PT-2026-152513

CVE-2026-99633

·

Publicado

2026-11-20

·

Atualizado

2026-11-20

Descrição por

Pontuação
N/A
Não disponível

Não há classificações de severidade ou métricas disponíveis. Quando houver, atualizaremos as informações correspondentes na página.

In the Linux kernel, the following vulnerability has been resolved:
drm/amdgpu: keep the CS user fence BO in GTT
The CS ioctl hands the GPU the GART address of the user fence BO, but since the BO keeps its own preferred and allowed domains, it no longer makes sure the BO is in GTT. A user fence in a VRAM BO, which only needs to be a page in size, gets its address from amdgpu bo gpu offset(), which warns about non-contiguous VRAM. On navi31, a CS with a user fence in a 4K VRAM BO, or in one created with AMDGPU GEM CREATE VM ALWAYS VALID:
WARNING: amdgpu object.c:1531 at amdgpu bo gpu offset+0x86/0xb0 [amdgpu], CPU#11: cs gem checks/18393 RIP: 0010:amdgpu bo gpu offset+0x86/0xb0 [amdgpu] Call Trace: amdgpu cs parser bos.isra.0+0x4eb/0x850 [amdgpu] amdgpu cs ioctl+0x97e/0x2510 [amdgpu] drm ioctl kernel+0xac/0x110 [drm] drm ioctl+0x2c6/0x5b0 [drm] x64 sys ioctl+0xa0/0xf0 do syscall 64+0xa3/0x4d0 entry SYSCALL 64 after hwframe+0x76/0x7e
and BOs which can't be in GTT at all, like GWS or OA ones of a page in size, pass the checks as well.
Only take BOs which can't be anywhere else than in GTT as user fence, which the CS then validates there along with its other BOs. While at it, detect userptrs with amdgpu ttm tt is userptr(); amdgpu ttm tt get usermm() returns NULL once the task which created the userptr is gone.
v2: only take BOs which can only be in GTT instead of special casing the user fence BO in the validation, use amdgpu ttm tt is userptr() (Christian)
(cherry picked from commit f16989def9561d04d0f2e77216f5319112d29bf1)
Fornecedores
Produtos afetados
Identificadores relacionados
Encontrou algum problema na descrição? Tem algo a acrescentar? Fique à vontade para nos escrever 👾