linux-kernel - Re: [PATCH] fs: mark lookup

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20251126-vermachen-sahne-c4f243016180@brauner>
Date: Wed, 26 Nov 2025 11:08:10 +0100
From: Christian Brauner <brauner@...nel.org>
To: Mateusz Guzik <mjguzik@...il.com>
Cc: viro@...iv.linux.org.uk, jack@...e.cz, linux-kernel@...r.kernel.org, 
	linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH] fs: mark lookup_slow() as noinline

On Tue, Nov 25, 2025 at 10:54:25AM +0100, Mateusz Guzik wrote:
> On Tue, Nov 25, 2025 at 10:05 AM Christian Brauner <brauner@...nel.org> wrote:
> >
> > On Wed, Nov 19, 2025 at 03:49:30PM +0100, Mateusz Guzik wrote:
> > > Otherwise it gets inlined notably in walk_component(), which convinces
> > > the compiler to push/pop additional registers in the fast path to
> > > accomodate existence of the inlined version.
> > >
> > > Shortens the fast path of that routine from 87 to 71 bytes.
> > >
> > > Signed-off-by: Mateusz Guzik <mjguzik@...il.com>
> > > ---
> >
> > Fwiw, the biggest problem is that we need to end up with something
> > obvious so that we don't accidently destroy any potential performance
> > gain in say 2 years because everyone forgot why we did things this way.
> >
> 
> I don't think there is a way around reviewing patches with an eye on
> what they are doing to the fast path, on top of periodically checking
> if whatever is considered the current compiler decided to start doing
> something funky with it.
> 
> I'm going to save a rant about benchmarking changes like these in the
> current kernel for another day.

Without knocking any tooling that we currently have but I don't think we
have _meaningful_ performance testing - especially not automated.