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 559E03858002; Fri, 31 Mar 2023 14:38:19 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 559E03858002
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1680273499;
	bh=O9Dc5DuuTtWBjTcwAJ2PJnGvGnhZ/AU27OEakevHyAQ=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=gKACk0HG/fuSca7XAVG5/WduqDSAfo2/iT5ylFNEAre/iDSiYJj5/Ba3Fwe4nqZLZ
	 OBKEMz1KAwGsOEtTMNV7sBvLB6XPQz+QUMh60H6bHYIUtM8k+dZFuHWYZgmr3Y0IYy
	 ixSVyNbWl0NbMuejMBY5yu0b36BIey70gGAXOfvw=
From: "barry.revzin at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/109362] codegen adds unnecessary extra add when reading
 atomic member
Date: Fri, 31 Mar 2023 14:38:19 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 12.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: barry.revzin at gmail dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Resolution: 
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-109362-4-poyOC9R9us@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-109362-4@http.gcc.gnu.org/bugzilla/>
References: <bug-109362-4@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=3D109362

--- Comment #1 from Barry Revzin <barry.revzin at gmail dot com> ---
Sorry, in this reduced example, it doesn't actually consume an extra regist=
er -
only rdi is used.=20

In this slightly less reduced example:

#include <atomic>

struct S {
    std::atomic<long> size;
    std::atomic<char*> read_ptr;

    auto peek() const -> const char* {=20=20=20=20
        auto const s =3D size.load(std::memory_order_acquire);
        return read_ptr.load(std::memory_order_acquire);
    }
};

auto with_atomic(S const& v) {
    while (true) {
        if (v.peek()) {
            return true;
        }
    }
}


the codegen is:

with_atomic(WithAtomic const&):
        lea     rdx, [rdi+8]
.L2:
        mov     rax, QWORD PTR [rdi]
        mov     rax, QWORD PTR [rdx]
        test    rax, rax
        je      .L2
        mov     eax, 1
        ret

which now does consume rdx unnecessarily.=