Generally, MCA information for an error is gathered on the CPU that
reported the error. In this case, CPU-specific information from the
running CPU will be correct.

However, this will be incorrect if the MCA information is gathered while
running on a CPU that didn't report the error. One example is creating
an MCA record using mce_prep_record() for errors reported from ACPI.

Split mce_prep_record() so that there is a helper function to gather
common, i.e. not CPU-specific, information and another helper for
CPU-specific information.

Leave mce_prep_record() defined as-is for the common case when running
on the reporting CPU.

Get MCG_CAP in the global helper even though the register is per-CPU.
This value is not already cached per-CPU like other values. And it does
not assist with any per-CPU decoding or handling.

Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
---
Link:
https://lkml.kernel.org/r/20240521125434.1555845-3-yazen.ghannam@amd.com

v1->v2:
* No change.

 arch/x86/kernel/cpu/mce/core.c     | 34 ++++++++++++++++++++----------
 arch/x86/kernel/cpu/mce/internal.h |  2 ++
 2 files changed, 25 insertions(+), 11 deletions(-)

diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index dd5192ef52e0..0133f88dfffb 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -117,20 +117,32 @@ static struct irq_work mce_irq_work;
  */
 BLOCKING_NOTIFIER_HEAD(x86_mce_decoder_chain);
 
-/* Do initial initialization of a struct mce */
-void mce_prep_record(struct mce *m)
+void mce_prep_record_common(struct mce *m)
 {
 	memset(m, 0, sizeof(struct mce));
-	m->cpu = m->extcpu = smp_processor_id();
+
+	m->cpuid	= cpuid_eax(1);
+	m->cpuvendor	= boot_cpu_data.x86_vendor;
+	m->mcgcap	= __rdmsr(MSR_IA32_MCG_CAP);
 	/* need the internal __ version to avoid deadlocks */
-	m->time = __ktime_get_real_seconds();
-	m->cpuvendor = boot_cpu_data.x86_vendor;
-	m->cpuid = cpuid_eax(1);
-	m->socketid = cpu_data(m->extcpu).topo.pkg_id;
-	m->apicid = cpu_data(m->extcpu).topo.initial_apicid;
-	m->mcgcap = __rdmsr(MSR_IA32_MCG_CAP);
-	m->ppin = cpu_data(m->extcpu).ppin;
-	m->microcode = boot_cpu_data.microcode;
+	m->time		= __ktime_get_real_seconds();
+}
+
+void mce_prep_record_per_cpu(unsigned int cpu, struct mce *m)
+{
+	m->cpu		= cpu;
+	m->extcpu	= cpu;
+	m->apicid	= cpu_data(m->extcpu).topo.initial_apicid;
+	m->microcode	= cpu_data(m->extcpu).microcode;
+	m->ppin		= cpu_data(m->extcpu).ppin;
+	m->socketid	= cpu_data(m->extcpu).topo.pkg_id;
+}
+
+/* Do initial initialization of a struct mce */
+void mce_prep_record(struct mce *m)
+{
+	mce_prep_record_common(m);
+	mce_prep_record_per_cpu(smp_processor_id(), m);
 }
 
 DEFINE_PER_CPU(struct mce, injectm);
diff --git a/arch/x86/kernel/cpu/mce/internal.h b/arch/x86/kernel/cpu/mce/internal.h
index 01f8f03969e6..43c7f3b71df5 100644
--- a/arch/x86/kernel/cpu/mce/internal.h
+++ b/arch/x86/kernel/cpu/mce/internal.h
@@ -261,6 +261,8 @@ enum mca_msr {
 
 /* Decide whether to add MCE record to MCE event pool or filter it out. */
 extern bool filter_mce(struct mce *m);
+void mce_prep_record_common(struct mce *m);
+void mce_prep_record_per_cpu(unsigned int cpu, struct mce *m);
 
 #ifdef CONFIG_X86_MCE_AMD
 extern bool amd_filter_mce(struct mce *m);
-- 
2.34.1

On 25.06.24 г. 0:20 ч., Yazen Ghannam wrote:
> Generally, MCA information for an error is gathered on the CPU that
> reported the error. In this case, CPU-specific information from the
> running CPU will be correct.
> 
> However, this will be incorrect if the MCA information is gathered while
> running on a CPU that didn't report the error. One example is creating
> an MCA record using mce_prep_record() for errors reported from ACPI.
> 
> Split mce_prep_record() so that there is a helper function to gather
> common, i.e. not CPU-specific, information and another helper for
> CPU-specific information.
> 
> Leave mce_prep_record() defined as-is for the common case when running
> on the reporting CPU.
> 
> Get MCG_CAP in the global helper even though the register is per-CPU.
> This value is not already cached per-CPU like other values. And it does
> not assist with any per-CPU decoding or handling.
> 
> Signed-off-by: Yazen Ghannam <yazen.ghannam@amd.com>
> ---
> Link:
> https://lkml.kernel.org/r/20240521125434.1555845-3-yazen.ghannam@amd.com
> 
> v1->v2:
> * No change.
> 
>   arch/x86/kernel/cpu/mce/core.c     | 34 ++++++++++++++++++++----------
>   arch/x86/kernel/cpu/mce/internal.h |  2 ++
>   2 files changed, 25 insertions(+), 11 deletions(-)
> 
> diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
> index dd5192ef52e0..0133f88dfffb 100644
> --- a/arch/x86/kernel/cpu/mce/core.c
> +++ b/arch/x86/kernel/cpu/mce/core.c
> @@ -117,20 +117,32 @@ static struct irq_work mce_irq_work;
>    */
>   BLOCKING_NOTIFIER_HEAD(x86_mce_decoder_chain);
>   
> -/* Do initial initialization of a struct mce */
> -void mce_prep_record(struct mce *m)
> +void mce_prep_record_common(struct mce *m)
>   {
>   	memset(m, 0, sizeof(struct mce));
> -	m->cpu = m->extcpu = smp_processor_id();
> +
> +	m->cpuid	= cpuid_eax(1);
> +	m->cpuvendor	= boot_cpu_data.x86_vendor;
> +	m->mcgcap	= __rdmsr(MSR_IA32_MCG_CAP);
>   	/* need the internal __ version to avoid deadlocks */
> -	m->time = __ktime_get_real_seconds();
> -	m->cpuvendor = boot_cpu_data.x86_vendor;
> -	m->cpuid = cpuid_eax(1);
> -	m->socketid = cpu_data(m->extcpu).topo.pkg_id;
> -	m->apicid = cpu_data(m->extcpu).topo.initial_apicid;
> -	m->mcgcap = __rdmsr(MSR_IA32_MCG_CAP);
> -	m->ppin = cpu_data(m->extcpu).ppin;
> -	m->microcode = boot_cpu_data.microcode;
> +	m->time		= __ktime_get_real_seconds();
> +}
> +
> +void mce_prep_record_per_cpu(unsigned int cpu, struct mce *m)
> +{
> +	m->cpu		= cpu;
> +	m->extcpu	= cpu;
> +	m->apicid	= cpu_data(m->extcpu).topo.initial_apicid;
> +	m->microcode	= cpu_data(m->extcpu).microcode;
> +	m->ppin		= cpu_data(m->extcpu).ppin;

nit: Similar to tglx's feedback for patch 2 you could use topology_ppin()

> +	m->socketid	= cpu_data(m->extcpu).topo.pkg_id;
nit: topology_physical_package_id()


> +}
> +


<snip>

On Tue, Jun 25, 2024 at 04:19:40PM +0300, Nikolay Borisov wrote:

[...]

> > +
> > +void mce_prep_record_per_cpu(unsigned int cpu, struct mce *m)
> > +{
> > +	m->cpu		= cpu;
> > +	m->extcpu	= cpu;
> > +	m->apicid	= cpu_data(m->extcpu).topo.initial_apicid;
> > +	m->microcode	= cpu_data(m->extcpu).microcode;
> > +	m->ppin		= cpu_data(m->extcpu).ppin;
> 
> nit: Similar to tglx's feedback for patch 2 you could use topology_ppin()
> 
> > +	m->socketid	= cpu_data(m->extcpu).topo.pkg_id;
> nit: topology_physical_package_id()
> 
>

Yes, will update both.

Thanks,
Yazen