Re: [PATCH] Fix a race in pid generation that causes pids to bereused immediately.

[Peter Zijlstra]

On Mon, 2010-06-14 at 18:55 -0700, Andrew Morton wrote:
> > kernel/sched_clock.c:   if (cmpxchg64(&scd->clock, old_clock, clock) != old_cloc
> 
> I guess that'll flush out any stragglers.

cmpxchg64() is special, at the time i386 didn't handle the 8 byte
cmpxchg(), although we could easily make it do today.

> I suspect sched_clock.c might be generating fair amounts of code which
> UP builds don't need. 

Only sched_clock_remote() and its caller, something like the below, not
much code..

UP machines can still have utterly sucky TSC, although the
inter-cpu-drift thing isn't much of an issue ;-)

---
 kernel/sched_clock.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/kernel/sched_clock.c b/kernel/sched_clock.c
index 52f1a14..7ff5b56 100644
--- a/kernel/sched_clock.c
+++ b/kernel/sched_clock.c
@@ -170,6 +170,7 @@ again:
 	return clock;
 }
 
+#ifdef CONFIG_SMP
 static u64 sched_clock_remote(struct sched_clock_data *scd)
 {
 	struct sched_clock_data *my_scd = this_scd();
@@ -205,6 +206,7 @@ again:
 
 	return val;
 }
+#endif
 
 /*
  * Similar to cpu_clock(), but requires local IRQs to be disabled.
@@ -226,9 +228,11 @@ u64 sched_clock_cpu(int cpu)
 
 	scd = cpu_sdc(cpu);
 
+#ifdef CONFIG_SMP
 	if (cpu != smp_processor_id())
 		clock = sched_clock_remote(scd);
 	else
+#endif
 		clock = sched_clock_local(scd);
 
 	return clock;

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/