linux-kernel - [PATCH] [PATCH RESEND] drm/virtio: fix memory leak of vbuf

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>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20240509015225.38669-1-liweishi@kylinos.cn>
Date: Thu,  9 May 2024 09:52:25 +0800
From: Weishi Li <liweishi@...inos.cn>
To: airlied@...hat.com,
	kraxel@...hat.com,
	gurchetansingh@...omium.org,
	olvaffe@...il.com,
	maarten.lankhorst@...ux.intel.com,
	mripard@...nel.org,
	tzimmermann@...e.de,
	daniel@...ll.ch
Cc: dri-devel@...ts.freedesktop.org,
	virtualization@...ts.linux.dev,
	linux-kernel@...r.kernel.org,
	liweishi@...inos.cn
Subject: [PATCH] [PATCH RESEND] drm/virtio: fix memory leak of vbuf

Both virtio_gpu_queue_ctrl_buffer and virtio_gpu_queue_cursor use
virtqueue_add_sgs to upload the structure virtio_gpu_vbuffer * vbuf
to virtqueue. However, when the vbuf fails to upload and virtqueue_add_sgs
returns -EIO or -ENOMEM, the vbuf will not be able to be free by
virtio_gpu_dequeue_*_func, resulting in a continuous increase
in memory allocated to vgdev ->vbufs.

Therefore, make virtio_gpu_queue_ctrl_sgs and virtio_gpu_queue_cursor
free vbuf directly after virtqueue_add_sgs returns -EIO or -ENOMEM.

Signed-off-by: Weishi Li <liweishi@...inos.cn>
---
 drivers/gpu/drm/virtio/virtgpu_vq.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
index b1a00c0c25a7..6701ce9d0ee8 100644
--- a/drivers/gpu/drm/virtio/virtgpu_vq.c
+++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
@@ -356,12 +356,14 @@ static int virtio_gpu_queue_ctrl_sgs(struct virtio_gpu_device *vgdev,
 
 	ret = virtqueue_add_sgs(vq, sgs, outcnt, incnt, vbuf, GFP_ATOMIC);
 	WARN_ON(ret);
+	if (ret < 0 && ret != -ENOSPC) {
+		free_vbuf(vgdev, vbuf);
+	} else {
+		vbuf->seqno = ++vgdev->ctrlq.seqno;
+		trace_virtio_gpu_cmd_queue(vq, virtio_gpu_vbuf_ctrl_hdr(vbuf), vbuf->seqno);
 
-	vbuf->seqno = ++vgdev->ctrlq.seqno;
-	trace_virtio_gpu_cmd_queue(vq, virtio_gpu_vbuf_ctrl_hdr(vbuf), vbuf->seqno);
-
-	atomic_inc(&vgdev->pending_commands);
-
+		atomic_inc(&vgdev->pending_commands);
+	}
 	spin_unlock(&vgdev->ctrlq.qlock);
 
 	drm_dev_exit(idx);
@@ -469,6 +471,9 @@ static void virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev,
 		wait_event(vgdev->cursorq.ack_queue, vq->num_free >= outcnt);
 		spin_lock(&vgdev->cursorq.qlock);
 		goto retry;
+	} else if (ret < 0) {
+		free_vbuf(vgdev, vbuf);
+		notify = false;
 	} else {
 		vbuf->seqno = ++vgdev->cursorq.seqno;
 		trace_virtio_gpu_cmd_queue(vq,
-- 
2.25.1