NetBSD Problem Report #11195

Received: (qmail 11140 invoked from network); 11 Oct 2000 19:02:46 -0000
Message-Id: <20001011190149.D7ECF1A9@beverly.kleinbus.org>
Date: Wed, 11 Oct 2000 21:01:49 +0200 (MEST)
From: is@beverly.kleinbus.org
Reply-To: is@netbsd.org
To: gnats-bugs@gnats.netbsd.org
Subject: m68k internal compiler error compiling tcl-8.3.2
X-Send-Pr-Version: 3.95

>Number:         11195
>Category:       port-m68k
>Synopsis:       m68k internal compiler error on tcl-8.3.2
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-m68k-maintainer
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Oct 11 19:03:00 +0000 2000
>Closed-Date:    Fri Feb 08 23:09:57 +0000 2002
>Last-Modified:  Fri Feb 08 23:09:57 +0000 2002
>Originator:     Ignatios Souvatzis
>Release:        1.5_ALPHA2
>Organization:
	just a humble port maintainer
>Environment:
System: NetBSD jocelyn 1.5B NetBSD 1.5B (ISDN20) #3: Wed Aug 30 21:28:46 MEST 2000 is@jocelyn:/build2/i4b/ISDN20 amiga

But only the kernel is 1.5B. Base system is 1.5_ALPHA2 as of end-September.
gcc is:

Using builtin specs.
gcc version egcs-2.91.66 19990314 (egcs-1.1.2 release)

>Description:
Compiling tcl as in the pkg system of today results in an internal compiler
error.
>How-To-Repeat:
jocelyn tcl !% cd /usr/pkgsrc/net/tcl; make
===> Building for tcl-8.3.2
/usr/pkg/bin/libtool  --mode=compile cc -c -O -Wall -Wconversion -Wno-implicit-int -fPIC  -I./../generic -I.  -DHAVE_UNISTD_H=1 -DHAVE_LIMITS_H=1 -DHAVE_GETCWD=1 -DHAVE_OPENDIR=1 -DHAVE_STRSTR=1 -DHAVE_STRTOL=1 -DHAVE_TMPNAM=1 -DHAVE_WAITPID=1 -DNO_VALUES_H=1 -DHAVE_UNISTD_H=1 -DHAVE_SYS_PARAM_H=1 -DUSE_TERMIOS=1 -DHAVE_SYS_TIME_H=1 -DTIME_WITH_SYS_TIME=1 -DHAVE_TM_ZONE=1 -DHAVE_TM_GMTOFF=1 -DHAVE_ST_BLKSIZE=1 -DSTDC_HEADERS=1 -DNEED_MATHERR=1 -DHAVE_SIGNED_CHAR=1 -DHAVE_SYS_IOCTL_H=1 -DHAVE_SYS_FILIO_H=1          -DTCL_SHLIB_EXT=\".so\" ./../generic/tclExecute.c
rm -f .libs/tclExecute.lo
cc -c -O -Wall -Wconversion -Wno-implicit-int -fPIC -I./../generic -I. -DHAVE_UNISTD_H=1 -DHAVE_LIMITS_H=1 -DHAVE_GETCWD=1 -DHAVE_OPENDIR=1 -DHAVE_STRSTR=1 -DHAVE_STRTOL=1 -DHAVE_TMPNAM=1 -DHAVE_WAITPID=1 -DNO_VALUES_H=1 -DHAVE_UNISTD_H=1 -DHAVE_SYS_PARAM_H=1 -DUSE_TERMIOS=1 -DHAVE_SYS_TIME_H=1 -DTIME_WITH_SYS_TIME=1 -DHAVE_TM_ZONE=1 -DHAVE_TM_GMTOFF=1 -DHAVE_ST_BLKSIZE=1 -DSTDC_HEADERS=1 -DNEED_MATHERR=1 -DHAVE_SIGNED_CHAR=1 -DHAVE_SYS_IOCTL_H=1 -DHAVE_SYS_FILIO_H=1 -DTCL_SHLIB_EXT=\".so\" ./../generic/tclExecute.c  -fPIC -DPIC -o .libs/tclExecute.lo
./../generic/tclExecute.c: In function `TclExecuteByteCode':
./../generic/tclExecute.c:2920: internal error--unrecognizable insn:
(insn 8723 11842 8725 (parallel[ 
            (set (reg:SI 1477)
                (div:SI (mem:SI (reg:SI 1952))
                    (mem:SI (reg:SI 1961))))
            (set (reg:SI 1478)
                (mod:SI (mem:SI (reg:SI 1954))
                    (mem:SI (reg:SI 1963))))
        ] ) -1 (insn_list 11824 (insn_list 11827 (insn_list 11839 (insn_list 11842 (nil)))))
    (expr_list:REG_DEAD (reg:SI 1954)
        (expr_list:REG_DEAD (reg:SI 1963)
            (expr_list:REG_DEAD (reg:SI 1952)
                (expr_list:REG_DEAD (reg:SI 1961)
                    (expr_list:REG_UNUSED (reg:SI 1478)
                        (nil)))))))
/usr/local/netbsdwork/src/gnu/usr.bin/egcs/common/../../../dist/gcc/toplev.c:1367: Internal compiler error in function fatal_insn
*** Error code 1

Stop.

>Fix:
>Release-Note:
>Audit-Trail:

From: Ignatios Souvatzis <is@beverly.kleinbus.org>
To: gnats-bugs@netbsd.org
Cc:  
Subject: port-m68k/11195
Date: Thu, 12 Oct 2000 00:05:29 +0200

 I found a workaround:

 force the tcl pkg to use -O2 (like all the system build) instead of -O.

 	-is
State-Changed-From-To: open->feedback 
State-Changed-By: fair 
State-Changed-When: Sat Feb 2 13:23:22 PST 2002 
State-Changed-Why:  
Two questions: 

1. NetBSD has updated its toolchain; does this still happen? 

2. If it does still happen, isn't this more properly worked-around 
in the pkgsrc (which would make this a "pkg" category PR)? 

State-Changed-From-To: feedback->closed 
State-Changed-By: scw 
State-Changed-When: Fri Feb 8 23:06:07 GMT 2002 
State-Changed-Why:  

I've done a test-build of the tcl package using new-toolchain on 
mvme68k without problem. 

Looks like it's an `old' toolchain issue only, and therefore can 
be closed. 
>Unformatted:

Home	PR Database Search