Re: Segfault on syscall

On 1/2/07, Florian Delizy <[email protected]> wrote:
> Shuveb Hussain wrote:
> > Hi,
> >
> >> Well, it's not, but as I said there still is this bug in the syscalls
> >> thing, and since this is a 'random'  bug, I need as much full detailed
> >> bug report (see the wiki :
> >> http://howto.x-tend.be/openMosixWiki/index.php/OpenMosix2.6#Bug_Reporting
> >>
> >> ) to solve it faster
> >
> > Doesn't seem to be random. I have a simple program that does  a long
> > loop followed by an exit(). The loop gives me time to initiate
> > migration. Once the loop completes on remote and does exit(), it hits
> > a seg fault. I fugured the address of the segfault from the page fault
> > handler in mm/fault.c, it seems to be coming from the C library pages
> > that are copied.
> >
> > main()
> > {
> >    int a,b;
> >
> >    for(a=0;a<100000;a++)
> >        for(b=0;b<100000;b++);
> >
> >    // compiler inserts exit() here.
> > }
> >
> > The interesting thing is that this program does not segfault when
> > compiled with '-static'. This fact confused me, since we are copying
> > the process memory pages verbatim. Certain system calls always
> > segfault, like exit(), nanosleep(), etc.
> >
> > My setup is 2 QEMUs running the Dec 22 release with TUN/TAP and
> > etherbridge setup for netwoking between them. The QEMUs are running
> > Debian Etch. Hope this is a good enough setup for debugging, or do I
> > need 2 real machines?
> >
>
>
> Well, it sounds like a problem we had, what arch are you running ?
>

I'm running QEMU 0.8.0 and OpenMosix 2.6.17 with the Dec 22 patch.
Arch is x86(32 bit).

-- 
Shuveb Hussain.
I Blog at http://binarykarma.org
Spread the Karma.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV