On 07/11/2018 12:39 PM, Marco Felsch wrote: > Hi Javier, > > On 18-07-08 15:11, Javier Martinez Canillas wrote: >> [adding Marco Felsch since he has been working on this driver] >> >> On 07/05/2018 03:12 PM, Hans Verkuil wrote: >>> On 05/07/18 14:44, Mauro Carvalho Chehab wrote: >>>> Javier, >>>> >>>> How standard settings work with the current OMAP3 drivers with tvp5150? >>> >>> It looks like tvp5150 uses autodetect of the standard, which in general is >> >> That's correct, the driver uses standard autodetect. >> >>> not a good thing to do since different standards have different buffer >>> sizes. But this chip can scale, so it might scale PAL to NTSC or vice versa >>> if the standard switches mid-stream. Or it only detects the standard when >>> it starts streaming, I'm not sure. >>> >> >> Not sure about this either, IIUC switching the standard mid-stream won't work. > > As far as I know, the detection happens after a sync lost event. > Ah, good to know. >>> In any case, this is not normal behavior, for almost all analog video >>> receivers you need to be able to set the std explicitly. >>> >> >> Indeed. I see that Marco's recent series [0] add supports for the .querystd [1] >> and .g_std [2] callbacks to the tvp5150 driver, so that way user-space can get >> back the detected standard. >> >> [0]: https://www.spinics.net/lists/linux-media/msg136869.html >> [1]: https://www.spinics.net/lists/linux-media/msg136872.html >> [2]: https://www.spinics.net/lists/linux-media/msg136875.html > > I tought the std will be set by the v4l2_subdev_video_ops.s_std() > operation. If the user change the std manually, the autodection is > disabled. > Yes, what I tried to say is that user-space won't have a way to know which std to set without a .querystd, or know what std was autodetected withou a .g_std. Best regards, -- Javier Martinez Canillas Software Engineer - Desktop Hardware Enablement Red Hat
