On Tue, Aug 22, 2023 at 11:46:38AM +0200, Christian Brauner wrote:
> > Christian: I've been planning to merge the {freeze,thaw}_super @who
> > changes for 6.6; do you think more 'cooperating with the maintainer' is
> > needed, or shall I simply push my branch to Linus with a note that
> > s/down_write/super_lock_excl/ s/up_write/super_unlock_excl is needed to
> > resolve the merge the conflict?
>
> Hm, that's not a pleasant merge conflict given that it's locking
> changes. It would probably be fine to just bring it up the way it is but
> it looks needlessly messy/uncoordinated. I'm wonder why this isn't just
> all in vfs.super since it's core vfs infra change anyway. Maybe I just
> missed the patches if so then sorry about that.
Ah, I had wondered about that. :)
> That's the two infrastructure patches in the kernel-fsfreeze
> branch/kernel-fsfreeze_2023-07-27 tag?:
>
> ad0164493b81 ("fs: distinguish between user initiated freeze and kernel initiated freeze")
> 53f65fd7a3d5 ("fs: wait for partially frozen filesystemskernel-fsfreeze_2023-07-27kernel-fsfreeze")
>
> If you give me a tag with your description and just the two commits or I
> just cherry pick them and cite your description in my pr that would be
> my preferred solution. How do you feel about that?
I'm happy to have you pull my xfs-linux tags into your vfs tree. :)
Here's a tag with just the two vfs patches:
https://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git/tag/?h=vfs-6.6-merge-2
This second tag builds on that, by adding the first actual user of
FREEZE_HOLDER_KERNEL:
https://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git/tag/?h=vfs-6.6-merge-3
There will be more for 6.7(+?) if Luis manages to get back to his
auto-fsfreeze during suspend, or if Shiyang finishes the series to
handle pmem media error reporting in xfs.
Thanks!
--D
