[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20260111185139.GI3634291@ZenIV>
Date: Sun, 11 Jan 2026 18:51:39 +0000
From: Al Viro <viro@...iv.linux.org.uk>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Eric Dumazet <edumazet@...gle.com>, oe-kbuild-all@...ts.linux.dev,
linux-kernel@...r.kernel.org, Jakub Kicinski <kuba@...nel.org>,
Maciej Żenczykowski <maze@...gle.com>,
Will Deacon <will@...nel.org>,
"Paul E. McKenney" <paulmck@...nel.org>,
Peter Zijlstra <peterz@...radead.org>
Subject: Re: include/net/sock.h:2100:16: sparse: sparse: cast to non-scalar
On Sun, Jan 11, 2026 at 06:20:10PM +0000, Al Viro wrote:
> Guys, do you have any problems with replacing __unqual_scalar_typeof()
> uses with that thing?
>
> As in,#ifndef __smp_load_acquire
> #define __smp_load_acquire(p) \
> ({ \
> unqual_non_array(__typeof__(*p)) ___p1 = READ_ONCE(*p); \
> ...
>
> Objections? IMO it's more palatable than current __unqual_scalar_typeof()...
Argh... gcc prior to 8.4 fucks it up, and our minimum is 8.1 ;-/
Pity, that... I still think it's worth keeping in mind for when
we bump the minimal gcc revision, though.
FWIW, with clang it works all way back at least to 3.1; icc - at least
to 13.0.1...
Powered by blists - more mailing lists