[Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem

public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed

* [Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem
@ 2004-10-16 14:49 gcc-bugzilla at gcc dot gnu dot org
  2004-10-16 15:01 ` [Bug target/18034] " pinskia at gcc dot gnu dot org
                   ` (3 more replies)
  0 siblings, 4 replies; 5+ messages in thread
From: gcc-bugzilla at gcc dot gnu dot org @ 2004-10-16 14:49 UTC (permalink / raw)
  To: gcc-bugs

NOTE: Defaulting component because reported component no longer exists
The bootstrap process fails during stage 3 when the
stage 2 xgcc is used to compile the root file of the
Ada Standard library, the package Ada itself:

stage2/xgcc -Bstage2/ -B/usr/local2/mips-sgi-irix6.2/bin/ -c -g -O2      -gnatpg -gnata -I- -I. -Iada -I../../srcdir/gcc/ada ../../srcdir/gcc/ada/ada.ads -o ada/ada.o
xgcc: Internal error: Segmentation fault (program gnat1)
Please submit a full bug report.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
make[2]: *** [ada/ada.o] Error 1
make[2]: Leaving directory `/usr/people/franke/install/objdir/gcc'
make[1]: *** [stage3_build] Error 2
make[1]: Leaving directory `/usr/people/franke/install/objdir/gcc'
make: *** [bootstrap] Error 2

A core file is generated:

(glanzer 504) $ file install/objdir/gcc/core 
install/objdir/gcc/core:        IRIX N32 core dump of 'gnat1'
(glanzer 505) $ strings install/objdir/gcc/core | head
gnat1
stage2/gnat1 -I- -I. -Iada -I../../srcdir/gcc/adstage2/gnat1 -I- -I. -Iada -I../../srcdir/gcc/ada -quiet -dumpbase ada.ads -g -

Environment:
System: IRIX64 glanzer 6.2 06101031 IP28


	
host: mips-sgi-irix6.2
build: mips-sgi-irix6.2
target: mips-sgi-irix6.2
configured with: ../srcdir/configure --enable-languages=c,c++,ada --prefix=/usr/local2 --disable-nls --disable-multilib

How-To-Repeat:
Rerun the bootstrap.
------- Additional Comments From franke at euro-telematik dot de  2004-10-16 14:49 -------
Fix:
Fix unknown, but behaviour observed in 3.4.0 and 3.4.1, too.

-- 
           Summary: GCC 3.4.2/GNAT bootstrap problem
           Product: gcc
           Version: 3.4.2
            Status: UNCONFIRMED
          Severity: critical
          Priority: P2
         Component: pending
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: franke at euro-telematik dot de
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: mips-sgi-irix6.2
  GCC host triplet: mips-sgi-irix6.2
GCC target triplet: mips-sgi-irix6.2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18034


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug target/18034] GCC 3.4.2/GNAT bootstrap problem
  2004-10-16 14:49 [Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem gcc-bugzilla at gcc dot gnu dot org
@ 2004-10-16 15:01 ` pinskia at gcc dot gnu dot org
  2004-10-26 19:02 ` rsandifo at gcc dot gnu dot org
                   ` (2 subsequent siblings)
  3 siblings, 0 replies; 5+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-10-16 15:01 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-16 15:01 -------
Try looking into PR 6552 and PR 6669.  Those two might tell you what the problem is.
It might be stage1 is being miscompile which miscompiles stage2 which causes this problem.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|critical                    |normal
          Component|pending                     |target
           Keywords|                            |build, wrong-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18034


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug target/18034] GCC 3.4.2/GNAT bootstrap problem
  2004-10-16 14:49 [Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem gcc-bugzilla at gcc dot gnu dot org
  2004-10-16 15:01 ` [Bug target/18034] " pinskia at gcc dot gnu dot org
@ 2004-10-26 19:02 ` rsandifo at gcc dot gnu dot org
  2004-11-27 17:25 ` franke at euro-telematik dot de
  2005-04-05 21:35 ` echristo at redhat dot com
  3 siblings, 0 replies; 5+ messages in thread
From: rsandifo at gcc dot gnu dot org @ 2004-10-26 19:02 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From rsandifo at gcc dot gnu dot org  2004-10-26 19:02 -------
Andrew's explanation certainly sounds plausible.  Could you
try bootstrapping using Rainer Orth's GNAT binaries:

   ftp://ftp.Uni-Bielefeld.DE/pub/gnu/gnat/binaries/

I've had no problems bootstrapping 3.4 on our IRIX 6.5
when using these binaries as the starting point.

Richard


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18034


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug target/18034] GCC 3.4.2/GNAT bootstrap problem
  2004-10-16 14:49 [Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem gcc-bugzilla at gcc dot gnu dot org
  2004-10-16 15:01 ` [Bug target/18034] " pinskia at gcc dot gnu dot org
  2004-10-26 19:02 ` rsandifo at gcc dot gnu dot org
@ 2004-11-27 17:25 ` franke at euro-telematik dot de
  2005-04-05 21:35 ` echristo at redhat dot com
  3 siblings, 0 replies; 5+ messages in thread
From: franke at euro-telematik dot de @ 2004-11-27 17:25 UTC (permalink / raw)
  To: gcc-bugs

------- Additional Comments From franke at euro-telematik dot de  2004-11-27 17:25 -------
Subject: Re:  GCC 3.4.2/GNAT bootstrap problem

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> Try  looking into PR 6552 and PR 6669.  Those two might tell you
> what the problem is.  It might be  stage1  is  being  miscompile
> which miscompiles stage2 which causes this problem.

I  did, but it brought no alleviation.  Removing -g from various
CFLAGS, as proposed in PR  6669,  doesn't  change  the  outcome,
while removing -O2 (PR 6552) makes it even worse, as some of the
little helper utilities that get built along the  way  (gentype,
genmode, and friends) produce a bus error when executed.

> Andrew's explanation certainly sounds plausible.  Could you
> try bootstrapping using Rainer Orth's GNAT binaries:

>    ftp://ftp.Uni-Bielefeld.DE/pub/gnu/gnat/binaries/

> I've had no problems bootstrapping 3.4 on our IRIX 6.5
> when using these binaries as the starting point.

Tried that, too.  No betterment.  And I only have IRIX 6.2 at my
disposal.

   I  have, meanwhile, upgraded to GCC 3.4.3, but it didn't make
a difference.  Out of sheer desparation, I have adopted a brute-
force  method  to  get  rid  of  the symptoms (not the problems,
though!).  See the transcript below.  I have finally  been  able
to  produce  a GCC 3.4.3 but I am all too aware of the fact that
it is a botched one.  Not surprisingly, it could not be used  to
bootstrap itself.

   Any ideas are appreciated.

Regards,

Frank

- ---- Transcript of the brute-force session -----

gcc says> stage2/xgcc -Bstage2/ -B/usr/local2/mips-sgi-irix6.2/bin/ -c -g -O2      -gnatpg -gnata -I- -I. -Iada -I../../srcdir/gcc/ada ../../srcdir/gcc/ada/ada.ads -o ada/ada.o
gcc says> xgcc: Internal error: Segmentation fault (program gnat1)
gcc says> Please submit a full bug report.
gcc says> See <URL:http://gcc.gnu.org/bugs.htmlgcc-says> for instructions.
gcc says> make[2]: *** [ada/ada.o] Error 1
gcc says> make[2]: Leaving directory `/usr/people/franke/install/objdir/gcc'
gcc says> make[1]: *** [stage3_build] Error 2
gcc says> make[1]: Leaving directory `/usr/people/franke/install/objdir/gcc'
gcc says> make: *** [bootstrap] Error 2

Used gnat1 from stage1.

gcc says> Stage comparison failed.

Replaced the call to cmp with a call to /bin/true in
objdir/gcc/Makefile.

gcc says> /usr/people/franke/install/objdir/gcc/xgcc -B/usr/people/franke/install/objdir/gcc/ -B/usr/local2/mips-sgi-irix6.2/bin/ -B/usr/local2/mips-sgi-irix6.2/lib/ -isystem /usr/local2/mips-sgi-irix6.2/include -DHAVE_CONFIG_H -I. -I../../../../srcdir/libstdc++-v3/libmath -I.. -O2 -g -O2 -c ../../../../srcdir/libstdc++-v3/libmath/stubs.c   -DPIC -o .libs/stubs.o
gcc says> ../../../../srcdir/libstdc++-v3/libmath/stubs.c:268: error: conflicting types for `hypotl'
gcc says> /usr/people/franke/install/objdir/gcc/include/math.h:506: error: previous declaration of `hypotl'
gcc says> make[4]: *** [stubs.lo] Error 1
gcc says> make[4]: Leaving directory `/usr/people/franke/install/objdir/mips-sgi-irix6.2/libstdc++-v3/libmath'
gcc says> make[3]: *** [all-recursive] Error 1
gcc says> make[3]: Leaving directory `/usr/people/franke/install/objdir/mips-sgi-irix6.2/libstdc++-v3'
gcc says> make[2]: *** [all-recursive-am] Error 2
gcc says> make[2]: Leaving directory `/usr/people/franke/install/objdir/mips-sgi-irix6.2/libstdc++-v3'
gcc says> make[1]: *** [all-target-libstdc++-v3] Error 2
gcc says> make[1]: Leaving directory `/usr/people/franke/install/objdir'
gcc says> make: *** [bootstrap] Error 2

Commented out srcdir/libstdc++-v3/libmath/stubs.c:266--276

gcc says> ../../xgcc -B../../ -c -g -O2      -W -Wall -gnatpg  a-caldel.adb -o a-caldel.o
gcc says> xgcc: Internal error: Segmentation fault (program gnat1)
gcc says> Please submit a full bug report.
gcc says> See <URL:http://gcc.gnu.org/bugs.htmlgcc-says> for instructions.
gcc says> make[2]: *** [a-caldel.o] Error 1
gcc says> make[2]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada/rts'
gcc says> make[1]: *** [gnatlib] Error 2
gcc says> make[1]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada'
gcc says> make: *** [gnatlib] Error 2

Used gnat1 from stage1.

gcc says> ../../xgcc -B../../ -c -g -O2      -W -Wall -gnatpg  a-llftio.ads -o a-llftio.o
gcc says> a-tiflio.adb: In function `Ada.Long_Long_Float_Text_Io.Get':
gcc says> a-tiflio.adb:44: error: unrecognizable insn:
gcc says> (insn/i:HI 185 184 186 16 s-fatgen.adb:840 (set (subreg:DI (reg:TF 245) 0)
gcc says>         (mem/u/f:TF (symbol_ref/u:SI ("*.LC4") [flags 0x2]) [25 S16 A128])) -1 (insn_list 184 (nil))
gcc says>     (nil))
gcc says> +===========================GNAT BUG DETECTED==============================+
gcc says> | 3.4.3 (mips-sgi-irix6.2) in extract_insn, at recog.c:2083                |
gcc says> | Error detected at a-tiflio.adb:65:8 [a-llftio.ads:18:1]                  |
gcc says> | Please submit a bug report; see http://gcc.gnu.org/bugs.html.            |
gcc says> | Include the entire contents of this bug box in the report.               |
gcc says> | Include the exact gcc or gnatmake command that you entered.              |
gcc says> | Also include sources listed below in gnatchop format                     |
gcc says> | (concatenated together with no headers between files).                   |
gcc says> +==========================================================================+
gcc says>
gcc says> Please include these source files with error report
gcc says> Note that list may not be accurate in some cases,
gcc says> so please double check that the problem can still
gcc says> be reproduced with the set of files listed.
gcc says>
gcc says>
gcc says> compilation abandoned
gcc says> make[2]: *** [a-llftio.o] Error 1
gcc says> make[2]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada/rts'
gcc says> make[1]: *** [gnatlib] Error 2
gcc says> make[1]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada'
gcc says> make: *** [gnatlib] Error 2

Temporarily (i.e., for a-llftio.ads only) made s-fatgen.adb:840
return True.

gcc says> ../../xgcc -B../../ -c -g -O2      -W -Wall -gnatpg  s-fatllf.ads -o s-fatllf.o
gcc says> s-fatgen.adb: In function `System.Fat_Llf.Fat_Long_Long_Float.Unaligned_Valid':
gcc says> s-fatgen.adb:848: error: unrecognizable insn:
gcc says> (insn/i:HI 104 103 105 4 s-fatgen.adb:840 (set (subreg:DI (reg:TF 223) 0)
gcc says>         (mem/u/f:TF (symbol_ref/u:SI ("*.LC21") [flags 0x2]) [8 S16 A128])) -1 (insn_list 103 (nil))
gcc says>     (nil))
gcc says> +===========================GNAT BUG DETECTED==============================+
gcc says> | 3.4.3 (mips-sgi-irix6.2) in extract_insn, at recog.c:2083                |
gcc says> | Error detected at s-fatgen.adb:861:1 [s-fatllf.ads:47:4]                 |
gcc says> | Please submit a bug report; see http://gcc.gnu.org/bugs.html.            |
gcc says> | Include the entire contents of this bug box in the report.               |
gcc says> | Include the exact gcc or gnatmake command that you entered.              |
gcc says> | Also include sources listed below in gnatchop format                     |
gcc says> | (concatenated together with no headers between files).                   |
gcc says> +==========================================================================+
gcc says>
gcc says> Please include these source files with error report
gcc says> Note that list may not be accurate in some cases,
gcc says> so please double check that the problem can still
gcc says> be reproduced with the set of files listed.
gcc says>
gcc says>
gcc says>
gcc says> raised TYPES.UNRECOVERABLE_ERROR : comperr.adb:376
gcc says> make[2]: *** [s-fatllf.o] Error 1
gcc says> make[2]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada/rts'
gcc says> make[1]: *** [gnatlib] Error 2
gcc says> make[1]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada'
gcc says> make: *** [gnatlib] Error 2

Temporarily (i.e., for s-fatllf.ads only) made s-fatgen.adb:840
return True.

gcc says> ../../xgcc -B../../ -c -g -O2      -W -Wall -gnatpg  s-imgrea.adb -o s-imgrea.o
gcc says> s-imgrea.adb: In function `System.Img_Real.Set_Image_Real':
gcc says> s-imgrea.adb:136: error: unrecognizable insn:
gcc says> (insn/i:HI 165 164 166 6 s-fatgen.adb:840 (set (subreg:DI (reg:TF 262) 0)
gcc says>         (mem/u/f:TF (symbol_ref/u:SI ("*.LC9") [flags 0x2]) [7 S16 A128])) -1 (insn_list 164 (nil))
gcc says>     (nil))
gcc says> +===========================GNAT BUG DETECTED==============================+
gcc says> | 3.4.3 (mips-sgi-irix6.2) in extract_insn, at recog.c:2083                |
gcc says> | Error detected at s-imgrea.adb:700:8                                     |
gcc says> | Please submit a bug report; see http://gcc.gnu.org/bugs.html.            |
gcc says> | Include the entire contents of this bug box in the report.               |
gcc says> | Include the exact gcc or gnatmake command that you entered.              |
gcc says> | Also include sources listed below in gnatchop format                     |
gcc says> | (concatenated together with no headers between files).                   |
gcc says> +==========================================================================+
gcc says>
gcc says> Please include these source files with error report
gcc says> Note that list may not be accurate in some cases,
gcc says> so please double check that the problem can still
gcc says> be reproduced with the set of files listed.
gcc says>
gcc says>
gcc says> compilation abandoned
gcc says> make[2]: *** [s-imgrea.o] Error 1
gcc says> make[2]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada/rts'
gcc says> make[1]: *** [gnatlib] Error 2
gcc says> make[1]: Leaving directory `/usr/people/franke/install/objdir/gcc/ada'
gcc says> make: *** [gnatlib] Error 2

Temporarily (i.e., for s-imgrea.adb only) made s-fatgen.adb:840
return True.

- ---> BUILT! <---

(glanzer 512) $ gnatmake -v hello.adb -cargs -v -bargs -v -largs -v

GNATMAKE 3.4.3 Copyright 1995-2004 Free Software Foundation, Inc.
  "hello.ali" being checked ...
  -> "hello.ali" missing.
gcc -c -v hello.adb
Reading specs from /usr/local2/lib/gcc/mips-sgi-irix6.2/3.4.3/specs
Configured with: ../srcdir/configure --enable-languages=c,c++,ada --prefix=/usr/local2 --disable-nls --disable-multilib --enable-threads=gnat
Thread model: gnat
gcc version 3.4.3
 /usr/local2/libexec/gcc/mips-sgi-irix6.2/3.4.3/gnat1 -quiet -dumpbase hello.adb hello.adb -o /var/tmp//ccF3aH5a.s
 /usr/bin/as -EB -O0 -g0 -nocpp -show -G 0 -w -n32 -mips3 -o hello.o /var/tmp//ccF3aH5a.s
/usr/bin/../lib32/cmplrs/as -EB -O0 -g0 -nocpp -show -G 0 -w -n32 -mips3 -o hello.o /var/tmp//ccF3aH5a.s
/usr/lib32/cmplrs/asm -pic2 -elf -EB -O0 -g0 -G0 -w -mips3 -n32 /var/tmp//ccF3aH5a.s -o hello.o
End of compilation
gnatbind -aO./ -v -I- -x hello.ali

GNATBIND 3.4.3 Copyright 1995-2004 Free Software Foundation, Inc.

Binding: hello.ali

No errors
gnatlink hello.ali -v

GNATLINK 3.4.3 Copyright 1995-2004 Free Software Foundation, Inc
gcc -c -gnatA -gnatWb -gnatiw -gnatws b~hello.adb
/usr/local2/bin/gcc b~hello.o ./hello.o -o hello -L./ -L/usr/local2/lib/gcc/mips-sgi-irix6.2/3.4.3/adalib/ /usr/local2/lib/gcc/mips-sgi-irix6.2/3.4.3/adalib/libgnat.a -lexc
(glanzer 513) $

- ---- End of transcript of the brute-force session -----

- --

Frank Ecke <franke@euro-telematik.de>, see pgpkeys.mit.edu:11371
Fingerprint = 2CCD 7D50 E07E 557E EA2F  7454 068C B20A 20FF C83D
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (IRIX64)

iD8DBQFBqLboBoyyCiD/yD0RAnOvAJ9aAECxQNxzBDYg6DV8PZafZoJyrACg1CIY
4I4xLMDOSK+aK8BZ9jZqn9Y=
=c0bh
-----END PGP SIGNATURE-----

-- 

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18034

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug target/18034] GCC 3.4.2/GNAT bootstrap problem
  2004-10-16 14:49 [Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem gcc-bugzilla at gcc dot gnu dot org
                   ` (2 preceding siblings ...)
  2004-11-27 17:25 ` franke at euro-telematik dot de
@ 2005-04-05 21:35 ` echristo at redhat dot com
  3 siblings, 0 replies; 5+ messages in thread
From: echristo at redhat dot com @ 2005-04-05 21:35 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From echristo at redhat dot com  2005-04-05 21:35 -------
Since I don't have a 6.2 box (and neither does Richard afaik), I'm not certain
there's anything we can do about this. You'll either need to debug the
differences on your platform between this and 6.5 or something else. Sorry.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |WORKSFORME


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18034


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2005-04-05 21:35 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-10-16 14:49 [Bug pending/18034] New: GCC 3.4.2/GNAT bootstrap problem gcc-bugzilla at gcc dot gnu dot org
2004-10-16 15:01 ` [Bug target/18034] " pinskia at gcc dot gnu dot org
2004-10-26 19:02 ` rsandifo at gcc dot gnu dot org
2004-11-27 17:25 ` franke at euro-telematik dot de
2005-04-05 21:35 ` echristo at redhat dot com

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).