From: Andrea A. <an...@su...> - 2002-07-29 20:24:34
|
On Mon, Jul 29, 2002 at 02:57:46PM -0500, Paul Larson wrote: > This is a fix for the problem where get_pid will hang the machine if you > request a new pid when all available pids are in use. This also adds > the appropriate checking for p->tgid in get_pid that was somehow > overlooked before. This patch has been in 2.4 since 2.4.19-pre9. no please, this below patch is superior and it fixes as well fixes a longstanding race condition in fork() (see the semaphore): http://www.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.19rc3aa3/10_get_pid-no-deadlock-and-boosted-4 it may not apply cleanly but porting should be trivial. Andrea |