[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <cc4b91e4-13f8-c489-e2ce-445770ddf876@xs4all.nl>
Date: Mon, 13 May 2019 15:56:34 +0200
From: Hans Verkuil <hverkuil-cisco@...all.nl>
To: Naresh Kamboju <naresh.kamboju@...aro.org>,
linux-media@...r.kernel.org
Cc: open list <linux-kernel@...r.kernel.org>,
Sami Tolvanen <samitolvanen@...gle.com>,
paul.kocialkowski@...tlin.com, ezequiel@...labora.com,
treding@...dia.com, niklas.soderlund+renesas@...natech.se,
sakari.ailus@...ux.intel.com,
Hans Verkuil <hans.verkuil@...co.com>, mchehab@...nel.org,
lkft-triage@...ts.linaro.org
Subject: Re: Format ioctls (Input 1) test Cropping failed: 4.19, 4.14, 4.9 and
4.4
On 5/13/19 3:36 PM, Naresh Kamboju wrote:
> Format ioctls (Input 1) test Cropping failed on all devices running
> 4.19, 4.14, 4.9 and 4.4 kernel branches.
>
> Format ioctls (Input 1):
>
> ...
> test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
> fail: ../../../v4l-utils-1.16.0/utils/v4l2-compliance/v4l2-test-formats.cpp(1425):
> doioctl(node, VIDIOC_CROPCAP, &cap)
> fail: ../../../v4l-utils-1.16.0/utils/v4l2-compliance/v4l2-test-formats.cpp(1447):
> testLegacyCrop(node)
> test Cropping: FAIL
There is no easy patch you can backport to fix this and I wouldn't bother with this.
>
> Test passes on mainline, -next and 5.0
> Test failed on 4.19, 4.14, 4.9 and 4.4 for all devices
>
> steps to reproduce:
>
> # boot 4.19/4.14/4.9 kernel on x86_64 / Juno / hikey / device
> # install v4l-utils package
> # modprobe vivid.ko no_error_inj=1
> # v4l2-compliance -v -d /dev/video0
>
> Test results comparison on all branches,
> https://qa-reports.linaro.org/_/comparetest/?project=22&project=6&project=58&project=135&project=40&project=23&project=159&suite=v4l2-compliance&test=Cropping
>
> Test output log,
> https://lkft.validation.linaro.org/scheduler/job/708755#L2236
>
> Best regards
> Naresh Kamboju
>
Be aware that v4l2-compliance is meant to be compiled directly from the
v4l-utils master repo, and run against the latest media-tree master
repository. It's not meant to be backwards compatible with older kernels.
If you want to do regression testing with v4l2-compliance against older
kernels, then it would actually be better to pick a v4l2-compliance version
from around the time that kernel version was first released.
Regards,
Hans
Powered by blists - more mailing lists