Re: [GIT PULL] vfs fixes

From: Christian Brauner <brauner@kernel.org>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [GIT PULL] vfs fixes
Date: Fri, 20 Oct 2023 13:14:32 +0200	[thread overview]
Message-ID: <20231020-abringen-gesurft-5691e5a997ec@brauner> (raw)
In-Reply-To: <CAHk-=whBXdLJ=QDpYmDEH-Tn71dXasGJSX4Jz4qMo8V4-7vYkg@mail.gmail.com>

> Ouch. That filename ref by audit was always supposed to be
> thread-local in a "for this system call" kind of sense.

Yeah, I wasn't happy when that bug showed up.

> That said, using atomics for reference counting is our default
> behavior and should be normal, so the patch isn't wrong, it's just
> annoying since getname/putname is very much in the critical path of
> filename handling.

Yeah.

> That said, the extra atomics are hopefully not really noticeable.
> 
> Some people might want to use the non-refcounted version (ie we have
> getname/putname used by ksmbd too, for example), if they really care.
> 
> It already exists, as __getname/__putname.
> 
> But the normal open/stat/etc system call paths are obviously now going
> to hit those extra atomics. Not lovely, but I guess it's the best we
> can do.

I didn't spend too much time on this issue because it's -rc7 and the
straightforward seemed ok, if annoying.

But if we really really really really cared we could probably do a
deranged thing and massage both audit and io_uring to allows us to
separate the regular system call getname from the io_uring getname. But
I think that would be ugly and likely error prone.