On Tue, Mar 10, 2020 at 11:17:10AM +0000, Lad, Prabhakar wrote: > Hi Sakari, > > On Fri, Mar 6, 2020 at 10:18 AM Lad, Prabhakar > <prabhakar.csengg@xxxxxxxxx> wrote: > > > > Hi Sakari, > > > > On Thu, Mar 5, 2020 at 11:43 AM Sakari Ailus > > <sakari.ailus@xxxxxxxxxxxxxxx> wrote: > > > > > > Hi Prabhakar, > > > > > > On Fri, Feb 28, 2020 at 04:41:26PM +0000, Lad Prabhakar wrote: > > > > OV5645 can operate in virtual channel 0-3 in CSI2 interfaces, this patch > > > > adds support for module parameter virtual_channel to select the required > > > > channel. By default OV5645 operates in virtual channel 0. > > > > > > What's your use case for different virtual channels? > > > > > Just ability to switch to different virtual channel, based on ov5640 > > driver. The rcar-csi2 > > has capability to capture from multiple channels so that we can > > capture simultaneously > > from two sensors. > > > Any thoughts on how this could be handled ? A module parameter to support sending the pixel data on virtual channels is certainly a hack. You couldn't support the same kind of sensors with different virtual channel configuration in a deterministic way nor the receiver would have an ability to verify the hardware is properly configured. The multiplexed streams patchset (subject "[PATCH v3 00/31] v4l: add support for multiplexed streams" on LMML) adds support for CSI-2 virtual channels and data types. It's been a while since a version of that was posted though. Jacopo, what are your plans regarding that set? -- Kind regards, Sakari Ailus
