Re: [PATCH] fs/ceph/io: make ceph_start_io_*() killable

[Max Kellermann]

On Fri, Jun 6, 2025 at 7:15 PM Viacheslav Dubeyko <Slava.Dubeyko@xxxxxxx> wrote:
> I see the point. Our last discussion has finished with statement that Max
> doesn't care about this patch set and we don't need to pick it up. If he changed
> his mind, then I can return to the review of the patch. :) My understanding was
> that he prefers another person for the review. :) This is why I keep silence.

I do care, always did. I answered your questions, but they were not
really about my patch but about whether error handling is necessary.
Well, yes, of course! The whole point of my patch is to add an error
condition that did not exist before. If locking can fail, of course
you have to check that and propagate the error to the caller (and
unlocking after a failed lock of course leads to sorrow). That is so
trivial, I don't even know where to start to explain this if that
isn't already obvious enough.

If you keep questioning that, are you really qualified to do a code review?

Max