Re: [GIT pull] timers/core for v6.13-rc1

[Ingo Molnar <mingo@kernel.org>]

* Christian Brauner <brauner@kernel.org> wrote:

> On Wed, Nov 20, 2024 at 01:12:28PM +0100, Ingo Molnar wrote:
> > 
> > * Christian Brauner <brauner@kernel.org> wrote:
> > 
> > > On Tue, Nov 19, 2024 at 04:33:45PM -0800, Linus Torvalds wrote:
> > > > On Mon, 18 Nov 2024 at 11:22, Thomas Gleixner <tglx@linutronix.de> wrote:
> > > > >
> > > > 
> > > > >   - Core infrastructure for VFS multigrain timestamping
> > > > >
> > > > >     This is required to allow the kernel to use coarse grained time stamps
> > > > >     by default and switch to fine grained time stamps when inode attributes
> > > > >     are actively observed via getattr().
> > > > >
> > > > >     These changes have been provided to the VFS tree as well, so that the
> > > > >     VFS specific infrastructure could be built on top.
> > > > 
> > > > Bah. Except the vfs tree didn't take it as a shared branch, but
> > > > instead cherry-picked the commits and as a result they are duplicate
> > > > and caused a (trivial) merge conflict.
> > > 
> > > Wait, I'm confused. I definitely pulled that branch the day after Thomas
> > > gave it to me and in my vfs.mgtime branch I clearly see:
> > > 
> > > commit d7c898a73f875bd205df53074c1d542766171da1
> > > Merge: 8cf0b93919e1 2a15385742c6
> > > Author:     Christian Brauner <brauner@kernel.org>
> > > AuthorDate: Mon Oct 7 12:47:19 2024 +0200
> > > Commit:     Christian Brauner <brauner@kernel.org>
> > > CommitDate: Thu Oct 10 10:20:57 2024 +0200
> > > 
> > >     Merge tag 'timers-core-for-vfs' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/tip/tip into vfs.mgtime
> > > 
> > >     Timekeeping interfaces for consumption by the VFS tree.
> > > 
> > >     Signed-off-by: Christian Brauner <brauner@kernel.org>
> > > 
> > > Unless I did something odd during the pull?
> > 
> > The problem was caused by two commits which got rebased in the VFS 
> > tree:
> > 
> > Commit 1:
> > 
> >   ee3283c608df ("timekeeping: Add interfaces for handling timestamps with a floor value")
> > 
> >   commit ee3283c608dfa21251b0821d7bb198c7ae3189f6
> >   Author:     Jeff Layton <jlayton@kernel.org>
> >   AuthorDate: Wed Oct 2 17:27:16 2024 -0400
> >   Commit:     Christian Brauner <brauner@kernel.org>
> >   CommitDate: Thu Oct 10 10:20:46 2024 +0200
> >               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > 
> > Which is a rebase of Thomas's original commit:
> > 
> >   70c8fd00a9bd ("timekeeping: Add interfaces for handling timestamps with a floor value")
> > 
> >   commit 70c8fd00a9bd0509bbf7bccd9baea8bbd5ddc756
> >   Author:     Jeff Layton <jlayton@kernel.org>
> >   AuthorDate: Wed Oct 2 17:27:16 2024 -0400
> >   Commit:     Thomas Gleixner <tglx@linutronix.de>
> >   CommitDate: Sun Oct 6 20:56:07 2024 +0200
> > 
> > And commit 2:
> > 
> >   2a15385742c6 ("timekeeping: Add percpu counter for tracking floor swap events")
> > 
> >   commit 2a15385742c689a271345dcbb4c28b9c568bc7ce
> >   Author:     Jeff Layton <jlayton@kernel.org>
> >   AuthorDate: Wed Oct 2 17:27:17 2024 -0400
> >   Commit:     Christian Brauner <brauner@kernel.org>
> >   CommitDate: Thu Oct 10 10:20:46 2024 +0200
> >               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > 
> > Which is a rebase of Thomas's original commit:
> > 
> >   96f9a366ec8a timekeeping: Add percpu counter for tracking floor swap events
> > 
> >   commit 96f9a366ec8abe027326d7aab84d64370019f0f1 (tag: timers-core-for-vfs)
> >   Author:     Jeff Layton <jlayton@kernel.org>
> >   AuthorDate: Wed Oct 2 17:27:17 2024 -0400
> >   Commit:     Thomas Gleixner <tglx@linutronix.de>
> >   CommitDate: Sun Oct 6 20:56:07 2024 +0200
> 
> I was just looking through my reflog and I realised that I did a rebase
> onto v6.12-rc2 (v6.12-rc1 had a horribly virtqueue bug that made testing
> in a vm a giant pain). Sorry about that I should've noticed that
> earlier.

So, so sh1t happens, but I think there could also be a workflow bug 
here: somehow your tooling added your Signed-off-by during the rebase, 
to a commit not committed by you originally.

> Though I'm confused why -next didn't catch this. When that happens -next
> usually reports duplicate commits. Hm... Did I miss the report?

I haven't seen duplicate commit warnings from -next for some time - but 
it can detect accidental rebases, because without your SOB added -next 
would have detected the incorrect SOB chain I believe. So I'd 
investigate how your SOB got there. I don't think vanilla rebase is 
adding it automatically?

Thanks,

	Ingo