On 10/7/19 6:24 PM, Maxime Jourdan wrote:
On 07/10/2019 17:12, Hans Verkuil wrote:
On 10/7/19 4:59 PM, Maxime Jourdan wrote:
Hello,
This patch series aims to bring H.264 support as well as compliance update
to the amlogic stateful video decoder driver.
There is 1 issue that remains currently:
 - The following codepath had to be commented out from v4l2-compliance as
it led to stalling:
if (node->codec_mask & STATEFUL_DECODER) {
ÂÂÂÂstruct v4l2_decoder_cmd cmd;
ÂÂÂÂbuffer buf_cap(m2m_q);
ÂÂÂÂmemset(&cmd, 0, sizeof(cmd));
ÂÂÂÂcmd.cmd = V4L2_DEC_CMD_STOP;
ÂÂÂÂ/* No buffers are queued, call STREAMON, then STOP */
ÂÂÂÂfail_on_test(node->streamon(q.g_type()));
ÂÂÂÂfail_on_test(node->streamon(m2m_q.g_type()));
ÂÂÂÂfail_on_test(doioctl(node, VIDIOC_DECODER_CMD, &cmd));
ÂÂÂÂfail_on_test(buf_cap.querybuf(node, 0));
ÂÂÂÂfail_on_test(buf_cap.qbuf(node));
ÂÂÂÂfail_on_test(buf_cap.dqbuf(node));
ÂÂÂÂfail_on_test(!(buf_cap.g_flags() & V4L2_BUF_FLAG_LAST));
ÂÂÂÂfor (unsigned p = 0; p < buf_cap.g_num_planes(); p++)
ÂÂÂÂÂÂÂ fail_on_test(buf_cap.g_bytesused(p));
ÂÂÂÂfail_on_test(node->streamoff(q.g_type()));
ÂÂÂÂfail_on_test(node->streamoff(m2m_q.g_type()));
ÂÂÂÂ/* Call STREAMON, queue one CAPTURE buffer, then STOP */
ÂÂÂÂfail_on_test(node->streamon(q.g_type()));
ÂÂÂÂfail_on_test(node->streamon(m2m_q.g_type()));
ÂÂÂÂfail_on_test(buf_cap.querybuf(node, 0));
ÂÂÂÂfail_on_test(buf_cap.qbuf(node));
ÂÂÂÂfail_on_test(doioctl(node, VIDIOC_DECODER_CMD, &cmd));
ÂÂÂÂfail_on_test(buf_cap.dqbuf(node));
ÂÂÂÂfail_on_test(!(buf_cap.g_flags() & V4L2_BUF_FLAG_LAST));
ÂÂÂÂfor (unsigned p = 0; p < buf_cap.g_num_planes(); p++)
ÂÂÂÂÂÂÂ fail_on_test(buf_cap.g_bytesused(p));
ÂÂÂÂfail_on_test(node->streamoff(q.g_type()));
ÂÂÂÂfail_on_test(node->streamoff(m2m_q.g_type()));
}
The reason for this is because the driver has a limitation where all
capturebuffers must be queued to the driver before STREAMON is effective.
The firmware needs to know in advance what all the buffers are before
starting to decode.
This limitation is enforced via q->min_buffers_needed.
As such, in this compliance codepath, STREAMON is never actually called
driver-side and there is a stall on fail_on_test(buf_cap.dqbuf(node));
That's interesting. I will have to look more closely at this.
One last detail: V4L2_FMT_FLAG_DYN_RESOLUTION is currently not recognized
by v4l2-compliance, so it was left out for the test. However, it is
present in the patch series.
It is definitely recognized by v4l2-compliance.
The second patch has 3 "Alignment should match open parenthesis" lines
where I preferred to keep them that way.
Thanks Stanimir for sharing your HDR file creation tools, this was very
helpful :).
Maxime
# v4l2-compliance --stream-from-hdr test-25fps.h264.hdr -s250
v4l2-compliance SHA: a162244d47d4bb01d0692da879dce5a070f118e7, 64 bits
But this SHA isn't in the v4l-utils repo, so this makes me wonder where you
got this repo from.
I am based off the hverkuil/vicodec branch. The SHA I am on is actually 05387265053bc6f9 ("test-media: add vicodec tests"), but it wasn't updated as I found out it requires a new bootstrap to refresh
the SHA. Maybe some rebasing at some point got rid of a162244d.
Don't use the hverkuil/vicodec branch. Everything there has been merged into the
regular v4l-utils repo some time ago. So just clone git://linuxtv.org/v4l-utils.git
and use that.
Regards,
Hans