Re: [PATCH] simplify update_times (avoid jiffies/jiffies_64aliasing problem)

[Atsushi Nemoto]

>>>>> On Fri, 03 Mar 2006 12:17:28 -0800, john stultz <johnstul@xxxxxxxxxx> said:

john> I'm not opposed to queuing it up as it seems like a logical
john> cleanup. I'd be fine with it going in before my patch, however
john> it still needs to address i386 lost tick compensation.  I worry
john> that addressing that issue before my patchset (which makes the
john> lost tick compensation unnecessary) might be a bit more
john> complex. I think it would be easier going in after my patch. I
john> do think the barrier fix (with a comment) is a good short term
john> fix.

john> Atsushi: Your thoughts?

I agree.  I missed i386 lost tick case and it seems more complex than
x86_64 case.  Your patchset looks to make this cleanup very easy.
Then, here is an updated barrier fix patch.


Add an optimization barrier to prevent prefetching jiffies before
incrementing jiffies_64.

Signed-off-by: Atsushi Nemoto <anemo@xxxxxxxxxxxxx>

diff --git a/kernel/timer.c b/kernel/timer.c
index fc6646f..decd19e 100644
--- a/kernel/timer.c
+++ b/kernel/timer.c
@@ -925,6 +925,8 @@ static inline void update_times(void)
 void do_timer(struct pt_regs *regs)
 {
 	jiffies_64++;
+	/* prevent loading jiffies before storing new jiffies_64 value. */
+	barrier();
 	update_times();
 	softlockup_tick(regs);
 }
-
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/