| Message ID | 20200421135743.1381930-1-niklas.soderlund+renesas@ragnatech.se (mailing list archive) |
|---|---|
| Headers | show |
| Series | v4l2-dev/ioctl: Add V4L2_CAP_IO_MC | expand |
Hi Niklas, (With a question for Hans below) Thank you for the patches. On Tue, Apr 21, 2020 at 03:57:37PM +0200, Niklas Söderlund wrote: > Hi, > > This series aims to reduce the amount of boiler plate code in video > device drivers who's inputs and/or outputs are controlled by the Media > Controller instead of the V4L2 API. > > Patch 1/6 adds a new video device capability flag V4L2_CAP_IO_MC which > if set provides helper implementations for the enum inputs and outputs > ioctls freeing the video device driver from the need to implement them. > > Patch 2/6 fix initialization of reserved fields in the cx18 and ivtv > drivers which becomes a problem in 3/6 where Laurent adds mbus filters to > VIDIOC_ENUM_FMT. > > Patch 4/6, 5/6 and 6/6 converts the R-Car VIN, Intel IPU3 and VIMC > drivers to use the new default handlers and capability flag and delete > the now redundant boiler plate code. I'm sure more video device drivers > can make use of this new flag but as I can only test on these two > platforms I have limited my changes to those. > > A separate patch to v4l-utils have been posted as [1] to add a > test for this feature in v4l2-compliance. > > This version have been rebased to latest media-tree to account for the > large shuffles of files. It has also replaced patch 2/6 with a different > version that addresses both cx18 and ivtv instead of only ivtv. It looks like everything has been reviewed. Hans, do you see any remaining blocker, or will you take it in your tree and send a pull request ? > 1. [PATCH 0/2] v4l2-compliance: add tests for V4L2_CAP_IO_MC > > Laurent Pinchart (2): > media: pci: Fill v4l2_fmtdesc with designated initializers > media: v4l2: Extend VIDIOC_ENUM_FMT to support MC-centric devices > > Niklas Söderlund (4): > v4l2-dev/ioctl: Add V4L2_CAP_IO_MC > rcar-vin: Make use of V4L2_CAP_IO_MC > staging/intel-ipu3: Make use of V4L2_CAP_IO_MC > vimc: Make use of V4L2_CAP_IO_MC > > .../media/v4l/vidioc-enum-fmt.rst | 16 ++++- > .../media/v4l/vidioc-querycap.rst | 6 ++ > .../media/videodev2.h.rst.exceptions | 1 + > drivers/media/pci/cx18/cx18-ioctl.c | 22 ++++-- > drivers/media/pci/ivtv/ivtv-ioctl.c | 26 +++---- > drivers/media/platform/rcar-vin/rcar-v4l2.c | 40 ++++++----- > .../media/test-drivers/vimc/vimc-capture.c | 14 +++- > drivers/media/v4l2-core/v4l2-dev.c | 25 +++++-- > drivers/media/v4l2-core/v4l2-ioctl.c | 70 +++++++++++++++++-- > drivers/staging/media/ipu3/ipu3-v4l2.c | 64 ++--------------- > include/uapi/linux/videodev2.h | 5 +- > 11 files changed, 180 insertions(+), 109 deletions(-) >
On 22/04/2020 19:22, Laurent Pinchart wrote: > Hi Niklas, > > (With a question for Hans below) > > Thank you for the patches. > > On Tue, Apr 21, 2020 at 03:57:37PM +0200, Niklas Söderlund wrote: >> Hi, >> >> This series aims to reduce the amount of boiler plate code in video >> device drivers who's inputs and/or outputs are controlled by the Media >> Controller instead of the V4L2 API. >> >> Patch 1/6 adds a new video device capability flag V4L2_CAP_IO_MC which >> if set provides helper implementations for the enum inputs and outputs >> ioctls freeing the video device driver from the need to implement them. >> >> Patch 2/6 fix initialization of reserved fields in the cx18 and ivtv >> drivers which becomes a problem in 3/6 where Laurent adds mbus filters to >> VIDIOC_ENUM_FMT. >> >> Patch 4/6, 5/6 and 6/6 converts the R-Car VIN, Intel IPU3 and VIMC >> drivers to use the new default handlers and capability flag and delete >> the now redundant boiler plate code. I'm sure more video device drivers >> can make use of this new flag but as I can only test on these two >> platforms I have limited my changes to those. >> >> A separate patch to v4l-utils have been posted as [1] to add a >> test for this feature in v4l2-compliance. >> >> This version have been rebased to latest media-tree to account for the >> large shuffles of files. It has also replaced patch 2/6 with a different >> version that addresses both cx18 and ivtv instead of only ivtv. > > It looks like everything has been reviewed. Hans, do you see any > remaining blocker, or will you take it in your tree and send a pull > request ? Hmm, I'm getting v4l2-compliance errors when testing with vivid et al. Looking into that now. Regards, Hans > >> 1. [PATCH 0/2] v4l2-compliance: add tests for V4L2_CAP_IO_MC >> >> Laurent Pinchart (2): >> media: pci: Fill v4l2_fmtdesc with designated initializers >> media: v4l2: Extend VIDIOC_ENUM_FMT to support MC-centric devices >> >> Niklas Söderlund (4): >> v4l2-dev/ioctl: Add V4L2_CAP_IO_MC >> rcar-vin: Make use of V4L2_CAP_IO_MC >> staging/intel-ipu3: Make use of V4L2_CAP_IO_MC >> vimc: Make use of V4L2_CAP_IO_MC >> >> .../media/v4l/vidioc-enum-fmt.rst | 16 ++++- >> .../media/v4l/vidioc-querycap.rst | 6 ++ >> .../media/videodev2.h.rst.exceptions | 1 + >> drivers/media/pci/cx18/cx18-ioctl.c | 22 ++++-- >> drivers/media/pci/ivtv/ivtv-ioctl.c | 26 +++---- >> drivers/media/platform/rcar-vin/rcar-v4l2.c | 40 ++++++----- >> .../media/test-drivers/vimc/vimc-capture.c | 14 +++- >> drivers/media/v4l2-core/v4l2-dev.c | 25 +++++-- >> drivers/media/v4l2-core/v4l2-ioctl.c | 70 +++++++++++++++++-- >> drivers/staging/media/ipu3/ipu3-v4l2.c | 64 ++--------------- >> include/uapi/linux/videodev2.h | 5 +- >> 11 files changed, 180 insertions(+), 109 deletions(-) >> >