When a machine error is graded as PANIC by AMD grading logic, the MCE
handler calls mce_panic(). The notification chain does not come into effect
so the AMD EDAC driver does not decode the errors. In these cases, the
messages displayed to the user are more cryptic and miss information
that might be relevant, like the context in which the error took place.
Fix the above issue including messages on AMD's grading logic for machine
errors graded as PANIC.
Signed-off-by: Carlos Bilbao <[email protected]>
---
arch/x86/kernel/cpu/mce/severity.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/arch/x86/kernel/cpu/mce/severity.c b/arch/x86/kernel/cpu/mce/severity.c
index 4d52eef21230..ea4b9407bbad 100644
--- a/arch/x86/kernel/cpu/mce/severity.c
+++ b/arch/x86/kernel/cpu/mce/severity.c
@@ -307,6 +307,7 @@ static noinstr int error_context(struct mce *m, struct pt_regs *regs)
static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **msg, bool is_excp)
{
int ret;
+ char *panic_msg;
/*
* Default return value: Action required, the error must be handled
@@ -316,6 +317,7 @@ static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **
/* Processor Context Corrupt, no need to fumble too much, die! */
if (m->status & MCI_STATUS_PCC) {
+ panic_msg = "Processor Context Corrupt";
ret = MCE_PANIC_SEVERITY;
goto amd_severity;
}
@@ -339,16 +341,21 @@ static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **
if (((m->status & MCI_STATUS_OVER) && !mce_flags.overflow_recov)
|| !mce_flags.succor) {
+ panic_msg = "Uncorrected unrecoverable error";
ret = MCE_PANIC_SEVERITY;
goto amd_severity;
}
if (error_context(m, regs) == IN_KERNEL) {
+ panic_msg = "Uncorrected error in kernel context";
ret = MCE_PANIC_SEVERITY;
}
amd_severity:
+ if (msg && panic_msg)
+ *msg = panic_msg;
+
return ret;
}
--
2.31.1
[AMD Official Use Only]
-----Original Message-----
From: Bilbao, Carlos <[email protected]>
Sent: Thursday, March 31, 2022 12:39 PM
To: [email protected]
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Ghannam, Yazen <[email protected]>; [email protected]; [email protected]; [email protected]; Bilbao, Carlos <[email protected]>
Subject: [PATCH v2 2/2] x86/mce: Add messages to describe panic machine errors on AMD's MCEs grading
When a machine error is graded as PANIC by AMD grading logic, the MCE handler calls mce_panic(). The notification chain does not come into effect so the AMD EDAC driver does not decode the errors. In these cases, the messages displayed to the user are more cryptic and miss information that might be relevant, like the context in which the error took place.
Fix the above issue including messages on AMD's grading logic for machine errors graded as PANIC.
Signed-off-by: Carlos Bilbao <[email protected]>
---
arch/x86/kernel/cpu/mce/severity.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/arch/x86/kernel/cpu/mce/severity.c b/arch/x86/kernel/cpu/mce/severity.c
index 4d52eef21230..ea4b9407bbad 100644
--- a/arch/x86/kernel/cpu/mce/severity.c
+++ b/arch/x86/kernel/cpu/mce/severity.c
@@ -307,6 +307,7 @@ static noinstr int error_context(struct mce *m, struct pt_regs *regs) static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **msg, bool is_excp) {
int ret;
+ char *panic_msg;
I suggest char *panic_msg = NULL;
Simply because its stack-based and a non-null value may get assigned to *msg after the amd_severity: label.
/*
* Default return value: Action required, the error must be handled @@ -316,6 +317,7 @@ static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **
/* Processor Context Corrupt, no need to fumble too much, die! */
if (m->status & MCI_STATUS_PCC) {
+ panic_msg = "Processor Context Corrupt";
ret = MCE_PANIC_SEVERITY;
goto amd_severity;
}
@@ -339,16 +341,21 @@ static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **
if (((m->status & MCI_STATUS_OVER) && !mce_flags.overflow_recov)
|| !mce_flags.succor) {
+ panic_msg = "Uncorrected unrecoverable error";
ret = MCE_PANIC_SEVERITY;
goto amd_severity;
}
if (error_context(m, regs) == IN_KERNEL) {
+ panic_msg = "Uncorrected error in kernel context";
ret = MCE_PANIC_SEVERITY;
}
amd_severity:
+ if (msg && panic_msg)
+ *msg = panic_msg;
+
return ret;
}
--
2.31.1
On Thu, Mar 31, 2022 at 11:38:50AM -0500, Carlos Bilbao wrote:
...
> static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **msg, bool is_excp)
> {
> int ret;
> + char *panic_msg;
Order variable lines from longest to shortest.
And the pointer should be initiliazed to NULL like Mike said also.
>
> /*
> * Default return value: Action required, the error must be handled
> @@ -316,6 +317,7 @@ static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **
>
> /* Processor Context Corrupt, no need to fumble too much, die! */
> if (m->status & MCI_STATUS_PCC) {
> + panic_msg = "Processor Context Corrupt";
> ret = MCE_PANIC_SEVERITY;
> goto amd_severity;
> }
> @@ -339,16 +341,21 @@ static noinstr int mce_severity_amd(struct mce *m, struct pt_regs *regs, char **
>
> if (((m->status & MCI_STATUS_OVER) && !mce_flags.overflow_recov)
> || !mce_flags.succor) {
> + panic_msg = "Uncorrected unrecoverable error";
So these two cases should definitely be separate. One is "Overflowed
uncorrected error without MCA Overflow Recovery", and the other is
"Uncorrected error without MCA Recovery".
> ret = MCE_PANIC_SEVERITY;
> goto amd_severity;
> }
>
> if (error_context(m, regs) == IN_KERNEL) {
> + panic_msg = "Uncorrected error in kernel context";
This should be "Uncorrected unrecoverable error in kernel context". There is
the IN_KERNEL_RECOV error context for a recoverable error in kernel context.
Thanks,
Yazen