Svante Signell wrote:

> ON, now I have made the bisection:
>
> 2.6.33-1~experimental.5 works OK, however with the address space
> collision warning reported earlier.
>
> 2.6.34-1~experimental.1 shows the same problems ar later kernels.
>
> Do you need dmesg output or anything else to make progress?

Would it be possible to bisect between those versions, to find the
exact change that introduced this problem? It works like this:

$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
$ cd linux-2.6
$ git bisect start v2.6.34 v2.6.33

Git checks out a version half-way between. So:

$ make localmodconfig; # minimal configuration
$ make deb-pkg; # builds a .deb
# dpkg -i ../<whatever>.deb; reboot
$ git bisect good; # if it works
$ git bisect bad; # if it shows the same problem
$ git bisect skip; # if another problem makes it impossible to test

Then:
$ make silentoldconfig; # reuse configuration
$ make deb-pkg; # build a .deb
# dpkg -i ../<whatever>.deb; reboot
$ git bisect good / bad / skip
... rinse and repeat until it shows the "first bad commit" ...

If the gitk package is installed, you can use "git bisect visualize"
at any point to see the regression range narrowing. The point of this
exercise is to find out what subsystem in particular broke. For more
information on "git bisect", see [1] if interested.

Thanks for reporting, and sorry for the trouble.
Jonathan

[1] http://www.kernel.org/pub/software/scm/git/docs/git-bisect-lk2009.html



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110713163347.GA747@elie">http://lists.debian.org/20110713163347.GA747@elie

On Wed, 2011-07-13 at 11:33 -0500, Jonathan Nieder wrote:
> $ make deb-pkg; # builds a .deb

Build fails: gcc-4.6, 4.5, 4.4.
/tmp/ccY6xbIg.s: Error: .size expression for do_hypervisor_callback does
not evaluate to a constant

gcc-4.2 and 4.1:
/usr/bin/ld: cannot find crt1.o: No such file or directory
/usr/bin/ld: cannot find crti.o: No such file or directory
/usr/bin/ld: cannot find -lgcc_s




--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 1310632628.2611.1.camel@amd64.my.own.domain">http://lists.debian.org/1310632628.2611.1.camel@amd64.my.own.domain

On Thu, 2011-07-14 at 10:37 +0200, Svante Signell wrote:
> On Wed, 2011-07-13 at 11:33 -0500, Jonathan Nieder wrote:
> > $ make deb-pkg; # builds a .deb
>
> Build fails: gcc-4.6, 4.5, 4.4.
> /tmp/ccY6xbIg.s: Error: .size expression for do_hypervisor_callback does
> not evaluate to a constant

Solved by: .config
CONFIG_XEN=y->n

Next problem:
CC arch/x86/vdso/vgetcpu.o
CC arch/x86/vdso/vvar.o
VDSO arch/x86/vdso/vdso.so.dbg
gcc: error: elf_x86_64: No such file or directory
make[4]: *** [arch/x86/vdso/vdso.so.dbg] Error 1





--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 1310633939.2611.3.camel@amd64.my.own.domain">http://lists.debian.org/1310633939.2611.3.camel@amd64.my.own.domain

Svante Signell wrote:

> Next problem:
> CC arch/x86/vdso/vgetcpu.o
> CC arch/x86/vdso/vvar.o
> VDSO arch/x86/vdso/vdso.so.dbg
> gcc: error: elf_x86_64: No such file or directory
> make[4]: *** [arch/x86/vdso/vdso.so.dbg] Error 1

Yeah, you'll want "git cherry-pick --no-commit de2a8cf9" in order to
fix that one. Sorry for the fuss.



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110714090118.GB15271@elie">http://lists.debian.org/20110714090118.GB15271@elie

On Thu, 2011-07-14 at 04:01 -0500, Jonathan Nieder wrote:
> Svante Signell wrote:
>
> > Next problem:
> > CC arch/x86/vdso/vgetcpu.o
> > CC arch/x86/vdso/vvar.o
> > VDSO arch/x86/vdso/vdso.so.dbg
> > gcc: error: elf_x86_64: No such file or directory
> > make[4]: *** [arch/x86/vdso/vdso.so.dbg] Error 1
>
> Yeah, you'll want "git cherry-pick --no-commit de2a8cf9" in order to
> fix that one. Sorry for the fuss.

Now the bisect is finished, after building 12 kernels, rebooting, etc.
The bad commit is (phew):

Bisecting: 0 revisions left to test after this (roughly 0 steps)
[b74fd238a9cf39a81d94152f375b756bf795b4af] x86/pci: Use resource_size_t
in update_res

Hope you can help fixing this problem, both in Debian 2.6.39 kernels,
and upstream.

Thanks for your help!




--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 1310658668.3021.3.camel@amd64.my.own.domain">http://lists.debian.org/1310658668.3021.3.camel@amd64.my.own.domain