linux-kernel - Re: [RFC PATCH] fs: elide the smp_rmb fence in fd

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: <CAGudoHHRGrQc5ezOLytq1dwmpGkXYyysBut0SqGveuLwrGaTRg@mail.gmail.com>
Date: Thu, 5 Dec 2024 20:47:24 +0100
From: Mateusz Guzik <mjguzik@...il.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: paulmck@...nel.org, Al Viro <viro@...iv.linux.org.uk>, brauner@...nel.org, 
	jack@...e.cz, linux-fsdevel@...r.kernel.org, edumazet@...gle.com, 
	linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH] fs: elide the smp_rmb fence in fd_install()

On Thu, Dec 5, 2024 at 8:26 PM Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> On Thu, 5 Dec 2024 at 10:41, Paul E. McKenney <paulmck@...nel.org> wrote:
> >
> > > To my understanding this is the idiomatic way of spelling out the
> > > non-existent in Linux smp_consume_load, for the resize_in_progress
> > > flag.
> >
> > In Linus, "smp_consume_load()" is named rcu_dereference().
>
> Linux.
>
> But yes and no.
>
> It's worth making it really really clear that "rcu_dereference()" is
> *not* just a different name for some "smp_consume_load()" operation.
>
> Why? Because a true smp_consume_load() would work with any random kind
> of flags etc. And rcu_dereference() works only because it's a pointer,
> and there's an inherent data dependency to what the result points to.
>
> Paul obviously knows this, but let's make it very clear in this
> discussion, because if somebody decided "I want a smp_consume_load(),
> and I'll use rcu_dereference() to do that", the end result would
> simply not work for arbitrary data, like a flags field or something,
> where comparing it against a value will only result in a control
> dependency, not an actual data dependency.
>

So I checked for kicks and rcu_dereference comes with type checking,
as in passing something which is not a pointer even fails to compile.

I'll note thought that a smp_load_consume_ptr or similarly named
routine would be nice and I'm rather confused why it was not added
given smp_load_acquire and smp_store_release being there.

One immediate user would be mnt_idmap(), like so:
iff --git a/include/linux/mount.h b/include/linux/mount.h
index 33f17b6e8732..4d3486ff67ed 100644
--- a/include/linux/mount.h
+++ b/include/linux/mount.h
@@ -76,7 +76,7 @@ struct vfsmount {
 static inline struct mnt_idmap *mnt_idmap(const struct vfsmount *mnt)
 {
        /* Pairs with smp_store_release() in do_idmap_mount(). */
-       return READ_ONCE(mnt->mnt_idmap);
+       return smp_load_consume_ptr(mnt->mnt_idmap);
 }

 extern int mnt_want_write(struct vfsmount *mnt);


-- 
Mateusz Guzik <mjguzik gmail.com>