On Fri, Jun 11, 2021 at 08:04:09PM +0530, Tejas Upadhyay wrote:
> w/a on gen11 platforms not working as expected and causing
> more harm on the RC6 flow. Because of subslice steering
> disturbance w/a read is failing. By using ffs we can default
> steering of slice/sublice to minconfig hence w/a will pass
> and any warns will go away.
>
> Fixes: fb899dd8ea9c ("drm/i915: Apply Wa_1406680159:icl,ehl as an engine
> workaround")
> Cc: Mika Kuoppala <[email protected]>
> Cc: Matt Roper <[email protected]>
> Signed-off-by: Tejas Upadhyay <[email protected]>
> ---
> drivers/gpu/drm/i915/gt/intel_workarounds.c | 14 +++++++++++---
> drivers/gpu/drm/i915/intel_pm.c | 10 +++++++---
> 2 files changed, 18 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c
> b/drivers/gpu/drm/i915/gt/intel_workarounds.c
> index b62d1e31a645..68b14141088a 100644
> --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c
> +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c
> @@ -991,13 +991,21 @@ wa_init_mcr(struct drm_i915_private *i915, struct
> i915_wa_list *wal)
> l3_en = ~0;
> }
>
> - slice = fls(sseu->slice_mask) - 1;
> - subslice = fls(l3_en & intel_sseu_get_subslices(sseu, slice));
> + if (GRAPHICS_VER(i915) == 11) {
> + slice = ffs(sseu->slice_mask) - 1;
> + subslice = ffs(l3_en & intel_sseu_get_subslices(sseu, slice));
> + } else {
> + slice = fls(sseu->slice_mask) - 1;
> + subslice = fls(l3_en & intel_sseu_get_subslices(sseu, slice));
> + }
> if (!subslice) {
> drm_warn(&i915->drm,
> "No common index found between subslice mask %x and L3
> bank mask %x!\n",
> intel_sseu_get_subslices(sseu, slice), l3_en);
> - subslice = fls(l3_en);
> + if (GRAPHICS_VER(i915) == 11)
> + subslice = ffs(l3_en);
> + else
> + subslice = fls(l3_en);
> drm_WARN_ON(&i915->drm, !subslice);
> }
> subslice--;
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 45fefa0ed160..d1352ec3546a 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -4049,9 +4049,13 @@ skl_ddb_entry_for_slices(struct drm_i915_private
> *dev_priv, u8 slice_mask,
> ddb->end = 0;
> return;
> }
> -
> - ddb->start = (ffs(slice_mask) - 1) * slice_size;
> - ddb->end = fls(slice_mask) * slice_size;
> + if (GRAPHICS_VER(dev_priv) == 11) {
> + ddb->start = (fls(slice_mask) - 1) * slice_size;
> + ddb->end = ffs(slice_mask) * slice_size;
> + } else {
> + ddb->start = (ffs(slice_mask) - 1) * slice_size;
> + ddb->end = fls(slice_mask) * slice_size;
> + }
This code has nothing to do with GT slices.
>
> WARN_ON(ddb->start >= ddb->end);
> WARN_ON(ddb->end > INTEL_INFO(dev_priv)->dbuf.size);
> --
> 2.31.1
>
> _______________________________________________
> Intel-gfx mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
