I don't see anything wrong on the surface.<br>You might rebuild your eglibc with the code at _dl_important_hwcaps instrumented for debug.<br>Some kind of minimal logging could be added. <br>Anything to give you more visiblity as to what data sequence leads to the bug.<br>
<br>The log data could be written to a 'shared' mmap area in memory. A separate program could mmap the same area and dump the log for you after the segfault occurs. The log buffer should 'wrap' so you don't end up with too much data.<br>
<br>Rex<br><br><br><div class="gmail_quote">On Tue, Jun 22, 2010 at 11:59 AM, Enrico Scholz <span dir="ltr"><<a href="mailto:enrico.scholz@sigma-chemnitz.de" target="_blank">enrico.scholz@sigma-chemnitz.de</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div>Rex Ashbaugh <<a href="mailto:rexa@xeratech.com" target="_blank">rexa@xeratech.com</a>> writes:<br>
<br>
> Do you have IWMMXT support enabled in your kernel config?<br>
<br>
</div>yes; the mentioned tarball<br>
<div><br>
>> [3] <a href="https://www.cvg.de/people/ensc/iwmmxt-ld.tar.gz" target="_blank">https://www.cvg.de/people/ensc/iwmmxt-ld.tar.gz</a><br>
<br>
</div>contains full .config (and some extra information).<br>
<font color="#888888"><br>
<br>
Enrico<br>
</font></blockquote></div><br>