ada/7208: build of gnatlib failed with SIGILL

ada/7208: build of gnatlib failed with SIGILL

[adi]

>Number:         7208
>Category:       ada
>Synopsis:       build of gnatlib failed with SIGILL
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jul 04 11:46:01 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     Adrian Knoth
>Release:        3.2 20020704 (experimental)
>Organization:
Thueringen Netz e.V. (www.erfurt.thur.de)
>Environment:
System: Linux drcomp 2.4.18 #13 Sun Mar 3 13:22:15 CET 2002 i586 unknown
Architecture: i586

	
host: i586-pc-linux-gnu
build: i586-pc-linux-gnu
target: i586-pc-linux-gnu
configured with: ../gcc/configure --enable-languages=c,c++,ada --enable-threads --enable-shared
>Description:
	Compiling s-auxdec.adb causes xgcc to receive SIGILL, afterwards
	leaking much memory which will end in a kill with out-of-memory
	(I've interrupted here interactively)

../../xgcc -B../../ -c -g -O2   -g -O2   -W -Wall -gnatpg -I. -I/bigspace/build/gcc/gcc/ada s-auxdec.adb
+===========================GNAT BUG DETECTED==============================+
| 3.2 20020704 (experimental) (i586-pc-linux-gnu) GCC error:               |
| internal error: Illegal instruction                                      |
| No source file position information available                            |
| Please submit bug report by email to gcc-bugs@gcc.gnu.org.               |
| Use a subject line meaningful to you and us to track the bug.            |
| Include the entire contents of this bug box in the report.               |
| Include the exact gcc or gnatmake command that you entered.              |
| Also include sources listed below in gnatchop format                     |
| (concatenated together with no headers between files).                   |
| (use plain ASCII or MIME attachment).                                    |
| See gnatinfo.txt for full info on procedure for submitting bugs.         |
+==========================================================================+

Please include these source files with error report


make[2]: *** [s-auxdec.o] Interrupt

strace -f shows:
[...]
[pid 12067] open("s-stoele.ads", O_RDONLY) = 4
[pid 12067] fstat64(4, {st_mode=S_IFREG|0644, st_size=3810, ...}) = 0
[pid 12067] read(4, "--------------------------------"..., 3810) = 3810
[pid 12067] close(4)                    = 0
[pid 12067] brk(0x8a9c000)              = 0x8a9c000
[pid 12067] --- SIGILL (Illegal instruction) ---



>How-To-Repeat:
	Do "make gnatlib" for this release
>Fix:
	No idea.
>Release-Note:
>Audit-Trail:
>Unformatted:

Re: ada/7208: build of gnatlib failed with SIGILL

[fw]

Synopsis: build of gnatlib failed with SIGILL

State-Changed-From-To: open->closed
State-Changed-By: fw
State-Changed-When: Mon Nov 11 09:51:11 2002
State-Changed-Why:
    closed after request from original submitter, no longer reproducible

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=7208

Re: ada/7208: build of gnatlib failed with SIGILL

[Florian Weimer]

The following reply was made to PR ada/7208; it has been noted by GNATS.

From: Florian Weimer <fw@deneb.enyo.de>
To: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: ada/7208: build of gnatlib failed with SIGILL
Date: Sun, 14 Jul 2002 12:00:05 +0200

 Adrian Knoth <adi@drcomp.erfurt.thur.de> writes:
 
 >> > Yes. I've tried it with the actual cvs co, here is the result, though
 >> > SIGILL changed to segfault:
 >> We need a GDB backtrace of the gnat1 process when it receives SIGILL,
 >> and a disassembly of the function corresponding to the topmost frame.
 >
 > Is it ok for you if I'll send you a statically linked version of my xgcc
 > and the corefile?
 
 The core file does not help much, unless it has been generated by
 SIGILL.
 
 > (to your personal address? I guess the gcc-list doesn't want it)
 
 I can try it on my i586-pc-linux-gnu machine (maybe GCC really
 generates an instruction invalid on i586), but I'd rather like to see
 the backtrace on your machine.
 
 > I've now deleted everything, cvs co will follow and the static
 > built, too.  I guess building without threads is easier for you to
 > debug, isn't it?
 
 It shouldn't matter; the compiler itself does not use the
 multi-tasking run-time library.

Re: ada/7208: build of gnatlib failed with SIGILL

[Adrian Knoth]

The following reply was made to PR ada/7208; it has been noted by GNATS.

From: Adrian Knoth <adi@drcomp.erfurt.thur.de>
To: Florian Weimer <fw@deneb.enyo.de>
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: ada/7208: build of gnatlib failed with SIGILL
Date: Sun, 14 Jul 2002 11:56:20 +0200

 On Sun, Jul 14, 2002 at 10:46:57AM +0200, Florian Weimer wrote:
 
 > > Yes. I've tried it with the actual cvs co, here is the result, though
 > > SIGILL changed to segfault:
 > We need a GDB backtrace of the gnat1 process when it receives SIGILL,
 > and a disassembly of the function corresponding to the topmost frame.
 
 Is it ok for you if I'll send you a statically linked version of my xgcc
 and the corefile? (to your personal address? I guess the gcc-list doesn't
 want it)
 
 I've now deleted everything, cvs co will follow and the static built, too.
 I guess building without threads is easier for you to debug, isn't it?
 
 -- 
 mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver
 
 Ist die Katze schoen rund, freut sich der Hund!

Re: ada/7208: build of gnatlib failed with SIGILL

[Florian Weimer]

The following reply was made to PR ada/7208; it has been noted by GNATS.

From: Florian Weimer <fw@deneb.enyo.de>
To: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: ada/7208: build of gnatlib failed with SIGILL
Date: Sun, 14 Jul 2002 10:46:57 +0200

 Adrian Knoth <adi@drcomp.erfurt.thur.de> writes:
 
 > On Sat, Jul 13, 2002 at 07:00:37PM +0200, Florian Weimer wrote:
 >
 >> > 	Compiling s-auxdec.adb causes xgcc to receive SIGILL, afterwards
 >> > 	leaking much memory which will end in a kill with out-of-memory
 >> > 	(I've interrupted here interactively)
 >> I'm sorry, but I can't reproduce this with the current CVS.  
 >> Do you still see this problem?
 >
 > Yes. I've tried it with the actual cvs co, here is the result, though
 > SIGILL changed to segfault:
 
 We need a GDB backtrace of the gnat1 process when it receives SIGILL,
 and a disassembly of the function corresponding to the topmost frame.
 
 (Maybe you now get a segmentation fault because your VM scenario is
 different, so it still could be the very same problem.)

Re: ada/7208: build of gnatlib failed with SIGILL

[Adrian Knoth]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1645 bytes --]

The following reply was made to PR ada/7208; it has been noted by GNATS.

From: Adrian Knoth <adi@drcomp.erfurt.thur.de>
To: Florian Weimer <fw@deneb.enyo.de>
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: ada/7208: build of gnatlib failed with SIGILL
Date: Sun, 14 Jul 2002 02:38:54 +0200

 On Sat, Jul 13, 2002 at 07:00:37PM +0200, Florian Weimer wrote:
 
 > > 	Compiling s-auxdec.adb causes xgcc to receive SIGILL, afterwards
 > > 	leaking much memory which will end in a kill with out-of-memory
 > > 	(I've interrupted here interactively)
 > I'm sorry, but I can't reproduce this with the current CVS.  
 > Do you still see this problem?
 
 Yes. I've tried it with the actual cvs co, here is the result, though
 SIGILL changed to segfault:
 
 ../../xgcc -B../../ -c -g -O2   -g -O2   -W -Wall -gnatpg -I. -I/bigspace/build/gcc/gcc/ada s-auxdec.adb
 +===========================GNAT BUG DETECTED==============================+
 | 3.2 20020713 (experimental) (i586-pc-linux-gnu) GCC error:               |
 | internal error: Segmentation fault                                       |
 | No source file position information available                            |
 [...]
 xgcc: Internal error: Terminated (program gnat1)
 Please submit a full bug report.
 See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
 make[2]: *** [s-auxdec.o] Error 1
 make[2]: Leaving directory `/bigspace/build/compile/gcc/ada/rts'
 make[1]: *** [gnatlib] Error 2
 make[1]: Leaving directory `/bigspace/build/compile/gcc/ada'
 make: *** [gnatlib] Error 2
 
 -- 
 mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver
 
 Frauen sind Männersache!

Re: ada/7208: build of gnatlib failed with SIGILL

[Florian Weimer]

The following reply was made to PR ada/7208; it has been noted by GNATS.

From: Florian Weimer <fw@deneb.enyo.de>
To: adi@thur.de
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: ada/7208: build of gnatlib failed with SIGILL
Date: Sat, 13 Jul 2002 19:00:37 +0200

 adi@thur.de writes:
 
 > System: Linux drcomp 2.4.18 #13 Sun Mar 3 13:22:15 CET 2002 i586 unknown
 > Architecture: i586
 >
 > 	
 > host: i586-pc-linux-gnu
 > build: i586-pc-linux-gnu
 > target: i586-pc-linux-gnu
 > configured with: ../gcc/configure --enable-languages=c,c++,ada --enable-threads --enable-shared
 >>Description:
 > 	Compiling s-auxdec.adb causes xgcc to receive SIGILL, afterwards
 > 	leaking much memory which will end in a kill with out-of-memory
 > 	(I've interrupted here interactively)
 >
 > ../../xgcc -B../../ -c -g -O2   -g -O2   -W -Wall -gnatpg -I. -I/bigspace/build/gcc/gcc/ada s-auxdec.adb
 > +===========================GNAT BUG DETECTED==============================+
 > | 3.2 20020704 (experimental) (i586-pc-linux-gnu) GCC error:               |
 > | internal error: Illegal instruction                                      |
 
 I'm sorry, but I can't reproduce this with the current CVS.  Do you
 still see this problem?