intel_feature_init(c);
- mce_adjust_timer = cmci_intel_adjust_timer;
break;
case X86_VENDOR_AMD: {
@@ -2678,8 +2662,6 @@ static void mce_reenable_cpu(void)
static int mce_cpu_dead(unsigned int cpu)
{
- mce_intel_hcpu_update(cpu);
-
/* intentionally ignoring frozen here */
if (!cpuhp_tasks_frozen)
cmci_rediscover();
diff --git a/arch/x86/kernel/cpu/mce/intel.c b/arch/x86/kernel/cpu/mce/intel.c
index 52bce53..fc4ffc4 100644
--- a/arch/x86/kernel/cpu/mce/intel.c
+++ b/arch/x86/kernel/cpu/mce/intel.c
@@ -42,15 +42,6 @@
static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned);
/*
- * CMCI storm detection backoff counter
- *
- * During storm, we reset this counter to INITIAL_CHECK_INTERVAL in case we've
- * encountered an error. If not, we decrement it by one. We signal the end of
- * the CMCI storm when it reaches 0.
- */
-static DEFINE_PER_CPU(int, cmci_backoff_cnt);
-
-/*
* cmci_discover_lock protects against parallel discovery attempts
* which could race against each other.
*/
@@ -64,21 +55,6 @@ static DEFINE_RAW_SPINLOCK(cmci_discover_lock);
static DEFINE_SPINLOCK(cmci_poll_lock);
#define CMCI_THRESHOLD 1
-#define CMCI_POLL_INTERVAL (30 * HZ)
-#define CMCI_STORM_INTERVAL (HZ)
-#define CMCI_STORM_THRESHOLD 15
-
-static DEFINE_PER_CPU(unsigned long, cmci_time_stamp);
-static DEFINE_PER_CPU(unsigned int, cmci_storm_cnt);
-static DEFINE_PER_CPU(unsigned int, cmci_storm_state);
-
-enum {
- CMCI_STORM_NONE,
- CMCI_STORM_ACTIVE,
- CMCI_STORM_SUBSIDED,
-};
-
-static atomic_t cmci_storm_on_cpus;
static int cmci_supported(int *banks)
{
@@ -134,124 +110,6 @@ static bool lmce_supported(void)
return tmp & FEAT_CTL_LMCE_ENABLED;
}
-bool mce_intel_cmci_poll(void)
-{
- if (__this_cpu_read(cmci_storm_state) == CMCI_STORM_NONE)
- return false;
-
- /*
- * Reset the counter if we've logged an error in the last poll
- * during the storm.
- */
- if (machine_check_poll(0, this_cpu_ptr(&mce_banks_owned)))
- this_cpu_write(cmci_backoff_cnt, INITIAL_CHECK_INTERVAL);
- else
- this_cpu_dec(cmci_backoff_cnt);
-
- return true;
-}
-
-void mce_intel_hcpu_update(unsigned long cpu)
-{
- if (per_cpu(cmci_storm_state, cpu) == CMCI_STORM_ACTIVE)
- atomic_dec(&cmci_storm_on_cpus);
-
- per_cpu(cmci_storm_state, cpu) = CMCI_STORM_NONE;
-}
-
-static void cmci_toggle_interrupt_mode(bool on)
-{
- unsigned long flags, *owned;
- int bank;
- u64 val;
-
- raw_spin_lock_irqsave(&cmci_discover_lock, flags);
- owned = this_cpu_ptr(mce_banks_owned);
- for_each_set_bit(bank, owned, MAX_NR_BANKS) {
- rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
-
- if (on)
- val |= MCI_CTL2_CMCI_EN;
- else
- val &= ~MCI_CTL2_CMCI_EN;
-
- wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
- }
- raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
-}
-
-unsigned long cmci_intel_adjust_timer(unsigned long interval)
-{
- if ((this_cpu_read(cmci_backoff_cnt) > 0) &&
- (__this_cpu_read(cmci_storm_state) == CMCI_STORM_ACTIVE)) {
- mce_notify_irq();
- return CMCI_STORM_INTERVAL;
- }
-
- switch (__this_cpu_read(cmci_storm_state)) {
- case CMCI_STORM_ACTIVE:
-
- /*
- * We switch back to interrupt mode once the poll timer has
- * silenced itself. That means no events recorded and the timer
- * interval is back to our poll interval.
- */
- __this_cpu_write(cmci_storm_state, CMCI_STORM_SUBSIDED);
- if (!atomic_sub_return(1, &cmci_storm_on_cpus))
- pr_notice("CMCI storm subsided: switching to interrupt mode\n");
-
- fallthrough;
-
- case CMCI_STORM_SUBSIDED:
- /*
- * We wait for all CPUs to go back to SUBSIDED state. When that
- * happens we switch back to interrupt mode.
- */
- if (!atomic_read(&cmci_storm_on_cpus)) {
- __this_cpu_write(cmci_storm_state, CMCI_STORM_NONE);
- cmci_toggle_interrupt_mode(true);
- cmci_recheck();
- }
- return CMCI_POLL_INTERVAL;
- default:
-
- /* We have shiny weather. Let the poll do whatever it thinks. */
- return interval;
- }
-}
-
-static bool cmci_storm_detect(void)
-{
- unsigned int cnt = __this_cpu_read(cmci_storm_cnt);
- unsigned long ts = __this_cpu_read(cmci_time_stamp);
- unsigned long now = jiffies;
- int r;
-
- if (__this_cpu_read(cmci_storm_state) != CMCI_STORM_NONE)
- return true;
-
- if (time_before_eq(now, ts + CMCI_STORM_INTERVAL)) {
- cnt++;
- } else {
- cnt = 1;
- __this_cpu_write(cmci_time_stamp, now);
- }
- __this_cpu_write(cmci_storm_cnt, cnt);
-
- if (cnt <= CMCI_STORM_THRESHOLD)
- return false;
-
- cmci_toggle_interrupt_mode(false);
- __this_cpu_write(cmci_storm_state, CMCI_STORM_ACTIVE);
- r = atomic_add_return(1, &cmci_storm_on_cpus);
- mce_timer_kick(CMCI_STORM_INTERVAL);
- this_cpu_write(cmci_backoff_cnt, INITIAL_CHECK_INTERVAL);
-
- if (r == 1)
- pr_notice("CMCI storm detected: switching to poll mode\n");
- return true;
-}
-
/*
* The interrupt handler. This is called on every event.
* Just call the poller directly to log any events.
@@ -260,9 +118,6 @@ static bool cmci_storm_detect(void)
*/
static void intel_threshold_interrupt(void)
{
- if (cmci_storm_detect())
- return;
-
machine_check_poll(MCP_TIMESTAMP, this_cpu_ptr(&mce_banks_owned));
}
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index e13a26c..b18e990 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -41,9 +41,6 @@ struct dentry *mce_get_debugfs_dir(void);
extern mce_banks_t mce_banks_ce_disabled;
#ifdef CONFIG_X86_MCE_INTEL
-unsigned long cmci_intel_adjust_timer(unsigned long interval);
-bool mce_intel_cmci_poll(void);
-void mce_intel_hcpu_update(unsigned long cpu);
void cmci_disable_bank(int bank);
void intel_init_cmci(void);
void intel_init_lmce(void);
@@ -51,9 +48,6 @@ void intel_clear_lmce(void);
bool intel_filter_mce(struct mce *m);
bool intel_mce_usable_address(struct mce *m);
#else
-# define cmci_intel_adjust_timer mce_adjust_timer_default
-static inline bool mce_intel_cmci_poll(void) { return false; }
-static inline void mce_intel_hcpu_update(unsigned long cpu) { }
static inline void cmci_disable_bank(int bank) { }
static inline void intel_init_cmci(void) { }
static inline void intel_init_lmce(void) { }This is the core functionality to track CMCI storms at the
machine check bank granularity. Subsequent patches will add
the vendor specific hooks to supply input to the storm
detection and take actions on the start/end of a storm.
machine_check_poll() is called both by the CMCI interrupt code,
and for periodic polls from a timer. Add a hook in this routine
to maintain a bitmap history for each bank showing whether the bank
logged an corrected error or not each time it is polled.
In normal operation the interval between polls of this banks
determines how far to shift the history. The 64 bit width corresponds
to about one second.
When a storm is observed a CPU vendor specific action is taken to reduce
or stop CMCI from the bank that is the source of the storm. The bank
is added to the bitmap of banks for this CPU to poll. The polling rate
is increased to once per second. During a storm each bit in the history
indicates the status of the bank each time it is polled. Thus the history
covers just over a minute.
Declare a storm for that bank if the number of corrected interrupts
seen in that history is above some threshold (defined as 5 in this
series, could be tuned later if there is data to suggest a better
value).
A storm on a bank ends if enough consecutive polls of the bank show
no corrected errors (defined as 30, may also change). That calls the
CPU vendor specific function to revert to normal operational mode,
and changes the polling rate back to the default.
[Changes made based on Boris' comments 23 Oct 2023]
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
arch/x86/kernel/cpu/mce/internal.h | 58 +++++++++++++-
arch/x86/kernel/cpu/mce/core.c | 33 ++++++--
arch/x86/kernel/cpu/mce/threshold.c | 112 ++++++++++++++++++++++++++++
3 files changed, 194 insertions(+), 9 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index b18e99016ce5..e55676f096d8 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -56,7 +56,63 @@ static inline bool intel_filter_mce(struct mce *m) { return false; }
static inline bool intel_mce_usable_address(struct mce *m) { return false; }
#endif
-void mce_timer_kick(unsigned long interval);
+void mce_timer_kick(bool storm);
+
+#ifdef CONFIG_X86_MCE_THRESHOLD
+void cmci_storm_begin(unsigned int bank);
+void cmci_storm_end(unsigned int bank);
+void mce_track_storm(struct mce *mce);
+void mce_inherit_storm(unsigned int bank);
+bool mce_get_storm_mode(void);
+void mce_set_storm_mode(bool storm);
+#else
+static inline void cmci_storm_begin(unsigned int bank) {}
+static inline void cmci_storm_end(unsigned int bank) {}
+static inline void mce_track_storm(struct mce *mce) {}
+static inline void mce_inherit_storm(unsigned int bank) {}
+static inline bool mce_get_storm_mode(void) { return false; }
+static inline void mce_set_storm_mode(bool storm) {}
+#endif
+
+/*
+ * history: bitmask tracking whether errors were seen or not seen in
+ * the most recent polls of a bank. Each '1' bit represents
+ * an error seen.
+ * timestamp: last time (in jiffies) that the bank was polled
+ * in_storm_mode: Is this bank in storm mode?
+ * poll_only: Bank does not support CMCI, skip storm tracking
+ */
+struct storm_bank {
+ u64 history;
+ u64 timestamp;
+ bool in_storm_mode;
+ bool poll_only;
+};
+
+#define NUM_HISTORY_BITS (sizeof(u64) * BITS_PER_BYTE)
+
+/* How many errors within the history buffer mark the start of a storm. */
+#define STORM_BEGIN_THRESHOLD 5
+
+/*
+ * How many polls of machine check bank without an error before declaring
+ * the storm is over. Since it is tracked by the bitmaks in the history
+ * field of struct storm_bank the mask is 30 bits [0 ... 29].
+ */
+#define STORM_END_POLL_THRESHOLD 29
+
+/*
+ * banks: per-cpu, per-bank details
+ * stormy_bank_count: count of MC banks in storm state
+ * poll_mode: CPU is in poll mode
+ */
+struct mca_storm_desc {
+ struct storm_bank banks[MAX_NR_BANKS];
+ u8 stormy_bank_count;
+ bool poll_mode;
+};
+
+DECLARE_PER_CPU(struct mca_storm_desc, storm_desc);
#ifdef CONFIG_ACPI_APEI
int apei_write_mce(struct mce *m);
diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 117848a63aff..820bd7d448c1 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -670,6 +670,16 @@ bool machine_check_poll(enum mcp_flags flags, mce_banks_t *b)
barrier();
m.status = mce_rdmsrl(mca_msr_reg(i, MCA_STATUS));
+ /*
+ * Update storm tracking here, before checking for the
+ * MCI_STATUS_VAL bit. Valid corrected errors count
+ * towards declaring, or maintaining, storm status. No
+ * error in a bank counts towards avoiding, or ending,
+ * storm status.
+ */
+ if (!mca_cfg.cmci_disabled)
+ mce_track_storm(&m);
+
/* If this entry is not valid, ignore it */
if (!(m.status & MCI_STATUS_VAL))
continue;
@@ -1642,22 +1652,29 @@ static void mce_timer_fn(struct timer_list *t)
else
iv = min(iv * 2, round_jiffies_relative(check_interval * HZ));
- __this_cpu_write(mce_next_interval, iv);
- __start_timer(t, iv);
+ if (mce_get_storm_mode()) {
+ __start_timer(t, HZ);
+ } else {
+ __this_cpu_write(mce_next_interval, iv);
+ __start_timer(t, iv);
+ }
}
/*
- * Ensure that the timer is firing in @interval from now.
+ * When a storm starts on any bank on this CPU, switch to polling
+ * once per second. When the storm ends, revert to the default
+ * polling interval.
*/
-void mce_timer_kick(unsigned long interval)
+void mce_timer_kick(bool storm)
{
struct timer_list *t = this_cpu_ptr(&mce_timer);
- unsigned long iv = __this_cpu_read(mce_next_interval);
- __start_timer(t, interval);
+ mce_set_storm_mode(storm);
- if (interval < iv)
- __this_cpu_write(mce_next_interval, interval);
+ if (storm)
+ __start_timer(t, HZ);
+ else
+ __this_cpu_write(mce_next_interval, check_interval * HZ);
}
/* Must not be called in IRQ context where del_timer_sync() can deadlock */
diff --git a/arch/x86/kernel/cpu/mce/threshold.c b/arch/x86/kernel/cpu/mce/threshold.c
index ef4e7bb5fd88..0e1988468ee4 100644
--- a/arch/x86/kernel/cpu/mce/threshold.c
+++ b/arch/x86/kernel/cpu/mce/threshold.c
@@ -29,3 +29,115 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_threshold)
trace_threshold_apic_exit(THRESHOLD_APIC_VECTOR);
apic_eoi();
}
+
+DEFINE_PER_CPU(struct mca_storm_desc, storm_desc);
+
+void mce_inherit_storm(unsigned int bank)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ /*
+ * Previous CPU owning this bank had put it into storm mode,
+ * but the precise history of that storm is unknown. Assume
+ * the worst (all recent polls of the bank found a valid error
+ * logged). This will avoid the new owner prematurely declaring
+ * the storm has ended.
+ */
+ storm->banks[bank].history = ~0ull;
+ storm->banks[bank].timestamp = jiffies;
+}
+
+bool mce_get_storm_mode(void)
+{
+ return __this_cpu_read(storm_desc.poll_mode);
+}
+
+void mce_set_storm_mode(bool storm)
+{
+ __this_cpu_write(storm_desc.poll_mode, storm);
+}
+
+static void mce_handle_storm(unsigned int bank, bool on)
+{
+ switch (boot_cpu_data.x86_vendor) {
+ }
+}
+
+void cmci_storm_begin(unsigned int bank)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ __set_bit(bank, this_cpu_ptr(mce_poll_banks));
+ storm->banks[bank].in_storm_mode = true;
+
+ /*
+ * If this is the first bank on this CPU to enter storm mode
+ * start polling.
+ */
+ if (++storm->stormy_bank_count == 1)
+ mce_timer_kick(true);
+}
+
+void cmci_storm_end(unsigned int bank)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ __clear_bit(bank, this_cpu_ptr(mce_poll_banks));
+ storm->banks[bank].history = 0;
+ storm->banks[bank].in_storm_mode = false;
+
+ /* If no banks left in storm mode, stop polling. */
+ if (!this_cpu_dec_return(storm_desc.stormy_bank_count))
+ mce_timer_kick(false);
+}
+
+void mce_track_storm(struct mce *mce)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+ unsigned long now = jiffies, delta;
+ unsigned int shift = 1;
+ u64 history = 0;
+
+ /* No tracking needed for banks that do not support CMCI */
+ if (storm->banks[mce->bank].poll_only)
+ return;
+
+ /*
+ * When a bank is in storm mode it is polled once per second and
+ * the history mask will record about the last minute of poll results.
+ * If it is not in storm mode, then the bank is only checked when
+ * there is a CMCI interrupt. Check how long it has been since
+ * this bank was last checked, and adjust the amount of "shift"
+ * to apply to history.
+ */
+ if (!storm->banks[mce->bank].in_storm_mode) {
+ delta = now - storm->banks[mce->bank].timestamp;
+ shift = (delta + HZ) / HZ;
+ }
+
+ /* If it has been a long time since the last poll, clear history. */
+ if (shift < NUM_HISTORY_BITS)
+ history = storm->banks[mce->bank].history << shift;
+
+ storm->banks[mce->bank].timestamp = now;
+
+ /* History keeps track of corrected errors. VAL=1 && UC=0 */
+ if ((mce->status & MCI_STATUS_VAL) && mce_is_correctable(mce))
+ history |= 1;
+
+ storm->banks[mce->bank].history = history;
+
+ if (storm->banks[mce->bank].in_storm_mode) {
+ if (history & GENMASK_ULL(STORM_END_POLL_THRESHOLD, 0))
+ return;
+ printk_deferred(KERN_NOTICE "CPU%d BANK%d CMCI storm subsided\n", smp_processor_id(), mce->bank);
+ mce_handle_storm(mce->bank, false);
+ cmci_storm_end(mce->bank);
+ } else {
+ if (hweight64(history) < STORM_BEGIN_THRESHOLD)
+ return;
+ printk_deferred(KERN_NOTICE "CPU%d BANK%d CMCI storm detected\n", smp_processor_id(), mce->bank);
+ mce_handle_storm(mce->bank, true);
+ cmci_storm_begin(mce->bank);
+ }
+}
--
2.41.0The following commit has been merged into the ras/core branch of tip:
Commit-ID: 7eae17c4add5de46efcca45356388f480103e6d9
Gitweb: https://git.kernel.org/tip/7eae17c4add5de46efcca45356388f480103e6d9
Author: Tony Luck <tony.luck@intel.com>
AuthorDate: Wed, 15 Nov 2023 11:54:49 -08:00
Committer: Borislav Petkov (AMD) <bp@alien8.de>
CommitterDate: Fri, 15 Dec 2023 14:52:01 +01:00
x86/mce: Add per-bank CMCI storm mitigation
This is the core functionality to track CMCI storms at the machine check
bank granularity. Subsequent patches will add the vendor specific hooks
to supply input to the storm detection and take actions on the start/end
of a storm.
machine_check_poll() is called both by the CMCI interrupt code, and for
periodic polls from a timer. Add a hook in this routine to maintain
a bitmap history for each bank showing whether the bank logged an
corrected error or not each time it is polled.
In normal operation the interval between polls of these banks determines
how far to shift the history. The 64 bit width corresponds to about one
second.
When a storm is observed a CPU vendor specific action is taken to reduce
or stop CMCI from the bank that is the source of the storm. The bank is
added to the bitmap of banks for this CPU to poll. The polling rate is
increased to once per second. During a storm each bit in the history
indicates the status of the bank each time it is polled. Thus the
history covers just over a minute.
Declare a storm for that bank if the number of corrected interrupts seen
in that history is above some threshold (defined as 5 in this series,
could be tuned later if there is data to suggest a better value).
A storm on a bank ends if enough consecutive polls of the bank show no
corrected errors (defined as 30, may also change). That calls the CPU
vendor specific function to revert to normal operational mode, and
changes the polling rate back to the default.
[ bp: Massage. ]
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20231115195450.12963-3-tony.luck@intel.com
---
arch/x86/kernel/cpu/mce/core.c | 33 ++++++--
arch/x86/kernel/cpu/mce/internal.h | 58 +++++++++++++-
arch/x86/kernel/cpu/mce/threshold.c | 112 +++++++++++++++++++++++++++-
3 files changed, 194 insertions(+), 9 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index b2ef487..fd5ce12 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -686,6 +686,16 @@ bool machine_check_poll(enum mcp_flags flags, mce_banks_t *b)
barrier();
m.status = mce_rdmsrl(mca_msr_reg(i, MCA_STATUS));
+ /*
+ * Update storm tracking here, before checking for the
+ * MCI_STATUS_VAL bit. Valid corrected errors count
+ * towards declaring, or maintaining, storm status. No
+ * error in a bank counts towards avoiding, or ending,
+ * storm status.
+ */
+ if (!mca_cfg.cmci_disabled)
+ mce_track_storm(&m);
+
/* If this entry is not valid, ignore it */
if (!(m.status & MCI_STATUS_VAL))
continue;
@@ -1658,22 +1668,29 @@ static void mce_timer_fn(struct timer_list *t)
else
iv = min(iv * 2, round_jiffies_relative(check_interval * HZ));
- __this_cpu_write(mce_next_interval, iv);
- __start_timer(t, iv);
+ if (mce_get_storm_mode()) {
+ __start_timer(t, HZ);
+ } else {
+ __this_cpu_write(mce_next_interval, iv);
+ __start_timer(t, iv);
+ }
}
/*
- * Ensure that the timer is firing in @interval from now.
+ * When a storm starts on any bank on this CPU, switch to polling
+ * once per second. When the storm ends, revert to the default
+ * polling interval.
*/
-void mce_timer_kick(unsigned long interval)
+void mce_timer_kick(bool storm)
{
struct timer_list *t = this_cpu_ptr(&mce_timer);
- unsigned long iv = __this_cpu_read(mce_next_interval);
- __start_timer(t, interval);
+ mce_set_storm_mode(storm);
- if (interval < iv)
- __this_cpu_write(mce_next_interval, interval);
+ if (storm)
+ __start_timer(t, HZ);
+ else
+ __this_cpu_write(mce_next_interval, check_interval * HZ);
}
/* Must not be called in IRQ context where del_timer_sync() can deadlock */
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index b18e990..157b2f2 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -56,7 +56,63 @@ static inline bool intel_filter_mce(struct mce *m) { return false; }
static inline bool intel_mce_usable_address(struct mce *m) { return false; }
#endif
-void mce_timer_kick(unsigned long interval);
+void mce_timer_kick(bool storm);
+
+#ifdef CONFIG_X86_MCE_THRESHOLD
+void cmci_storm_begin(unsigned int bank);
+void cmci_storm_end(unsigned int bank);
+void mce_track_storm(struct mce *mce);
+void mce_inherit_storm(unsigned int bank);
+bool mce_get_storm_mode(void);
+void mce_set_storm_mode(bool storm);
+#else
+static inline void cmci_storm_begin(unsigned int bank) {}
+static inline void cmci_storm_end(unsigned int bank) {}
+static inline void mce_track_storm(struct mce *mce) {}
+static inline void mce_inherit_storm(unsigned int bank) {}
+static inline bool mce_get_storm_mode(void) { return false; }
+static inline void mce_set_storm_mode(bool storm) {}
+#endif
+
+/*
+ * history: Bitmask tracking errors occurrence. Each set bit
+ * represents an error seen.
+ *
+ * timestamp: Last time (in jiffies) that the bank was polled.
+ * in_storm_mode: Is this bank in storm mode?
+ * poll_only: Bank does not support CMCI, skip storm tracking.
+ */
+struct storm_bank {
+ u64 history;
+ u64 timestamp;
+ bool in_storm_mode;
+ bool poll_only;
+};
+
+#define NUM_HISTORY_BITS (sizeof(u64) * BITS_PER_BYTE)
+
+/* How many errors within the history buffer mark the start of a storm. */
+#define STORM_BEGIN_THRESHOLD 5
+
+/*
+ * How many polls of machine check bank without an error before declaring
+ * the storm is over. Since it is tracked by the bitmasks in the history
+ * field of struct storm_bank the mask is 30 bits [0 ... 29].
+ */
+#define STORM_END_POLL_THRESHOLD 29
+
+/*
+ * banks: per-cpu, per-bank details
+ * stormy_bank_count: count of MC banks in storm state
+ * poll_mode: CPU is in poll mode
+ */
+struct mca_storm_desc {
+ struct storm_bank banks[MAX_NR_BANKS];
+ u8 stormy_bank_count;
+ bool poll_mode;
+};
+
+DECLARE_PER_CPU(struct mca_storm_desc, storm_desc);
#ifdef CONFIG_ACPI_APEI
int apei_write_mce(struct mce *m);
diff --git a/arch/x86/kernel/cpu/mce/threshold.c b/arch/x86/kernel/cpu/mce/threshold.c
index ef4e7bb..0e19884 100644
--- a/arch/x86/kernel/cpu/mce/threshold.c
+++ b/arch/x86/kernel/cpu/mce/threshold.c
@@ -29,3 +29,115 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_threshold)
trace_threshold_apic_exit(THRESHOLD_APIC_VECTOR);
apic_eoi();
}
+
+DEFINE_PER_CPU(struct mca_storm_desc, storm_desc);
+
+void mce_inherit_storm(unsigned int bank)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ /*
+ * Previous CPU owning this bank had put it into storm mode,
+ * but the precise history of that storm is unknown. Assume
+ * the worst (all recent polls of the bank found a valid error
+ * logged). This will avoid the new owner prematurely declaring
+ * the storm has ended.
+ */
+ storm->banks[bank].history = ~0ull;
+ storm->banks[bank].timestamp = jiffies;
+}
+
+bool mce_get_storm_mode(void)
+{
+ return __this_cpu_read(storm_desc.poll_mode);
+}
+
+void mce_set_storm_mode(bool storm)
+{
+ __this_cpu_write(storm_desc.poll_mode, storm);
+}
+
+static void mce_handle_storm(unsigned int bank, bool on)
+{
+ switch (boot_cpu_data.x86_vendor) {
+ }
+}
+
+void cmci_storm_begin(unsigned int bank)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ __set_bit(bank, this_cpu_ptr(mce_poll_banks));
+ storm->banks[bank].in_storm_mode = true;
+
+ /*
+ * If this is the first bank on this CPU to enter storm mode
+ * start polling.
+ */
+ if (++storm->stormy_bank_count == 1)
+ mce_timer_kick(true);
+}
+
+void cmci_storm_end(unsigned int bank)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ __clear_bit(bank, this_cpu_ptr(mce_poll_banks));
+ storm->banks[bank].history = 0;
+ storm->banks[bank].in_storm_mode = false;
+
+ /* If no banks left in storm mode, stop polling. */
+ if (!this_cpu_dec_return(storm_desc.stormy_bank_count))
+ mce_timer_kick(false);
+}
+
+void mce_track_storm(struct mce *mce)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+ unsigned long now = jiffies, delta;
+ unsigned int shift = 1;
+ u64 history = 0;
+
+ /* No tracking needed for banks that do not support CMCI */
+ if (storm->banks[mce->bank].poll_only)
+ return;
+
+ /*
+ * When a bank is in storm mode it is polled once per second and
+ * the history mask will record about the last minute of poll results.
+ * If it is not in storm mode, then the bank is only checked when
+ * there is a CMCI interrupt. Check how long it has been since
+ * this bank was last checked, and adjust the amount of "shift"
+ * to apply to history.
+ */
+ if (!storm->banks[mce->bank].in_storm_mode) {
+ delta = now - storm->banks[mce->bank].timestamp;
+ shift = (delta + HZ) / HZ;
+ }
+
+ /* If it has been a long time since the last poll, clear history. */
+ if (shift < NUM_HISTORY_BITS)
+ history = storm->banks[mce->bank].history << shift;
+
+ storm->banks[mce->bank].timestamp = now;
+
+ /* History keeps track of corrected errors. VAL=1 && UC=0 */
+ if ((mce->status & MCI_STATUS_VAL) && mce_is_correctable(mce))
+ history |= 1;
+
+ storm->banks[mce->bank].history = history;
+
+ if (storm->banks[mce->bank].in_storm_mode) {
+ if (history & GENMASK_ULL(STORM_END_POLL_THRESHOLD, 0))
+ return;
+ printk_deferred(KERN_NOTICE "CPU%d BANK%d CMCI storm subsided\n", smp_processor_id(), mce->bank);
+ mce_handle_storm(mce->bank, false);
+ cmci_storm_end(mce->bank);
+ } else {
+ if (hweight64(history) < STORM_BEGIN_THRESHOLD)
+ return;
+ printk_deferred(KERN_NOTICE "CPU%d BANK%d CMCI storm detected\n", smp_processor_id(), mce->bank);
+ mce_handle_storm(mce->bank, true);
+ cmci_storm_begin(mce->bank);
+ }
+}Add an Intel specific hook into machine_check_poll() to keep track
of per-CPU, per-bank corrected error logs (with a stub for the
CONFIG_MCE_INTEL=n case).
When a storm is observed the rate of interrupts is reduced by setting
a large threshold value for this bank in IA32_MCi_CTL2. This bank is
added to the bitmap of banks for this CPU to poll. The polling rate
is increased to once per second.
When a storm ends reset the threshold in IA32_MCi_CTL2 back to 1, remove
the bank from the bitmap for polling, and change the polling rate back
to the default.
If a CPU with banks in storm mode is taken offline, the new CPU
that inherits ownership of those banks takes over management of
storm(s) in the inherited bank(s).
The cmci_discover() function was already very large. These changes
pushed it well over the top. Refactor with three helper functions
to bring it back under control.
Updated with review comments from Yazen.
Link: https://lore.kernel.org/r/6ae4df67-ba0b-4b50-8c1d-a5d382105ad2@amd.com
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
arch/x86/kernel/cpu/mce/internal.h | 2 +
arch/x86/kernel/cpu/mce/intel.c | 205 +++++++++++++++++++++-------
arch/x86/kernel/cpu/mce/threshold.c | 3 +
3 files changed, 160 insertions(+), 50 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index e55676f096d8..6315dbf58146 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -41,6 +41,7 @@ struct dentry *mce_get_debugfs_dir(void);
extern mce_banks_t mce_banks_ce_disabled;
#ifdef CONFIG_X86_MCE_INTEL
+void mce_intel_handle_storm(int bank, bool on);
void cmci_disable_bank(int bank);
void intel_init_cmci(void);
void intel_init_lmce(void);
@@ -48,6 +49,7 @@ void intel_clear_lmce(void);
bool intel_filter_mce(struct mce *m);
bool intel_mce_usable_address(struct mce *m);
#else
+static inline void mce_intel_handle_storm(int bank, bool on) { }
static inline void cmci_disable_bank(int bank) { }
static inline void intel_init_cmci(void) { }
static inline void intel_init_lmce(void) { }
diff --git a/arch/x86/kernel/cpu/mce/intel.c b/arch/x86/kernel/cpu/mce/intel.c
index fc4ffc434023..399b62e223d2 100644
--- a/arch/x86/kernel/cpu/mce/intel.c
+++ b/arch/x86/kernel/cpu/mce/intel.c
@@ -54,8 +54,27 @@ static DEFINE_RAW_SPINLOCK(cmci_discover_lock);
*/
static DEFINE_SPINLOCK(cmci_poll_lock);
+/* Linux non-storm CMCI threshold (may be overridden by BIOS) */
#define CMCI_THRESHOLD 1
+/*
+ * MCi_CTL2 threshold for each bank when there is no storm.
+ * Default value for each bank may have been set by BIOS.
+ */
+static u16 cmci_threshold[MAX_NR_BANKS];
+
+/*
+ * High threshold to limit CMCI rate during storms. Max supported is
+ * 0x7FFF. Use this slightly smaller value so it has a distinctive
+ * signature when some asks "Why am I not seeing all corrected errors?"
+ * A high threshold is used instead of just disabling CMCI for a
+ * bank because both corrected and uncorrected errors may be logged
+ * in the same bank and signalled with CMCI. The threshold only applies
+ * to corrected errors, so keeping CMCI enabled means that uncorrected
+ * errors will still be processed in a timely fashion.
+ */
+#define CMCI_STORM_THRESHOLD 32749
+
static int cmci_supported(int *banks)
{
u64 cap;
@@ -110,6 +129,31 @@ static bool lmce_supported(void)
return tmp & FEAT_CTL_LMCE_ENABLED;
}
+/*
+ * Set a new CMCI threshold value. Preserve the state of the
+ * MCI_CTL2_CMCI_EN bit in case this happens during a
+ * cmci_rediscover() operation.
+ */
+static void cmci_set_threshold(int bank, int thresh)
+{
+ unsigned long flags;
+ u64 val;
+
+ raw_spin_lock_irqsave(&cmci_discover_lock, flags);
+ rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
+ val &= ~MCI_CTL2_CMCI_THRESHOLD_MASK;
+ wrmsrl(MSR_IA32_MCx_CTL2(bank), val | thresh);
+ raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
+}
+
+void mce_intel_handle_storm(int bank, bool on)
+{
+ if (on)
+ cmci_set_threshold(bank, CMCI_STORM_THRESHOLD);
+ else
+ cmci_set_threshold(bank, cmci_threshold[bank]);
+}
+
/*
* The interrupt handler. This is called on every event.
* Just call the poller directly to log any events.
@@ -121,72 +165,130 @@ static void intel_threshold_interrupt(void)
machine_check_poll(MCP_TIMESTAMP, this_cpu_ptr(&mce_banks_owned));
}
+/*
+ * Check all the reasons why current CPU cannot claim
+ * ownership of a bank.
+ * 1: CPU already owns this bank
+ * 2: BIOS owns this bank
+ * 3: Some other CPU owns this bank
+ */
+static bool cmci_skip_bank(int bank, u64 *val)
+{
+ unsigned long *owned = (void *)this_cpu_ptr(&mce_banks_owned);
+
+ if (test_bit(bank, owned))
+ return true;
+
+ /* Skip banks in firmware first mode */
+ if (test_bit(bank, mce_banks_ce_disabled))
+ return true;
+
+ rdmsrl(MSR_IA32_MCx_CTL2(bank), *val);
+
+ /* Already owned by someone else? */
+ if (*val & MCI_CTL2_CMCI_EN) {
+ clear_bit(bank, owned);
+ __clear_bit(bank, this_cpu_ptr(mce_poll_banks));
+ return true;
+ }
+
+ return false;
+}
+
+/*
+ * Decide which CMCI interrupt threshold to use:
+ * 1: If this bank is in storm mode from whichever CPU was
+ * the previous owner, stay in storm mode.
+ * 2: If ignoring any threshold set by BIOS, set Linux default
+ * 3: Try to honor BIOS threshold (unless buggy BIOS set it at zero).
+ */
+static u64 cmci_pick_threshold(u64 val, int *bios_zero_thresh)
+{
+ if ((val & MCI_CTL2_CMCI_THRESHOLD_MASK) == CMCI_STORM_THRESHOLD)
+ return val;
+
+ if (!mca_cfg.bios_cmci_threshold) {
+ val &= ~MCI_CTL2_CMCI_THRESHOLD_MASK;
+ val |= CMCI_THRESHOLD;
+ } else if (!(val & MCI_CTL2_CMCI_THRESHOLD_MASK)) {
+ /*
+ * If bios_cmci_threshold boot option was specified
+ * but the threshold is zero, we'll try to initialize
+ * it to 1.
+ */
+ *bios_zero_thresh = 1;
+ val |= CMCI_THRESHOLD;
+ }
+
+ return val;
+}
+
+/*
+ * Try to claim ownership of a bank.
+ */
+static void cmci_claim_bank(int bank, u64 val, int bios_zero_thresh, int *bios_wrong_thresh)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ val |= MCI_CTL2_CMCI_EN;
+ wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
+ rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
+
+ /* If the enable bit did not stick, this bank should be polled. */
+ if (!(val & MCI_CTL2_CMCI_EN)) {
+ WARN_ON(!test_bit(bank, this_cpu_ptr(mce_poll_banks)));
+ storm->banks[bank].poll_only = true;
+ return;
+ }
+
+ /* This CPU successfully set the enable bit. */
+ set_bit(bank, (void *)this_cpu_ptr(&mce_banks_owned));
+
+ if ((val & MCI_CTL2_CMCI_THRESHOLD_MASK) == CMCI_STORM_THRESHOLD) {
+ pr_notice("CPU%d BANK%d CMCI inherited storm\n", smp_processor_id(), bank);
+ mce_inherit_storm(bank);
+ cmci_storm_begin(bank);
+ } else {
+ __clear_bit(bank, this_cpu_ptr(mce_poll_banks));
+ }
+
+ /*
+ * We are able to set thresholds for some banks that
+ * had a threshold of 0. This means the BIOS has not
+ * set the thresholds properly or does not work with
+ * this boot option. Note down now and report later.
+ */
+ if (mca_cfg.bios_cmci_threshold && bios_zero_thresh &&
+ (val & MCI_CTL2_CMCI_THRESHOLD_MASK))
+ *bios_wrong_thresh = 1;
+
+ /* Save default threshold for each bank */
+ if (cmci_threshold[bank] == 0)
+ cmci_threshold[bank] = val & MCI_CTL2_CMCI_THRESHOLD_MASK;
+}
+
/*
* Enable CMCI (Corrected Machine Check Interrupt) for available MCE banks
* on this CPU. Use the algorithm recommended in the SDM to discover shared
- * banks.
+ * banks. Called during initial bootstrap, and also for hotplug CPU operations
+ * to rediscover/reassign machine check banks.
*/
static void cmci_discover(int banks)
{
- unsigned long *owned = (void *)this_cpu_ptr(&mce_banks_owned);
+ int bios_wrong_thresh = 0;
unsigned long flags;
int i;
- int bios_wrong_thresh = 0;
raw_spin_lock_irqsave(&cmci_discover_lock, flags);
for (i = 0; i < banks; i++) {
u64 val;
int bios_zero_thresh = 0;
- if (test_bit(i, owned))
- continue;
-
- /* Skip banks in firmware first mode */
- if (test_bit(i, mce_banks_ce_disabled))
+ if (cmci_skip_bank(i, &val))
continue;
- rdmsrl(MSR_IA32_MCx_CTL2(i), val);
-
- /* Already owned by someone else? */
- if (val & MCI_CTL2_CMCI_EN) {
- clear_bit(i, owned);
- __clear_bit(i, this_cpu_ptr(mce_poll_banks));
- continue;
- }
-
- if (!mca_cfg.bios_cmci_threshold) {
- val &= ~MCI_CTL2_CMCI_THRESHOLD_MASK;
- val |= CMCI_THRESHOLD;
- } else if (!(val & MCI_CTL2_CMCI_THRESHOLD_MASK)) {
- /*
- * If bios_cmci_threshold boot option was specified
- * but the threshold is zero, we'll try to initialize
- * it to 1.
- */
- bios_zero_thresh = 1;
- val |= CMCI_THRESHOLD;
- }
-
- val |= MCI_CTL2_CMCI_EN;
- wrmsrl(MSR_IA32_MCx_CTL2(i), val);
- rdmsrl(MSR_IA32_MCx_CTL2(i), val);
-
- /* Did the enable bit stick? -- the bank supports CMCI */
- if (val & MCI_CTL2_CMCI_EN) {
- set_bit(i, owned);
- __clear_bit(i, this_cpu_ptr(mce_poll_banks));
- /*
- * We are able to set thresholds for some banks that
- * had a threshold of 0. This means the BIOS has not
- * set the thresholds properly or does not work with
- * this boot option. Note down now and report later.
- */
- if (mca_cfg.bios_cmci_threshold && bios_zero_thresh &&
- (val & MCI_CTL2_CMCI_THRESHOLD_MASK))
- bios_wrong_thresh = 1;
- } else {
- WARN_ON(!test_bit(i, this_cpu_ptr(mce_poll_banks)));
- }
+ val = cmci_pick_threshold(val, &bios_zero_thresh);
+ cmci_claim_bank(i, val, bios_zero_thresh, &bios_wrong_thresh);
}
raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
if (mca_cfg.bios_cmci_threshold && bios_wrong_thresh) {
@@ -225,6 +327,9 @@ static void __cmci_disable_bank(int bank)
val &= ~MCI_CTL2_CMCI_EN;
wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
__clear_bit(bank, this_cpu_ptr(mce_banks_owned));
+
+ if ((val & MCI_CTL2_CMCI_THRESHOLD_MASK) == CMCI_STORM_THRESHOLD)
+ cmci_storm_end(bank);
}
/*
diff --git a/arch/x86/kernel/cpu/mce/threshold.c b/arch/x86/kernel/cpu/mce/threshold.c
index 0e1988468ee4..89e31e1e5c9c 100644
--- a/arch/x86/kernel/cpu/mce/threshold.c
+++ b/arch/x86/kernel/cpu/mce/threshold.c
@@ -60,6 +60,9 @@ void mce_set_storm_mode(bool storm)
static void mce_handle_storm(unsigned int bank, bool on)
{
switch (boot_cpu_data.x86_vendor) {
+ case X86_VENDOR_INTEL:
+ mce_intel_handle_storm(bank, on);
+ break;
}
}
--
2.41.0The following commit has been merged into the ras/core branch of tip:
Commit-ID: 1f68ce2a027250aeeb1756391110cdc4dc97c797
Gitweb: https://git.kernel.org/tip/1f68ce2a027250aeeb1756391110cdc4dc97c797
Author: Tony Luck <tony.luck@intel.com>
AuthorDate: Wed, 15 Nov 2023 11:54:50 -08:00
Committer: Borislav Petkov (AMD) <bp@alien8.de>
CommitterDate: Fri, 15 Dec 2023 14:53:42 +01:00
x86/mce: Handle Intel threshold interrupt storms
Add an Intel specific hook into machine_check_poll() to keep track of
per-CPU, per-bank corrected error logs (with a stub for the
CONFIG_MCE_INTEL=n case).
When a storm is observed the rate of interrupts is reduced by setting
a large threshold value for this bank in IA32_MCi_CTL2. This bank is
added to the bitmap of banks for this CPU to poll. The polling rate is
increased to once per second.
When a storm ends reset the threshold in IA32_MCi_CTL2 back to 1, remove
the bank from the bitmap for polling, and change the polling rate back
to the default.
If a CPU with banks in storm mode is taken offline, the new CPU that
inherits ownership of those banks takes over management of storm(s) in
the inherited bank(s).
The cmci_discover() function was already very large. These changes
pushed it well over the top. Refactor with three helper functions to
bring it back under control.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20231115195450.12963-4-tony.luck@intel.com
---
arch/x86/kernel/cpu/mce/intel.c | 205 ++++++++++++++++++++-------
arch/x86/kernel/cpu/mce/internal.h | 2 +-
arch/x86/kernel/cpu/mce/threshold.c | 3 +-
3 files changed, 160 insertions(+), 50 deletions(-)
diff --git a/arch/x86/kernel/cpu/mce/intel.c b/arch/x86/kernel/cpu/mce/intel.c
index fc4ffc4..399b62e 100644
--- a/arch/x86/kernel/cpu/mce/intel.c
+++ b/arch/x86/kernel/cpu/mce/intel.c
@@ -54,8 +54,27 @@ static DEFINE_RAW_SPINLOCK(cmci_discover_lock);
*/
static DEFINE_SPINLOCK(cmci_poll_lock);
+/* Linux non-storm CMCI threshold (may be overridden by BIOS) */
#define CMCI_THRESHOLD 1
+/*
+ * MCi_CTL2 threshold for each bank when there is no storm.
+ * Default value for each bank may have been set by BIOS.
+ */
+static u16 cmci_threshold[MAX_NR_BANKS];
+
+/*
+ * High threshold to limit CMCI rate during storms. Max supported is
+ * 0x7FFF. Use this slightly smaller value so it has a distinctive
+ * signature when some asks "Why am I not seeing all corrected errors?"
+ * A high threshold is used instead of just disabling CMCI for a
+ * bank because both corrected and uncorrected errors may be logged
+ * in the same bank and signalled with CMCI. The threshold only applies
+ * to corrected errors, so keeping CMCI enabled means that uncorrected
+ * errors will still be processed in a timely fashion.
+ */
+#define CMCI_STORM_THRESHOLD 32749
+
static int cmci_supported(int *banks)
{
u64 cap;
@@ -111,6 +130,31 @@ static bool lmce_supported(void)
}
/*
+ * Set a new CMCI threshold value. Preserve the state of the
+ * MCI_CTL2_CMCI_EN bit in case this happens during a
+ * cmci_rediscover() operation.
+ */
+static void cmci_set_threshold(int bank, int thresh)
+{
+ unsigned long flags;
+ u64 val;
+
+ raw_spin_lock_irqsave(&cmci_discover_lock, flags);
+ rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
+ val &= ~MCI_CTL2_CMCI_THRESHOLD_MASK;
+ wrmsrl(MSR_IA32_MCx_CTL2(bank), val | thresh);
+ raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
+}
+
+void mce_intel_handle_storm(int bank, bool on)
+{
+ if (on)
+ cmci_set_threshold(bank, CMCI_STORM_THRESHOLD);
+ else
+ cmci_set_threshold(bank, cmci_threshold[bank]);
+}
+
+/*
* The interrupt handler. This is called on every event.
* Just call the poller directly to log any events.
* This could in theory increase the threshold under high load,
@@ -122,71 +166,129 @@ static void intel_threshold_interrupt(void)
}
/*
+ * Check all the reasons why current CPU cannot claim
+ * ownership of a bank.
+ * 1: CPU already owns this bank
+ * 2: BIOS owns this bank
+ * 3: Some other CPU owns this bank
+ */
+static bool cmci_skip_bank(int bank, u64 *val)
+{
+ unsigned long *owned = (void *)this_cpu_ptr(&mce_banks_owned);
+
+ if (test_bit(bank, owned))
+ return true;
+
+ /* Skip banks in firmware first mode */
+ if (test_bit(bank, mce_banks_ce_disabled))
+ return true;
+
+ rdmsrl(MSR_IA32_MCx_CTL2(bank), *val);
+
+ /* Already owned by someone else? */
+ if (*val & MCI_CTL2_CMCI_EN) {
+ clear_bit(bank, owned);
+ __clear_bit(bank, this_cpu_ptr(mce_poll_banks));
+ return true;
+ }
+
+ return false;
+}
+
+/*
+ * Decide which CMCI interrupt threshold to use:
+ * 1: If this bank is in storm mode from whichever CPU was
+ * the previous owner, stay in storm mode.
+ * 2: If ignoring any threshold set by BIOS, set Linux default
+ * 3: Try to honor BIOS threshold (unless buggy BIOS set it at zero).
+ */
+static u64 cmci_pick_threshold(u64 val, int *bios_zero_thresh)
+{
+ if ((val & MCI_CTL2_CMCI_THRESHOLD_MASK) == CMCI_STORM_THRESHOLD)
+ return val;
+
+ if (!mca_cfg.bios_cmci_threshold) {
+ val &= ~MCI_CTL2_CMCI_THRESHOLD_MASK;
+ val |= CMCI_THRESHOLD;
+ } else if (!(val & MCI_CTL2_CMCI_THRESHOLD_MASK)) {
+ /*
+ * If bios_cmci_threshold boot option was specified
+ * but the threshold is zero, we'll try to initialize
+ * it to 1.
+ */
+ *bios_zero_thresh = 1;
+ val |= CMCI_THRESHOLD;
+ }
+
+ return val;
+}
+
+/*
+ * Try to claim ownership of a bank.
+ */
+static void cmci_claim_bank(int bank, u64 val, int bios_zero_thresh, int *bios_wrong_thresh)
+{
+ struct mca_storm_desc *storm = this_cpu_ptr(&storm_desc);
+
+ val |= MCI_CTL2_CMCI_EN;
+ wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
+ rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
+
+ /* If the enable bit did not stick, this bank should be polled. */
+ if (!(val & MCI_CTL2_CMCI_EN)) {
+ WARN_ON(!test_bit(bank, this_cpu_ptr(mce_poll_banks)));
+ storm->banks[bank].poll_only = true;
+ return;
+ }
+
+ /* This CPU successfully set the enable bit. */
+ set_bit(bank, (void *)this_cpu_ptr(&mce_banks_owned));
+
+ if ((val & MCI_CTL2_CMCI_THRESHOLD_MASK) == CMCI_STORM_THRESHOLD) {
+ pr_notice("CPU%d BANK%d CMCI inherited storm\n", smp_processor_id(), bank);
+ mce_inherit_storm(bank);
+ cmci_storm_begin(bank);
+ } else {
+ __clear_bit(bank, this_cpu_ptr(mce_poll_banks));
+ }
+
+ /*
+ * We are able to set thresholds for some banks that
+ * had a threshold of 0. This means the BIOS has not
+ * set the thresholds properly or does not work with
+ * this boot option. Note down now and report later.
+ */
+ if (mca_cfg.bios_cmci_threshold && bios_zero_thresh &&
+ (val & MCI_CTL2_CMCI_THRESHOLD_MASK))
+ *bios_wrong_thresh = 1;
+
+ /* Save default threshold for each bank */
+ if (cmci_threshold[bank] == 0)
+ cmci_threshold[bank] = val & MCI_CTL2_CMCI_THRESHOLD_MASK;
+}
+
+/*
* Enable CMCI (Corrected Machine Check Interrupt) for available MCE banks
* on this CPU. Use the algorithm recommended in the SDM to discover shared
- * banks.
+ * banks. Called during initial bootstrap, and also for hotplug CPU operations
+ * to rediscover/reassign machine check banks.
*/
static void cmci_discover(int banks)
{
- unsigned long *owned = (void *)this_cpu_ptr(&mce_banks_owned);
+ int bios_wrong_thresh = 0;
unsigned long flags;
int i;
- int bios_wrong_thresh = 0;
raw_spin_lock_irqsave(&cmci_discover_lock, flags);
for (i = 0; i < banks; i++) {
u64 val;
int bios_zero_thresh = 0;
- if (test_bit(i, owned))
- continue;
-
- /* Skip banks in firmware first mode */
- if (test_bit(i, mce_banks_ce_disabled))
+ if (cmci_skip_bank(i, &val))
continue;
- rdmsrl(MSR_IA32_MCx_CTL2(i), val);
-
- /* Already owned by someone else? */
- if (val & MCI_CTL2_CMCI_EN) {
- clear_bit(i, owned);
- __clear_bit(i, this_cpu_ptr(mce_poll_banks));
- continue;
- }
-
- if (!mca_cfg.bios_cmci_threshold) {
- val &= ~MCI_CTL2_CMCI_THRESHOLD_MASK;
- val |= CMCI_THRESHOLD;
- } else if (!(val & MCI_CTL2_CMCI_THRESHOLD_MASK)) {
- /*
- * If bios_cmci_threshold boot option was specified
- * but the threshold is zero, we'll try to initialize
- * it to 1.
- */
- bios_zero_thresh = 1;
- val |= CMCI_THRESHOLD;
- }
-
- val |= MCI_CTL2_CMCI_EN;
- wrmsrl(MSR_IA32_MCx_CTL2(i), val);
- rdmsrl(MSR_IA32_MCx_CTL2(i), val);
-
- /* Did the enable bit stick? -- the bank supports CMCI */
- if (val & MCI_CTL2_CMCI_EN) {
- set_bit(i, owned);
- __clear_bit(i, this_cpu_ptr(mce_poll_banks));
- /*
- * We are able to set thresholds for some banks that
- * had a threshold of 0. This means the BIOS has not
- * set the thresholds properly or does not work with
- * this boot option. Note down now and report later.
- */
- if (mca_cfg.bios_cmci_threshold && bios_zero_thresh &&
- (val & MCI_CTL2_CMCI_THRESHOLD_MASK))
- bios_wrong_thresh = 1;
- } else {
- WARN_ON(!test_bit(i, this_cpu_ptr(mce_poll_banks)));
- }
+ val = cmci_pick_threshold(val, &bios_zero_thresh);
+ cmci_claim_bank(i, val, bios_zero_thresh, &bios_wrong_thresh);
}
raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
if (mca_cfg.bios_cmci_threshold && bios_wrong_thresh) {
@@ -225,6 +327,9 @@ static void __cmci_disable_bank(int bank)
val &= ~MCI_CTL2_CMCI_EN;
wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
__clear_bit(bank, this_cpu_ptr(mce_banks_owned));
+
+ if ((val & MCI_CTL2_CMCI_THRESHOLD_MASK) == CMCI_STORM_THRESHOLD)
+ cmci_storm_end(bank);
}
/*
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index 157b2f2..01f8f03 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -41,6 +41,7 @@ struct dentry *mce_get_debugfs_dir(void);
extern mce_banks_t mce_banks_ce_disabled;
#ifdef CONFIG_X86_MCE_INTEL
+void mce_intel_handle_storm(int bank, bool on);
void cmci_disable_bank(int bank);
void intel_init_cmci(void);
void intel_init_lmce(void);
@@ -48,6 +49,7 @@ void intel_clear_lmce(void);
bool intel_filter_mce(struct mce *m);
bool intel_mce_usable_address(struct mce *m);
#else
+static inline void mce_intel_handle_storm(int bank, bool on) { }
static inline void cmci_disable_bank(int bank) { }
static inline void intel_init_cmci(void) { }
static inline void intel_init_lmce(void) { }
diff --git a/arch/x86/kernel/cpu/mce/threshold.c b/arch/x86/kernel/cpu/mce/threshold.c
index 0e19884..89e31e1 100644
--- a/arch/x86/kernel/cpu/mce/threshold.c
+++ b/arch/x86/kernel/cpu/mce/threshold.c
@@ -60,6 +60,9 @@ void mce_set_storm_mode(bool storm)
static void mce_handle_storm(unsigned int bank, bool on)
{
switch (boot_cpu_data.x86_vendor) {
+ case X86_VENDOR_INTEL:
+ mce_intel_handle_storm(bank, on);
+ break;
}
}> The following commit has been merged into the ras/core branch of tip: > > Commit-ID: 1f68ce2a027250aeeb1756391110cdc4dc97c797 > Gitweb: https://git.kernel.org/tip/1f68ce2a027250aeeb1756391110cdc4dc97c797 > Author: Tony Luck <tony.luck@intel.com> > AuthorDate: Wed, 15 Nov 2023 11:54:50 -08:00 > Committer: Borislav Petkov (AMD) <bp@alien8.de> > CommitterDate: Fri, 15 Dec 2023 14:53:42 +01:00 Early X-Mas present for me! Thanks Boris. -Tony
On Fri, Dec 15, 2023 at 05:21:12PM +0000, Luck, Tony wrote:
> > The following commit has been merged into the ras/core branch of tip:
> >
> > Commit-ID: 1f68ce2a027250aeeb1756391110cdc4dc97c797
> > Gitweb: https://git.kernel.org/tip/1f68ce2a027250aeeb1756391110cdc4dc97c797
> > Author: Tony Luck <tony.luck@intel.com>
> > AuthorDate: Wed, 15 Nov 2023 11:54:50 -08:00
> > Committer: Borislav Petkov (AMD) <bp@alien8.de>
> > CommitterDate: Fri, 15 Dec 2023 14:53:42 +01:00
>
> Early X-Mas present for me! Thanks Boris.
:-)
You're welcome - thanks for answering my silly questions.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
© 2016 - 2026 Red Hat, Inc.