On Fri, Jun 07, 2024 at 02:01:54PM +0200, Geert Uytterhoeven wrote:
> Hi Greg,
>
> On Fri, Jun 7, 2024 at 12:25 PM Greg Kroah-Hartman
> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > On Fri, Jun 07, 2024 at 09:16:17AM +0200, Geert Uytterhoeven wrote:
> > > On Thu, Jun 6, 2024 at 4:45 PM Arnd Bergmann <arnd@xxxxxxxx> wrote:
> > > > On Thu, Jun 6, 2024, at 15:04, gregkh@xxxxxxxxxxxxxxxxxxx wrote:
> > > > > From 8f7f115596d3dccedc06f5813e0269734f5cc534 Mon Sep 17 00:00:00 2001
> > > > > From: Geert Uytterhoeven <geert+renesas@xxxxxxxxx>
> > > > > Date: Mon, 22 Apr 2024 12:30:38 +0200
> > > > > Subject: Revert "drm: Make drivers depends on DRM_DW_HDMI"
> > > > >
> > > > > From: Geert Uytterhoeven <geert+renesas@xxxxxxxxx>
> > > > >
> > > > > commit 8f7f115596d3dccedc06f5813e0269734f5cc534 upstream.
> > > > >
> > > > > This reverts commit c0e0f139354c01e0213204e4a96e7076e5a3e396, as helper
> > > >
> > > > As far as I can tell, the original commit never made it into
> > > > stable/linux-6.9.y, so maybe remove c0e0f139354c ("drm: Make
> > > > drivers depends on DRM_DW_HDMI") instead of reverting it?
> > >
> > > TBH, I do not understand why this commit was backported: it was reverted
> > > in the same pull request that introduced the bad commit, and thus both
> > > the bad commit and the revert ended up upstream together.
> > > Looks like a good opportunity to improve the backporting scripts...
> >
> > This was done on purpose, otherwise we would be going "oh look, we
> > missed that one!" and adding the broken commit again and then noticing
> > that we forgot the revert.
>
> OK, makes (some) sense...
>
> > Doing both makes accounting for things like
> > this much simpler by both us, and other people who would report the
> > original commit as being missed as well.
> >
> > Yes, it's churn, but it seems to work better so we have been doing it
> > this way for a few years now.
>
> However, that is not what seems to have happened?
> According to Mark's report[1], initially the revert was not picked up
> by the scripting.
>
> [1] https://lore.kernel.org/all/dc0c4e9d-e37c-442d-8b75-72f0e2927802@xxxxxxxxxxxxx
Initially, yes, but by the time the -rc went out, it was :)
And yes, sometimes we miss reverts, I try to catch them as part of the
"find the fixes for the fixes" pass I make right before I do the -rcs,
but sometimes I forget.
For example, this -rc round, I missed a bunch of them for at least one
of the branches and will be doing a new -rc2 release "soon" with them
included. I blame the my past two weeks of constant travel and meetings
on this, sorry about that, next week will be better.
thanks,
greg k-h
