'ip route' kernel problem w/ gcc-4.0

[Debian Bug Importer]

Automatically imported from Debian bug report #322723 http://bugs.debian.org/322723

[Joey Hess (joeyh)]

Here's an image to use to test if it's busybox or the kernel that's
causing the problem. This netboot image uses busybox-cvs:

http://kitenet.net/~joey/tmp/bb-cvs

--
see shy jo

[Frans Pop (aragorn)]

reassign 322723 linux-2.6
retitle 322723 D-I: 'id route add' fails w/ "Network is unreachable"
thanks

We have just tested this issue using an installer with the busybox version
from Sarge and the error shows up then as well.
We feel this makes the kernel the most likely source of the problem.

[Frans Pop (aragorn)]

On Tuesday 16 August 2005 19:35, Andres Salomon wrote:
> > Regarding this problem with iproute:
> >
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=282492
> >
> > Would it be possible to include a fix for this, or at least make
> > CONFIG_ATM_CLIP a module instead of a yes in the stock kernel?
> > I'm suffering it right now, and it is the only reason that forces
> > me to compile my own kernels...
> > Thanks,
>
> This should probably be the case anyways, as it appears to be marked
> EXPERIMENTAL, and the config does support tristate. Not only that, but
> it's configured differently across different archs; definitely a
> candidate for a global config.

Could this also be the cause of #322723 (d-i installs failing because of
failing ip route command)? It seems somewhat similar.
If so, please implement the change soonest.

Cheers,
FJP

[Andres Salomon (dilinger)]

On Wed, 2005-08-17 at 02:13 +0200, Frans Pop wrote:
> On Tuesday 16 August 2005 19:35, Andres Salomon wrote:
[...]
> >
> > This should probably be the case anyways, as it appears to be marked
> > EXPERIMENTAL, and the config does support tristate. Not only that, but
> > it's configured differently across different archs; definitely a
> > candidate for a global config.
>
> Could this also be the cause of #322723 (d-i installs failing because of
> failing ip route command)? It seems somewhat similar.
> If so, please implement the change soonest.

This has been done for all archs, and will show up in 2.6.12-6 (unless
someone has a reason why ATM and ATM_CLIPS shouldn't be modular).

[Joey Hess (joeyh)]

Here is a strace -s 4096 -xx as requested.

--
see shy jo

[Frans Pop (aragorn)]

With help from waldi I have done some additional tests for this issue
yesterday and today. Here are the main results:

- the problem is still there if Sarge's 'ip' is used instead of the
  busybox version
- the problem is still there if the new linux-image-2.6.12-1-386
  (2.6.12-5) kernel is used in the installer

- the problem is *gone* if the kernel is compiled with gcc-3.3 from
  linux-source-2.6.12 (2.6.12-5)

- the problem is also gone if 'route add default gw ...' is used instead
  of 'ip route add default via ...'; this could be used as a workaround

Cheers,
FJP

[Frans-pop (frans-pop)]

retitle 322723 D-I: 'ip route add' fails w/ "Network is unreachable"
thanks

********************* DISCLAIMER *********************
De informatie in dit e-mail bericht is uitsluitend
bestemd voor de geadresseerde. Verstrekking aan
en gebruik door anderen is niet toegestaan.
Door de electronische verzending van het bericht
kunnen er geen rechten worden ontleend aan de
informatie.
************************************************************

[Bastian Blank (waldi)]

On Fri, Aug 19, 2005 at 12:56:34AM +0200, Frans Pop wrote:
> - the problem is still there if the new linux-image-2.6.12-1-386
> (2.6.12-5) kernel is used in the installer
- the problem is not there in the 686 image.

Bastian

--
Warp 7 -- It's a law we can live with.

[Frans Pop (aragorn)]

On Friday 19 August 2005 21:01, Bastian Blank wrote:
> - the problem is not there in the 686 image.

On request from waldi, I've compiled a -386 kernel (using gcc-4.0) with
one change in ./debian/arch/i386/config.386:
- CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+ # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set

The problem was solved when installing from a mini.iso built using this
kernel.

[Simon Horman (horms)]

On Sat, Aug 20, 2005 at 06:03:59PM +0200, Frans Pop wrote:
> On Friday 19 August 2005 21:01, Bastian Blank wrote:
> > - the problem is not there in the 686 image.
>
> On request from waldi, I've compiled a -386 kernel (using gcc-4.0) with
> one change in ./debian/arch/i386/config.386:
> - CONFIG_CC_OPTIMIZE_FOR_SIZE=y
> + # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
>
> The problem was solved when installing from a mini.iso built using this
> kernel.

Funny, I ran into this bug over the weekend too.
Damn pain to track down too.

I've put the following in SVN, so this
should be resolved in the next release.

Thanks for tracking this down, I was dreading having to.

--
Horms

Index: arch/i386/config.386
===================================================================
--- arch/i386/config.386 (revision 3967)
+++ arch/i386/config.386 (working copy)
@@ -1,4 +1,4 @@
-CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
 CONFIG_M386=y
 # CONFIG_M686 is not set
 # CONFIG_MK7 is not set

[Frans Pop (aragorn)]

On Sunday 21 August 2005 03:56, Horms wrote:
> I've put the following in SVN, so this
> should be resolved in the next release.

I noticed on IRC that you put a "closes: #322723" in the changelog.
I'm wondering if that is correct as I would say there still is a bad bug
in gcc-4.0 causing the kernel to break when compiled with optimization.

You may want to leave the bug open at reduced priority so that
optimization may be reenabled when the problem in gcc-4.0 is fixed.

For the same reason, maybe this bug should be cloned to gcc-4.0?

[Simon Horman (horms)]

On Sun, Aug 21, 2005 at 01:14:45PM +0200, Frans Pop wrote:
> On Sunday 21 August 2005 03:56, Horms wrote:
> > I've put the following in SVN, so this
> > should be resolved in the next release.
>
> I noticed on IRC that you put a "closes: #322723" in the changelog.
> I'm wondering if that is correct as I would say there still is a bad bug
> in gcc-4.0 causing the kernel to break when compiled with optimization.
>
> You may want to leave the bug open at reduced priority so that
> optimization may be reenabled when the problem in gcc-4.0 is fixed.
>
> For the same reason, maybe this bug should be cloned to gcc-4.0?

Good point, though it might be best just to reassign it to gcc-4.0.
I've seen this flag break stuff many time before - I actually
wonder if it ever works as intended.

--
Horms

[Andres Salomon (dilinger)]

Source: linux-2.6
Source-Version: 2.6.12-6

We believe that the bug you reported is fixed in the latest version of
linux-2.6, which is due to be installed in the Debian FTP archive:

kernel-image-2.6-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-386_2.6.12-6_i386.deb
kernel-image-2.6-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-686-smp_2.6.12-6_i386.deb
kernel-image-2.6-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-686_2.6.12-6_i386.deb
kernel-image-2.6-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-k7-smp_2.6.12-6_i386.deb
kernel-image-2.6-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-k7_2.6.12-6_i386.deb
linux-2.6_2.6.12-6.diff.gz
  to pool/main/l/linux-2.6/linux-2.6_2.6.12-6.diff.gz
linux-2.6_2.6.12-6.dsc
  to pool/main/l/linux-2.6/linux-2.6_2.6.12-6.dsc
linux-doc-2.6.12_2.6.12-6_all.deb
  to pool/main/l/linux-2.6/linux-doc-2.6.12_2.6.12-6_all.deb
linux-headers-2.6-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-386_2.6.12-6_i386.deb
linux-headers-2.6-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-686-smp_2.6.12-6_i386.deb
linux-headers-2.6-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-686_2.6.12-6_i386.deb
linux-headers-2.6-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-k7-smp_2.6.12-6_i386.deb
linux-headers-2.6-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-k7_2.6.12-6_i386.deb
linux-headers-2.6.12-1-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-386_2.6.12-6_i386.deb
linux-headers-2.6.12-1-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-686-smp_2.6.12-6_i386.deb
linux-headers-2.6.12-1-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-686_2.6.12-6_i386.deb
linux-headers-2.6.12-1-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-k7-smp_2.6.12-6_i386.deb
linux-headers-2.6.12-1-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-k7_2.6.12-6_i386.deb
linux-headers-2.6.12-1_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1_2.6.12-6_i386.deb
linux-image-2.6-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-386_2.6.12-6_i386.deb
linux-image-2.6-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-686-smp_2.6.12-6_i386.deb
linux-image-2.6-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-686_2.6.12-6_i386.deb
linux-image-2.6-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-k7-smp_2.6.12-6_i386.deb
linux-image-2.6-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-k7_2.6.12-6_i386.deb
linux-image-2.6.12-1-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6.12-1-386_2.6.12-6_i386.deb
linux-image-2.6.12-1-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6.12-1-686-smp_2.6.12-6_i386.deb
linux-image-2.6.12-1-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6.12-1-686_2.6.12-6_i386.deb
linux-image-2.6.12-1-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6.12-1-k7-smp_2.6...

[Frans Pop (aragorn)]

reassign 322723 gcc-4.0
found 322723 4.0.1-3
retitle 322723 'ip route' kernel problem w/ gcc-4.0
thanks

[Frans Pop (aragorn)]

reopen 322723
thanks

"found" by itself does not seem to work; let's try good old reopen...

[Frans Pop (aragorn)]

I've reassigned this bug from the kernel to gcc-4.0 as we feel that the
solution chosen in the kernel packaging is not really a fix, but a
workaround.
As tests have shown that the problem does not exist when the same kernel
is compiled with gcc-3.3, the real bug is likely in gcc-4.0.

I've "found" the bug for 4.0.1-3 after looking at the dates for relevant
uploads, but as uploads were quite frequent, I could be off by a version
either way.

Please see the bug history for details.

Cheers,
FJP

[Falk Hueffner (falk)]

Frans Pop <email address hidden> writes:

> I've reassigned this bug from the kernel to gcc-4.0 as we feel that the
> solution chosen in the kernel packaging is not really a fix, but a
> workaround.
> As tests have shown that the problem does not exist when the same kernel
> is compiled with gcc-3.3, the real bug is likely in gcc-4.0.

It would be extremely helpful if you could find a way to reproduce
this that does not require root... any ideas?

--
 Falk

[Frans Pop (aragorn)]

On Wednesday 07 September 2005 23:21, you wrote:
> It would be extremely helpful if you could find a way to reproduce
> this that does not require root... any ideas?

Well, I think the problem can also be reproduced outside a
debian-installer environment (getting linux-image-2.6.12-1-386 (2.6.12-2)
out of the archives and running 'ip route add' should reproduce it), but
I've seen no other commands than 'ip route add' fail and that does
require root...

Unfortunately I have hardly any C background and feel I've already passed
the limits of what I can do (a lot of the info was gathered with help of
joeyh and waldi).
Maybe the best thing would be to work together with the kernel team on
this issue.

Hmm. waldi just pointed me at #325670 on IRC. Looks like it may be
related.

[Falk Hueffner (falk)]

retitle 284777 [fixed in 4.0.2] ICE in uses_template_parms at cp/pt.c:4860
tags 284777 + fixed-upstream
tags 322723 + wontfix # says upstream
retitle 228645 [fixed in 4.1.0] call of overloaded `llabs(int)' is ambiguous
tags 228645 + fixed-upstream
thanks

--
 Falk

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #322723 http://bugs.debian.org/322723

[Falk Hueffner (falk)]

tags 322723 - upstream
tags 322723 - wontfix

[Fabio Massimo Di Nitto (fabbione)]

This bug was caused by the kernel compiled with gcc-4.0. This is not the case
for Breezy, but
we will need to look at it if/when we will switch to gcc-4.0 in the kernel.
(NOTE that we share the same kernel option with debian
CONFIG_CC_OPTIMIZE_FOR_SIZE=y that's causing the problem)

Fabio

[Matthias Klose (doko-cs)]

severity 322723 important
thanks

there's a workaround, and gcc-3.4 is known to work as well. Is there
any reason to use gcc-4.0 for kernel builds on all architectures?

Frans Pop writes:
> I've reassigned this bug from the kernel to gcc-4.0 as we feel that the
> solution chosen in the kernel packaging is not really a fix, but a
> workaround.
> As tests have shown that the problem does not exist when the same kernel
> is compiled with gcc-3.3, the real bug is likely in gcc-4.0.
>
> I've "found" the bug for 4.0.1-3 after looking at the dates for relevant
> uploads, but as uploads were quite frequent, I could be off by a version
> either way.
>
> Please see the bug history for details.
>
> Cheers,
> FJP

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 15 Aug 2005 13:54:18 -0400
From: Joey Hess <email address hidden>
To: <email address hidden>, <email address hidden>
Subject: test if this is busybox or kernel

--huq684BweRXVnRxX
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Here's an image to use to test if it's busybox or the kernel that's
causing the problem. This netboot image uses busybox-cvs:

http://kitenet.net/~joey/tmp/bb-cvs

--=20
see shy jo

--huq684BweRXVnRxX
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFDANbJd8HHehbQuO8RAgd7AKDAQ/knnSZw9n7s98nxteBu621yywCgn7Vw
RSEwGEsch7IfvX9MJho/e34=
=xP87
-----END PGP SIGNATURE-----

--huq684BweRXVnRxX--

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Mon, 15 Aug 2005 22:31:39 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>,
 <email address hidden>
Subject: Reassign #322723

--nextPart1368340.MKTSHK4ZIT
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

reassign 322723 linux-2.6
retitle 322723 D-I: 'id route add' fails w/ "Network is unreachable"
thanks

We have just tested this issue using an installer with the busybox version=
=20
from Sarge and the error shows up then as well.
We feel this makes the kernel the most likely source of the problem.

--nextPart1368340.MKTSHK4ZIT
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDAPurgm/Kwh6ICoQRApzYAJ92MScBTAOzYJ0DOuDhzd+dorJAUQCggdYu
bGPbhhuMZHkp2nuIhIZpMlo=
=JzDH
-----END PGP SIGNATURE-----

--nextPart1368340.MKTSHK4ZIT--

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Wed, 17 Aug 2005 02:13:46 +0200
From: Frans Pop <email address hidden>
To: Andres Salomon <email address hidden>,
 <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#323143: ip neigh flush (iproute) hangs because CONFIG_ATM_CLIP=y

--nextPart224851607.GIenB0L8mc
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On Tuesday 16 August 2005 19:35, Andres Salomon wrote:
> > Regarding this problem with iproute:
> >
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D282492
> >
> > Would it be possible to include a fix for this, or at least make
> > CONFIG_ATM_CLIP a module instead of a yes in the stock kernel?
> > I'm suffering it right now, and it is the only reason that forces
> > me to compile my own kernels...
> > Thanks,
>
> This should probably be the case anyways, as it appears to be marked
> EXPERIMENTAL, and the config does support tristate. Not only that, but
> it's configured differently across different archs; definitely a
> candidate for a global config.

Could this also be the cause of #322723 (d-i installs failing because of=20
failing ip route command)? It seems somewhat similar.
If so, please implement the change soonest.

Cheers,
=46JP

--nextPart224851607.GIenB0L8mc
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDAoE6gm/Kwh6ICoQRAjhCAJ94q9kmZayVbRRhiRukuCP0c0id1wCcDikO
f2Jc+FVbYcpFsXNiyKeoWnA=
=PFCi
-----END PGP SIGNATURE-----

--nextPart224851607.GIenB0L8mc--

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Wed, 17 Aug 2005 00:48:25 -0400
From: Andres Salomon <email address hidden>
To: Frans Pop <email address hidden>
Cc: <email address hidden>, <email address hidden>
Subject: Re: Bug#323143: ip neigh flush (iproute) hangs because
 CONFIG_ATM_CLIP=y

On Wed, 2005-08-17 at 02:13 +0200, Frans Pop wrote:
> On Tuesday 16 August 2005 19:35, Andres Salomon wrote:
[...]
> >
> > This should probably be the case anyways, as it appears to be marked
> > EXPERIMENTAL, and the config does support tristate. Not only that, but
> > it's configured differently across different archs; definitely a
> > candidate for a global config.
>
> Could this also be the cause of #322723 (d-i installs failing because of
> failing ip route command)? It seems somewhat similar.
> If so, please implement the change soonest.

This has been done for all archs, and will show up in 2.6.12-6 (unless
someone has a reason why ATM and ATM_CLIPS shouldn't be modular).

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 17 Aug 2005 12:42:26 -0400
From: Joey Hess <email address hidden>
To: <email address hidden>, <email address hidden>
Subject: another strace

--tjCHc7DPkfUGtrlw
Content-Type: multipart/mixed; boundary="YiEDa0DAkWCtVeE4"
Content-Disposition: inline

--YiEDa0DAkWCtVeE4
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Here is a strace -s 4096 -xx as requested.

--=20
see shy jo

--YiEDa0DAkWCtVeE4
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename="ip.strace.2"
Content-Transfer-Encoding: quoted-printable

execve("/bin/ip", ["\x69\x70", "\x72\x6f\x75\x74\x65", "\x61\x64\x64", "\x6=
4\x65\x66\x61\x75\x6c\x74", "\x76\x69\x61", "\x31\x39\x32\x2e\x31\x36\x38\x=
2e\x32\x2e\x31"], [/* 10 vars */]) =3D 0
uname({sys=3D"Linux", node=3D"", ...}) =3D 0
brk(0) =3D 0x80f2000
access("/etc/ld.so.nohwcap", F_OK) =3D -1 ENOENT (No such file or dire=
ctory)
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0=
) =3D 0xb7fe6000
access("/etc/ld.so.preload", R_OK) =3D -1 ENOENT (No such file or dire=
ctory)
open("/etc/ld.so.cache", O_RDONLY) =3D -1 ENOENT (No such file or dire=
ctory)
open("/lib/tls/i686/sse2/cmov/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such =
file or directory)
stat64("/lib/tls/i686/sse2/cmov", 0xbfafae34) =3D -1 ENOENT (No such file o=
r directory)
open("/lib/tls/i686/sse2/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file =
or directory)
stat64("/lib/tls/i686/sse2", 0xbfafae34) =3D -1 ENOENT (No such file or dir=
ectory)
open("/lib/tls/i686/cmov/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file =
or directory)
stat64("/lib/tls/i686/cmov", 0xbfafae34) =3D -1 ENOENT (No such file or dir=
ectory)
open("/lib/tls/i686/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file or di=
rectory)
stat64("/lib/tls/i686", 0xbfafae34) =3D -1 ENOENT (No such file or dire=
ctory)
open("/lib/tls/sse2/cmov/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file =
or directory)
stat64("/lib/tls/sse2/cmov", 0xbfafae34) =3D -1 ENOENT (No such file or dir=
ectory)
open("/lib/tls/sse2/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file or di=
rectory)
stat64("/lib/tls/sse2", 0xbfafae34) =3D -1 ENOENT (No such file or dire=
ctory)
open("/lib/tls/cmov/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file or di=
rectory)
stat64("/lib/tls/cmov", 0xbfafae34) =3D -1 ENOENT (No such file or dire=
ctory)
open("/lib/tls/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file or dire=
ctory)
stat64("/lib/tls", 0xbfafae34) =3D -1 ENOENT (No such file or dire=
ctory)
open("/lib/i686/sse2/cmov/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file=
 or directory)
stat64("/lib/i686/sse2/cmov", 0xbfafae34) =3D -1 ENOENT (No such file or di=
rectory)
open("/lib/i686/sse2/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file or d=
irectory)
stat64("/lib/i686/sse2", 0xbfafae34) =3D -1 ENOENT (No such file or dire=
ctory)
open("/lib/i686/cmov/libc.so.6", O_RDONLY) =3D -1 ENOENT (No such file or d=
irectory)
stat64("/lib/i686/cmov", 0xbfafae34) =3D -1 ENOENT (No such file o...

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Fri, 19 Aug 2005 00:56:34 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>
Cc: <email address hidden>
Subject: #322723 D-I: 'id route add' fails w/ "Network is unreachable"

--nextPart1295031.Q39jX0ZO4q
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

With help from waldi I have done some additional tests for this issue=20
yesterday and today. Here are the main results:

=2D the problem is still there if Sarge's 'ip' is used instead of the
  busybox version
=2D the problem is still there if the new linux-image-2.6.12-1-386
  (2.6.12-5) kernel is used in the installer

=2D the problem is *gone* if the kernel is compiled with gcc-3.3 from
  linux-source-2.6.12 (2.6.12-5)

=2D the problem is also gone if 'route add default gw ...' is used instead
  of 'ip route add default via ...'; this could be used as a workaround

Cheers,
=46JP

--nextPart1295031.Q39jX0ZO4q
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDBRIlgm/Kwh6ICoQRAtdZAKCWHa1Pf3Ywrz+XyGSUJwqsK8VyTwCfUA8f
dxydx+25cbMpcPT+a69c2RM=
=mK4t
-----END PGP SIGNATURE-----

--nextPart1295031.Q39jX0ZO4q--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 19 Aug 2005 10:53:52 +0200
From: <email address hidden>
To: <email address hidden>
Subject: Fix typo in bug title

retitle 322723 D-I: 'ip route add' fails w/ "Network is unreachable"
thanks

********************* DISCLAIMER *********************
De informatie in dit e-mail bericht is uitsluitend
bestemd voor de geadresseerde. Verstrekking aan
en gebruik door anderen is niet toegestaan.
Door de electronische verzending van het bericht
kunnen er geen rechten worden ontleend aan de
informatie.
************************************************************

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 19 Aug 2005 21:01:43 +0200
From: Bastian Blank <email address hidden>
To: Frans Pop <email address hidden>, <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#322723: #322723 D-I: 'id route add' fails w/ "Network is unreachable"

--MGYHOYXEY6WxJCY8
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Aug 19, 2005 at 12:56:34AM +0200, Frans Pop wrote:
> - the problem is still there if the new linux-image-2.6.12-1-386
> (2.6.12-5) kernel is used in the installer
- the problem is not there in the 686 image.

Bastian

--=20
Warp 7 -- It's a law we can live with.

--MGYHOYXEY6WxJCY8
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iEYEARECAAYFAkMGLJcACgkQnw66O/MvCNHbjQCePBivQ8g65X5UF8kCYxLGz+EM
MtwAn0FPyly/Y4ytMaz91AHUZBDazSVh
=949b
-----END PGP SIGNATURE-----

--MGYHOYXEY6WxJCY8--

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Sat, 20 Aug 2005 18:03:59 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>,
 <email address hidden>
Subject: Re: Bug#322723: #322723 D-I: 'id route add' fails w/ "Network is unreachable"

--nextPart2525855.jzfDPV0CVe
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On Friday 19 August 2005 21:01, Bastian Blank wrote:
> - the problem is not there in the 686 image.

On request from waldi, I've compiled a -386 kernel (using gcc-4.0) with=20
one change in ./debian/arch/i386/config.386:
=2D CONFIG_CC_OPTIMIZE_FOR_SIZE=3Dy
+ # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set

The problem was solved when installing from a mini.iso built using this=20
kernel.

--nextPart2525855.jzfDPV0CVe
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDB1R3gm/Kwh6ICoQRAjvaAKCbMwGeOK8yFdohvgJZAY2gpTvBlgCfTshh
BMyPNl7ag6z14eplG5BpyKs=
=gMd3
-----END PGP SIGNATURE-----

--nextPart2525855.jzfDPV0CVe--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 21 Aug 2005 10:56:13 +0900
From: Horms <email address hidden>
To: Frans Pop <email address hidden>, <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#322723: #322723 D-I: 'id route add' fails w/ "Network is unreachable"

On Sat, Aug 20, 2005 at 06:03:59PM +0200, Frans Pop wrote:
> On Friday 19 August 2005 21:01, Bastian Blank wrote:
> > - the problem is not there in the 686 image.
>
> On request from waldi, I've compiled a -386 kernel (using gcc-4.0) with
> one change in ./debian/arch/i386/config.386:
> - CONFIG_CC_OPTIMIZE_FOR_SIZE=y
> + # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
>
> The problem was solved when installing from a mini.iso built using this
> kernel.

Funny, I ran into this bug over the weekend too.
Damn pain to track down too.

I've put the following in SVN, so this
should be resolved in the next release.

Thanks for tracking this down, I was dreading having to.

--
Horms

Index: arch/i386/config.386
===================================================================
--- arch/i386/config.386 (revision 3967)
+++ arch/i386/config.386 (working copy)
@@ -1,4 +1,4 @@
-CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
 CONFIG_M386=y
 # CONFIG_M686 is not set
 # CONFIG_MK7 is not set

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Sun, 21 Aug 2005 13:14:45 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>
Subject: Re: Bug#322723: #322723 D-I: 'ip route add' fails w/ "Network is unreachable"

--nextPart8524305.CoJNsrtVVA
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On Sunday 21 August 2005 03:56, Horms wrote:
> I've put the following in SVN, so this
> should be resolved in the next release.

I noticed on IRC that you put a "closes: #322723" in the changelog.
I'm wondering if that is correct as I would say there still is a bad bug=20
in gcc-4.0 causing the kernel to break when compiled with optimization.

You may want to leave the bug open at reduced priority so that=20
optimization may be reenabled when the problem in gcc-4.0 is fixed.

=46or the same reason, maybe this bug should be cloned to gcc-4.0?

--nextPart8524305.CoJNsrtVVA
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDCGImgm/Kwh6ICoQRAvDcAJ96u3EHnvNMWXKWm28HSz+Us0agLACeKM8H
dnOf2UY7ESB0pNDWlpHogC8=
=Cs/O
-----END PGP SIGNATURE-----

--nextPart8524305.CoJNsrtVVA--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 22 Aug 2005 10:50:44 +0900
From: Horms <email address hidden>
To: Frans Pop <email address hidden>, <email address hidden>
Subject: Re: Bug#322723: #322723 D-I: 'ip route add' fails w/ "Network is unreachable"

On Sun, Aug 21, 2005 at 01:14:45PM +0200, Frans Pop wrote:
> On Sunday 21 August 2005 03:56, Horms wrote:
> > I've put the following in SVN, so this
> > should be resolved in the next release.
>
> I noticed on IRC that you put a "closes: #322723" in the changelog.
> I'm wondering if that is correct as I would say there still is a bad bug
> in gcc-4.0 causing the kernel to break when compiled with optimization.
>
> You may want to leave the bug open at reduced priority so that
> optimization may be reenabled when the problem in gcc-4.0 is fixed.
>
> For the same reason, maybe this bug should be cloned to gcc-4.0?

Good point, though it might be best just to reassign it to gcc-4.0.
I've seen this flag break stuff many time before - I actually
wonder if it ever works as intended.

--
Horms

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Tue, 06 Sep 2005 14:17:24 -0700
From: Andres Salomon <email address hidden>
To: <email address hidden>
Subject: Bug#322723: fixed in linux-2.6 2.6.12-6

Source: linux-2.6
Source-Version: 2.6.12-6

We believe that the bug you reported is fixed in the latest version of
linux-2.6, which is due to be installed in the Debian FTP archive:

kernel-image-2.6-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-386_2.6.12-6_i386.deb
kernel-image-2.6-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-686-smp_2.6.12-6_i386.deb
kernel-image-2.6-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-686_2.6.12-6_i386.deb
kernel-image-2.6-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-k7-smp_2.6.12-6_i386.deb
kernel-image-2.6-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/kernel-image-2.6-k7_2.6.12-6_i386.deb
linux-2.6_2.6.12-6.diff.gz
  to pool/main/l/linux-2.6/linux-2.6_2.6.12-6.diff.gz
linux-2.6_2.6.12-6.dsc
  to pool/main/l/linux-2.6/linux-2.6_2.6.12-6.dsc
linux-doc-2.6.12_2.6.12-6_all.deb
  to pool/main/l/linux-2.6/linux-doc-2.6.12_2.6.12-6_all.deb
linux-headers-2.6-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-386_2.6.12-6_i386.deb
linux-headers-2.6-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-686-smp_2.6.12-6_i386.deb
linux-headers-2.6-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-686_2.6.12-6_i386.deb
linux-headers-2.6-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-k7-smp_2.6.12-6_i386.deb
linux-headers-2.6-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6-k7_2.6.12-6_i386.deb
linux-headers-2.6.12-1-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-386_2.6.12-6_i386.deb
linux-headers-2.6.12-1-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-686-smp_2.6.12-6_i386.deb
linux-headers-2.6.12-1-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-686_2.6.12-6_i386.deb
linux-headers-2.6.12-1-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-k7-smp_2.6.12-6_i386.deb
linux-headers-2.6.12-1-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1-k7_2.6.12-6_i386.deb
linux-headers-2.6.12-1_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-headers-2.6.12-1_2.6.12-6_i386.deb
linux-image-2.6-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-386_2.6.12-6_i386.deb
linux-image-2.6-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-686-smp_2.6.12-6_i386.deb
linux-image-2.6-686_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-686_2.6.12-6_i386.deb
linux-image-2.6-k7-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-k7-smp_2.6.12-6_i386.deb
linux-image-2.6-k7_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6-k7_2.6.12-6_i386.deb
linux-image-2.6.12-1-386_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6.12-1-386_2.6.12-6_i386.deb
linux-image-2.6.12-1-686-smp_2.6.12-6_i386.deb
  to pool/main/l/linux-2.6/linux-image-2.6.12-1-686-smp_2.6.12-6_i386.deb
linu...

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Wed, 7 Sep 2005 22:27:10 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>
Subject: Reopen/reassign #322723: 'ip route' kernel problem w/ gcc-4.0

reassign 322723 gcc-4.0
found 322723 4.0.1-3
retitle 322723 'ip route' kernel problem w/ gcc-4.0
thanks

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Wed, 7 Sep 2005 22:45:58 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>
Subject: Reopen/reassign #322723: 'ip route' kernel problem w/ gcc-4.0

reopen 322723
thanks

"found" by itself does not seem to work; let's try good old reopen...

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Wed, 7 Sep 2005 22:55:20 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>
Subject: Bug#322723: 'ip route' kernel problem w/ gcc-4.0

--nextPart2558284.reTYGQfyOX
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

I've reassigned this bug from the kernel to gcc-4.0 as we feel that the=20
solution chosen in the kernel packaging is not really a fix, but a=20
workaround.
As tests have shown that the problem does not exist when the same kernel=20
is compiled with gcc-3.3, the real bug is likely in gcc-4.0.

I've "found" the bug for 4.0.1-3 after looking at the dates for relevant=20
uploads, but as uploads were quite frequent, I could be off by a version=20
either way.

Please see the bug history for details.

Cheers,
=46JP

--nextPart2558284.reTYGQfyOX
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDH1O4gm/Kwh6ICoQRAjD4AJ9eXFLRkY1VALorlcEwuPYEG6dgBwCgmXbD
w25SpJycs6hkeFk8WQ3Wthc=
=Hffb
-----END PGP SIGNATURE-----

--nextPart2558284.reTYGQfyOX--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 07 Sep 2005 23:21:52 +0200
From: Falk Hueffner <email address hidden>
To: Frans Pop <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#322723: 'ip route' kernel problem w/ gcc-4.0

Frans Pop <email address hidden> writes:

> I've reassigned this bug from the kernel to gcc-4.0 as we feel that the
> solution chosen in the kernel packaging is not really a fix, but a
> workaround.
> As tests have shown that the problem does not exist when the same kernel
> is compiled with gcc-3.3, the real bug is likely in gcc-4.0.

It would be extremely helpful if you could find a way to reproduce
this that does not require root... any ideas?

--
 Falk

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Wed, 7 Sep 2005 23:53:57 +0200
From: Frans Pop <email address hidden>
To: <email address hidden>
Subject: Re: Bug#322723: 'ip route' kernel problem w/ gcc-4.0

--nextPart4677448.r64pmDypJp
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On Wednesday 07 September 2005 23:21, you wrote:
> It would be extremely helpful if you could find a way to reproduce
> this that does not require root... any ideas?

Well, I think the problem can also be reproduced outside a=20
debian-installer environment (getting linux-image-2.6.12-1-386 (2.6.12-2)=20
out of the archives and running 'ip route add' should reproduce it), but=20
I've seen no other commands than 'ip route add' fail and that does=20
require root...

Unfortunately I have hardly any C background and feel I've already passed=20
the limits of what I can do (a lot of the info was gathered with help of=20
joeyh and waldi).
Maybe the best thing would be to work together with the kernel team on=20
this issue.

Hmm. waldi just pointed me at #325670 on IRC. Looks like it may be=20
related.

--nextPart4677448.r64pmDypJp
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDH2F1gm/Kwh6ICoQRAmFsAKC+RWY5gdlWZsSJVzYxjnrSkWYafgCeIHk5
X3rO8lmMk/rO6vJ4P+Am+Yk=
=O+X+
-----END PGP SIGNATURE-----

--nextPart4677448.r64pmDypJp--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Thu, 08 Sep 2005 00:14:25 +0200
From: Falk Hueffner <email address hidden>
To: <email address hidden>
Subject: gcc stuff

retitle 284777 [fixed in 4.0.2] ICE in uses_template_parms at cp/pt.c:4860
tags 284777 + fixed-upstream
tags 322723 + wontfix # says upstream
retitle 228645 [fixed in 4.1.0] call of overloaded `llabs(int)' is ambiguous
tags 228645 + fixed-upstream
thanks

--
 Falk

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Thu, 08 Sep 2005 17:06:29 +0200
From: Falk Hueffner <email address hidden>
To: <email address hidden>
Subject: fix #322723 tags

tags 322723 - upstream
tags 322723 - wontfix

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 12 Sep 2005 09:33:57 +0200
From: Matthias Klose <email address hidden>
To: Frans Pop <email address hidden>, <email address hidden>,
        <email address hidden>
CC: <email address hidden>
Subject: Re: Bug#322723: 'ip route' kernel problem w/ gcc-4.0

severity 322723 important
thanks

there's a workaround, and gcc-3.4 is known to work as well. Is there
any reason to use gcc-4.0 for kernel builds on all architectures?

Frans Pop writes:
> I've reassigned this bug from the kernel to gcc-4.0 as we feel that the
> solution chosen in the kernel packaging is not really a fix, but a
> workaround.
> As tests have shown that the problem does not exist when the same kernel
> is compiled with gcc-3.3, the real bug is likely in gcc-4.0.
>
> I've "found" the bug for 4.0.1-3 after looking at the dates for relevant
> uploads, but as uploads were quite frequent, I could be off by a version
> either way.
>
> Please see the bug history for details.
>
> Cheers,
> FJP

[Falk Hueffner (falk)]

tags 322723 + moreinfo
tags 322723 + unreproducible
thanks

Hi,

is this still reproducible with current 4.0 or 4.1? If so, how?

--
 Falk

[Frans Pop (aragorn)]

On Saturday 25 March 2006 00:30, Falk Hueffner wrote:
> is this still reproducible with current 4.0 or 4.1? If so, how?

No, it looks like the problem has been resolved. Closing the BR.
I've reverted the workaround we introduced in the installer for the issue.
Thanks for the reminder.

Cheers,
FJP

[ccmail111 (ccmail111)]

Hi,

I am looking for fix to the problem:

ip route add 0.0.0.0/0 via 10.130.1.145 dev eth0
ip: RTNETLINK answers: Network is unreachable

I understand it is fixed.
Can you please email me the source code diff ?

I am unable to upgrade due to reasons. I would
like to patch it to my kernel.

Thanks,

Sunil.

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

[ccmail111 (ccmail111)]

Note: forwarded message attached.

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

Hi,

I am looking for fix to the problem:

ip route add 0.0.0.0/0 via 10.130.1.145 dev eth0
ip: RTNETLINK answers: Network is unreachable

I understand it is fixed.
Can you please email me the source code diff ?

I am unable to upgrade due to reasons. I would
like to patch it to my kernel.

Thanks,

Sunil.

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com