Floating point exception ICE on mips and mipsel
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
gcc-4.0 (Debian) |
Fix Released
|
Unknown
|
|||
gcc-4.0 (Ubuntu) |
Invalid
|
High
|
Matthias Klose |
Bug Description
Automatically imported from Debian bug report #336463 http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#1 |
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#2 |
reassign 336463 g++-4.0
severity 336463 important
quit
* Matthias Klose
> well, check it before uploading, if you think it's critical. please
> ask the debian admins to install what's needed on the developer
> machines. please reassign after providing the complete information.
Hello again Matthias. I reproduced the problem in Vaughan's change
root and am attaching the preprocessed source file that triggers the
bug. It seems it only happens if I'm using optimization, so I'm
lowering the severity as I can probably work around it by not disabling
optimization on those two architectures (I doubt there's many users
of ScummVM on mips[el] in any case). But of course, it would be more
preferable to have this bug fixed properly in g++.
Anyway, to reproduce the ICE, use a mips machine with g++ 4.0.2-3
installed (such as Vaughan's change root), and run "g++ -O -o /dev/null
thumbnail.ii". If I remove the "-O", the ICE is gone. I did test with
g++ 3.4 too, by the way, and it had no problems compiling the file with
optimization enabled.
Do you have sufficient information from me now?
Thanks
--
Tore Anderson
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#3 |
kismet has a similar problem.
Preprocessed source attached.
djpig@vaughan:
Using built-in specs.
Target: mipsel-linux-gnu
Configured with: ../src/configure -v
--enable-
--enable-shared --with-system-zlib --libexecdir=
--without-
--program-
--enable-
--enable-
--enable-
Thread model: posix
gcc version 4.0.3 20051023 (prerelease) (Debian 4.0.2-3)
Doesn't appear with -01 or g++-3.4. I've asked debian-admin to install
gcc-snapshot on the machine, too and will give the results once
available.
Gruesse,
--
Frank Lichtenheld <email address hidden>
www: http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#4 |
Package: g++-4.0
Version: 4.0.2-3
Followup-For: Bug #336463
My latest upload of fltk1.1 (1.1.6-9) runs into identical lossage on
mips and mipsel, on sources that did not change since before my
previous upload (which built fine), so G++ definitely appears to be
the culprit. Logs:
http://
http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#5 |
Hi,
Very likely, this is the same as PR 24514, which has a patch here:
http://
--
Falk
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#6 |
reassign 336463 gcc-4.0 4.0.2-3
severity 336463 grave
merge 336463 336167
tags 336167 +fixed-upstream
thanks
Aaron M. Ucko wrote:
> Package: g++-4.0
> Version: 4.0.2-3
> Followup-For: Bug #336463
>
> My latest upload of fltk1.1 (1.1.6-9) runs into identical lossage on
> mips and mipsel, on sources that did not change since before my
> previous upload (which built fine), so G++ definitely appears to be
> the culprit. Logs:
>
> http://
> http://
It's around a dozen packages now, and affects gcc as well as g++.
#336167 has a patch which fixes it.
(According to Sven Luther the patch fixes also the powerpc kernel build
mentioned at the start of #336167.)
Thiemo
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #7 |
Automatically imported from Debian bug report #336463 http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #8 |
Message-Id: <1130679087.
Date: Sun, 30 Oct 2005 14:31:27 +0100
From: Tore Anderson <email address hidden>
To: <email address hidden>
Subject: Floating point exception ICE on mips and mipsel
Package: g++-4.0
Severity: critical
Version: 4.0.2-3
Hi. I just uploaded ScummVM 0.8.0, which fails to build from source
due to an ICE:
scumm/
scumm/
You can find the buildd logs here:
<http://
<http://
I'm afraid this is all the debug information I can supply, as I don't
own any mips[el] machines. The sources that makes it fail should be
available in the archive shortly, though.
Kind regards
--
Tore Anderson
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #9 |
Message-ID: <email address hidden>
Date: Sun, 30 Oct 2005 17:39:26 +0100
From: Matthias Klose <email address hidden>
To: Tore Anderson <email address hidden>, <email address hidden>
CC: <email address hidden>
Subject: Re: Bug#336463: Floating point exception ICE on mips and mipsel
reassign 336463 scummvm
thanks
well, check it before uploading, if you think it's critical. please
ask the debian admins to install what's needed on the developer
machines. please reassign after providing the complete information.
Matthias
Tore Anderson writes:
>
> Package: g++-4.0
> Severity: critical
> Version: 4.0.2-3
>
> Hi. I just uploaded ScummVM 0.8.0, which fails to build from source
> due to an ICE:
>
> scumm/thumbnail
> scumm/thumbnail
>
> You can find the buildd logs here:
> <http://
> <http://
>
> I'm afraid this is all the debug information I can supply, as I don't
> own any mips[el] machines. The sources that makes it fail should be
> available in the archive shortly, though.
>
> Kind regards
> --
> Tore Anderson
>
>
>
> --
> To UNSUBSCRIBE, email to <email address hidden>
> with a subject of "unsubscribe". Trouble? Contact <email address hidden>
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #10 |
Message-Id: <email address hidden>
Date: Sat, 05 Nov 2005 14:39:50 -0500
From: "Aaron M. Ucko" <email address hidden>
To: Debian Bug Tracking System <email address hidden>
Subject: g++-4.0: FPE ICE on mipsen also affects fltk1.1
Package: g++-4.0
Version: 4.0.2-3
Followup-For: Bug #336463
My latest upload of fltk1.1 (1.1.6-9) runs into identical lossage on
mips and mipsel, on sources that did not change since before my
previous upload (which built fine), so G++ definitely appears to be
the culprit. Logs:
http://
http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #11 |
Message-ID: <email address hidden>
Date: Sun, 06 Nov 2005 02:09:21 +0100
From: Falk Hueffner <email address hidden>
To: <email address hidden>
Subject: Bug#336463: g++-4.0: FPE ICE on mipsen also affects fltk1.1
Hi,
Very likely, this is the same as PR 24514, which has a patch here:
http://
--
Falk
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #12 |
Message-ID: <email address hidden>
Date: Sun, 6 Nov 2005 10:39:20 +0100
From: Thiemo Seufer <email address hidden>
To: "Aaron M. Ucko" <email address hidden>, <email address hidden>
Subject: Re: Bug#336463: g++-4.0: FPE ICE on mipsen also affects fltk1.1
reassign 336463 gcc-4.0 4.0.2-3
severity 336463 grave
merge 336463 336167
tags 336167 +fixed-upstream
thanks
Aaron M. Ucko wrote:
> Package: g++-4.0
> Version: 4.0.2-3
> Followup-For: Bug #336463
>
> My latest upload of fltk1.1 (1.1.6-9) runs into identical lossage on
> mips and mipsel, on sources that did not change since before my
> previous upload (which built fine), so G++ definitely appears to be
> the culprit. Logs:
>
> http://
> http://
It's around a dozen packages now, and affects gcc as well as g++.
#336167 has a patch which fixes it.
(According to Sven Luther the patch fixes also the powerpc kernel build
mentioned at the start of #336167.)
Thiemo
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Dennis Kaarsemaker (dennis) wrote : | #13 |
-EMIPSNOTSUPPORTED
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#14 |
block 337864 with 336167
thanks
Aníbal Monsalve Salazar
--
.''`. Debian GNU/Linux
: :' : Free Operating System
`. `' http://
`- http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #15 |
Message-ID: <email address hidden>
Date: Mon, 7 Nov 2005 16:31:18 +1100
From: =?iso-8859-
To: <email address hidden>
Subject: ncpfs: block 337864 with 336167
--mCFSFPgrfLsDD2US
Content-Type: text/plain; charset=iso-8859-1
Content-
Content-
block 337864 with 336167
thanks
An=EDbal Monsalve Salazar
--
.''`. Debian GNU/Linux
: :' : Free Operating System
`. `' http://
`- http://
--mCFSFPgrfLsDD2US
Content-Type: application/
Content-
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iD8DBQFDbuamgY5
s/hM6yOiAoUf+
=hNEu
-----END PGP SIGNATURE-----
--mCFSFPgrfLsDD
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#16 |
tags 336463 + pending
tags 330857 + pending
tags 325050 + pending
thanks
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #17 |
Message-ID: <email address hidden>
Date: Mon, 7 Nov 2005 19:38:23 +0100
From: Matthias Klose <email address hidden>
To: <email address hidden>
Subject: tagging gcc reports
tags 336463 + pending
tags 330857 + pending
tags 325050 + pending
thanks
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#18 |
block 338062 with 336167
thanks
Aníbal Monsalve Salazar
--
.''`. Debian GNU/Linux
: :' : Free Operating System
`. `' http://
`- http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #19 |
Message-ID: <email address hidden>
Date: Tue, 8 Nov 2005 12:42:32 +1100
From: =?iso-8859-
To: <email address hidden>
Subject: hercules: block 338062 with 336167
--c/LjENROlAtlNO85
Content-Type: text/plain; charset=iso-8859-1
Content-
Content-
block 338062 with 336167
thanks
An=EDbal Monsalve Salazar
--
.''`. Debian GNU/Linux
: :' : Free Operating System
`. `' http://
`- http://
--c/LjENROlAtlNO85
Content-Type: application/
Content-
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iD8DBQFDcAKIgY5
sF6WnfVrCq4Bmv5
=I16B
-----END PGP SIGNATURE-----
--c/LjENROlAtlN
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#20 |
block 338220 with 336167
thanks
Aníbal Monsalve Salazar
--
.''`. Debian GNU/Linux
: :' : Free Operating System
`. `' http://
`- http://
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #21 |
Message-ID: <email address hidden>
Date: Wed, 9 Nov 2005 09:07:00 +1100
From: =?iso-8859-
To: <email address hidden>
Subject: asterisk: block 338220 with 336167
--4mN1FYI7gnXysABs
Content-Type: text/plain; charset=iso-8859-1
Content-
Content-
block 338220 with 336167
thanks
An=EDbal Monsalve Salazar
--
.''`. Debian GNU/Linux
: :' : Free Operating System
`. `' http://
`- http://
--4mN1FYI7gnXysABs
Content-Type: application/
Content-
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
iD8DBQFDcSGEgY5
GMiJdUGwvZqZyXN
=G+Tb
-----END PGP SIGNATURE-----
--4mN1FYI7gnXys
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
|
#22 |
Source: gcc-4.0
Source-Version: 4.0.2-4
We believe that the bug you reported is fixed in the latest version of
gcc-4.0, which is due to be installed in the Debian FTP archive:
cpp-4.0-
to pool/main/
cpp-4.0_
to pool/main/
fastjar_
to pool/main/
fixincludes_
to pool/main/
g++-4.0_
to pool/main/
gcc-4.0-
to pool/main/
gcc-4.0-
to pool/main/
gcc-4.0-
to pool/main/
gcc-4.0_
to pool/main/
gcc-4.0_4.0.2-4.dsc
to pool/main/
gcc-4.0_
to pool/main/
gcj-4.0_
to pool/main/
gfortran-
to pool/main/
gfortran-
to pool/main/
gij-4.0_
to pool/main/
gnat-4.
to pool/main/
gnat-4.
to pool/main/
gobjc-4.
to pool/main/
lib64gcc1_
to pool/main/
lib64gfortran0_
to pool/main/
lib64objc1_
to pool/main/
lib64stdc+
to pool/main/
lib64stdc+
to pool/main/
libffi4-
to pool/main/
libffi4_
to pool/main/
libgcc1_
to pool/main/
libgcj-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6_
to pool/main/
libgfortran0-
to pool/main/
libgfortran0_
to pool/main/
libgnat-
to pool/main/
libmudflap0-
to pool/main/
lib...
![](/+icing/build/overlay/assets/skins/sam/images/close.gif)
Debian Bug Importer (debzilla) wrote : | #23 |
Message-Id: <email address hidden>
Date: Wed, 16 Nov 2005 10:32:19 -0800
From: Matthias Klose <email address hidden>
To: <email address hidden>
Subject: Bug#336463: fixed in gcc-4.0 4.0.2-4
Source: gcc-4.0
Source-Version: 4.0.2-4
We believe that the bug you reported is fixed in the latest version of
gcc-4.0, which is due to be installed in the Debian FTP archive:
cpp-4.0-
to pool/main/
cpp-4.0_
to pool/main/
fastjar_
to pool/main/
fixincludes_
to pool/main/
g++-4.0_
to pool/main/
gcc-4.0-
to pool/main/
gcc-4.0-
to pool/main/
gcc-4.0-
to pool/main/
gcc-4.0_
to pool/main/
gcc-4.0_4.0.2-4.dsc
to pool/main/
gcc-4.0_
to pool/main/
gcj-4.0_
to pool/main/
gfortran-
to pool/main/
gfortran-
to pool/main/
gij-4.0_
to pool/main/
gnat-4.
to pool/main/
gnat-4.
to pool/main/
gobjc-4.
to pool/main/
lib64gcc1_
to pool/main/
lib64gfortran0_
to pool/main/
lib64objc1_
to pool/main/
lib64stdc+
to pool/main/
lib64stdc+
to pool/main/
libffi4-
to pool/main/
libffi4_
to pool/main/
libgcc1_
to pool/main/
libgcj-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6-
to pool/main/
libgcj6_
to pool/main/
libgfortran0-
to pool/main/
libgfortran0_
to pool/main/
reassign 336463 scummvm
thanks
well, check it before uploading, if you think it's critical. please
ask the debian admins to install what's needed on the developer
machines. please reassign after providing the complete information.
Matthias
Tore Anderson writes: .cpp: In member function 'Graphics::Surface* Scumm:: ScummEngine: :loadThumbnail( Common: :InSaveFile* )': .cpp:100: internal compiler error: Floating point exception buildd. debian. org/fetch. php?&pkg= scummvm& ver=0.8. 0-1&arch= mips&stamp= 1130622019& file=log& as=raw> buildd. debian. org/fetch. php?&pkg= scummvm& ver=0.8. 0-1&arch= mipsel& stamp=113062274 5&file= log&as= raw>
>
> Package: g++-4.0
> Severity: critical
> Version: 4.0.2-3
>
> Hi. I just uploaded ScummVM 0.8.0, which fails to build from source
> due to an ICE:
>
> scumm/thumbnail
> scumm/thumbnail
>
> You can find the buildd logs here:
> <http://
> <http://
>
> I'm afraid this is all the debug information I can supply, as I don't
> own any mips[el] machines. The sources that makes it fail should be
> available in the archive shortly, though.
>
> Kind regards
> --
> Tore Anderson
>
>
>
> --
> To UNSUBSCRIBE, email to <email address hidden>
> with a subject of "unsubscribe". Trouble? Contact <email address hidden>