[libcamera-devel] [PATCH v6 4/5] android: camera_device: Get properties from configuration
Laurent Pinchart
laurent.pinchart at ideasonboard.com
Sun May 23 17:21:17 CEST 2021
On Sun, May 23, 2021 at 03:56:54PM +0300, Laurent Pinchart wrote:
> On Sun, May 23, 2021 at 03:32:08AM +0300, Laurent Pinchart wrote:
> > On Fri, May 21, 2021 at 05:56:41PM +0200, Jacopo Mondi wrote:
> > > On Fri, May 21, 2021 at 03:41:10PM +0800, Han-lin Chen wrote:
> > > > Hi Jacopo,
> > > > Thanks for the efforts. The commits should be contained in 13971 and
> > > > any later versions.
> > >
> > > Great, thanks, I have now verified my series works on an
> > > out-of-the-box image.
> > >
> > > However the manifests for the SDK in manifest-version have not yet
> > > catch up with the images and are stuck to R91. We'll have to wait
> > > before merging the series for the right manifest to land, so we can
> > > have both the image and the SDK at the same version.
> >
> > Seems like it's available now. I'll kick off a build with
> > full/buildspecs/92/13971.0.0-rc1.xml and report the results.
>
> I had a build error in the chromeos-base/regions packages:
>
> 14:49:49 >>> Compiling source in /build/soraka-libcamera/tmp/portage/chromeos-base/regions-0.0.1-r2020/work/regions-0.0.1/regions ...
> usage: regions.py [-h] [--format {human-readable,csv,json,yaml}] [--all] [--notes] [--include_pseudolocales] [--output OUTPUT] [--overlay OVERLAY]
> regions.py: error: unrecognized arguments:
>
> This was caused by the ebuild passing an empty argument to the
> regions.py script (a `print(args)` in regions.py clearly shows this).
> I've fixed it with the following patch.
>
> diff --git a/chromeos-base/regions/regions-0.0.1-r2020.ebuild b/chromeos-base/regions/regions-0.0.1-r2020.ebuild
> index 22684c2bf7e8..06b1a82f027c 100644
> --- a/chromeos-base/regions/regions-0.0.1-r2020.ebuild
> +++ b/chromeos-base/regions/regions-0.0.1-r2020.ebuild
> @@ -31,7 +31,7 @@ src_unpack() {
> }
>
> src_compile() {
> - ./regions.py --format=json --output "${WORKDIR}/cros-regions.json" "$(usex cros-debug "--include_pseudolocales" "")"
> + ./regions.py --format=json --output "${WORKDIR}/cros-regions.json" $(usex cros-debug "--include_pseudolocales" "")
> }
>
> src_test() {
>
> I'm curious, as the offending code was added in
>
> commit a05940690157c100c902f41e8118400f2183eb3b
> Author: Matt Stark <msta at google.com>
> AuthorDate: Mon Apr 19 11:16:04 2021 +1000
> Commit: Commit Bot <commit-bot at chromium.org>
> CommitDate: Wed May 5 05:06:19 2021 +0000
>
> When creating a chrome OS debug build, add pseudolocales to the build.
>
> which is more than two weeks old, and I would have expected the issue to
> be caught. I can't see any fix in neither the chromeos-overlay (for the
> build) nor the platform2 (for regions.py) main branches.
>
> I can't rule out that it could be specific to my environment, as I
> haven( recreated the SDK from scratch, I've removed the build directory
> and updated the chroot with update_chroot.
>
> Could anyone on the Chrome OS site which if they can reproduce this
> issue when the cros-debug use flag isn't set ? I've added a comment to
> https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/2879390.
>
> Other than that, the build is still ongoing, I'll report when it
> completes.
Or when it fails again :-S
The next issue is related to cryptohome.
cryptohome-0.0.1-r3658: 17:36:31 >>> Install chromeos-base/cryptohome-0.0.1-r3658 into /build/soraka-libcamera/tmp/portage/chromeos-base/cryptohome-0.0.1-r3658/image
cryptohome-0.0.1-r3658: * ERROR: chromeos-base/cryptohome-0.0.1-r3658::chromiumos failed (install phase):
cryptohome-0.0.1-r3658: * direncription_allow_v2 is enabled where it shouldn't be. Do you need to change the board overlay? Note, uprev boards should have it disabled!
cryptohome-0.0.1-r3658: *
cryptohome-0.0.1-r3658: * Call stack:
cryptohome-0.0.1-r3658: * ebuild.sh, line 125: Called src_install
cryptohome-0.0.1-r3658: * environment, line 4321: Called die
cryptohome-0.0.1-r3658: * The specific snippet of code:
cryptohome-0.0.1-r3658: * die "direncription_allow_v2 is enabled where it shouldn't be. Do you need to change the board overlay? Note, uprev boards should have it disabled!";
The error check was added in the ebuild in
commit c9cbca71d43158c5dc1bd366224087e34e7905b7
Author: Daniil Lunev <dlunev at chromium.org>
AuthorDate: Fri Feb 12 10:17:01 2021 +1100
Commit: Commit Bot <commit-bot at chromium.org>
CommitDate: Mon Mar 8 23:30:29 2021 +0000
cryptohome: rollout control for fscrypt_v2
The uprev-4-to-5 USE flag got enabled for Soraka (and Nautilus) in
commit 14bb32f0477d5956f86802f80e8e4d0ab8c7b9bc
Author: Daniil Lunev <dlunev at chromium.org>
AuthorDate: Wed Mar 10 09:22:06 2021 +1100
Commit: Commit Bot <commit-bot at chromium.org>
CommitDate: Tue Mar 9 23:56:23 2021 +0000
overlays: add a special tag for 4->5 uprevs
The direncription_allow_v2 USE flag got enabled for soraka-libcamera in
commit 6b846af8ee5752878fc2e125c612719b3467cc3b
Author: Daniil Lunev <dlunev at chromium.org>
AuthorDate: Mon Mar 1 09:28:11 2021 +1100
Commit: Commit Bot <commit-bot at chromium.org>
CommitDate: Tue Mar 2 06:14:31 2021 +0000
overlays: enable fscrypt v2 attempt on post 5.4 boards
Disabling direncription_allow_v2 manually allows the build to continue,
but I don't think that's a very good idea.
> > > > On Fri, May 21, 2021 at 3:29 PM Jacopo Mondi wrote:
> > > > > On Fri, May 21, 2021 at 02:15:51PM +0900, Hirokazu Honda wrote:
> > > > > > On Thu, Apr 15, 2021 at 10:51 PM Jacopo Mondi <jacopo at jmondi.org> wrote:
> > > > > >
> > > > > > > Open the HAL configuration file in the Camera HAL manager and get
> > > > > > > the camera properties for each created CameraDevice and initialize it
> > > > > > > with them.
> > > > > > >
> > > > > > > Reviewed-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> > > > > >
> > > > > > Reviewed-by: Hirokazu Honda <hiroh at chromium.org>
> > > > > >
> > > > > > When would you merge this series?
> > > > >
> > > > > My plan was to update my SDK to the latest version which contains the
> > > > > CL to the libcamera and libcamera-configs ebuild, so I could have
> > > > > tested one more time on a fresh image. Am I too paranoid ?
> > > > >
> > > > > Could you or Han-lin tell me which is the id of the first CPFE image
> > > > > that contains:
> > > > > https://chromium-review.googlesource.com/c/chromiumos/overlays/board-overlays/+/2887093
> > > > > https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/2886535
> > > > >
> > > > > I've tested those patches applied on my rather ancient SDK version,
> > > > > but I would like to try with an out-of-the-box image
> > > > >
> > > > > > > Signed-off-by: Jacopo Mondi <jacopo at jmondi.org>
--
Regards,
Laurent Pinchart
More information about the libcamera-devel
mailing list