Re: [PATCH] wifi: ath11k: rely on mac80211 debugfs handling for vif

Re: [PATCH] wifi: ath11k: rely on mac80211 debugfs handling for vif

[Kalle Valo]

benjamin@sipsolutions.net writes:

> From: Benjamin Berg <benjamin.berg@intel.com>
>
> mac80211 started to delete debugfs entries in certain cases, causing a
> conflict between ath11k also trying to delete them later on. Fix this by
> relying on mac80211 to delete the entries when appropriate and adding
> them from the vif_add_debugfs handler.
>
> Fixes: 0a3d898ee9a8 ("wifi: mac80211: add/remove driver debugfs entries as appropriate")
> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=218364
> Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>

Adding ath11k list so that the whole team sees this.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches

Re: [PATCH] wifi: ath11k: rely on mac80211 debugfs handling for vif

[Kalle Valo]

Kalle Valo <kvalo@kernel.org> writes:

> benjamin@sipsolutions.net writes:
>
>> From: Benjamin Berg <benjamin.berg@intel.com>
>>
>> mac80211 started to delete debugfs entries in certain cases, causing a
>> conflict between ath11k also trying to delete them later on. Fix this by
>> relying on mac80211 to delete the entries when appropriate and adding
>> them from the vif_add_debugfs handler.
>>
>> Fixes: 0a3d898ee9a8 ("wifi: mac80211: add/remove driver debugfs entries as appropriate")
>> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=218364
>> Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
>
> Adding ath11k list so that the whole team sees this.

Thanks, this patch passes my ath11k tests and I don't see crashes
anymore. But what about other drivers, can we trust that they don't have
similar problems? I'm just wondering should we consider reverting the
mac80211 commit for the time being?

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches

Re: [PATCH] wifi: ath11k: rely on mac80211 debugfs handling for vif

[Kalle Valo]

Kalle Valo <kvalo@kernel.org> writes:

> benjamin@sipsolutions.net writes:
>
>> From: Benjamin Berg <benjamin.berg@intel.com>
>>
>> mac80211 started to delete debugfs entries in certain cases, causing a
>> conflict between ath11k also trying to delete them later on.

Instead of calling it a conflict I would prefer to clearly document that
ath11k was crashing so badly that it was unusable.

>>
>> Fixes: 0a3d898ee9a8 ("wifi: mac80211: add/remove driver debugfs entries as appropriate")
>> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=218364

s/Bugzilla:/Closes:/

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches

Re: [PATCH] wifi: ath11k: rely on mac80211 debugfs handling for vif

[Benjamin Berg]

On Fri, 2024-01-12 at 10:24 +0200, Kalle Valo wrote:
> Kalle Valo <kvalo@kernel.org> writes:
> 
> > benjamin@sipsolutions.net writes:
> > 
> > > From: Benjamin Berg <benjamin.berg@intel.com>
> > > 
> > > mac80211 started to delete debugfs entries in certain cases,
> > > causing a
> > > conflict between ath11k also trying to delete them later on. Fix
> > > this by
> > > relying on mac80211 to delete the entries when appropriate and
> > > adding
> > > them from the vif_add_debugfs handler.
> > > 
> > > Fixes: 0a3d898ee9a8 ("wifi: mac80211: add/remove driver debugfs
> > > entries as appropriate")
> > > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=218364
> > > Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
> > 
> > Adding ath11k list so that the whole team sees this.
> 
> Thanks, this patch passes my ath11k tests and I don't see crashes
> anymore. But what about other drivers, can we trust that they don't
> have
> similar problems? I'm just wondering should we consider reverting the
> mac80211 commit for the time being?

Good question.

I did have a look originally and I just went over all drivers (i.e.
debugfs_remove calls). None of these seem to be vif related. As such, I
would say we could just go ahead with this patch and leave in the fix.

That said, we could also revert 0a3d898ee9a8 for now. I don't think
that has any severe side effects other than a warning from debugfs
reappearing again for iwlwifi.

Benjamin