Re: [RFC] vmstat: Avoid waking up idle-cpu to service shepherd work

[Viresh Kumar]

On 30 March 2015 at 19:29, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> Yeah, so that _should_ not trigger (obviously), and while I agree with
> the sentiment of sanity checks, I'm not sure its worth keeping that
> variable around just for that.

I read it as I can remove it then ? :)

> Anyway, while I'm looking at struct tvec_base I notice the cpu member
> should be second after the lock, that'll save 8 bytes on the structure
> on 64bit machines.

Hmm, I tried it on my macbook-pro.

$ uname -a
Linux vireshk 3.13.0-46-generic #79-Ubuntu SMP Tue Mar 10 20:06:50 UTC
2015 x86_64 x86_64 x86_64 GNU/Linux

$ gcc --version
gcc (Ubuntu 4.8.2-19ubuntu1) 4.8.2

config: arch/x86/configs/x86_64_defconfig

And all I get it is 8256 bytes, with or without the change.

diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 2d3f5c504939..afc5d74678df 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -77,12 +77,12 @@ struct tvec_root {

 struct tvec_base {
        spinlock_t lock;
+       int cpu;
        struct timer_list *running_timer;
        unsigned long timer_jiffies;
        unsigned long next_timer;
        unsigned long active_timers;
        unsigned long all_timers;
-       int cpu;
        struct tvec_root tv1;
        struct tvec tv2;
        struct tvec tv3;
--
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/