Debian Bug report logs - #521712
linux-headers-2.6.29-1-686: Makefile_32.cpu file missing from header files for x86 architecture

Display info messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: hildebrand victor <hildebrand.us@gmail.com>

To: submit@bugs.debian.org

Subject: linux-headers-2.6.29-1-686: Makefile_32.cpu file missing from header files for x86 architecture

Date: Sun, 29 Mar 2009 17:39:34 +0100

[Message part 1 (text/plain, inline)]

Package: linux-headers-2.6.29-1-686
Version: 2.6.29-1
Severity: normal

On installation of the package there should be a file installed
/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu
This file has traditionally contained the cpu tuning information. An extra
driver which I use (virtualbox) when compiled with
this header refused to compile when I installed 2.6.29 kernel from sid and
the related headers. On inspection of the output I found this file is
missing.
The output on compiling was as follows:-
** Compiling vboxdrv
make KBUILD_VERBOSE= -C /lib/modules/2.6.29-1-686/build SUBDIRS=/tmp/vbox.1
SRCROOT=/tmp/vbox.1 modules
make -C /usr/src/linux-headers-2.6.29-1-686 \
KBUILD_SRC=/usr/src/linux-headers-2.6.29-1-common \
KBUILD_EXTMOD="/tmp/vbox.1" -f
/usr/src/linux-headers-2.6.29-1-common/Makefile \
modules
/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile:41:
/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu: No such
file or directory
make[3]: *** No rule to make target
`/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu'. Stop.
make[2]: *** [sub-make] Error 2
make[1]: *** [all] Error 2
make: *** [vboxdrv] Error 2

My module seems to compile when I put equivalent file from the stock 2.6.26
kernel headers which
ships with Lenny release and it seemed to work. Request the maintainers to
please put the file so that other people like me
building drivers needing the above mentioned file are able to do so. It
seems to be missing from the this kernel header package.
Thanks for your attention and keep up the excellent work being done by your
team.

-- System Information:
Debian Release: 5.0
  APT prefers stable
  APT policy: (500, 'stable')
Architecture: i386 (i686)

Kernel: Linux 2.6.29-1-686 (SMP w/2 CPU cores)
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages linux-headers-2.6.29-1-686 depends on:
ii  gcc-4.3                       4.3.2-1.1  The GNU C compiler
ii  linux-headers-2.6.29-1-common 2.6.29-1   Common header files for Linux
2.6.
ii  linux-kbuild-2.6.29           2.6.29-1   Kbuild infrastructure for Linux
2.

linux-headers-2.6.29-1-686 recommends no packages.

linux-headers-2.6.29-1-686 suggests no packages.

-- no debconf information

[Message part 2 (text/html, inline)]

Message #10 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Don Pellegrino <don@drexel.edu>

To: 521712@bugs.debian.org

Subject: Confirmation

Date: Sun, 29 Mar 2009 22:15:24 -0400

I can confirm that /usr/src/linux-headers-2.6.29-1-common/arch/x86:41
makes references to a non-existent Makefile_32.cpu.  I bumped into this
problem while attempting to build the fglrx module (ATI Closed Graphcs
Driver).

Message #15 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Jelle de Jong <jelledejong@powercraft.nl>

To: 521712@bugs.debian.org

Subject: confirmation

Date: Mon, 30 Mar 2009 22:14:33 +0200

I can also confirm this bug when compiling the not so nice mainstream
em28xx code for my usb video devices:

$ make

running ./build.sh build

make[1]: Entering directory `/home/jelle/em28xx-new'
rm -rf Module.symvers;
make -C /lib/modules/`if [ -d /lib/modules/2.6.21.4-eeepc ]; then echo
2.6.21.4-eeepc; else uname -r; fi`/build SUBDIRS=`pwd` modules
make[2]: Entering directory `/usr/src/linux-headers-2.6.29-1-686'
/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile:41:
/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu: No such
file or directory
make[4]: *** No rule to make target
`/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu'.  Stop.
make[3]: *** [sub-make] Error 2
make[2]: *** [all] Error 2
make[2]: Leaving directory `/usr/src/linux-headers-2.6.29-1-686'
make[1]: *** [default] Error 2
make[1]: Leaving directory `/home/jelle/em28xx-new'

Any ideas when this will be fixed?

Best regards,

Jelle de Jong

Message #20 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Encolpe Degoute <encolpe.degoute@free.fr>

To: 521712@bugs.debian.org

Subject: [linux-headers-2.6.29-1-686] Short resolution

Date: Mon, 30 Mar 2009 23:28:42 +0200

Package: linux-headers-2.6.29-1-686
Version: 2.6.29-1

--- Please enter the report below this line. ---

As Google says it seems to be a recurrent error :
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501995


A quicker fix is to rebuild kernel-headers on your computer from sources:

# apt-get install linux-source-2.6.29
# cd /usr/src
# tar jxf linux-source-2.6.29.tar.bz2
# cd linux-source-2.6.29
# make-kpkg build kernel_headers
##long long time

# dpkg -i linux-header*.deb
# cd /usr/src/linux-headers-2.6.26/include/
# rm asm && ln -s asm-x86 asm

Then you can rebuild you modules.

--- System information. ---
Architecture: i386
Kernel: Linux 2.6.29-1-686

Debian Release: squeeze/sid
500 unstable ftp.fr.debian.org

--- Package information. ---
Depends (Version) | Installed
==============================================-+-=============
linux-headers-2.6.29-1-common (= 2.6.29-1) | 2.6.29-1
linux-kbuild-2.6.29 | 2.6.29-1
gcc-4.3 | 4.3.3-5



-- 
Encolpe DEGOUTE
http://encolpe.degoute.free.fr/
Logiciels libres, hockey sur glace et autres activités cérébrales

Message #25 received at 521712-done@bugs.debian.org (full text, mbox, reply):

From: maximilian attems <max@stro.at>

To: 521712-done@bugs.debian.org

Subject: Re: Bug#521712: [linux-headers-2.6.29-1-686] Short resolution

Date: Tue, 31 Mar 2009 00:11:45 +0200

On Mon, Mar 30, 2009 at 11:28:42PM +0200, Encolpe Degoute wrote:
> Package: linux-headers-2.6.29-1-686
> Version: 2.6.29-1
> 
> --- Please enter the report below this line. ---
> 
> As Google says it seems to be a recurrent error :
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501995
> 
> 
> A quicker fix is to rebuild kernel-headers on your computer from sources:
> 
> # apt-get install linux-source-2.6.29
> # cd /usr/src
> # tar jxf linux-source-2.6.29.tar.bz2
> # cd linux-source-2.6.29
> # make-kpkg build kernel_headers
> ##long long time
> 
> # dpkg -i linux-header*.deb
> # cd /usr/src/linux-headers-2.6.26/include/
> # rm asm && ln -s asm-x86 asm
> 
> Then you can rebuild you modules.

closing as duplicate useless submisson.
post to a mailinglist if you want to give an howto.

kernel-package use is deprecated.

Message #30 received at 521712@bugs.debian.org (full text, mbox, reply):

From: lsorense@csclub.uwaterloo.ca (Lennart Sorensen)

To: 521712@bugs.debian.org

Subject: This really is a bug.

Date: Mon, 30 Mar 2009 19:36:10 -0400

linux-headers-2.6.29-1-686 is missing all the symlinks to
linux-headers-2.6.29-1-common.

For example:
# ls -l /lib/modules/2.6.26-2-686/build/include/linux/agp_backend.h
lrwxrwxrwx 1 root root 66 Mar 30 19:30 /lib/modules/2.6.26-2-686/build/include/linux/agp_backend.h -> ../../../linux-headers-2.6.26-2-common/include/linux/agp_backend.h

# ls -l /lib/modules/2.6.29-1-686/build/include/linux/agp_backend.h
ls: cannot access /lib/modules/2.6.29-1-686/build/include/linux/agp_backend.h: No such file or directory

# ls -l /usr/src/linux-headers-2.6.29-1-common/include/linux/agp_backend.h
-rw-r--r-- 1 root root 3579 Mar 23 19:12 /usr/src/linux-headers-2.6.29-1-common/include/linux/agp_backend.h

This breaks anything that tries to access the kernel headers that the
/lib/modules/<version>/build is supposed to point to.  The linux-headers
for 2.6.29 are severely broken it seems.

Just because someone posts a potential workaround (which really had
nothing to do with it) doesn't mean the original bug isn't in fact a bug.

-- 
Len Sorensen

Message #35 received at 521712@bugs.debian.org (full text, mbox, reply):

From: lsorense@csclub.uwaterloo.ca (Lennart Sorensen)

To: 521712@bugs.debian.org

Subject: More info

Date: Mon, 30 Mar 2009 20:07:00 -0400

OK, so I see now that the symlinks from the past release are no longer
used due to new magic Makefile stuff.  Rather neat.

However as the original bug says, Makefile_32.cpu is in fact missing
from linux-headers-2.6.29-1-common, which breaks i386.  amd64 doesn't
need it and hence works fine.  Copying in the missing file makes
drm-modules compile perfectly using module-assistant.  Without it,
you get an error from make about the missing include for Makefile_32.cpu

-- 
Len Sorensen

Message #40 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Ivan Vershinin <ivan@vershinin.net>

To: 521712@bugs.debian.org

Subject: Makefile_32.cpu

Date: Sun, 5 Apr 2009 15:48:57 +0300

I tried to build madwifi module, using linux-headers-2.6.29-1, and had such error:

/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu: No such
file or directory
make[3]: *** No rule to make target
`/usr/src/linux-headers-2.6.29-1-common/arch/x86/Makefile_32.cpu'. Stop.
make[2]: *** [sub-make] Error 2

Message #45 received at 521712-done@bugs.debian.org (full text, mbox, reply):

From: Daniel Baumann <daniel@debian.org>

To: Ivan Vershinin <ivan@vershinin.net>, 521712-done@bugs.debian.org

Subject: Re: Bug#521712: Makefile_32.cpu

Date: Sun, 05 Apr 2009 15:38:50 +0200

Version: 2.6.29-2

Ivan Vershinin wrote:
> I tried to build madwifi module, using linux-headers-2.6.29-1, and had such error:

this was fixed in 2.6.29-2.

-- 
Address:        Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email:          daniel.baumann@panthera-systems.net
Internet:       http://people.panthera-systems.net/~daniel-baumann/

Message #50 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Ivan Vershinin <ivan@vershinin.net>

To: 521712@bugs.debian.org

Subject: Makefile_32.cpu

Date: Mon, 6 Apr 2009 00:56:06 +0300

But there are still not linux-headers-2.6.29-2 package in repositories.

Message #55 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Michel Loos <loos.michel-external@gemalto.com>

To: <521712@bugs.debian.org>

Date: Wed, 8 Apr 2009 11:56:34 -0300

[Message part 1 (text/plain, inline)]

Still Nothing

Michel.

[Message part 2 (text/html, inline)]

[signature.asc (application/pgp-signature, inline)]

Message #60 received at 521712@bugs.debian.org (full text, mbox, reply):

From: maximilian attems <max@stro.at>

To: Michel Loos <loos.michel-external@gemalto.com>, 521712@bugs.debian.org

Subject: Re: Bug#521712:

Date: Wed, 8 Apr 2009 17:11:14 +0200

On Wed, Apr 08, 2009 at 11:56:34AM -0300, Michel Loos wrote:
> 
> Still Nothing
> 
> Michel.

learn to write a *meaningful* subject next time
if you have anything to say!
otherwise don't waste others time.

Message #65 received at 521712@bugs.debian.org (full text, mbox, reply):

From: MICHEL-EXTERNAL loos <loos.MICHEL-EXTERNAL@gemalto.com>

To: maximilian attems <max@stro.at>

Cc: "521712@bugs.debian.org" <521712@bugs.debian.org>

Subject: RE: Bug#521712:

Date: Wed, 8 Apr 2009 18:08:00 +0200

Could you please provide the missing file!

In the last few weeks we had a LOT of problems with your packages:

dependence on linux-kbuild and now linux-headers
which does not allow us to use the newer kernel.

And every time you give very unhelpfull and unpolite  answers and no solutions



Michel Loos
Integration Engineer LATAM South
Phones: +55 11 5105-7661 / +55 11 3638-7661
Mobile: +55 11 9885-1477
Fax: +55 11 5105-7610
Av. das Nacoes Unidas, 12495 - 8º andar
CEP 04571-011 - Brooklin - São Paulo
loos.michel-external@gemalto.com
www.gemalto.com

________________________________________
From: maximilian attems [max@stro.at]
Sent: Wednesday, April 08, 2009 12:11
To: MICHEL-EXTERNAL loos; 521712@bugs.debian.org
Subject: Re: Bug#521712:

On Wed, Apr 08, 2009 at 11:56:34AM -0300, Michel Loos wrote:
>
> Still Nothing
>
> Michel.

learn to write a *meaningful* subject next time
if you have anything to say!
otherwise don't waste others time.

Message #70 received at 521712@bugs.debian.org (full text, mbox, reply):

From: Daniel Baumann <daniel@debian.org>

To: MICHEL-EXTERNAL loos <loos.MICHEL-EXTERNAL@gemalto.com>, 521712@bugs.debian.org

Cc: maximilian attems <max@stro.at>

Subject: Re: Bug#521712:

Date: Wed, 08 Apr 2009 18:30:47 +0200

MICHEL-EXTERNAL loos wrote:
> Could you please provide the missing file!

could you please calm down? thanks.

> In the last few weeks we had a LOT of problems with your packages:

first of all, the origin of the problem is that the i386 buildd is
fucked up. it build the package but did not upload it yet (to be
precise, it did finally upload it today). you can see this here:
https://buildd.debian.org/pkg.cgi?pkg=linux-2.6

which is linked from: http://packages.qa.debian.org/l/linux-2.6.html

note that running the build architecture is *NOT* the responsiblity of
the kernel team, there is no point in bitching the wrong people about it.

second, linux-2.6 uploads always gets delayed by having them passing NEW
all the time for each release and abi bump. this causes problems with
having linux-2.6 and linux-kbuild-2.6 in sync. however, this is *NOT* a
problem caused by the kernel team and cannot be fixed by us. there is
also no point bitching the wrong people about it.

third, all the packages you want are in svn and tagged accordingly. if
you really depend on having these packages at *any* given point in time,
you an always build it yourself. UTSL, fwiw.

and yes, it really sucks that people are not getting this although it is
always happening the same way all over again and again, and because
people running the *unstable* distribution which is targeted for
*developers* and clueful people only, that are able to help themselfs,
lamenting about it all the time makes no sense, doesn't help, is boring,
and *wastes* our precious time.

Regards,
Daniel

-- 
Address:        Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email:          daniel.baumann@panthera-systems.net
Internet:       http://people.panthera-systems.net/~daniel-baumann/

Message #75 received at 521712@bugs.debian.org (full text, mbox, reply):

From: MICHEL-EXTERNAL loos <loos.MICHEL-EXTERNAL@gemalto.com>

To: "daniel@debian.org" <daniel@debian.org>, "521712@bugs.debian.org" <521712@bugs.debian.org>

Cc: maximilian attems <max@stro.at>

Subject: RE: Bug#521712:

Date: Wed, 8 Apr 2009 19:08:36 +0200

Daniel,

Thanks for your answer

In fact I have no problems with the bug in itself,
I am a develloper and I am running unstable since slink was unstable.
I perfectly understand that some things can get overseen,
some monthes ago, it was mostly amd64 that was with problems last weeks it became more i386,
clearly the maintaners don't test on both those architectures, but again this is not a problem.

I have problems with the kind of answers that Maximilian gave the community
those last few weeks.

I have problems with him closing bugs while the problem is in fact still waiting for a real solution for the end user.


Respectfully,

Michel Loos
Integration Engineer LATAM South
Phones: +55 11 5105-7661 / +55 11 3638-7661
Mobile: +55 11 9885-1477
Fax: +55 11 5105-7610
Av. das Nacoes Unidas, 12495 - 8º andar
CEP 04571-011 - Brooklin - São Paulo
loos.michel-external@gemalto.com
www.gemalto.com

________________________________________
From: Daniel Baumann [daniel@debian.org]
Sent: Wednesday, April 08, 2009 13:30
To: MICHEL-EXTERNAL loos; 521712@bugs.debian.org
Cc: maximilian attems
Subject: Re: Bug#521712:

MICHEL-EXTERNAL loos wrote:
> Could you please provide the missing file!

could you please calm down? thanks.

> In the last few weeks we had a LOT of problems with your packages:

first of all, the origin of the problem is that the i386 buildd is
fucked up. it build the package but did not upload it yet (to be
precise, it did finally upload it today). you can see this here:
https://buildd.debian.org/pkg.cgi?pkg=linux-2.6

which is linked from: http://packages.qa.debian.org/l/linux-2.6.html

note that running the build architecture is *NOT* the responsiblity of
the kernel team, there is no point in bitching the wrong people about it.

second, linux-2.6 uploads always gets delayed by having them passing NEW
all the time for each release and abi bump. this causes problems with
having linux-2.6 and linux-kbuild-2.6 in sync. however, this is *NOT* a
problem caused by the kernel team and cannot be fixed by us. there is
also no point bitching the wrong people about it.

third, all the packages you want are in svn and tagged accordingly. if
you really depend on having these packages at *any* given point in time,
you an always build it yourself. UTSL, fwiw.

and yes, it really sucks that people are not getting this although it is
always happening the same way all over again and again, and because
people running the *unstable* distribution which is targeted for
*developers* and clueful people only, that are able to help themselfs,
lamenting about it all the time makes no sense, doesn't help, is boring,
and *wastes* our precious time.

Regards,
Daniel

--
Address:        Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email:          daniel.baumann@panthera-systems.net
Internet:       http://people.panthera-systems.net/~daniel-baumann/

Message #80 received at 521712@bugs.debian.org (full text, mbox, reply):

From: maximilian attems <max@stro.at>

To: MICHEL-EXTERNAL loos <loos.MICHEL-EXTERNAL@gemalto.com>

Cc: "daniel@debian.org" <daniel@debian.org>, "521712@bugs.debian.org" <521712@bugs.debian.org>

Subject: Re: Bug#521712:

Date: Thu, 9 Apr 2009 12:20:19 +0200

On Wed, Apr 08, 2009 at 07:08:36PM +0200, MICHEL-EXTERNAL loos wrote:
> In fact I have no problems with the bug in itself,
> I am a develloper and I am running unstable since slink was unstable.
> I perfectly understand that some things can get overseen,
> some monthes ago, it was mostly amd64 that was with problems last weeks it became more i386,
> clearly the maintaners don't test on both those architectures, but again this is not a problem.

this is pretty wrong.

a trouble on the buildd has nothing to do with testing.
yes i test every image on x86_32 and x86_64 in our lab.

due to good buying politics hardware in our institute doesn't depend
on external modules. i haven't touched linux-kbuild and don't plan.
you are blaming the wrong guy.

Send a report that this bug log contains spam.