SIGSEGV after infinite recursion in __vfprintf_chk?

[Luca Ceresoli]

Yann E. MORIN wrote:
> Care to send your .config, please?

Attached.

I did the tests with 2 configs, apparently with the same symptom.

The first is arm-cortex_a8-linux-gnueabi with a few (supposedly
irrelevant) changes: prefix dir, number of parallel jobs and
tuple's alias.

The second has a few changes, most notably I set the kernel headers
version to 2.6.32 since they are closer to the kernel I'm using
(http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git
at 4fa42e4679).

> > So I wrote a minimal C program
> 
> Care to send your test program, please ?

Attached, along with the Makefile.

> Basically, __vfprintf_chk is called in place of vfprintf. That's a linker
> trick, where vfprintf is in fact an alias to __vfprintf_chk. Its purpose
> is to check for stack overflow (I don't dare look at how it does so).
> 
> The gst-doc package has had the same issue, and they fixed their code:
> https://bugs.launchpad.net/ubuntu/+source/gnu-smalltalk/+bug/410263

Thanks for the links. Anyway I'm not doing dirty tricks as passing name[-1]
as gst-doc used to, so perhaps my problem is not the same.

Thanks,
Luca

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ct-ng-configs.tar.bz2
Type: application/x-bzip
Size: 3413 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/crossgcc/attachments/20100226/30a74fa2/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: helloworld.tar.bz2
Type: application/x-bzip
Size: 578 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/crossgcc/attachments/20100226/30a74fa2/attachment-0001.bin>
-------------- next part --------------
--
For unsubscribe information see http://sourceware.org/lists.html#faq