Re: [PATCH v16 1/3] ACPI: APEI: send SIGBUS to current task if synchronous memory error not recovered

["Jarkko Sakkinen" <jarkko@kernel.org>]

On Tue Nov 5, 2024 at 5:09 PM EET, Yazen Ghannam wrote:
> > diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
> > index ada93cfde9ba..af3339dd3817 100644
> > --- a/drivers/acpi/apei/ghes.c
> > +++ b/drivers/acpi/apei/ghes.c
> > @@ -801,6 +801,16 @@ static bool ghes_do_proc(struct ghes *ghes,
> >  		}
> >  	}
> >  
> > +	/*
> > +	 * If no memory failure work is queued for abnormal synchronous
> > +	 * errors, do a force kill.
> > +	 */
> > +	if (sync && !queued) {
> > +		pr_err(HW_ERR GHES_PFX "%s:%d: hardware memory corruption (SIGBUS)\n",
>
> Is this always a memory error? The code flow above implies that an
> unrecoverable ARM processor error can all be !queued. So should the
> message be more generic like "synchronous unrecoverable error" or
> similar?
>
> In any case, this is just a minor nit if this is not an issue in
> practice.

One minor thing that came to mind after reading your response: wouldn't
it be a better idea to use dev_err() against ghes->dev, rather than raw
pr_err()?

That would better context information and also I just (re-)checked the
file and also ghes_remove() is using dev_err().

>
> Reviewed-by: Yazen Ghannam <yazen.ghannam@amd.com>
>
> Thanks,
> Yazen

BR, Jarkko