From: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
To: intel-xe@lists.freedesktop.org
Subject: [CI 12/14] drm/xe: Force flush system memory AuxCCS framebuffers before scan out
Date: Wed, 22 Oct 2025 08:32:37 +0100 [thread overview]
Message-ID: <20251022073241.71401-13-tvrtko.ursulin@igalia.com> (raw)
In-Reply-To: <20251022073241.71401-1-tvrtko.ursulin@igalia.com>
Even though frame buffer objects are created as write-combined, in
practice, on top of all the ring buffer flushing, an additional clflush
seems to be needed before display engine can coherently scan out the
AuxCCS compressed data without transient artifacts.
If for comparison we look at how i915 handles things (where AuxCCS works
fine), as it happens it has this same clflush before a frame buffer is
pinned for display for the first time, courtesy the dynamic tracking of
the buffer cache mode and setting the latter to uncached before handing
to display.
Since xe considers the buffer object caching mode as static we can
implement the same approach by adding a flag telling us if the buffer
was ever pinned for display and flush on the first pin. Subsequent re-pins
will not repeat the clflush but so far I have not observed any glitching
after the first pin.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
---
drivers/gpu/drm/xe/display/xe_fb_pin.c | 14 +++++++++++++-
drivers/gpu/drm/xe/xe_bo_types.h | 14 +++++++++-----
2 files changed, 22 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/xe/display/xe_fb_pin.c b/drivers/gpu/drm/xe/display/xe_fb_pin.c
index 8217abc3a20d..02e069df5b73 100644
--- a/drivers/gpu/drm/xe/display/xe_fb_pin.c
+++ b/drivers/gpu/drm/xe/display/xe_fb_pin.c
@@ -345,6 +345,7 @@ static struct i915_vma *__xe_pin_fb_vma(const struct intel_framebuffer *fb,
struct xe_bo *bo = gem_to_xe_bo(obj);
struct xe_validation_ctx ctx;
struct drm_exec exec;
+ bool first_pin;
int ret = 0;
if (!vma)
@@ -385,8 +386,11 @@ static struct i915_vma *__xe_pin_fb_vma(const struct intel_framebuffer *fb,
ret = xe_bo_validate(bo, NULL, true, &exec);
drm_exec_retry_on_contention(&exec);
xe_validation_retry_on_oom(&ctx, &ret);
- if (!ret)
+ if (!ret) {
ttm_bo_pin(&bo->ttm);
+ first_pin = !bo->display_pin;
+ bo->display_pin = true;
+ }
}
if (ret)
goto err;
@@ -399,6 +403,14 @@ static struct i915_vma *__xe_pin_fb_vma(const struct intel_framebuffer *fb,
if (ret)
goto err_unpin;
+ /*
+ * Force flush frame buffer data for non-coherent display access when
+ * AuxCCS formats are used.
+ */
+ if (first_pin && !xe_bo_is_vram(bo) && !xe_bo_is_stolen(bo) &&
+ intel_fb_is_ccs_modifier(fb->base.modifier))
+ drm_clflush_sg(xe_bo_sg(bo));
+
return vma;
err_unpin:
diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
index d4fe3c8dca5b..8ebe14a76cee 100644
--- a/drivers/gpu/drm/xe/xe_bo_types.h
+++ b/drivers/gpu/drm/xe/xe_bo_types.h
@@ -81,11 +81,6 @@ struct xe_bo {
struct llist_node freed;
/** @update_index: Update index if PT BO */
int update_index;
- /** @created: Whether the bo has passed initial creation */
- bool created;
-
- /** @ccs_cleared: true means that CCS region of BO is already cleared */
- bool ccs_cleared;
/** @bb_ccs: BB instructions of CCS read/write. Valid only for VF */
struct xe_bb *bb_ccs[XE_SRIOV_VF_CCS_CTX_COUNT];
@@ -97,6 +92,15 @@ struct xe_bo {
*/
u16 cpu_caching;
+ /** @created: Whether the bo has passed initial creation */
+ bool created : 1;
+
+ /** @ccs_cleared: true means that CCS region of BO is already cleared */
+ bool ccs_cleared : 1;
+
+ /** @display_pin: Was it ever pinned to display */
+ bool display_pin : 1;
+
/** @devmem_allocation: SVM device memory allocation */
struct drm_pagemap_devmem devmem_allocation;
--
2.48.0
next prev parent reply other threads:[~2025-10-22 7:33 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-22 7:32 [CI 00/14] auxccs ci run no stolen with flush Tvrtko Ursulin
2025-10-22 7:32 ` [CI 01/14] drm/xe: Fix ggtt fb alignment Tvrtko Ursulin
2025-10-22 7:32 ` [CI 02/14] drm/xe/xelpg: Flush CCS when flushing caches Tvrtko Ursulin
2025-10-22 7:32 ` [CI 03/14] drm/xe/xelp: Quiesce memory traffic before invalidating AuxCCS Tvrtko Ursulin
2025-10-22 7:32 ` [CI 04/14] drm/xe/xelp: Support auxccs invalidation on blitter Tvrtko Ursulin
2025-10-22 7:32 ` [CI 05/14] drm/xe/xelp: Use MI_FLUSH_DW_CCS on auxccs platforms Tvrtko Ursulin
2025-10-22 7:32 ` [CI 06/14] drm/xe/xelp: Wait for AuxCCS invalidation to complete Tvrtko Ursulin
2025-10-22 7:32 ` [CI 07/14] drm/xe: Export xe_emit_aux_table_inv Tvrtko Ursulin
2025-10-22 7:32 ` [CI 08/14] drm/xe/xelp: Add AuxCCS invalidation to the indirect context workarounds Tvrtko Ursulin
2025-10-22 7:32 ` [CI 09/14] drm/xe: Flush GGTT writes after populating DPT Tvrtko Ursulin
2025-10-22 7:32 ` [CI 10/14] drm/xe: Handle DPT in system memory Tvrtko Ursulin
2025-10-22 7:32 ` [CI 11/14] drm/xe: Do not use stolen memory for DPT on IGFX and AuxCCS Tvrtko Ursulin
2025-10-22 7:32 ` Tvrtko Ursulin [this message]
2025-10-22 7:32 ` [CI 13/14] drm/xe/display: Add support for AuxCCS Tvrtko Ursulin
2025-10-22 7:32 ` [CI 14/14] drm/i915/display: Expose AuxCCS frame buffer modifiers for Xe Tvrtko Ursulin
2025-10-22 8:53 ` ✗ CI.checkpatch: warning for auxccs ci run no stolen with flush Patchwork
2025-10-22 8:54 ` ✓ CI.KUnit: success " Patchwork
2025-10-22 9:36 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-22 11:39 ` ✗ Xe.CI.Full: failure " Patchwork
2025-10-22 17:16 ` ✗ CI.checkpatch: warning for auxccs ci run no stolen with flush (rev2) Patchwork
2025-10-22 17:17 ` ✓ CI.KUnit: success " Patchwork
2025-10-22 17:55 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-22 20:47 ` ✗ Xe.CI.Full: failure " Patchwork
2025-10-23 9:31 ` ✗ CI.checkpatch: warning for auxccs ci run no stolen with flush (rev3) Patchwork
2025-10-23 9:32 ` ✓ CI.KUnit: success " Patchwork
2025-10-23 10:42 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-23 16:11 ` ✗ Xe.CI.Full: failure " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20251022073241.71401-13-tvrtko.ursulin@igalia.com \
--to=tvrtko.ursulin@igalia.com \
--cc=intel-xe@lists.freedesktop.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox