Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
To: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
Cc: intel-xe@lists.freedesktop.org, kernel-dev@igalia.com
Subject: Re: [PATCH v13 08/12] drm/xe/xelp: Add AuxCCS invalidation to the indirect context workarounds
Date: Mon, 20 Oct 2025 16:13:11 +0100	[thread overview]
Message-ID: <328e531b-e0cc-480a-a3a1-4ebf27f0c5a5@igalia.com> (raw)
In-Reply-To: <aPYdlzzeMLg4bQuJ@intel.com>


On 20/10/2025 12:31, Ville Syrjälä wrote:
> On Mon, Oct 20, 2025 at 08:58:26AM +0100, Tvrtko Ursulin wrote:
>> Following from the i915 reference implementation, we add the AuxCCS
>> invalidation to the indirect context workarounds page.
>>
>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
>> ---
>>   drivers/gpu/drm/xe/xe_lrc.c | 38 +++++++++++++++++++++++++++++++++++++
>>   1 file changed, 38 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_lrc.c b/drivers/gpu/drm/xe/xe_lrc.c
>> index b5083c99dd50..c77a166a2742 100644
>> --- a/drivers/gpu/drm/xe/xe_lrc.c
>> +++ b/drivers/gpu/drm/xe/xe_lrc.c
>> @@ -27,6 +27,7 @@
>>   #include "xe_map.h"
>>   #include "xe_memirq.h"
>>   #include "xe_mmio.h"
>> +#include "xe_ring_ops.h"
>>   #include "xe_sriov.h"
>>   #include "xe_trace_lrc.h"
>>   #include "xe_vm.h"
>> @@ -88,6 +89,16 @@ gt_engine_needs_indirect_ctx(struct xe_gt *gt, enum xe_engine_class class)
>>   					       class, NULL))
>>   		return true;
>>   
>> +	/* AuxCCS invalidation */
>> +	if (GRAPHICS_VERx100(xe) >= 1200 &&
>> +	    GRAPHICS_VERx100(xe) <= 1210 &&
>> +	    (class == XE_ENGINE_CLASS_RENDER ||
>> +	     class == XE_ENGINE_CLASS_COMPUTE ||
>> +	     class == XE_ENGINE_CLASS_COPY ||
>> +	     class == XE_ENGINE_CLASS_VIDEO_DECODE ||
>> +	     class == XE_ENGINE_CLASS_VIDEO_ENHANCE))
>> +		return true;
>> +
>>   	return false;
>>   }
>>   
>> @@ -1188,6 +1199,31 @@ static ssize_t setup_invalidate_state_cache_wa(struct xe_lrc *lrc,
>>   	return cmd - batch;
>>   }
>>   
>> +static ssize_t setup_invalidate_auxccs_wa(struct xe_lrc *lrc,
>> +					  struct xe_hw_engine *hwe,
>> +					  u32 *batch, size_t max_len)
>> +{
>> +	struct xe_gt *gt = lrc->gt;
>> +	struct xe_device *xe = gt_to_xe(gt);
>> +	const unsigned int class = hwe->class;
>> +	u32 *cmd;
>> +
>> +	if (GRAPHICS_VERx100(xe) < 1200 || GRAPHICS_VERx100(xe) > 1210 ||
>> +	    !(class == XE_ENGINE_CLASS_RENDER ||
>> +	      class == XE_ENGINE_CLASS_COMPUTE ||
>> +	      class == XE_ENGINE_CLASS_COPY ||
>> +	      class == XE_ENGINE_CLASS_VIDEO_DECODE ||
>> +	      class == XE_ENGINE_CLASS_VIDEO_ENHANCE))
>> +		return 0;
> 
> I would extract that into a function instead of repeating it twice.

Indeed, thanks! I consolidated it locally as:

	if (xe_engine_supports_auxccs(xe, class))
		return true;

And:

	if (!xe_engine_supports_auxccs(xe, class))
		return 0;

Waiting on more feedback or CI to get better before I send out the respin.

Regards,

Tvrtko

> 
>> +
>> +	if (xe_gt_WARN_ON(gt, max_len < 8))
>> +		return -ENOSPC;
>> +
>> +	cmd = xe_emit_aux_table_inv(hwe, batch);
>> +
>> +	return cmd - batch;
>> +}
>> +
>>   struct bo_setup {
>>   	ssize_t (*setup)(struct xe_lrc *lrc, struct xe_hw_engine *hwe,
>>   			 u32 *batch, size_t max_size);
>> @@ -1320,9 +1356,11 @@ setup_indirect_ctx(struct xe_lrc *lrc, struct xe_hw_engine *hwe)
>>   {
>>   	static const struct bo_setup rcs_funcs[] = {
>>   		{ .setup = setup_timestamp_wa },
>> +		{ .setup = setup_invalidate_auxccs_wa },
>>   		{ .setup = setup_configfs_mid_ctx_restore_bb },
>>   	};
>>   	static const struct bo_setup xcs_funcs[] = {
>> +		{ .setup = setup_invalidate_auxccs_wa },
>>   		{ .setup = setup_configfs_mid_ctx_restore_bb },
>>   	};
>>   	struct bo_setup_state state = {
>> -- 
>> 2.48.0
> 


  reply	other threads:[~2025-10-20 15:13 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-20  7:58 [PATCH v13 00/12] AuxCCS handling and render compression modifiers Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 01/12] drm/xe: Fix ggtt fb alignment Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 02/12] drm/xe/xelpg: Flush CCS when flushing caches Tvrtko Ursulin
2025-10-20 11:27   ` Ville Syrjälä
2025-10-20  7:58 ` [PATCH v13 03/12] drm/xe/xelp: Quiesce memory traffic before invalidating AuxCCS Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 04/12] drm/xe/xelp: Support auxccs invalidation on blitter Tvrtko Ursulin
2025-10-22 22:58   ` Matt Roper
2025-10-23  7:51     ` Tvrtko Ursulin
2025-10-23 17:10       ` Matt Roper
2025-10-24 10:50         ` Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 05/12] drm/xe/xelp: Use MI_FLUSH_DW_CCS on auxccs platforms Tvrtko Ursulin
2025-10-22 23:13   ` Matt Roper
2025-10-23  7:55     ` Tvrtko Ursulin
2025-10-23 17:14       ` Matt Roper
2025-10-20  7:58 ` [PATCH v13 06/12] drm/xe/xelp: Wait for AuxCCS invalidation to complete Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 07/12] drm/xe: Export xe_emit_aux_table_inv Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 08/12] drm/xe/xelp: Add AuxCCS invalidation to the indirect context workarounds Tvrtko Ursulin
2025-10-20 11:31   ` Ville Syrjälä
2025-10-20 15:13     ` Tvrtko Ursulin [this message]
2025-10-20  7:58 ` [PATCH v13 09/12] drm/xe: Flush GGTT writes after populating DPT Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 10/12] drm/xe: Handle DPT in system memory Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 11/12] drm/xe/display: Add support for AuxCCS Tvrtko Ursulin
2025-10-20  7:58 ` [PATCH v13 12/12] drm/i915/display: Expose AuxCCS frame buffer modifiers for Xe Tvrtko Ursulin
2025-10-20 11:46 ` ✗ CI.checkpatch: warning for AuxCCS handling and render compression modifiers Patchwork
2025-10-20 11:47 ` ✓ CI.KUnit: success " Patchwork
2025-10-20 12:32 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-20 14:52 ` ✗ Xe.CI.Full: failure " Patchwork
2025-10-20 15:03   ` Tvrtko Ursulin
2025-10-20 16:35 ` ✗ CI.checkpatch: warning for AuxCCS handling and render compression modifiers (rev2) Patchwork
2025-10-20 16:37 ` ✓ CI.KUnit: success " Patchwork
2025-10-20 22:10 ` ✗ Xe.CI.Full: failure " Patchwork
2025-10-21 16:02 ` ✗ CI.checkpatch: warning for AuxCCS handling and render compression modifiers (rev3) Patchwork
2025-10-21 16:03 ` ✓ CI.KUnit: success " Patchwork
2025-10-21 16:47 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-21 19:39 ` ✗ Xe.CI.Full: failure " Patchwork
2025-10-22  7:36   ` Tvrtko Ursulin
2025-10-22 18:01 ` ✗ CI.checkpatch: warning for AuxCCS handling and render compression modifiers (rev4) Patchwork
2025-10-22 18:02 ` ✓ CI.KUnit: success " Patchwork
2025-10-22 18:40 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-22 22:10 ` ✗ Xe.CI.Full: failure " Patchwork
2025-10-23 10:54 ` ✗ CI.checkpatch: warning for AuxCCS handling and render compression modifiers (rev5) Patchwork
2025-10-23 10:56 ` ✓ CI.KUnit: success " Patchwork
2025-10-23 12:15 ` ✓ Xe.CI.BAT: " Patchwork
2025-10-23 18:58 ` ✗ 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=328e531b-e0cc-480a-a3a1-4ebf27f0c5a5@igalia.com \
    --to=tvrtko.ursulin@igalia.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=kernel-dev@igalia.com \
    --cc=ville.syrjala@linux.intel.com \
    /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