5.10-stable review patch. If anyone has any objections, please let me know. ------------------ From: Dave Airlie [ Upstream commit 5e17160d41d92823f3379c1982e1369680c5ce4d ] In the non-uvmm path, there could be a race between the info lookup finding the vma, and the gem close path closing the vma leading to a use-after-free. Spotted with the help of Opus 4.6. Signed-off-by: Dave Airlie Fixes: e758a3111914 ("drm/nouveau: fixup gem_info ioctl to return client-specific bo virtual") Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260612020658.3176270-1-airlied@gmail.com Signed-off-by: Danilo Krummrich [ adapted VMM and UVMM handling to the legacy direct VMM selection and condition. ] Signed-off-by: Sasha Levin --- drivers/gpu/drm/nouveau/nouveau_gem.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index 6e84616311369..e8c02dea16b01 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -227,11 +227,20 @@ nouveau_gem_info(struct drm_file *file_priv, struct drm_gem_object *gem, rep->domain = NOUVEAU_GEM_DOMAIN_VRAM; rep->offset = nvbo->offset; if (vmm->vmm.object.oclass >= NVIF_CLASS_VMM_NV50) { + int ret; + + ret = ttm_bo_reserve(&nvbo->bo, false, false, NULL); + if (ret) + return ret; + vma = nouveau_vma_find(nvbo, vmm); - if (!vma) + if (!vma) { + ttm_bo_unreserve(&nvbo->bo); return -EINVAL; + } rep->offset = vma->addr; + ttm_bo_unreserve(&nvbo->bo); } rep->size = nvbo->bo.mem.num_pages << PAGE_SHIFT; -- 2.53.0