From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-prs-return-23433-listarch-gcc-prs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 8542 invoked by alias); 25 Oct 2002 21:56:01 -0000
Mailing-List: contact gcc-prs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc-prs/>
List-Post: <mailto:gcc-prs@gcc.gnu.org>
List-Help: <mailto:gcc-prs-help@gcc.gnu.org>
Sender: gcc-prs-owner@gcc.gnu.org
Received: (qmail 8505 invoked by uid 71); 25 Oct 2002 21:56:01 -0000
Resent-Date: 25 Oct 2002 21:56:01 -0000
Resent-Message-ID: <20021025215601.8503.qmail@sources.redhat.com>
Resent-From: gcc-gnats@gcc.gnu.org (GNATS Filer)
Resent-Cc: gcc-prs@gcc.gnu.org, gcc-bugs@gcc.gnu.org
Resent-Reply-To: gcc-gnats@gcc.gnu.org, dje@watson.ibm.com
Received: (qmail 7427 invoked by uid 61); 25 Oct 2002 21:48:11 -0000
Message-Id: <20021025214811.7426.qmail@sources.redhat.com>
Date: Fri, 25 Oct 2002 14:56:00 -0000
From: dje@watson.ibm.com
Reply-To: dje@watson.ibm.com
To: gcc-gnats@gcc.gnu.org
X-Send-Pr-Version: gnatsweb-2.9.3 (1.1.1.1.2.31)
Subject: bootstrap/8362: hard reg not substituted for all pseudos on PowerPC
X-SW-Source: 2002-10/txt/msg01039.txt.bz2
List-Id: <gcc-prs.sourceware.org>


>Number:         8362
>Category:       bootstrap
>Synopsis:       hard reg not substituted for all pseudos on PowerPC
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Oct 25 14:56:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     dje@watson.ibm.com
>Release:        gcc version 3.2.1 20021024 (prerelease)
>Organization:
>Environment:
powerpc-ibm-aix.4.3.3.0
>Description:
AIX bootstrap on the 3.2 branch ICEs while building
libstdc++-v3 abi_check.cc with -mcpu=power option.  The
failure turns out to be more general and more serious
because I can reproduce it with a PowerPC compiler using
the string instructions, not POWER-specific instructions.
The ICE is in reload1.c:reload_combine_note_use() testing
that a REG is a hard register and not still a pseudo.  greg
dump file shows:

Reloads for insn # 1856
Reload 0: reload_in (SI) = (plus:SI (reg/f:SI 1 r1)
                                                    (const_int 520 [0x208]))
        BASE_REGS, RELOAD_FOR_INPUT (opnum = 2)
        reload_in_reg: (plus:SI (reg/f:SI 1 r1)
                                                    (const_int 520 [0x208]))
        reload_reg_rtx: (reg:SI 11 r11)

(insn 16488 1853 1856 (set (reg:SI 11 r11)
        (plus:SI (reg/f:SI 1 r1)
            (const_int 520 [0x208]))) 36 {*addsi3_internal1} (nil)
    (nil))

(insn 1856 16488 16458 (parallel[ 
            (set (reg:SI 4 r4)
                (mem/s:SI (reg:SI 11 r11) [0 S4 A128]))
            (set (reg:SI 5 r5)
                (mem/s:SI (plus:SI (reg/f:SI 929)
                        (const_int 4 [0x4])) [0 S4 A32]))
            (set (reg:SI 6 r6)
                (mem/s:SI (plus:SI (reg/f:SI 929)
                        (const_int 8 [0x8])) [0 S4 A64]))
            (set (reg:SI 7 r7)
                (mem/s:SI (plus:SI (reg/f:SI 929)
                        (const_int 12 [0xc])) [0 S4 A32]))
        ] ) 322 {*rs6000.md:8737} (insn_list:REG_DEP_OUTPUT 1843 (insn_list:REG_
DEP_OUTPUT 1846 (insn_list 1842 (insn_list:REG_DEP_ANTI 1849 (nil)))))
    (nil))

r11 only was substituted for the pseudo in one location, not
the multiple locations in the instruction.
>How-To-Repeat:
cd powerpc-ibm-aix4.3.3.0/libstdc++-v3/testsuite
gcc -mstring -O2 -c abi_check.cc
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: