> -----Original Message-----
> From: Gupta, Anshuman
> Sent: Saturday, December 3, 2022 3:14 PM
> To: Dixit, Ashutosh <ashutosh.dixit@xxxxxxxxx>; intel-
> gfx@xxxxxxxxxxxxxxxxxxxxx
> Cc: Nilawar, Badal <badal.nilawar@xxxxxxxxx>; Vivi, Rodrigo
> <rodrigo.vivi@xxxxxxxxx>
> Subject: RE: [PATCH] drm/i915/hwmon: Silence "mailbox access failed"
> warning in snb_pcode_read
>
>
>
> > -----Original Message-----
> > From: Dixit, Ashutosh <ashutosh.dixit@xxxxxxxxx>
> > Sent: Saturday, December 3, 2022 8:45 AM
> > To: intel-gfx@xxxxxxxxxxxxxxxxxxxxx
> > Cc: Gupta, Anshuman <anshuman.gupta@xxxxxxxxx>; Nilawar, Badal
> > <badal.nilawar@xxxxxxxxx>
> > Subject: [PATCH] drm/i915/hwmon: Silence "mailbox access failed"
> > warning in snb_pcode_read
> >
> > hwm_pcode_read_i1 is called during i915 load. This results in the
> > following warning from snb_pcode_read because
> > POWER_SETUP_SUBCOMMAND_READ_I1 is unsupported on DG1/DG2.
> >
> > [drm:snb_pcode_read [i915]] warning: pcode (read from mbox 47c) \
> > mailbox access failed for snb_pcode_read_p
> > [i915]: -6
> >
> > The code handles the unsupported command but the warning in dmesg is a
> > red herring which has resulted in a couple of bugs being filed.
> > Therefore silence the warning by avoiding calling snb_pcode_read_p for
> DG1/DG2.
> >
> > Signed-off-by: Ashutosh Dixit <ashutosh.dixit@xxxxxxxxx>
> > ---
> > drivers/gpu/drm/i915/i915_hwmon.c | 4 ++++
> > 1 file changed, 4 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_hwmon.c
> > b/drivers/gpu/drm/i915/i915_hwmon.c
> > index c588a17f97e98..cca7a4350ec8f 100644
> > --- a/drivers/gpu/drm/i915/i915_hwmon.c
> > +++ b/drivers/gpu/drm/i915/i915_hwmon.c
> > @@ -293,6 +293,10 @@ static const struct hwmon_channel_info
> > *hwm_gt_info[] = {
> > /* I1 is exposed as power_crit or as curr_crit depending on bit 31 */
> > static int hwm_pcode_read_i1(struct drm_i915_private *i915, u32 *uval)
> > {
> > + /* Avoid ILLEGAL_SUBCOMMAND "mailbox access failed" warning in
> > snb_pcode_read */
> > + if (IS_DG1(i915) || IS_DG2(i915))
> > + return -ENXIO;
> AFAIK it is specific to client specific parts, how about declaring a is_client
> intel_runtime_info flag to distinguish between client and server part. That
* intel_device_info*
> will also cover any future platform as well.
>
> BR ,
> Anshuman
> > +
> > return snb_pcode_read_p(&i915->uncore, PCODE_POWER_SETUP,
> > POWER_SETUP_SUBCOMMAND_READ_I1, 0,
> uval); }
> > --
> > 2.38.0
