Re: [GIT PULL] Rockchip dts64 changes for 6.15 #3

["Heiko Stübner" <heiko@sntech.de>]

Am Dienstag, 8. April 2025, 16:01:21 Mitteleuropäische Sommerzeit schrieb Rob Herring:
> On Mon, Apr 7, 2025 at 11:49 AM Heiko Stübner <heiko@sntech.de> wrote:
> >
> > Am Montag, 7. April 2025, 18:43:54 Mitteleuropäische Sommerzeit schrieb Rob Herring:
> > > On Mon, Apr 7, 2025 at 11:40 AM Heiko Stübner <heiko@sntech.de> wrote:
> > > >
> > > > Hi Rob,
> > > >
> > > > Am Montag, 7. April 2025, 17:56:37 Mitteleuropäische Sommerzeit schrieb Rob Herring:
> > > > > On Wed, Mar 19, 2025 at 6:56 PM Heiko Stuebner <heiko@sntech.de> wrote:
> > > > > >
> > > > > > Hi soc maintainers,
> > > > > >
> > > > > > I made an error and accidentially applied a patch that was meant for
> > > > > > the mfd tree. Thankfully Stephen noticed that when the duplicate
> > > > > > commit appeared in linux-next.
> > > > >
> > > > > Both commits are in v6.15-rc1 now and the revert is not, so this
> > > > > should not get applied/pulled. Or you will need to revert the revert.
> > > >
> > > > yes, that was the intention.
> > > >
> > > > Back when I submitted this PR, I talked with Arnd on IRC the next day.
> > > >
> > > > As both commits are identical sans some Signed-off-by lines, he suggested
> > > > not trying to put a revert in, but instead let git solve it itself, because
> > > >
> > > > arnd on IRC:
> > > > > [...], but I worry that this would make things worse if 'git merge'
> > > > > ends up doing the revert on top of the original commit once it gets to
> > > > > torvalds. Not sure if that's still a problem in git these days, but
> > > > > I've seen it happen in the past.
> > > > > if two identical patches are in different branches, just leaving them
> > > > > there is usually easier
> > > >
> > > > So this PR was already marked as "superseeded" in patchwork back
> > > > on march 20th.
> > > >
> > > >
> > > > Nevertheless, thanks for making sure no funky revert happens now.
> > >
> > > The commit is still in linux-next though. That's how I happened upon this.
> >
> > Ah right. I've recreated my next branch earlier today, after the -rc1
> > release, so this should be gone with the next linux-next.
> 
> Still there today and now we have:

Not much luck for me this week it seems.

So I've now re-created the for-next branch again, especially without the
stuff from below. I guess the Revert slipped in again via that miss-merged
branch.

Looking at the rec-created
https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git/log/Documentation/devicetree?h=for-next

that revert should be finally gone now and not reappear.


> ['tsd,px30-cobra-ltk050h3146w', 'rockchip,px30-cobra', 'rockchip,px30']
> ['tsd,px30-cobra-ltk050h3146w-a2', 'rockchip,px30-cobra', 'rockchip,px30']
> ['tsd,px30-cobra-ltk500hd1829', 'tsd,px30-cobra', 'rockchip,px30']
> 
> Which appear on no list.

And I've also fixed my scripts to not miss-merge branches not actually
intended for next.

Heiko