Introduce mod_timer_noact() which for example is to replace the calls to del_timer()/add_timer() in __nf_ct_refresh_acct(). It works like mod_timer() but doesn't activate or modify the timeout of an inactive timer which is the behaviour we want in order to be able to use timers as a means of synchronization in nf_conntrack. A later patch will modify __nf_ct_refresh_acct() to use mod_timer_noact() which will then save one spin_lock_irqsave() / spin_lock_irqrestore() pair per conntrack timer update. This will also get rid of the race we currently have without adding more locking in nf_conntrack. Signed-off-by: Martin Josefsson <gandalf@wlug.westbo.se> --- include/linux/timer.h | 8 ++++++-- kernel/relay.c | 2 +- kernel/timer.c | 40 +++++++++++++++++++++++++++++++++++----- 3 files changed, 42 insertions(+), 8 deletions(-) --- a/include/linux/timer.h 2009-02-17 10:55:33.427785986 -0800 +++ b/include/linux/timer.h 2009-02-17 11:04:10.291844534 -0800 @@ -25,6 +25,9 @@ struct timer_list { extern struct tvec_base boot_tvec_bases; +#define TIMER_ACT 1 +#define TIMER_NOACT 0 + #define TIMER_INITIALIZER(_function, _expires, _data) { \ .entry = { .prev = TIMER_ENTRY_STATIC }, \ .function = (_function), \ @@ -86,8 +89,9 @@ static inline int timer_pending(const st extern void add_timer_on(struct timer_list *timer, int cpu); extern int del_timer(struct timer_list * timer); -extern int __mod_timer(struct timer_list *timer, unsigned long expires); +extern int __mod_timer(struct timer_list *timer, unsigned long expires, int activate); extern int mod_timer(struct timer_list *timer, unsigned long expires); +extern int mod_timer_noact(struct timer_list *timer, unsigned long expires); /* * The jiffies value which is added to now, when there is no timer @@ -163,7 +167,7 @@ static inline void timer_stats_timer_cle static inline void add_timer(struct timer_list *timer) { BUG_ON(timer_pending(timer)); - __mod_timer(timer, timer->expires); + __mod_timer(timer, timer->expires, TIMER_ACT); } #ifdef CONFIG_SMP --- a/kernel/timer.c 2009-02-17 10:55:33.403580297 -0800 +++ b/kernel/timer.c 2009-02-17 11:04:10.291844534 -0800 @@ -589,7 +589,7 @@ static struct tvec_base *lock_timer_base } } -int __mod_timer(struct timer_list *timer, unsigned long expires) +int __mod_timer(struct timer_list *timer, unsigned long expires, int activate) { struct tvec_base *base, *new_base; unsigned long flags; @@ -603,7 +603,8 @@ int __mod_timer(struct timer_list *timer if (timer_pending(timer)) { detach_timer(timer, 0); ret = 1; - } + } else if (activate == TIMER_NOACT) + goto out_unlock; debug_timer_activate(timer); @@ -629,8 +630,9 @@ int __mod_timer(struct timer_list *timer timer->expires = expires; internal_add_timer(base, timer); - spin_unlock_irqrestore(&base->lock, flags); +out_unlock: + spin_unlock_irqrestore(&base->lock, flags); return ret; } @@ -699,11 +701,39 @@ int mod_timer(struct timer_list *timer, if (timer->expires == expires && timer_pending(timer)) return 1; - return __mod_timer(timer, expires); + return __mod_timer(timer, expires, TIMER_ACT); } EXPORT_SYMBOL(mod_timer); +/*** + * mod_timer_noact - modify a timer's timeout + * @timer: the timer to be modified + * + * mod_timer_noact works like mod_timer except that it doesn't activate an + * inactive timer, instead it returns without updating timer->expires. + * + * The function returns whether it has modified a pending timer or not. + * (ie. mod_timer_noact() of an inactive timer returns 0, mod_timer_noact() of + * an active timer returns 1.) + */ +int mod_timer_noact(struct timer_list *timer, unsigned long expires) +{ + BUG_ON(!timer->function); + + /* + * This is a common optimization triggered by the + * networking code - if the timer is re-modified + * to be the same thing then just return: + */ + if (timer->expires == expires && timer_pending(timer)) + return 1; + + return __mod_timer(timer, expires, TIMER_NOACT); +} + +EXPORT_SYMBOL(mod_timer_noact); + /** * del_timer - deactive a timer. * @timer: the timer to be deactivated @@ -1268,7 +1298,7 @@ signed long __sched schedule_timeout(sig expire = timeout + jiffies; setup_timer_on_stack(&timer, process_timeout, (unsigned long)current); - __mod_timer(&timer, expire); + __mod_timer(&timer, expire, TIMER_ACT); schedule(); del_singleshot_timer_sync(&timer); --- a/kernel/relay.c 2009-02-17 10:55:33.416279439 -0800 +++ b/kernel/relay.c 2009-02-17 11:04:10.291844534 -0800 @@ -750,7 +750,7 @@ size_t relay_switch_subbuf(struct rchan_ * from the scheduler (trying to re-grab * rq->lock), so defer it. */ - __mod_timer(&buf->timer, jiffies + 1); + __mod_timer(&buf->timer, jiffies + 1, TIMER_NOACT); } old = buf->data; -- -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html