NAK: [SRU][J/allwinner-5.17][PATCH 0/4] Fix more Allwinner D1 drivers

[Emil Renner Berthing]

On Fri, 8 Jul 2022 at 17:04, Tim Gardner <tim.gardner at canonical.com> wrote:
>
> On 7/8/22 08:07, Emil Renner Berthing wrote:
> > On Fri, 8 Jul 2022 at 15:59, Tim Gardner <tim.gardner at canonical.com> wrote:
> >>
> >> On 7/8/22 07:41, Emil Renner Berthing wrote:
> >>> On Fri, 8 Jul 2022 at 14:40, Tim Gardner <tim.gardner at canonical.com> wrote:
> >>>>
> >>>> On 7/8/22 04:50, Emil Renner Berthing wrote:
> >>>>> [Impact]
> >>>>>
> >>>>>     * The drivers for the display engine, crypto acceleration and USB
> >>>>>       don't probe because of missing dependencies.
> >>>>>
> >>>>> [Test Plan]
> >>>>>
> >>>>>     * Boot the kernel in linux-image-5.17.0-1002-allwinner and see
> >>>>>       /sys/kerne/debug/devices_deferred
> >>>>>
> >>>>> [Where problems could occur]
> >>>>>
> >>>>>     * Fixing this may introduce now bugs if the now probing drivers
> >>>>>       turn out to be buggy.
> >>>>>
> >>>>> Andre Przywara (1):
> >>>>>      phy: sun4i-usb: Rework HCI PHY (aka "pmu_unk1") handling
> >>>>>
> >>>>> Emil Renner Berthing (1):
> >>>>>      UBUNTU: [Config] Enable additional Allwinner D1 options
> >>>>>
> >>>>> Samuel Holland (2):
> >>>>>      UBUNTU: SAUCE: phy: sun4i-usb: Add D1 variant
> >>>>>      UBUNTU: SAUCE: PM / devfreq: Add dummy R329/D1 MBUS driver
> >>>>>
> >>>>>     debian.allwinner/config/config.common.ubuntu |  5 ++-
> >>>>>     drivers/devfreq/Kconfig                      |  6 +++
> >>>>>     drivers/devfreq/Makefile                     |  1 +
> >>>>>     drivers/devfreq/sun50i-r329-mbus.c           | 27 +++++++++++++
> >>>>>     drivers/phy/allwinner/phy-sun4i-usb.c        | 40 +++++++++++---------
> >>>>>     5 files changed, 60 insertions(+), 19 deletions(-)
> >>>>>     create mode 100644 drivers/devfreq/sun50i-r329-mbus.c
> >>>>>
> >>>>
> >>>> When backporting, don't forget the description of the steps you took. In
> >>>> this case it looks like it was simple context adjustments in patch 1/1.
> >>>
> >>> I'm sorry I don't know what you mean by this. Do you mean describe the
> >>> steps I took produce the error? Eg. install the allwinner kernel and
> >>> look at /sys/kernel/debug/devices_deferred?
> >>>
> >>> Context adjustments? Patch 1/1 is a cherry-pick from mainline and it
> >>> applied cleanly.
> >>>
> >>
> >> Patch 1/1 says '(backported from commit
> >> 1743dea7f06b939f67ba258bab993fa5ff6e43fb)'. This implies you had to futz
> >> with the patch in order to get it to apply. If it was a clean cherry
> >> pick, then it should say 'cherry picked from commit
> >> 1743dea7f06b939f67ba258bab993fa5ff6e43fb'.
> >>
> >> The description of the steps you took to apply the backport typically
> >> follow on the next line, e.g.,
> >>
> >> [emil - applied some grease to the muffler bearings]
> >
> > Ah! Got it, thanks. Yeah, patch 1/1 should just have been
> > "cherry-picked from" then.
> >
> >>>> Where are the annotation updates for the config changes in patch 4/4 ?
> >>>
> >>> I ran cranky updateconfigs and it didn't complain, so I guess these
> >>> options weren't annotated. Should I add annotations for all of them or
> >>> just the option added in patch 3/4?
> >>>
> >>
> >> Do you have 'FORMAT: 3' with "do_enforce_all=true" ?
> >
> > Not in this tree unfortunately, but I'll be sure to add that when
> > creating the allwinner and starfive trees based on 5.19.
> >
> >> The annotations are there to remind us _why_ a config option was
> >> changed. At the very least you'll want an annotation that enforces the
> >> setting along with a note referencing the LP bug. Every config should
> >> have an annotation, either in the master annotation file or in the
> >> derivative annotation file.
> >
> > Right, that's of course the goal. I'll add annotations for the options
> > changed in this series, but do you want me to add annotations for
> > every other un-annotated option too?
> >
>
> For the purposes of this patch, adding annotations for the affected
> config options is sufficient. When you transition to 'FORMAT: 3' and
> do_enforce_all you'll have to have an annotation for every config option.

Of course. Thanks.
/Emil