drm/amdgpu: fix possible null-pointer dereference in amdgpu_ttm_tt_populate()
Subsystems (3)
PATCH 1/1 drm/amdgpu: fix possible null-pointer dereference in amdgpu_ttm_tt_populate()
The variable ttm is assigned to the variable gtt, and the variable gtt is checked in:
if (gtt && gtt->userptr)
This indicates that both ttm and gtt can be NULL.
If so, a null-pointer dereference will occur:
if (ttm->page_flags & TTM_PAGE_FLAG_SG)
Also, some null-pointer dereferences will occur in the function ttm_pool_alloc() which is called in:
return ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);
To fix these possible null-pointer dereferences, the function returns -EINVAL when ttm is NULL.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +4 -1
@@ -1120,8 +1120,11 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,struct amdgpu_device *adev =amdgpu_ttm_adev (bdev);struct amdgpu_ttm_tt *gtt = (void *)ttm; +if (ttm ==NULL ) +return -EINVAL ; +/* user pages are bound by amdgpu_ttm_tt_pin_userptr() */ -if (gtt && gtt->userptr) { +if (gtt->userptr) { ttm->sg =kzalloc (sizeof (struct sg_table ),GFP_KERNEL );if (!ttm->sg)return -ENOMEM ;
Signature · 1 line
2.25.1
Am 31.07.21 um 10:04 schrieb Tuo Li:
Quoted 6 lines
The variable ttm is assigned to the variable gtt, and the variable gtt is checked in:
if (gtt && gtt->userptr)This indicates that both ttm and gtt can be NULL.
If so, a null-pointer dereference will occur:
Quoted 8 lines
if (ttm->page_flags & TTM_PAGE_FLAG_SG)Also, some null-pointer dereferences will occur in the function
ttm_pool_alloc()which is called in:return ttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);To fix these possible null-pointer dereferences, the function returns
-EINVALwhen ttm is NULL.
NAK, the NULL test is just a leftover from when the objects where distinct.
Please remove the NULL test instead.
Regards,
Christian.
Quoted 24 lines
Reported-by TOTE Robot <[email protected]>Signed-off-by Tuo Li <[email protected]>--- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +4
-1@@ -1120,8 +1120,11 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,struct amdgpu_device *adev =amdgpu_ttm_adev (bdev);struct amdgpu_ttm_tt *gtt = (void *)ttm; +if (ttm ==NULL ) +return -EINVAL ; +/* user pages are bound by amdgpu_ttm_tt_pin_userptr() */ -if (gtt && gtt->userptr) { +if (gtt->userptr) { ttm->sg =kzalloc (sizeof (struct sg_table ),GFP_KERNEL );if (!ttm->sg)return -ENOMEM ;
Thanks for your feedback! We will remove the null tests according to your advice and prepare a V2 patch.
Best wishes,
Tuo Li
Quoted 48 lines · Christian König
Am 31.07.21 um 10:04 schrieb Tuo Li:
Quoted 8 lines
The variable ttm is assigned to the variable gtt, and the variable gtt is checked in:
if (gtt &>t->userptr)This indicates that both ttm and gtt can be NULL.
If so, a null-pointer dereference will occur:
if (ttm->page_flags&TTM_PAGE_FLAG_SG)Quoted 6 lines
Also, some null-pointer dereferences will occur in the function
ttm_pool_alloc()which is called in:
returnttm_pool_alloc(&adev->mman.bdev.pool, ttm, ctx);To fix these possible null-pointer dereferences, the function returns
-EINVALwhen ttm is NULL.NAK, the NULL test is just a leftover from when the objects where distinct.
Please remove the NULL test instead.
Regards,
Christian.Quoted 24 lines
Reported-by TOTE Robot <[email protected]>Signed-off-by Tuo Li <[email protected]>---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c| 5 ++++-1 file changed, 4 insertions(+), 1 deletion(-)drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +3
-1@@ -1120,8 +1120,11 @@ static int amdgpu_ttm_tt_populate(struct ttm_device *bdev, struct amdgpu_device *adev = amdgpu_ttm_adev(bdev); struct amdgpu_ttm_tt *gtt = (void *)ttm; + if (ttm == NULL) + return -EINVAL; + /* user pages are bound by amdgpu_ttm_tt_pin_userptr() */ - if (gtt && gtt->userptr) { + if (gtt->userptr) { ttm->sg = kzalloc(sizeof(struct sg_table), GFP_KERNEL); if (!ttm->sg)return
-ENOMEM;