[PATCH] drm/xe/xe3: Extend wa_15015404425 for xe3
Matt Atwood
matthew.s.atwood at intel.com
Thu Feb 6 17:12:49 UTC 2025
On Thu, Jan 30, 2025 at 09:30:30AM -0800, Matt Roper wrote:
> On Tue, Jan 28, 2025 at 04:55:51PM -0800, Matt Atwood wrote:
> > From: Tejas Upadhyay <tejas.upadhyay at intel.com>
> >
> > wa_15015404425 applies to xe3 A0 step as well.
>
> We really need a better commit message explaining how this workaround is
> different from our usual ones and why it needs to be handled in a
> special way that's different from all of our other workarounds,
> especially since someone who just quickly skims the workaround database
> without digging into the details is going to incorrectly think this just
> applies to Xe2 media and isn't relevant to PTL.
Ack.
>
> I was kind of hoping we'd add a more well-defined infrastructure for
> handling device / SoC workarounds. Adding quick hacks like this is
> going to get out of hand quickly if more of these start showing up.
Any particular ideas here?
+Lucas
>
>
> Matt
>
> >
> > v2: query based off SOC stepping
> >
> > Signed-off-by: Tejas Upadhyay <tejas.upadhyay at intel.com>
> > Signed-off-by: Matt Atwood <matthew.s.atwood at intel.com>
> > ---
> > drivers/gpu/drm/xe/xe_mmio.c | 13 +++++++------
> > drivers/gpu/drm/xe/xe_step.h | 2 ++
> > 2 files changed, 9 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/xe/xe_mmio.c b/drivers/gpu/drm/xe/xe_mmio.c
> > index d321a21aacf0..98d068ad33bb 100644
> > --- a/drivers/gpu/drm/xe/xe_mmio.c
> > +++ b/drivers/gpu/drm/xe/xe_mmio.c
> > @@ -150,14 +150,15 @@ int xe_mmio_init(struct xe_device *xe)
> > static void mmio_flush_pending_writes(struct xe_mmio *mmio)
> > {
> > #define DUMMY_REG_OFFSET 0x130030
> > + struct xe_device *xe = tile_to_xe(mmio->tile);
> > int i;
> >
> > - if (mmio->tile->xe->info.platform != XE_LUNARLAKE)
> > - return;
> > -
> > - /* 4 dummy writes */
> > - for (i = 0; i < 4; i++)
> > - writel(0, mmio->regs + DUMMY_REG_OFFSET);
> > + if (xe->info.platform == XE_LUNARLAKE ||
> > + (xe->info.platform == XE_PANTHERLAKE &&
> > + xe->info.revid == PTL_SOC_STEP_A0))
> > + /* 4 dummy writes */
> > + for (i = 0; i < 4; i++)
> > + writel(0, mmio->regs + DUMMY_REG_OFFSET);
> > }
> >
> > u8 xe_mmio_read8(struct xe_mmio *mmio, struct xe_reg reg)
> > diff --git a/drivers/gpu/drm/xe/xe_step.h b/drivers/gpu/drm/xe/xe_step.h
> > index 686cb59200c2..879486a818e9 100644
> > --- a/drivers/gpu/drm/xe/xe_step.h
> > +++ b/drivers/gpu/drm/xe/xe_step.h
> > @@ -20,4 +20,6 @@ static inline u32 xe_step_to_gmdid(enum xe_step step) { return step - STEP_A0; }
> >
> > const char *xe_step_name(enum xe_step step);
> >
> > +#define PTL_SOC_STEP_A0 0x0
> > +
> > #endif
> > --
> > 2.45.0
> >
>
> --
> Matt Roper
> Graphics Software Engineer
> Linux GPU Platform Enablement
> Intel Corporation
More information about the Intel-xe
mailing list