| Message ID | 1402675736-15379-19-git-send-email-p.zabel@pengutronix.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 06/13/2014 06:08 PM, Philipp Zabel wrote: > This disables forcing IDR frames at GOP size intervals on CODA7541 and CODA960, > which is only needed to work around a firmware bug on CodaDx6. > Instead, the V4L2_BUF_FLAG_KEYFRAME v4l2 buffer flag is cleared before marking > the source buffer done for dequeueing. Userspace can set it before queueing a > frame to force an IDR frame, to implement VFU (Video Fast Update). I'd like to see an RFC for this feature. Rather than 'misuse' it, I think this should be standardized. I have nothing against using KEYFRAME in order to implement VFU (in fact, I like it!), but it should be documented and well-defined. Regards, Hans > > Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> > --- > drivers/media/platform/coda.c | 18 ++++++++++-------- > 1 file changed, 10 insertions(+), 8 deletions(-) > > diff --git a/drivers/media/platform/coda.c b/drivers/media/platform/coda.c > index 11e059d..cf75112 100644 > --- a/drivers/media/platform/coda.c > +++ b/drivers/media/platform/coda.c > @@ -1264,22 +1264,22 @@ static void coda_prepare_encode(struct coda_ctx *ctx) > * frame as IDR. This is a problem for some decoders that can't > * recover when a frame is lost. > */ > - if (src_buf->v4l2_buf.sequence % ctx->params.gop_size) { > - src_buf->v4l2_buf.flags |= V4L2_BUF_FLAG_PFRAME; > - src_buf->v4l2_buf.flags &= ~V4L2_BUF_FLAG_KEYFRAME; > - } else { > + if ((src_buf->v4l2_buf.sequence % ctx->params.gop_size) == 0) > src_buf->v4l2_buf.flags |= V4L2_BUF_FLAG_KEYFRAME; > + if (src_buf->v4l2_buf.flags & V4L2_BUF_FLAG_KEYFRAME) > src_buf->v4l2_buf.flags &= ~V4L2_BUF_FLAG_PFRAME; > - } > + else > + src_buf->v4l2_buf.flags |= V4L2_BUF_FLAG_PFRAME; > > if (dev->devtype->product == CODA_960) > coda_set_gdi_regs(ctx); > > /* > - * Copy headers at the beginning of the first frame for H.264 only. > - * In MPEG4 they are already copied by the coda. > + * Copy headers in front of the first frame and forced I frames for > + * H.264 only. In MPEG4 they are already copied by the CODA. > */ > - if (src_buf->v4l2_buf.sequence == 0) { > + if (src_buf->v4l2_buf.sequence == 0 || > + src_buf->v4l2_buf.flags & V4L2_BUF_FLAG_KEYFRAME) { > pic_stream_buffer_addr = > vb2_dma_contig_plane_dma_addr(dst_buf, 0) + > ctx->vpu_header_size[0] + > @@ -3245,6 +3245,8 @@ static void coda_finish_encode(struct coda_ctx *ctx) > src_buf->v4l2_buf.flags & V4L2_BUF_FLAG_TSTAMP_SRC_MASK; > dst_buf->v4l2_buf.timecode = src_buf->v4l2_buf.timecode; > > + /* Clear keyframe flag so userspace can misuse it to force an IDR frame */ > + src_buf->v4l2_buf.flags &= ~V4L2_BUF_FLAG_KEYFRAME; > v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE); > > dst_buf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); > -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Am Montag, den 16.06.2014, 10:24 +0200 schrieb Hans Verkuil: > On 06/13/2014 06:08 PM, Philipp Zabel wrote: > > This disables forcing IDR frames at GOP size intervals on CODA7541 and CODA960, > > which is only needed to work around a firmware bug on CodaDx6. > > Instead, the V4L2_BUF_FLAG_KEYFRAME v4l2 buffer flag is cleared before marking > > the source buffer done for dequeueing. Userspace can set it before queueing a > > frame to force an IDR frame, to implement VFU (Video Fast Update). > > I'd like to see an RFC for this feature. Rather than 'misuse' it, I think this > should be standardized. I have nothing against using KEYFRAME in order to > implement VFU (in fact, I like it!), but it should be documented and well-defined. Thanks, I'll prepare a separate RFC for this. The other possibility would be to use a V4L2_ENC_CMD for this feature. One thing I'm not sure about is how to signal to userspace that this feature is available when using the KEYFRAME clearing. With encoder commands, TRY_COMMAND could be used. But when using the buffer KEYFRAME bits, I suppose a flag would have to be introduced somewhere? regards Philipp -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/media/platform/coda.c b/drivers/media/platform/coda.c index 11e059d..cf75112 100644 --- a/drivers/media/platform/coda.c +++ b/drivers/media/platform/coda.c @@ -1264,22 +1264,22 @@ static void coda_prepare_encode(struct coda_ctx *ctx) * frame as IDR. This is a problem for some decoders that can't * recover when a frame is lost. */ - if (src_buf->v4l2_buf.sequence % ctx->params.gop_size) { - src_buf->v4l2_buf.flags |= V4L2_BUF_FLAG_PFRAME; - src_buf->v4l2_buf.flags &= ~V4L2_BUF_FLAG_KEYFRAME; - } else { + if ((src_buf->v4l2_buf.sequence % ctx->params.gop_size) == 0) src_buf->v4l2_buf.flags |= V4L2_BUF_FLAG_KEYFRAME; + if (src_buf->v4l2_buf.flags & V4L2_BUF_FLAG_KEYFRAME) src_buf->v4l2_buf.flags &= ~V4L2_BUF_FLAG_PFRAME; - } + else + src_buf->v4l2_buf.flags |= V4L2_BUF_FLAG_PFRAME; if (dev->devtype->product == CODA_960) coda_set_gdi_regs(ctx); /* - * Copy headers at the beginning of the first frame for H.264 only. - * In MPEG4 they are already copied by the coda. + * Copy headers in front of the first frame and forced I frames for + * H.264 only. In MPEG4 they are already copied by the CODA. */ - if (src_buf->v4l2_buf.sequence == 0) { + if (src_buf->v4l2_buf.sequence == 0 || + src_buf->v4l2_buf.flags & V4L2_BUF_FLAG_KEYFRAME) { pic_stream_buffer_addr = vb2_dma_contig_plane_dma_addr(dst_buf, 0) + ctx->vpu_header_size[0] + @@ -3245,6 +3245,8 @@ static void coda_finish_encode(struct coda_ctx *ctx) src_buf->v4l2_buf.flags & V4L2_BUF_FLAG_TSTAMP_SRC_MASK; dst_buf->v4l2_buf.timecode = src_buf->v4l2_buf.timecode; + /* Clear keyframe flag so userspace can misuse it to force an IDR frame */ + src_buf->v4l2_buf.flags &= ~V4L2_BUF_FLAG_KEYFRAME; v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE); dst_buf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx);
This disables forcing IDR frames at GOP size intervals on CODA7541 and CODA960, which is only needed to work around a firmware bug on CodaDx6. Instead, the V4L2_BUF_FLAG_KEYFRAME v4l2 buffer flag is cleared before marking the source buffer done for dequeueing. Userspace can set it before queueing a frame to force an IDR frame, to implement VFU (Video Fast Update). Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> --- drivers/media/platform/coda.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-)