Re: [PATCH] powerpc/mce: log the error for all unrecoverable errors

[Mahesh J Salgaonkar <mahesh@linux.ibm.com>]

On 2022-11-14 13:26:07 Mon, Ganesh Goudar wrote:
> machine_check_log_err() is not getting called for all
> unrecoverable errors, And we are missing to log the error.
> 
> Raise irq work in save_mce_event() for unrecoverable errors,
> So that we log the error from MCE event handling block in
> timer handler.

Thanks for fixing this.
Reviewed-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>

> 
> Signed-off-by: Ganesh Goudar <ganeshgr@linux.ibm.com>
> ---
>  arch/powerpc/kernel/mce.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/powerpc/kernel/mce.c b/arch/powerpc/kernel/mce.c
> index 6c5d30fba766..a1cb2172eb7b 100644
> --- a/arch/powerpc/kernel/mce.c
> +++ b/arch/powerpc/kernel/mce.c
> @@ -131,6 +131,13 @@ void save_mce_event(struct pt_regs *regs, long handled,
>  	if (mce->error_type == MCE_ERROR_TYPE_UE)
>  		mce->u.ue_error.ignore_event = mce_err->ignore_event;
>  
> +	/*
> +	 * Raise irq work, So that we don't miss to log the error for
> +	 * unrecoverable errors.
> +	 */
> +	if (mce->disposition == MCE_DISPOSITION_NOT_RECOVERED)
> +		mce_irq_work_queue();
> +
>  	if (!addr)
>  		return;
>  
> @@ -235,7 +242,6 @@ static void machine_check_ue_event(struct machine_check_event *evt)
>  	       evt, sizeof(*evt));
>  
>  	/* Queue work to process this event later. */
> -	mce_irq_work_queue();
>  }

With your patch now we can see RTAS event logged for other
unrecoverable errors as well.

[  573.006337] Disabling lock debugging due to kernel taint
[  573.006357] MCE: CPU27: machine check (Severe)  Real address Load/Store (foreign/control memory) [Not recovered]
[  573.006362] MCE: CPU27: PID: 10580 Comm: inject-ra-err NIP: [0000000010000df4]
[  573.006366] MCE: CPU27: Initiator CPU
[  573.006369] MCE: CPU27: Unknown
[  573.006426] RTAS: event: 1, Type: Platform Error (224), Severity: 3

Tested-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>

Thanks,
-Mahesh.