[PATCH 1/2] drm/amdgpu: Warn when bad pages approaches 90% threshold

[Russell, Kent]

[AMD Official Use Only]



> -----Original Message-----
> From: Tuikov, Luben <Luben.Tuikov at amd.com>
> Sent: Thursday, October 21, 2021 12:45 PM
> To: Russell, Kent <Kent.Russell at amd.com>; amd-gfx at lists.freedesktop.org
> Cc: Joshi, Mukul <Mukul.Joshi at amd.com>
> Subject: Re: [PATCH 1/2] drm/amdgpu: Warn when bad pages approaches 90% threshold
> 
> On 2021-10-21 12:35, Luben Tuikov wrote:
> > On 2021-10-21 11:57, Kent Russell wrote:
> >> dmesg doesn't warn when the number of bad pages approaches the
> >> threshold for page retirement. WARN when the number of bad pages
> >> is at 90% or greater for easier checks and planning, instead of waiting
> >> until the GPU is full of bad pages.
> >>
> >> Cc: Luben Tuikov <luben.tuikov at amd.com>
> >> Cc: Mukul Joshi <Mukul.Joshi at amd.com>
> >> Signed-off-by: Kent Russell <kent.russell at amd.com>
> >> ---
> >>  drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c | 6 ++++++
> >>  1 file changed, 6 insertions(+)
> >>
> >> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> >> index f4c05ff4b26c..ce5089216474 100644
> >> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> >> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
> >> @@ -1077,6 +1077,12 @@ int amdgpu_ras_eeprom_init(struct
> amdgpu_ras_eeprom_control *control,
> >>  		if (res)
> >>  			DRM_ERROR("RAS table incorrect checksum or error:%d\n",
> >>  				  res);
> >> +
> >> +		/* Warn if we are at 90% of the threshold or above */
> > The kernel uses a couple of styles, this is one of them:
> >
> > /* Warn ...
> >  */
> > if (...)
> >
> > Please use this style as it is used extensively in the amdgpu_ras_eeprom.c file.
> >
> >> +		if ((10 * control->ras_num_recs) >= (ras->bad_page_cnt_threshold * 9))
> > You don't need the extra parenthesis around multiplication--it has higher precedence
> than relational operators--drop the extra parenthesis.
> >
> > Regards,
> > Luben
> >
> >> +			DRM_WARN("RAS records:%u exceeds 90%% of threshold:%d",
> >> +					control->ras_num_recs,
> >> +					ras->bad_page_cnt_threshold);
> 
> One more note: The code uses "dev_err()" for this very similar message:
> 
>             dev_err(adev->dev,
>                 "RAS records:%d exceed threshold:%d, "
>                 "GPU will not be initialized. Replace this GPU or increase the threshold",
>                 control->ras_num_recs, ras->bad_page_cnt_threshold);
> 
> Since your message is essentially the same, sans the "90% of threshold", perhaps you want
> to use dev_warn(), instead of "DRM_WARN()".

Agreed. Lijo had a similar comment. I may follow up with another patch to change all of these table-specific DRM_* messages to dev_*

 Kent

> 
> Regards,
> Luben
> 
> >>  	} else if (hdr->header == RAS_TABLE_HDR_BAD &&
> >>  		   amdgpu_bad_page_threshold != 0) {
> >>  		res = __verify_ras_table_checksum(control);