On Thu, Sep 25, 2025 at 11:31:13AM -0500, Govindapillai, Vinod wrote:
On Wed, 2025-09-24 at 15:46 -0500, Lucas De Marchi wrote:
On Wed, Sep 24, 2025 at 01:03:08PM +0300, Vinod Govindapillai wrote:
> wa_22014263786 is not applicable to the BMG and hence exclude it
> from the wa.
>
> Bspec: 74212, 66624
> Signed-off-by: Vinod Govindapillai <[email protected]>
> ---
> drivers/gpu/drm/i915/display/intel_fbc.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_fbc.c
b/drivers/gpu/drm/i915/display/intel_fbc.c
> index 0d380c825791..2938ee4b64d8 100644
> --- a/drivers/gpu/drm/i915/display/intel_fbc.c
> +++ b/drivers/gpu/drm/i915/display/intel_fbc.c
> @@ -934,7 +934,8 @@ static void intel_fbc_program_workarounds(struct
intel_fbc *fbc)
> 0, DPFC_CHICKEN_COMP_DUMMY_PIXEL);
>
> /* Wa_22014263786:icl,jsl,tgl,dg1,rkl,adls,adlp,mtl */
at this point it's pretty safe to drop these comments as they are not
being updated
> - if (DISPLAY_VER(display) >= 11 && !display->platform.dg2)
> + if (DISPLAY_VER(display) >= 11 && !display->platform.dg2 &&
> + !display->platform.battlemage)
is the plan really to keep adding exceptions? That doesn't seem to
scale.
According to WA database, dg2 should **not** be an exception. Commit
1d7426945716 ("drm/i915/display/fbc: Do not apply WA 22014263786 to
DG2") added that exception without proper follow up to update the
spec/db.
And besides BMG, it looks like LNL is also not affected.
Okay. I misinterpreted the "Project" column in bspec 66624. That's the reason I
decided to add bmg
as an exception to the list! I will fix that!
I recommend you switch to querying the WA db as it's much clearer (and
correct).
While at it, please migrate these to use
intel_display_wa(display, 22014263786)). Then you can probably just
check by IP version `1100 <= display_ver < 1401` seems to cover
everything, but DG2 situation will need to be figured out.
Okay i will use the intel_display_wa() and about the DG2 may be I will put that as
"Todo: assess DG2
need this wa" or something like that?
I think it would be better to follow the spec and test if things break.
If they do, then follow up with HW / spec owners about that.
Nobody will do anything if we just leave a TODO comment in the driver.
Lucas De Marchi
BR
Vinod
Lucas De Marchi
> intel_de_rmw(display, ILK_DPFC_CHICKEN(fbc->id),
> 0, DPFC_CHICKEN_FORCE_SLB_INVALIDATION);
>
> --
> 2.43.0
>
