[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231107122940.3762228-2-sashal@kernel.org>
Date: Tue, 7 Nov 2023 07:28:56 -0500
From: Sasha Levin <sashal@...nel.org>
To: linux-kernel@...r.kernel.org, stable@...r.kernel.org
Cc: Xiaogang Chen <xiaogang.chen@....com>,
Philip Yang <Philip.Yang@....com>,
Jesse Zhang <Jesse.Zhang@....com>,
Alex Deucher <alexander.deucher@....com>,
Sasha Levin <sashal@...nel.org>, Felix.Kuehling@....com,
christian.koenig@....com, Xinhui.Pan@....com, airlied@...il.com,
daniel@...ll.ch, amd-gfx@...ts.freedesktop.org,
dri-devel@...ts.freedesktop.org
Subject: [PATCH AUTOSEL 5.15 02/20] drm/amdkfd: Fix a race condition of vram buffer unref in svm code
From: Xiaogang Chen <xiaogang.chen@....com>
[ Upstream commit 709c348261618da7ed89d6c303e2ceb9e453ba74 ]
prange->svm_bo unref can happen in both mmu callback and a callback after
migrate to system ram. Both are async call in different tasks. Sync svm_bo
unref operation to avoid random "use-after-free".
Signed-off-by: Xiaogang Chen <xiaogang.chen@....com>
Reviewed-by: Philip Yang <Philip.Yang@....com>
Reviewed-by: Jesse Zhang <Jesse.Zhang@....com>
Tested-by: Jesse Zhang <Jesse.Zhang@....com>
Signed-off-by: Alex Deucher <alexander.deucher@....com>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 22a70aaccf13c..b7d32a5062b6c 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -550,8 +550,15 @@ svm_range_vram_node_new(struct amdgpu_device *adev, struct svm_range *prange,
void svm_range_vram_node_free(struct svm_range *prange)
{
- svm_range_bo_unref(prange->svm_bo);
- prange->ttm_res = NULL;
+ /* serialize prange->svm_bo unref */
+ mutex_lock(&prange->lock);
+ /* prange->svm_bo has not been unref */
+ if (prange->ttm_res) {
+ prange->ttm_res = NULL;
+ mutex_unlock(&prange->lock);
+ svm_range_bo_unref(prange->svm_bo);
+ } else
+ mutex_unlock(&prange->lock);
}
struct amdgpu_device *
--
2.42.0
Powered by blists - more mailing lists