提交 f103b5d6 编写于 作者: M Mauro Carvalho Chehab

[media] videobuf2: Fix CodingStyle

Changeset b18a8ff2 added a comment violating the 80cols max size,
with no good reason.

Fix it.

Cc: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>
上级 b18a8ff2
...@@ -1246,15 +1246,16 @@ static int __buf_prepare(struct vb2_buffer *vb, const struct v4l2_buffer *b) ...@@ -1246,15 +1246,16 @@ static int __buf_prepare(struct vb2_buffer *vb, const struct v4l2_buffer *b)
break; break;
case V4L2_MEMORY_USERPTR: case V4L2_MEMORY_USERPTR:
/* /*
* In case of user pointer buffers vb2 allocators need to get direct * In case of user pointer buffers vb2 allocators need to get
* access to userspace pages. This requires getting the mmap semaphore * direct access to userspace pages. This requires getting
* for read access in the current process structure. The same semaphore * the mmap semaphore for read access in the current process
* is taken before calling mmap operation, while both qbuf/prepare_buf * structure. The same semaphore is taken before calling mmap
* and mmap are called by the driver or v4l2 core with the driver's lock * operation, while both qbuf/prepare_buf and mmap are called
* held. To avoid an AB-BA deadlock (mmap_sem then driver's lock in mmap * by the driver or v4l2 core with the driver's lock held.
* and driver's lock then mmap_sem in qbuf/prepare_buf) the videobuf2 * To avoid an AB-BA deadlock (mmap_sem then driver's lock in
* core releases the driver's lock, takes mmap_sem and then takes the * mmap and driver's lock then mmap_sem in qbuf/prepare_buf),
* driver's lock again. * the videobuf2 core releases the driver's lock, takes
* mmap_sem and then takes the driver's lock again.
*/ */
mmap_sem = &current->mm->mmap_sem; mmap_sem = &current->mm->mmap_sem;
call_qop(q, wait_prepare, q); call_qop(q, wait_prepare, q);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册