* [PATCH] timer: optimize apply_slack()
@ 2012-03-17 21:31 Felix Fietkau
2012-03-17 22:06 ` Eric Dumazet
0 siblings, 1 reply; 2+ messages in thread
From: Felix Fietkau @ 2012-03-17 21:31 UTC (permalink / raw)
To: linux-kernel; +Cc: tglx
__fls(mask) is equivalent to find_last_bit(&mask, BITS_PER_LONG), but cheaper.
find_last_bit was showing up high on the list when I was profiling for stalls
on icache misses on a system with very small cache size (MIPS).
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
---
kernel/timer.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/kernel/timer.c b/kernel/timer.c
index a297ffc..659371e 100644
--- a/kernel/timer.c
+++ b/kernel/timer.c
@@ -809,7 +809,7 @@ unsigned long apply_slack(struct timer_list *timer, unsigned long expires)
if (mask == 0)
return expires;
- bit = find_last_bit(&mask, BITS_PER_LONG);
+ bit = __fls(mask);
mask = (1 << bit) - 1;
--
1.7.3.2
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] timer: optimize apply_slack()
2012-03-17 21:31 [PATCH] timer: optimize apply_slack() Felix Fietkau
@ 2012-03-17 22:06 ` Eric Dumazet
0 siblings, 0 replies; 2+ messages in thread
From: Eric Dumazet @ 2012-03-17 22:06 UTC (permalink / raw)
To: Felix Fietkau; +Cc: linux-kernel, tglx
On Sat, 2012-03-17 at 22:31 +0100, Felix Fietkau wrote:
> __fls(mask) is equivalent to find_last_bit(&mask, BITS_PER_LONG), but cheaper.
> find_last_bit was showing up high on the list when I was profiling for stalls
> on icache misses on a system with very small cache size (MIPS).
>
> Signed-off-by: Felix Fietkau <nbd@openwrt.org>
> ---
> kernel/timer.c | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/kernel/timer.c b/kernel/timer.c
> index a297ffc..659371e 100644
> --- a/kernel/timer.c
> +++ b/kernel/timer.c
> @@ -809,7 +809,7 @@ unsigned long apply_slack(struct timer_list *timer, unsigned long expires)
> if (mask == 0)
> return expires;
>
> - bit = find_last_bit(&mask, BITS_PER_LONG);
> + bit = __fls(mask);
>
> mask = (1 << bit) - 1;
>
Not a comment on your patch, but this code seems not correct.
This probably should be :
mask = (1UL << bit) - 1;
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2012-03-17 22:07 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-03-17 21:31 [PATCH] timer: optimize apply_slack() Felix Fietkau
2012-03-17 22:06 ` Eric Dumazet
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®