From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugs-return-432205-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 13645 invoked by alias); 19 Oct 2013 11:01:14 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Received: (qmail 13624 invoked by uid 48); 19 Oct 2013 11:01:10 -0000
From: "Jeremie.Detrey at loria dot fr" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug inline-asm/58805] New: Inline assembly wrongly optimized out when inside a conditional
Date: Sat, 19 Oct 2013 11:01:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: inline-asm
X-Bugzilla-Version: 4.8.1
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: Jeremie.Detrey at loria dot fr
X-Bugzilla-Status: UNCONFIRMED
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: bug_id short_desc product version bug_status bug_severity priority component assigned_to reporter attachments.created
Message-ID: <bug-58805-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2013-10/txt/msg01349.txt.bz2

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58805

            Bug ID: 58805
           Summary: Inline assembly wrongly optimized out when inside a
                    conditional
           Product: gcc
           Version: 4.8.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: inline-asm
          Assignee: unassigned at gcc dot gnu.org
          Reporter: Jeremie.Detrey at loria dot fr

Created attachment 31050
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31050&action=edit
Minimal example to reproduce the bug.

When compiling the attached example with GCC 4.8.1 (on an x86_64 ArchLinux)
using -O2 optimization, the `else' branch of the conditional in foo() is
optimized out: GCC only generates code for the first branch, and does not test
whether the value of `n' is zero or not. This can be seen in the following
assembly output:

$ gcc -O2 -S bug.c -o-
        .file   "bug.c"
        .text
        .p2align 4,,15
        .globl  foo
        .type   foo, @function
foo:
.LFB1:
        .cfi_startproc
#APP
# 4 "bug.c" 1
        movq $42, %rdx
        movq %rdx, %rax

# 0 "" 2
#NO_APP
        movq    %rax, (%rsi)
        ret
        .cfi_endproc
.LFE1:
        .size   foo, .-foo
        .ident  "GCC: (GNU) 4.8.1"
        .section        .note.GNU-stack,"",@progbits

This bug does not appear when compiling with GCC 4.7.3 (on an x86_64 Ubuntu
13.04, also with -O2 optimization): the two branches are generated, and the
test on `n' is performed to select which branch to jump to:

$ gcc -O2 -S bug.c -o-
        .file   "bug.c"
        .text
        .p2align 4,,15
        .globl  foo
        .type   foo, @function
foo:
.LFB1:
        .cfi_startproc
        testl   %edi, %edi
        je      .L5
#APP
# 4 "bug.c" 1
        movq $42, %rcx
        movq %rcx, %rax

# 0 "" 2
#NO_APP
        movq    %rax, (%rdx)
        ret
        .p2align 4,,10
        .p2align 3
.L5:
#APP
# 4 "bug.c" 1
        movq $42, %rdx
        movq %rdx, %rax

# 0 "" 2
#NO_APP
        movq    %rax, (%rsi)
        ret
        .cfi_endproc
.LFE1:
        .size   foo, .-foo
        .ident  "GCC: (Ubuntu/Linaro 4.7.3-1ubuntu1) 4.7.3"
        .section        .note.GNU-stack,"",@progbits

This bug disappears (in GCC 4.8.1) when either declaring the inline asm block
as `volatile', or removing the intermediate variable `t' and assigning 42
directly to `%[r]'.