Skip to contentSkip to conversation

AMD VM crashing on deferred memory error injection

PATCH Superseded “William Roche 2 participants 3 messages Feb 13 – Feb 17

Subsystems (3)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

cover AMD VM crashing on deferred memory error injection

From: William Roche <[email protected]>

After the integration of the following commit:
7cb735d7c0cb x86/mce: Unify AMD DFR handler with MCA Polling

AMD Qemu VM started to crash when dealing with deferred memory error injection with a stack trace like:

mce: MSR access error: WRMSR to 0xc0002098 (tried to write 0x0000000000000000) at rIP: 0xffffffff8229894d (mce_wrmsrq+0x1d/0x60)

  amd_clear_bank+0x6e/0x70
  machine_check_poll+0x228/0x2e0
  ? __pfx_mce_timer_fn+0x10/0x10
  mce_timer_fn+0xb1/0x130
  ? __pfx_mce_timer_fn+0x10/0x10
  call_timer_fn+0x26/0x120
  __run_timers+0x202/0x290
  run_timer_softirq+0x49/0x100
  handle_softirqs+0xeb/0x2c0
  __irq_exit_rcu+0xda/0x100
  sysvec_apic_timer_interrupt+0x71/0x90
[...]
 Kernel panic - not syncing: MCA architectural violation!

See the discussion at:
https://lore.kernel.org/all/[email protected]/

We identified a problem with SMCA specific registers access from non-SMCA platforms like a Qemu/KVM machine.

This patch is checkpatch.pl clean.
Unit test of memory error injection works fine with it.

The commit introducing this error has been integrated into the stable tree too, that's the reason why I added the Cc: stable... entry.

Thanks in advance for your feedback.

William Roche (1):
x86/mce: AMD deferred error handling crashes Qemu VMs

 arch/x86/kernel/cpu/mce/amd.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)
Signature · 1 line
2.47.3
#1 lore raw

PATCH 1/1 x86/mce: AMD deferred error handling crashes Qemu VMs

From: William Roche <[email protected]>

A non Scalable MCA system may prevent access to SMCA specific registers like MCA_DESTAT. This is the case of Qemu/kvm VMs, and the VM kernel needs to avoid accessing SMCA registers on non-SMCA platforms.

Fixes 7cb735d7c0cb ("x86/mce: Unify AMD DFR handler with MCA Polling")
---
 arch/x86/kernel/cpu/mce/amd.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)
arch/x86/kernel/cpu/mce/amd.c +11 -7
index 3f1dda355307..53c4b032ad35 100644
@@ -875,14 +875,18 @@ void amd_clear_bank(struct mce *m)
 {
 	amd_reset_thr_limit(m->bank);
 
-	/* Clear MCA_DESTAT for all deferred errors even those logged in MCA_STATUS. */
-	if (m->status & MCI_STATUS_DEFERRED)
-		mce_wrmsrq(MSR_AMD64_SMCA_MCx_DESTAT(m->bank), 0);
-
-	/* Don't clear MCA_STATUS if MCA_DESTAT was used exclusively. */
-	if (m->kflags & MCE_CHECK_DFR_REGS)
-		return;
+	if (mce_flags.smca) {
+		/*
+		 * Clear MCA_DESTAT for all deferred errors even those
+		 * logged in MCA_STATUS.
+		 */
+		if (m->status & MCI_STATUS_DEFERRED)
+			mce_wrmsrq(MSR_AMD64_SMCA_MCx_DESTAT(m->bank), 0);
 
+		/* Don't clear MCA_STATUS if MCA_DESTAT was used exclusively. */
+		if (m->kflags & MCE_CHECK_DFR_REGS)
+			return;
+	}
 	mce_wrmsrq(mca_msr_reg(m->bank, MCA_STATUS), 0);
 }
 
Signature · 1 line
2.47.3
Reviewed-by on v1 1/1
#2 lore raw

Re: [PATCH v1 1/1] x86/mce: AMD deferred error handling crashes Qemu VMs

Quoted 2 lines · “William Roche

From: William Roche <[email protected]>

Hi William, I agree with the fix, and feedback is mostly around style/convention.

Use "x86/mce/amd" as the $SUBJECT prefix.

Subject should be imperative.

Ex. "x86/mce/amd: Fix VM crash during deferred error handling"

Quoted 2 lines

A non Scalable MCA system may prevent access to SMCA specific registers like MCA_DESTAT. This is the case of Qemu/kvm VMs, and the VM kernel

I expect that QEMU and KVM should be capitalized.

Quoted 2 lines

needs to avoid accessing SMCA registers on non-SMCA platforms.

The commit message should include an imperative statement.

Ex. "Check for the SMCA feature before accessing MCA_DESTAT."

Quoted 3 lines
Fixes 7cb735d7c0cb ("x86/mce: Unify AMD DFR handler with MCA Polling")

"CC stable" tag should go after Signed-off-by.

Quoted 21 lines
---
 arch/x86/kernel/cpu/mce/amd.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)
arch/x86/kernel/cpu/mce/amd.c +2 -7
index 3f1dda355307..53c4b032ad35 100644
@@ -875,14 +875,18 @@ void amd_clear_bank(struct mce *m)
 {
 	amd_reset_thr_limit(m->bank);
 
-	/* Clear MCA_DESTAT for all deferred errors even those logged in MCA_STATUS. */
-	if (m->status & MCI_STATUS_DEFERRED)
-		mce_wrmsrq(MSR_AMD64_SMCA_MCx_DESTAT(m->bank), 0);
-
-	/* Don't clear MCA_STATUS if MCA_DESTAT was used exclusively. */
-	if (m->kflags & MCE_CHECK_DFR_REGS)
-		return;
+	if (mce_flags.smca) {
+		/*

arch/x86/kernel/cpu/mce/amd.c void amd_clear_bank(struct mce *m) · line 889

Quoted 10 lines
arch/x86/kernel/cpu/mce/amd.c +9 -0
+		 * Clear MCA_DESTAT for all deferred errors even those
+		 * logged in MCA_STATUS.
+		 */
+		if (m->status & MCI_STATUS_DEFERRED)
+			mce_wrmsrq(MSR_AMD64_SMCA_MCx_DESTAT(m->bank), 0);
 
+		/* Don't clear MCA_STATUS if MCA_DESTAT was used exclusively. */
+		if (m->kflags & MCE_CHECK_DFR_REGS)
+			return;
+	}

Please include a newline here.

arch/x86/kernel/cpu/mce/amd.c void amd_clear_bank(struct mce *m) · line 892

Quoted 3 lines
arch/x86/kernel/cpu/mce/amd.c +0 -0
 	mce_wrmsrq(mca_msr_reg(m->bank, MCA_STATUS), 0);
 }
 

Otherwise, looks good to me.

Thanks,
Yazen