Re: possible 6.6 regression: Deadlock involving super_lock()

[Christian Brauner]

On Thu, Mar 21, 2024 at 10:15:04AM +0100, Donald Buczek wrote:
> Hi,
> 
> we have a set of 6 systems with similar usage patterns which ran on
> 5.15 kernels for over a year.  Only two weeks after we've switched one
> of the systems from a 5.15 kernel to a 6.6 kernel, it went into a
> deadlock. I'm aware that I don't have enough information that this
> could be analyzed, but I though I drop it here anyway, because the
> deadlock seems to involve the locking of a superblock and I've seen
> that some changes in that area went into 6.6. Maybe someone has an
> idea or suggestions for further inspection if this happens the next
> time.

Ok, I'll take a look. It might take a little.