Question on glibc: free() API

[Tvrtko Ursulin]

On Wednesday 04 March 2009 15:39:04 Carlos O'Donell wrote:
> On Tue, Mar 3, 2009 at 7:12 PM, Bizhan Gholikhamseh (bgholikh)
>
> <bgholikh@cisco.com> wrote:
> > We see the suspended thread block for a long time (10 to 100 msec), is
> > this acceptable, is this scheduler problem? This is impacting our system
> > performance. Any idea how to work aroung this?
>
> Yes, upgrade to an NTPL enabled pthreads.
>
> The use of pthread_wait_for_restart_signal(...) in your pseudo-trace
> suggests you are still using linuxthreads.

Hi,

Please excuse me with jumping in but I have a related question.

If one had a dynamic executable built on a LinuxThreads system (glibc without 
NPTL support), will it have this problem when ran in a modern glibc 
environemnt? Or will it need a recompile against NPTL glibc to benefit from 
all changes?

Thanks,

Tvrtko