FAILED: patch "[PATCH] drm/amdgpu: hold a runtime PM reference for P2P dma-buf" failed to apply to 6.18-stable tree

PATCH New [email protected] 2 participants 3 messages Sep 22

Subsystems (3)

Stable branch DRM drivers Radeon and amdgpu DRM drivers

lore

Share & tools

lore

[email protected]

author maintainer

2026-09-22 15:12 UTC #0 lore raw

The patch below does not apply to the 6.18-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-6.18.y
git checkout FETCH_HEAD
git cherry-pick -x 636139603b99d2e3a18a46cf3f8d39313ce8042e
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026092254-cactus-employee-2903@gregkh' --subject-prefix 'PATCH 6.18.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

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

From 636139603b99d2e3a18a46cf3f8d39313ce8042e Mon Sep 17 00:00:00 2001
From: Mike Lothian <[email protected]>
Date: Sat, 12 Sep 2026 00:29:08 +0100
Subject: [PATCH] drm/amdgpu: hold a runtime PM reference for P2P dma-buf
attachments
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

amdgpu_dma_buf_map() adds VRAM to the allowed domains for a peer2peer attachment. GTT is only a fallback placement when VRAM is preferred, so ttm_bo_validate() migrates the buffer from GTT into VRAM. While the exporting device is runtime suspended its SDMA rings are down and the move fails:

amdgpu: Move buffer fallback to memcpy unavailable

An importer on a second GPU reaches this holding no runtime PM reference on the exporter, e.g. a compositor on the APU submitting a frame that references a buffer exported by an idle dGPU:

amdgpu_cs_ioctl -> amdgpu_cs_parser_bos -> amdgpu_cs_bo_validate

    -> ttm_bo_validate -> amdgpu_bo_move -> dma_buf_map_attachment
      -> amdgpu_dma_buf_map -> ttm_bo_validate -> amdgpu_bo_move

Pinning a dma-buf into VRAM has the same requirement, which commit 030631e97b20 ("drm/amdgpu: revert "take runtime pm reference when we attach a buffer" v2") called out as the one case that would need the reference back.

Take it in attach and drop it in detach. pm_runtime_get_if_active() never resumes the device, so it cannot deadlock against the reservation taken during resume, which is why the old pm_runtime_get_sync() had to go. If the device is not active, clear peer2peer instead: the buffer then stays in GTT, which remains accessible while the GPU is powered down. If runtime PM is disabled, take a plain reference so the put in detach stays balanced.

Fixes 030631e97b20 ("drm/amdgpu: revert "take runtime pm reference when we attach a buffer" v2")

Reviewed-by Christian König <[email protected]>

Signed-off-by Mike Lothian <[email protected]>

Assisted-by Claude:Opus-5 [Claude Code]

Signed-off-by Alex Deucher <[email protected]>

(cherry picked from commit 062ff15e30a48d14fb7d7558eba84f8dc97197f0)

Cc [email protected]

drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +42 -1

index b33c300e26e2..9adf3eed8822 100644
@@ -43,6 +43,7 @@
 #include <linux/dma-buf.h>
 #include <linux/dma-fence-array.h>
 #include <linux/pci-p2pdma.h>
+#include <linux/pm_runtime.h>
 
 static const struct dma_buf_attach_ops amdgpu_dma_buf_attach_ops;

@@ -100,15 +101,54 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	    pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
 		attach->peer2peer = false;
 
+	/*
+	 * Only allow P2P while the exporter is active, and keep it active
+	 * until detach.  With runtime PM disabled take a plain reference so
+	 * the put in detach stays balanced.
+	 */
+	if (attach->peer2peer) {
+		struct device *dev = adev_to_drm(adev)->dev;
+		int ret = pm_runtime_get_if_active(dev);
+
+		if (!ret)
+			attach->peer2peer = false;
+		else if (ret < 0)
+			pm_runtime_get_noresume(dev);
+	}
+
 	r = dma_resv_lock(bo->tbo.base.resv, NULL);
 	if (r)
-		return r;
+		goto err_pm_put;
 
 	amdgpu_vm_bo_update_shared(bo);
 
 	dma_resv_unlock(bo->tbo.base.resv);
 
 	return 0;
+
+err_pm_put:
+	if (attach->peer2peer)
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+	return r;
+}
+
+/**
+ * amdgpu_dma_buf_detach - &dma_buf_ops.detach implementation
+ *
+ * @dmabuf: DMA-buf where we remove the attachment from
+ * @attach: the attachment to remove
+ *
+ * Drop the runtime PM reference taken in amdgpu_dma_buf_attach().
+ */
+static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf,
+				  struct dma_buf_attachment *attach)
+{
+	struct drm_gem_object *obj = dmabuf->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+
+	if (attach->peer2peer)
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 }
 
 /**

@@ -350,6 +390,7 @@ static void amdgpu_dma_buf_vunmap(struct dma_buf *dma_buf, struct iosys_map *map
 
 const struct dma_buf_ops amdgpu_dmabuf_ops = {
 	.attach = amdgpu_dma_buf_attach,
+	.detach = amdgpu_dma_buf_detach,
 	.pin = amdgpu_dma_buf_pin,
 	.unpin = amdgpu_dma_buf_unpin,
 	.map_dma_buf = amdgpu_dma_buf_map,

Sasha Levin

maintainer

2026-09-22 18:51 UTC #1 lore raw

PATCH 1/2 drm/amdgpu: lock bo before calling amdgpu_vm_bo_update_shared

From: Pierre-Eric Pelloux-Prayer <[email protected]>

[ Upstream commit 36ffc58b8a8704e690a0ce679db26baa5759256f ]

BO's reservation object must be locked before using amdgpu_vm_bo_update_shared otherwise dma_resv_assert_held will complain in amdgpu_vm_update_shared.

Signed-off-by Pierre-Eric Pelloux-Prayer <[email protected]>

Reviewed-by Christian König <[email protected]>

Signed-off-by Alex Deucher <[email protected]>

Stable-dep-of: 636139603b99 ("drm/amdgpu: hold a runtime PM reference for P2P dma-buf attachments")

Signed-off-by Sasha Levin <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c | 7 +++++++
 1 file changed, 7 insertions(+)

drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +7 -0

index f20d4bc58904d..83fed04436ad7 100644
@@ -81,6 +81,7 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	struct drm_gem_object *obj = dmabuf->priv;
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	int r;
 
 	/*
 	 * Disable peer-to-peer access for DCC-enabled VRAM surfaces on GFX12+.

@@ -98,8 +99,14 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	    pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
 		attach->peer2peer = false;
 
+	r = dma_resv_lock(bo->tbo.base.resv, NULL);
+	if (r)
+		return r;
+
 	amdgpu_vm_bo_update_shared(bo);
 
+	dma_resv_unlock(bo->tbo.base.resv);
+
 	return 0;
 }

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-09-22 18:51 UTC #2 lore raw

PATCH 2/2 drm/amdgpu: hold a runtime PM reference for P2P dma-buf attachments

From: Mike Lothian <[email protected]>

[ Upstream commit 636139603b99d2e3a18a46cf3f8d39313ce8042e ]

amdgpu: Move buffer fallback to memcpy unavailable

An importer on a second GPU reaches this holding no runtime PM reference on the exporter, e.g. a compositor on the APU submitting a frame that references a buffer exported by an idle dGPU:

amdgpu_cs_ioctl -> amdgpu_cs_parser_bos -> amdgpu_cs_bo_validate

    -> ttm_bo_validate -> amdgpu_bo_move -> dma_buf_map_attachment
      -> amdgpu_dma_buf_map -> ttm_bo_validate -> amdgpu_bo_move

Fixes 030631e97b20 ("drm/amdgpu: revert "take runtime pm reference when we attach a buffer" v2")

Reviewed-by Christian König <[email protected]>

Signed-off-by Mike Lothian <[email protected]>

Assisted-by Claude:Opus-5 [Claude Code]

Signed-off-by Alex Deucher <[email protected]>

(cherry picked from commit 062ff15e30a48d14fb7d7558eba84f8dc97197f0)

Cc [email protected]

Signed-off-by Sasha Levin <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c | 43 ++++++++++++++++++++-
 1 file changed, 42 insertions(+), 1 deletion(-)

drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +42 -1

index 83fed04436ad7..efa6214eee315 100644
@@ -42,6 +42,7 @@
 #include <linux/dma-buf.h>
 #include <linux/dma-fence-array.h>
 #include <linux/pci-p2pdma.h>
+#include <linux/pm_runtime.h>
 
 static const struct dma_buf_attach_ops amdgpu_dma_buf_attach_ops;

@@ -99,15 +100,54 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
 	    pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
 		attach->peer2peer = false;
 
+	/*
+	 * Only allow P2P while the exporter is active, and keep it active
+	 * until detach.  With runtime PM disabled take a plain reference so
+	 * the put in detach stays balanced.
+	 */
+	if (attach->peer2peer) {
+		struct device *dev = adev_to_drm(adev)->dev;
+		int ret = pm_runtime_get_if_active(dev);
+
+		if (!ret)
+			attach->peer2peer = false;
+		else if (ret < 0)
+			pm_runtime_get_noresume(dev);
+	}
+
 	r = dma_resv_lock(bo->tbo.base.resv, NULL);
 	if (r)
-		return r;
+		goto err_pm_put;
 
 	amdgpu_vm_bo_update_shared(bo);
 
 	dma_resv_unlock(bo->tbo.base.resv);
 
 	return 0;
+
+err_pm_put:
+	if (attach->peer2peer)
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+	return r;
+}
+
+/**
+ * amdgpu_dma_buf_detach - &dma_buf_ops.detach implementation
+ *
+ * @dmabuf: DMA-buf where we remove the attachment from
+ * @attach: the attachment to remove
+ *
+ * Drop the runtime PM reference taken in amdgpu_dma_buf_attach().
+ */
+static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf,
+				  struct dma_buf_attachment *attach)
+{
+	struct drm_gem_object *obj = dmabuf->priv;
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+
+	if (attach->peer2peer)
+		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
 }
 
 /**

@@ -336,6 +376,7 @@ static void amdgpu_dma_buf_vunmap(struct dma_buf *dma_buf, struct iosys_map *map
 
 const struct dma_buf_ops amdgpu_dmabuf_ops = {
 	.attach = amdgpu_dma_buf_attach,
+	.detach = amdgpu_dma_buf_detach,
 	.pin = amdgpu_dma_buf_pin,
 	.unpin = amdgpu_dma_buf_unpin,
 	.map_dma_buf = amdgpu_dma_buf_map,

Signature · 1 line

2.53.0