From: jsg <jsg@openbsd.org>
Date: Fri, 11 Oct 2024 02:04:07 +0000 (+0000)
Subject: drm/amdgpu: disallow multiple BO_HANDLES chunks in one submit
X-Git-Url: http://artulab.com/gitweb/?a=commitdiff_plain;h=a5f3710f751f5dda9ef2c851df550d26490ff032;p=openbsd

drm/amdgpu: disallow multiple BO_HANDLES chunks in one submit

From Pierre-Eric Pelloux-Prayer
44e4aeaef96b2624b341597b1a8e595c89925892 in linux-6.6.y/6.6.55
fec5f8e8c6bcf83ed7a392801d7b44c5ecfc1e82 in mainline linux
---

diff --git a/sys/dev/pci/drm/amd/amdgpu/amdgpu_cs.c b/sys/dev/pci/drm/amd/amdgpu/amdgpu_cs.c
index dca5c07fa3b..784e66c1e21 100644
--- a/sys/dev/pci/drm/amd/amdgpu/amdgpu_cs.c
+++ b/sys/dev/pci/drm/amd/amdgpu/amdgpu_cs.c
@@ -263,6 +263,10 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p,
 			if (size < sizeof(struct drm_amdgpu_bo_list_in))
 				goto free_partial_kdata;
 
+			/* Only a single BO list is allowed to simplify handling. */
+			if (p->bo_list)
+				ret = -EINVAL;
+
 			ret = amdgpu_cs_p1_bo_handles(p, p->chunks[i].kdata);
 			if (ret)
 				goto free_partial_kdata;