Re: [PATCH v4 0/3] ceph: don't unregister an MDS session before removing its caps

[Max Kellermann]

On Thu, Oct 15, 2026 at 6:22 PM Alex Markuze <amarkuze@xxxxxxxxxx> wrote:
> Two context lines still call
>   doutc(), and testing calls boutc() there:
>     ceph_mdsc_open_export_target_session(): "to mds%d"
>     send_mds_reconnect(): "session %p state %s"
>   That logging change is already on testing.

You mean your new "BLOG" subsystem that I questioned, but you never
replied to my review?

This is one of the things I will revert in our kernel fork once we
jump to the next LTS. Everything about it feels wrong to me. I'm not
happy with where the Ceph kclient is heading.

>   Unfortunately, it's not scalable for me to rebase every patch sent
> to the mailing list, your cooperation is appreciated.

And it's not scalable for me to rebase every patch that gets delayed
by you until it doesn't apply cleanly anymore.

Really, this was only caused by you, and if you don't want that extra
work, don't ignore my patches, and don't merge conflicting patches 4
weeks later - it's that simple. And no, I won't "argue for any other
patches". Our kernel fork has all the patches we need, that's all that
counts for me. I have offered everything in the past, but most of it
wasn't accepted, so I stopped submitting a while ago (because upstream
lagged behind my work so much my patches weren't compatible anymore,
and I don't want to do extra work to maintain two branches).

In this case, it's really trivial. There isn't even a conflict. Just
some patch context changes. Apply my patches to a commit before your
BLOG merge, then "git rebase". There are a hundred other ways to
resolve this without editing anything manually.

--
Max Kellermann
Principal Architect
Hosting Technology

cm4all | Im Mediapark 6a | 50670 Köln | Germany
General information about the company can be found here:
https://www.cm4all.com/impressum
A member of the IONOS Group