From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id 1BB04385B530; Tue,  7 Mar 2023 17:44:07 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 1BB04385B530
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1678211047;
	bh=UUr4LjofMTgme8JtAkjGpH1KePSmcWcPv/eHf0FBN2g=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=HJPenOu8g/pZSKRrE6EDqxpzabKRjsyLfxenZ6mYPpu9+gGD2c58Eh2DG7/sEqlLX
	 SzvaPYIqNc76KXm8IzWkcGLRiE5I11BszsdR+P66NFCXSf/5FsCUC7qIt2MqbcdeQ4
	 VpfM+qQkSocZtf+bahXwKZUty9MkDUr14LUNFh5c=
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/109057] Does GCC interpret assembly when deciding to
 optimize away a variable?
Date: Tue, 07 Mar 2023 17:44:06 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: middle-end
X-Bugzilla-Version: unknown
X-Bugzilla-Keywords: inline-asm
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jakub at gcc dot gnu.org
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Resolution: INVALID
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-109057-4-2cxXwubvEQ@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-109057-4@http.gcc.gnu.org/bugzilla/>
References: <bug-109057-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D109057

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Henry from comment #9)
> Just to make it clear, I'm not saying this is a bug on GCC.=20
>=20
> Im just trying to understand what is happening since this is affecting so=
me
> of our benchmarks. Then we can counter with some wit.=20
>=20
> Perhaps there is an alternate venue for this type of clarification? I tri=
ed
> Reddit but no dice. The GCC IRC channel perhaps?

If you use
inline void DoNotOptimize( unsigned int value) {
  asm volatile("" : : "r,m"(value) : "memory");
}
static const unsigned char LUT[8] =3D {1,5,3,0,2,7,1,2};
void func1(unsigned int val) {
    DoNotOptimize(LUT[val]);=20
}
then obviously it can't choose the "m" variant for value, because value is
32-bit,
while LUT(%rdi) is 8-bit.  So it can choose only "r" variant and therefore =
it
needs
to emit an instruction that computes that (zero extends the value).
If you change LUT array to be const unsigned int LUT[8], then "m" variant c=
an
be selected (and is in my testing).=