Message ID | 20220719121003.1829916-1-paul.elder@ideasonboard.com |
---|---|
Headers | show |
Series |
|
Related | show |
Hi Paul, On Tue, Jul 19, 2022 at 09:10:01PM +0900, Paul Elder via libcamera-devel wrote: > This series makes the rkisp1 pipeline get the list of supported formats > from the V4L2 device, as opposed to hardcoding it as before. This lets > us add support for, notably, UYVY without having to worry about driver > versions. > > The current implementation of V4L2VideoDevice prevents this from working > as-is, so it is massaged in patch 1. > > This series also obviously deprecates the series that I sent earlier > today: v3 "pipeline: rkisp1: Support devices without self path". That Did you mean to mention another patch here ? Supporting devices without a self path still seems relevant to me. > series was meant to be a quick fix for the formats that we know are > supported for all versions of the driver. > > Paul Elder (2): > libcamera: v4l2_videodevice: Get formats even if no framesizes > pipeline: rkisp1: Query the driver for formats > > include/libcamera/internal/v4l2_videodevice.h | 2 +- > src/libcamera/pipeline/rkisp1/rkisp1_path.cpp | 50 +++++++++---------- > src/libcamera/pipeline/rkisp1/rkisp1_path.h | 6 ++- > src/libcamera/v4l2_videodevice.cpp | 12 +++-- > 4 files changed, 38 insertions(+), 32 deletions(-)
On Wed, Jul 20, 2022 at 01:43:03AM +0300, Laurent Pinchart wrote: > Hi Paul, > > On Tue, Jul 19, 2022 at 09:10:01PM +0900, Paul Elder via libcamera-devel wrote: > > This series makes the rkisp1 pipeline get the list of supported formats > > from the V4L2 device, as opposed to hardcoding it as before. This lets > > us add support for, notably, UYVY without having to worry about driver > > versions. > > > > The current implementation of V4L2VideoDevice prevents this from working > > as-is, so it is massaged in patch 1. > > > > This series also obviously deprecates the series that I sent earlier > > today: v3 "pipeline: rkisp1: Support devices without self path". That > > Did you mean to mention another patch here ? Supporting devices without > a self path still seems relevant to me. Yeah, I meant v2 of "pipeline: rkisp1: Add output support for YUV420 and YVU420" :/ Paul > > > series was meant to be a quick fix for the formats that we know are > > supported for all versions of the driver. > > > > Paul Elder (2): > > libcamera: v4l2_videodevice: Get formats even if no framesizes > > pipeline: rkisp1: Query the driver for formats > > > > include/libcamera/internal/v4l2_videodevice.h | 2 +- > > src/libcamera/pipeline/rkisp1/rkisp1_path.cpp | 50 +++++++++---------- > > src/libcamera/pipeline/rkisp1/rkisp1_path.h | 6 ++- > > src/libcamera/v4l2_videodevice.cpp | 12 +++-- > > 4 files changed, 38 insertions(+), 32 deletions(-)