ahh, I did it again...
---------- Forwarded message ----------
Date: Wed, 23 Aug 2000 19:04:38 +0100 (BST)
From: Tigran Aivazian <tigran@veritas.com>
To: Linus Torvalds <torvalds@transmeta.com>
Cc: linux-kernel@vger.rutgers.edu
Subject: [patch-2.4.0-test7-pre7] do_fork() optimization.
Hi Linus,
I was thinking if I could get a few more cycles out of do_fork() (the
recent SCO propaganda about their _lwp_create(2) being faster than our
clone(2) made me think). And I noticed that get_pid() doesn't actually
need the lastpid_lock because it is only ever called from do_fork() inside
the lock_kernel(). So, here is the obvious patch.
Regards.
Tigran
--- linux/kernel/fork.c Wed Aug 23 08:04:33 2000
+++ work/kernel/fork.c Wed Aug 23 19:01:04 2000
@@ -74,9 +74,6 @@
init_task.rlim[RLIMIT_NPROC].rlim_max = max_threads/2;
}
-/* Protects next_safe and last_pid. */
-spinlock_t lastpid_lock = SPIN_LOCK_UNLOCKED;
-
static int get_pid(unsigned long flags)
{
static int next_safe = PID_MAX;
@@ -85,7 +82,6 @@
if (flags & CLONE_PID)
return current->pid;
- spin_lock(&lastpid_lock);
if((++last_pid) & 0xffff8000) {
last_pid = 300; /* Skip daemons etc. */
goto inside;
@@ -115,8 +111,6 @@
}
read_unlock(&tasklist_lock);
}
- spin_unlock(&lastpid_lock);
-
return last_pid;
}
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Aug 23 2000 - 21:00:09 EST