Hi Hans, Thank you for the patch. On Thursday 24 December 2015 14:18:10 Hans Verkuil wrote: > A driver might detect a new standard or DV timings, but it will never > change to > those new timings automatically. Instead it will send an event and let > the application > take care of it. Your webmail seriously messes up patches :-/ > Make this explicit in the documentation. > > Signed-off-by: Hans Verkuil <hans.verkuil@xxxxxxxxx> > --- > .../DocBook/media/v4l/vidioc-query-dv-timings.xml | 14 > ++++++++++++-- > Documentation/DocBook/media/v4l/vidioc-querystd.xml | 10 > ++++++++++ > 2 files changed, 22 insertions(+), 2 deletions(-) > > diff --git a/Documentation/DocBook/media/v4l/vidioc-query-dv-timings.xml > b/Documentation/DocBook/media/v4l/vidioc-query-dv-timings.xml > index e9c70a8..eba0293 100644 > --- a/Documentation/DocBook/media/v4l/vidioc-query-dv-timings.xml > +++ b/Documentation/DocBook/media/v4l/vidioc-query-dv-timings.xml > @@ -60,9 +60,19 @@ input</refpurpose> > automatically, similar to sensing the video standard. To do so, > applications > call <constant>VIDIOC_QUERY_DV_TIMINGS</constant> with a pointer to a > &v4l2-dv-timings;. Once the hardware detects the timings, it will fill > in the > -timings structure. > +timings structure.</para> > > -If the timings could not be detected because there was no signal, then > +<para>Please note that drivers will <emphasis>never</emphasis> switch > timings automatically > +if new timings are detected. As the document is an API specification I'd phrase this as a requirement: "Note that driver must never switch timings automatically if new timings are detected." (or, if we want to comply with the IEEE Standards Style Manual, s/must never/shall not/) > Instead, drivers will send the > +<constant>V4L2_EVENT_SOURCE_CHANGE</constant> event (if they support > this) Similarly, "drivers should send". Apart from that and similar comments for VIDIOC_QUERYSTD, Acked-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> > and expect > +that userspace will take action by calling > <constant>VIDIOC_QUERY_DV_TIMINGS</constant>. > +The reason is that new timings usually mean different buffer sizes as > well, and you > +cannot change buffer sizes on the fly. In general, applications that > receive the > +Source Change event will have to call > <constant>VIDIOC_QUERY_DV_TIMINGS</constant>, > +and if the detected timings are valid they will have to stop streaming, > set the new > +timings, allocate new buffers and start streaming again.</para> > + > +<para>If the timings could not be detected because there was no signal, > then > <errorcode>ENOLINK</errorcode> is returned. If a signal was detected, > but > it was unstable and the receiver could not lock to the signal, then > <errorcode>ENOLCK</errorcode> is returned. If the receiver could lock > to the signal, > diff --git a/Documentation/DocBook/media/v4l/vidioc-querystd.xml > b/Documentation/DocBook/media/v4l/vidioc-querystd.xml > index 2223485..8efa917 100644 > --- a/Documentation/DocBook/media/v4l/vidioc-querystd.xml > +++ b/Documentation/DocBook/media/v4l/vidioc-querystd.xml > @@ -59,6 +59,16 @@ then the driver will return V4L2_STD_UNKNOWN. When > detection is not > possible or fails, the set must contain all standards supported by the > current video input or output.</para> > > +<para>Please note that drivers will <emphasis>never</emphasis> switch > the video standard > +automatically if a new video standard is detected. Instead, drivers > will send the > +<constant>V4L2_EVENT_SOURCE_CHANGE</constant> event (if they support > this) and expect > +that userspace will take action by calling > <constant>VIDIOC_QUERYSTD</constant>. > +The reason is that a new video standard can mean different buffer sizes > as well, and you > +cannot change buffer sizes on the fly. In general, applications that > receive the > +Source Change event will have to call > <constant>VIDIOC_QUERYSTD</constant>, > +and if the detected video standard is valid they will have to stop > streaming, set the new > +standard, allocate new buffers and start streaming again.</para> > + > </refsect1> > > <refsect1> -- Regards, Laurent Pinchart -- 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
