Skip to contentSkip to conversation

FAILED: patch "[PATCH] drm/amdgpu: Fix amdgpu_bo_move() when old_mem and new_mem are" failed to apply to 5.15-stable tree

PATCH New [email protected] 2 participants 2 messages Jul 29 – Aug 1

Subsystems (3)
Stable branch DRM drivers Radeon and amdgpu DRM drivers
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x ee94a65f192c05c543b4d3ad7137cd696b5c18fc
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026072905-unbundle-serving-0266@gregkh' --subject-prefix 'PATCH 5.15.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From ee94a65f192c05c543b4d3ad7137cd696b5c18fc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Timur=20Krist=C3=B3f?= <[email protected]>
Date: Mon, 25 May 2026 13:33:18 +0200
Subject: [PATCH] drm/amdgpu: Fix amdgpu_bo_move() when old_mem and new_mem are
both GTT
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The UVD code relies on GTT to GTT moves in order to ensure that its BOs don't cross 256M segments.

Fixes bfe5e585b44f ("drm/ttm: move last binding into the drivers.")
Signed-off-by Timur Kristóf <[email protected]>
Reviewed-by Christian König <[email protected]>
Signed-off-by Alex Deucher <[email protected]>
(cherry picked from commit 21fd45e5e2628d00b478590bcc3d14d3de5d45b6)
Cc [email protected]
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +18 -0
index 2740de94e93c..16c060badaee 100644
@@ -515,6 +515,15 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (new_mem->mem_type == TTM_PL_TT ||
 	    new_mem->mem_type == AMDGPU_PL_PREEMPT) {
+		if (old_mem && (old_mem->mem_type == TTM_PL_TT ||
+				old_mem->mem_type == AMDGPU_PL_PREEMPT)) {
+			r = ttm_bo_wait_ctx(bo, ctx);
+			if (r)
+				return r;
+
+			amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
+		}
+
 		r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, new_mem);
 		if (r)
 			return r;

@@ -549,6 +558,15 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 		ttm_bo_assign_mem(bo, new_mem);
 		return 0;
 	}
+	if ((old_mem->mem_type == TTM_PL_TT ||
+	     old_mem->mem_type == AMDGPU_PL_PREEMPT) &&
+	    (new_mem->mem_type == TTM_PL_TT ||
+	     new_mem->mem_type == AMDGPU_PL_PREEMPT)) {
+		amdgpu_bo_move_notify(bo, evict, new_mem);
+		ttm_resource_free(bo, &bo->resource);
+		ttm_bo_assign_mem(bo, new_mem);
+		return 0;
+	}
 
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
 	    old_mem->mem_type == AMDGPU_PL_GWS ||
maintainer
#1 lore raw

PATCH 1/1 drm/amdgpu: Fix amdgpu_bo_move() when old_mem and new_mem are both GTT

From: Timur Kristóf <[email protected]>

[ Upstream commit ee94a65f192c05c543b4d3ad7137cd696b5c18fc ]

The UVD code relies on GTT to GTT moves in order to ensure that its BOs don't cross 256M segments.

Fixes bfe5e585b44f ("drm/ttm: move last binding into the drivers.")
Signed-off-by Timur Kristóf <[email protected]>
Reviewed-by Christian König <[email protected]>
Signed-off-by Alex Deucher <[email protected]>
(cherry picked from commit 21fd45e5e2628d00b478590bcc3d14d3de5d45b6)
Cc [email protected]
[ dropped the third `new_mem` argument from `amdgpu_bo_move_notify()` and changed `return 0` to `goto out` to match 5.15's 2-arg prototype and shared epilogue ]
Signed-off-by Sasha Levin <[email protected]>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +17 -0
index 7b3293b37144..ef7b7586779a 100644
@@ -472,6 +472,15 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	if (new_mem->mem_type == TTM_PL_TT ||
 	    new_mem->mem_type == AMDGPU_PL_PREEMPT) {
+		if (old_mem && (old_mem->mem_type == TTM_PL_TT ||
+				old_mem->mem_type == AMDGPU_PL_PREEMPT)) {
+			r = ttm_bo_wait_ctx(bo, ctx);
+			if (r)
+				return r;
+
+			amdgpu_ttm_backend_unbind(bo->bdev, bo->ttm);
+		}
+
 		r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, new_mem);
 		if (r)
 			return r;

@@ -506,6 +515,14 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 		ttm_bo_assign_mem(bo, new_mem);
 		goto out;
 	}
+	if ((old_mem->mem_type == TTM_PL_TT ||
+	     old_mem->mem_type == AMDGPU_PL_PREEMPT) &&
+	    (new_mem->mem_type == TTM_PL_TT ||
+	     new_mem->mem_type == AMDGPU_PL_PREEMPT)) {
+		ttm_resource_free(bo, &bo->resource);
+		ttm_bo_assign_mem(bo, new_mem);
+		goto out;
+	}
 
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
 	    old_mem->mem_type == AMDGPU_PL_GWS ||
Signature · 1 line
2.53.0