From: Sasha Levin <sashal@kernel.org>
To: stable@vger.kernel.org
Cc: Wentao Liang <vulab@iscas.ac.cn>,
	Alex Deucher <alexander.deucher@amd.com>,
	Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.10.y] drm/amdgpu: Fix runtime PM leak in amdgpu_debugfs_test_ib_show()
Date: Tue, 29 Sep 2026 12:41:05 -0400
Message-ID: <20260929164105.122654-1-sashal@kernel.org>
In-Reply-To: <2026092929-vividness-vanquish-6e88@gregkh>
References: <2026092929-vividness-vanquish-6e88@gregkh>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

From: Wentao Liang <vulab@iscas.ac.cn>

[ Upstream commit 2b86ab1bd6673c525adda88819d7658ba9e784ec ]

amdgpu_debugfs_test_ib_show() resumes the device with
pm_runtime_get_sync() before taking the reset domain semaphore with
down_write_killable().  If the write lock acquisition is interrupted,
the function returns without calling pm_runtime_put_autosuspend(),
leaking the runtime PM reference acquired for the device and keeping
the GPU awake.

Drop the runtime PM reference on the interrupted down_write_killable()
error path before returning.

Fixes: 6049db43d6dd ("drm/amdgpu: change reset lock from mutex to rw_semaphore")
Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit ec30a576c2d4c0364549e6c04218f50704ef56c8)
Cc: stable@vger.kernel.org
[ Adapted `down_write_killable(&adev->reset_domain->sem)` to `down_read_killable(&adev->reset_sem)`. ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index 8a1cb1de2b13a..5ba4fd1e9342f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -1259,8 +1259,10 @@ static int amdgpu_debugfs_test_ib(struct seq_file *m, void *data)
 
 	/* Avoid accidently unparking the sched thread during GPU reset */
 	r = down_read_killable(&adev->reset_sem);
-	if (r)
+	if (r) {
+		pm_runtime_put_autosuspend(dev->dev);
 		return r;
+	}
 
 	/* hold on the scheduler */
 	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
-- 
2.53.0