[Intel-gfx] Regression on linux-next (next-20231130)

[Berg, Johannes]

Hi,

> [snip lockdep report]

> commit f4acfcd4deb158b96595250cc332901b282d15b0
> Author: Johannes Berg johannes.berg at intel.com
> Date:   Fri Nov 24 17:25:25 2023 +0100
> 
>     debugfs: annotate debugfs handlers vs. removal with lockdep

Yes, obviously, since before that there was no lockdep class "debugfs:i915_pipe" 😊

> We also verified that if we revert the patch the issue is not seen.
> 
> Could you please check why the patch causes this regression and provide a fix
> if necessary?

First off, I already sent a revert, which should be included in the next -rc. Anyway this patch shouldn't have been included in the -rc cycle, I just erroneously included it with some bugfixes (that patch-wise had a dependency).

Secondly, we did find a false positive in another case, and yours seems to be the same or similar, due to seq_file not differentiating between the file instances.

That's a bit unfortunate, because we _did_ have actual deadlocks in wireless with debugfs_remove() being called on a file while holding a lock that the file also acquires, which can lead to a deadlock. Unless we differentiate seq_file instances though, there doesn't seem to be a good way to annotate that in debugfs, as this report and others show.

johannes
-- 

Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de <http://www.intel.de>
Managing Directors: Christin Eisenschmid, Sharon Heck, Tiffany Doon Silva  
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928