linux-kernel - [PATCH] drm/vmwgfx: handle allocation failure of waiter context

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20250630211948.1556524-1-pchelkin@ispras.ru>
Date: Tue,  1 Jul 2025 00:19:47 +0300
From: Fedor Pchelkin <pchelkin@...ras.ru>
To: Zack Rusin <zack.rusin@...adcom.com>,
	Ian Forbes <ian.forbes@...adcom.com>
Cc: Fedor Pchelkin <pchelkin@...ras.ru>,
	Broadcom internal kernel review list <bcm-kernel-feedback-list@...adcom.com>,
	Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
	Maxime Ripard <mripard@...nel.org>,
	Thomas Zimmermann <tzimmermann@...e.de>,
	dri-devel@...ts.freedesktop.org,
	linux-kernel@...r.kernel.org,
	lvc-project@...uxtesting.org,
	stable@...r.kernel.org
Subject: [PATCH] drm/vmwgfx: handle allocation failure of waiter context

Handle result of kmalloc() to prevent possible NULL pointer dereference.
For the sake of not introducing additional layer of indentation, extract
seqno_waiter_rm_context creating code into a separate helper function.

Judging by commit 0039a3b35b10 ("drm/vmwgfx: Add seqno waiter for
sync_files"), possible errors in seqno waiting aren't fatal, thus just
skip this block if an allocation failure occurs - no need to propagate
upwards.

Found by Linux Verification Center (linuxtesting.org) with Svace static
analysis tool.

Fixes: 0039a3b35b10 ("drm/vmwgfx: Add seqno waiter for sync_files")
Cc: stable@...r.kernel.org
Signed-off-by: Fedor Pchelkin <pchelkin@...ras.ru>
---
 drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c | 27 ++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index e831e324e737..12d897eca410 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -4085,6 +4085,23 @@ static void seqno_waiter_rm_cb(struct dma_fence *f, struct dma_fence_cb *cb)
 	kfree(ctx);
 }
 
+static void seqno_waiter_create(struct dma_fence *f,
+				struct vmw_private *dev_priv)
+{
+	struct seqno_waiter_rm_context *ctx;
+
+	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return;
+
+	ctx->dev_priv = dev_priv;
+	vmw_seqno_waiter_add(dev_priv);
+	if (dma_fence_add_callback(f, &ctx->base, seqno_waiter_rm_cb) < 0) {
+		vmw_seqno_waiter_remove(dev_priv);
+		kfree(ctx);
+	}
+}
+
 int vmw_execbuf_process(struct drm_file *file_priv,
 			struct vmw_private *dev_priv,
 			void __user *user_commands, void *kernel_commands,
@@ -4265,15 +4282,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 		} else {
 			/* Link the fence with the FD created earlier */
 			fd_install(out_fence_fd, sync_file->file);
-			struct seqno_waiter_rm_context *ctx =
-				kmalloc(sizeof(*ctx), GFP_KERNEL);
-			ctx->dev_priv = dev_priv;
-			vmw_seqno_waiter_add(dev_priv);
-			if (dma_fence_add_callback(&fence->base, &ctx->base,
-						   seqno_waiter_rm_cb) < 0) {
-				vmw_seqno_waiter_remove(dev_priv);
-				kfree(ctx);
-			}
+			seqno_waiter_create(&fence->base, dev_priv);
 		}
 	}
 
-- 
2.50.0