This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: SIGSTOP weirdness
Sorry, forgot to mention the platform -- it's a stock Fedora 3 kernel,
x86, 2.6.11 series. I've also reproduced the same on much older Linux -
Redhat 9 and 7.3..
Cheers,
David
On Wed, 2005-06-01 at 17:41 +0100, David Lecomber wrote:
> Dear all,
>
> I am experiencing a problem with SIGSTOP and SIGCONT - when sent in
> large doses to a debugged program:
>
> Take for example the program:
>
> int main() {
> while (-1) ;
> }
>
> In a fairly recent CVS GDB (20050310), I set
>
> (gdb) handle SIGSTOP nostop noprint pass
> Signal Stop Print Pass to program Description
> SIGSTOP No No Yes Stopped (signal)
> (gdb) handle SIGCONT nostop noprint pass
> Signal Stop Print Pass to program Description
> SIGCONT No No Yes Continued
> (gdb) r
>
> Which means GDB should pretty much ignore them.. but look what happens
> when many signals are sent from the command line to the target process:
>
> while (true ) ; do kill -SIGSTOP 31639; kill -SIGCONT 31639; done
>
> We get this GDB error:
>
> Couldn't get registers: No such process.
> (gdb)
>
>
> Anyone know why?
>
> Regards
> David
--
David Lecomber <david@allinea.com>