ACK/Cmnt: [PULL V5][SRU][Jammy] Support Intel IPU6 MIPI camera on Alder Lake platforms

[Andrea Righi]

On Tue, Apr 12, 2022 at 11:39:17AM +0800, You-Sheng Yang wrote:
> BugLink: https://bugs.launchpad.net/bugs/1955383
> 
> [Impact]
> 
> To support Intel IPU6 MIPI camera on Alder Lake platforms.
> 
> [Fix]
> 
> Initial support for Intel IPU6 MIPI camera on Tiger Lake platforms has
> been addressed by bug 1921345 and 1939539. They are backported from
> https://github.com/intel/ipu6-drivers.
> 
> Further works to enable IPU6 camera on Alder Lake platforms depend on a
> few more fixes from same ipu6-drivers repository, as well as an extra
> https://github.com/intel/ivsc-driver for Intel Vision Sensing
> Controller(IVSC).
> 
> [Test Case]
> 
> This depends on an integral of enablement components inclusive of the
> kernel drivers that are being proposed, firmware, updates for the
> userspace camera hardware abstration layer library and a gstreamer
> element as what we have for Tiger Lake platforms.
> 
> [Where problems could occur]
> 
> It's confirmed Intel IPU6 MIPI camera doesn't support suspend at
> streaming.
> 
> [Other Info]
> 
> Before this, we did not submit IPU6 driver for generic kernel due to
> the lack of a solid commitment from the vendor to upstream this
> before kernel camera API is out. There is still no such commitment,
> but during the process we've been able to address version migration
> issues and we would continue to do so. We're getting to have the
> confidence that it's possible to maintain and share the support for
> generic Ubuntu users.
> 
> Fixes for oem-5.14 have been merged, so here we nominate for Jammy
> only. This patchset is almost identical to that for oem-5.14 except
> one additioanl fix for 5.15 kernel API change in "UBUNTU: SAUCE: Fix
> build error for kernel 5.15".
> 
> V2: no change rebase only.
> V3: build only on amd64 and add config annotations
> V4: drop obsoleted cflags manipulation and fix UBSAN warnings
> V5: correct wrong branch/repo generated in previous rev

Glad to see that the UBSAN warnings have been fixed. I've done a review
of the new patch set and the code looks sane to me, I didn't spot any
obvious bug.

I'd really like to see more stuff imported from upstream rather than
github.com/intel/ivsc-driver and I was wondering if all of this could be
provided as a dkms module, since the code is pretty much self-contained
as distinct kernel module(s), I think we may have mentioned this when we
reviewed the previous version of this patch.

If the dkms route isn't doable then I think we can accept this code in
the main kernel, so overall:

Acked-by: Andrea Righi <andrea.righi at canonical.com>