Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Michal Wajdeczko <michal.wajdeczko@intel.com>
To: intel-xe@lists.freedesktop.org
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>,
	John Harrison <John.C.Harrison@Intel.com>,
	Rodrigo Vivi <rodrigo.vivi@intel.com>
Subject: [PATCH 2/2] drm/xe/guc: Don't allocate managed BO for each policy change
Date: Tue, 13 May 2025 00:00:18 +0200	[thread overview]
Message-ID: <20250512220018.172-3-michal.wajdeczko@intel.com> (raw)
In-Reply-To: <20250512220018.172-1-michal.wajdeczko@intel.com>

We shouldn't use xe_managed_bo_create_from_data() to allocate
temporary BO, as it will be released only on unload and every
change in wedge_mode policy will consume resources (including
precious GGTT). Instead just switchover to GuC buffer cache.

Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/xe_guc_ads.c | 27 +++++++++++++--------------
 1 file changed, 13 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_guc_ads.c b/drivers/gpu/drm/xe/xe_guc_ads.c
index 44c1fa2fe7c8..07a027755627 100644
--- a/drivers/gpu/drm/xe/xe_guc_ads.c
+++ b/drivers/gpu/drm/xe/xe_guc_ads.c
@@ -20,6 +20,7 @@
 #include "xe_gt_ccs_mode.h"
 #include "xe_gt_printk.h"
 #include "xe_guc.h"
+#include "xe_guc_buf.h"
 #include "xe_guc_capture.h"
 #include "xe_guc_ct.h"
 #include "xe_hw_engine.h"
@@ -994,6 +995,16 @@ static int guc_ads_action_update_policies(struct xe_guc_ads *ads, u32 policy_off
 	return xe_guc_ct_send(ct, action, ARRAY_SIZE(action), 0, 0);
 }
 
+static int guc_ads_update_policies(struct xe_guc_ads *ads, const struct guc_policies *policies)
+{
+	CLASS(xe_guc_buf_from_data, buf)(&ads_to_guc(ads)->buf, policies, sizeof(*policies));
+
+	if (!xe_guc_buf_is_valid(buf))
+		return -ENOBUFS;
+
+	return guc_ads_action_update_policies(ads, xe_guc_buf_flush(buf));
+}
+
 /**
  * xe_guc_ads_scheduler_policy_toggle_reset - Toggle reset policy
  * @ads: Additional data structures object
@@ -1005,11 +1016,8 @@ static int guc_ads_action_update_policies(struct xe_guc_ads *ads, u32 policy_off
 int xe_guc_ads_scheduler_policy_toggle_reset(struct xe_guc_ads *ads)
 {
 	struct xe_device *xe = ads_to_xe(ads);
-	struct xe_gt *gt = ads_to_gt(ads);
-	struct xe_tile *tile = gt_to_tile(gt);
 	struct guc_policies *policies;
-	struct xe_bo *bo;
-	int ret = 0;
+	int ret;
 
 	policies = kmalloc(sizeof(*policies), GFP_KERNEL);
 	if (!policies)
@@ -1023,16 +1031,7 @@ int xe_guc_ads_scheduler_policy_toggle_reset(struct xe_guc_ads *ads)
 	else
 		policies->global_flags &= ~GLOBAL_POLICY_DISABLE_ENGINE_RESET;
 
-	bo = xe_managed_bo_create_from_data(xe, tile, policies, sizeof(struct guc_policies),
-					    XE_BO_FLAG_VRAM_IF_DGFX(tile) |
-					    XE_BO_FLAG_GGTT);
-	if (IS_ERR(bo)) {
-		ret = PTR_ERR(bo);
-		goto out;
-	}
-
-	ret = guc_ads_action_update_policies(ads, xe_bo_ggtt_addr(bo));
-out:
+	ret = guc_ads_update_policies(ads, policies);
 	kfree(policies);
 	return ret;
 }
-- 
2.47.1


  parent reply	other threads:[~2025-05-12 22:00 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-12 22:00 [PATCH 0/2] Don't allocate managed BO for each policy change Michal Wajdeczko
2025-05-12 22:00 ` [PATCH 1/2] drm/xe/guc: Unblock GuC buffer cache for all modes Michal Wajdeczko
2025-05-12 22:11   ` Matthew Brost
2025-05-13 20:57     ` Lucas De Marchi
2025-05-12 22:32   ` Daniele Ceraolo Spurio
2025-05-12 22:00 ` Michal Wajdeczko [this message]
2025-05-12 22:12   ` [PATCH 2/2] drm/xe/guc: Don't allocate managed BO for each policy change Matthew Brost
2025-05-12 22:15 ` ✓ CI.Patch_applied: success for " Patchwork
2025-05-12 22:15 ` ✓ CI.checkpatch: " Patchwork
2025-05-12 22:16 ` ✓ CI.KUnit: " Patchwork
2025-05-12 22:26 ` ✗ CI.Build: failure " Patchwork
2025-05-13 14:24 ` ✓ CI.Patch_applied: success " Patchwork
2025-05-13 14:24 ` ✓ CI.checkpatch: " Patchwork
2025-05-13 14:25 ` ✓ CI.KUnit: " Patchwork
2025-05-13 14:35 ` ✓ CI.Build: " Patchwork
2025-05-14 18:51 ` ✓ CI.Patch_applied: success for Don't allocate managed BO for each policy change (rev2) Patchwork
2025-05-14 18:51 ` ✓ CI.checkpatch: " Patchwork
2025-05-14 18:53 ` ✓ CI.KUnit: " Patchwork
2025-05-14 19:03 ` ✓ CI.Build: " Patchwork
2025-05-14 19:06 ` ✓ CI.Hooks: " Patchwork
2025-05-14 19:07 ` ✓ CI.checksparse: " Patchwork
2025-05-14 19:59 ` ✓ Xe.CI.BAT: " Patchwork
2025-05-15  3:31 ` ✗ Xe.CI.Full: failure " Patchwork
2025-05-15  6:29   ` Michal Wajdeczko
2025-05-27  0:49 ` ✗ CI.Patch_applied: " 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=20250512220018.172-3-michal.wajdeczko@intel.com \
    --to=michal.wajdeczko@intel.com \
    --cc=John.C.Harrison@Intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=rodrigo.vivi@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