Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: threaded programs report bogus process size ?

Hello, (Christos Zoulas) wrote:
> >Patch is wrong.  Apart from unnecessary re-lock in the loop and lock leak
> >after - consider the case when 'first_thread' after re-insertion into the
> >queue gets taken out by other pthread_create().  Infinite loop.
> Here's a simpler one.

Looking from tail is not good for cache, I think PTQ_FIRST() should be
attempted first.  Otherwise, loop looks OK to me.

Though _lwp_kill() with lock held might have small effect to concurrency,
but it's very unlikely case, especially if do unlocked check for PTQ_FIRST().


Home | Main Index | Thread Index | Old Index