Hello,
On Thursday, November 18, 2010 1:27 PM Andy Walls wrote:
> Hi Hans and Marek,
>
> Some meta-comments below... ;)
>
> On Thu, 2010-11-18 at 10:17 +0100, Hans Verkuil wrote:
> > Hi Marek!
> >
> > Some comments below...
> >
> > On Wednesday, November 17, 2010 09:39:28 Marek Szyprowski wrote:
> ...
> > > +/**
> > > + * vb2_reqbufs() - Initiate streaming
> > > + * @q: videobuf2 queue
> > > + * @req: struct passed from userspace to vidioc_reqbufs handler in driver
> > > + *
> > > + * Should be called from vidioc_reqbufs ioctl handler of a driver.
> > > + * This function:
> > > + * 1) verifies streaming parameters passed from the userspace,
> > > + * 2) sets up the queue,
> > > + * 3) negotiates number of buffers and planes per buffer with the driver
> > > + * to be used during streaming,
> > > + * 4) allocates internal buffer structures (struct vb2_buffer), according to
> > > + * the agreed parameters,
> > > + * 5) for MMAP memory type, allocates actual video memory, using the
> > > + * memory handling/allocation routines provided during queue initialization
> > > + *
> > > + * If req->count is 0, all the memory will be freed instead.
> > > + * If the queue has been allocated previously (by a previous vb2_reqbufs) call
> > > + * and the queue is not busy, memory will be reallocated.
> > > + *
> > > + * The return values from this function are intended to be directly returned
> > > + * from vidioc_reqbufs handler in driver.
> > > + */
> > > +int vb2_reqbufs(struct vb2_queue *q, struct v4l2_requestbuffers *req)
> > > +{
> > > + unsigned int num_buffers, num_planes;
> > > + int ret = 0;
> > > +
> > > + if (req->memory != V4L2_MEMORY_MMAP
> > > + && req->memory != V4L2_MEMORY_USERPTR) {
> > > + dprintk(1, "reqbufs: unsupported memory type\n");
> > > + return -EINVAL;
> > > + }
> > > +
> > > + if (req->type != q->type) {
> > > + dprintk(1, "reqbufs: queue type invalid\n");
> > > + return -EINVAL;
> > > + }
> > > +
> > > + if (q->streaming) {
> > > + dprintk(1, "reqbufs: streaming active\n");
> > > + return -EBUSY;
> > > + }
> > > +
> > > + if (req->count == 0) {
> > > + /* Free/release memory for count = 0, but only if unused */
> > > + if (q->memory == V4L2_MEMORY_MMAP && __buffers_in_use(q)) {
> > > + dprintk(1, "reqbufs: memory in use, cannot free\n");
> > > + ret = -EBUSY;
> > > + goto end;
> > > + }
> > > +
> > > + ret = __vb2_queue_free(q);
> >
> > OK, I have a problem here. How do I detect as an application whether a driver
> > supports MMAP and/or USERPTR?
> >
> > What I am using in qv4l2 (and it doesn't work properly with videobuf) is that
> > I call REQBUFS with count == 0 for MMAP and for USERPTR and I check whether it
> > returns 0 or -EINVAL.
>
>
> > It seems a reasonable test since it doesn't allocate anything. It would be nice
> > if vb2 can check for the memory field here based on what the driver supports.
> >
> > Ideally we should make this an official part of the spec (or have some other
> > mechanism to find out what it supported).
>
> Well it seems to be documented, just not clearly:
>
> First paragraph here:
> http://linuxtv.org/downloads/v4l-dvb-apis/mmap.html
>
> First & second Paragraph here:
> http://linuxtv.org/downloads/v4l-dvb-apis/userp.html
> (which mentions not allocating things.)
>
> And in the description here:
> http://linuxtv.org/downloads/v4l-dvb-apis/vidioc-reqbufs.html
>
>
> I suppose adding flags to VIDIOC_QUERYCAP results would remove all the
> hokey algorithmic probing of what the driver should already know.
> http://linuxtv.org/downloads/v4l-dvb-apis/vidioc-querycap.html
>
> V4L2_CAP_STREAMING 0x04000000
> V4L2_CAP_STREAMING_MMAP 0x08000000
> V4L2_CAP_STREAMING_USER_PTR 0x10000000
Good idea! This way querying memory access types will be much less intrusive :)
> ...
> >
> > Great job! Thanks for all the hard work!
>
> Ditto!
Thx :)
Best regards
--
Marek Szyprowski
Samsung Poland R&D Center
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
