All of lore.kernel.org
 help / color / mirror / Atom feed
* [RFC PATCH 0/2] drm/i915/color: 3D LUT fixes
@ 2026-11-18 19:15 Matthew Schwartz
  2026-11-18 19:15 ` [RFC PATCH 1/2] drm/i915/color: only let the owning plane program the 3D LUT Matthew Schwartz
                   ` (5 more replies)
  0 siblings, 6 replies; 7+ messages in thread
From: Matthew Schwartz @ 2026-11-18 19:15 UTC (permalink / raw)
  To: intel-gfx, intel-xe; +Cc: uma.shankar, chaitanya.kumar.borah, Matthew Schwartz

Two fixes for the 3D LUT support that came with the plane color
pipeline. I ran into both while bringing up the pipe-bound 3D LUT on
Lunar Lake for gamescope, using out-of-tree patches that expose it
through the CRTC color pipeline proposed in [1].

Patch 1 stops planes that do not own the 3D LUT from rewriting
LUT_3D_CTL. On Lunar Lake an enable bit set by hand is cleared by the
next plane update. Reading the code, the same write from any
non-primary plane on Xe3p_LPD disables the LUT the primary plane
enabled.

Patch 2 writes the table in the order the hardware reads it. The colorop
blob has red as the fastest axis and the hardware has blue. I measured
this on Lunar Lake only, with the LUT bound to the pipe, by writing
LUT_3D_INDEX and LUT_3D_DATA from userspace and decoding the pipe CRC of
solid color frames. I have no Nova Lake hardware, so the plane-bound
case is unchecked. If the plane binding indexes the table differently
this patch is wrong for it, and I would appreciate a pointer to what
Bspec says about the entry order.

[1] https://lore.kernel.org/all/20260716160556.227556-1-mwen@igalia.com/

Matthew Schwartz (2):
  drm/i915/color: only let the owning plane program the 3D LUT
  drm/i915/color: load the 3D LUT in the order the hardware indexes it

 drivers/gpu/drm/i915/display/intel_color.c    | 28 ++++++++++++++++---
 drivers/gpu/drm/i915/display/intel_color.h    |  5 ++++
 .../drm/i915/display/intel_color_pipeline.c   | 12 ++------
 3 files changed, 31 insertions(+), 14 deletions(-)


base-commit: fb078d17b966f7a24720b6e3a5de06d43c2e184a
-- 
2.56.1


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2026-11-19 22:50 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 19:15 [RFC PATCH 0/2] drm/i915/color: 3D LUT fixes Matthew Schwartz
2026-11-18 19:15 ` [RFC PATCH 1/2] drm/i915/color: only let the owning plane program the 3D LUT Matthew Schwartz
2026-11-18 19:15 ` [RFC PATCH 2/2] drm/i915/color: load the 3D LUT in the order the hardware indexes it Matthew Schwartz
2026-11-19 17:51 ` ✓ CI.KUnit: success for drm/i915/color: 3D LUT fixes Patchwork
2026-11-19 17:53 ` [RFC PATCH 0/2] " Borah, Chaitanya Kumar
2026-11-19 18:28 ` ✓ Xe.CI.BAT: success for " Patchwork
2026-11-19 22:50 ` ✓ i915.CI.BAT: " Patchwork

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.