Re: [PATCH] IPV4 : Move ip route cache flush (secret_rebuild) from softirq to workqueue

[Eric Dumazet <dada1@cosmosbay.com>]

Simon Horman a écrit :
> Is it ever neccessary to call cond_resched() if rt_hash_table[i].chain
> is NULL? If not, the following looks cleaner to my eyes:
> 
> 	for (i = 0; i <= rt_hash_mask; i++) {
> 		rth = rt_hash_table[i].chain;
> 		if (!rth)
> 			continue;
> 
> 		spin_lock_bh(rt_hash_lock_addr(i));
> 		rth = rt_hash_table[i].chain;
> 		rt_hash_table[i].chain = NULL;
> 		spin_unlock_bh(rt_hash_lock_addr(i));
> 
> 		...
> 

Well, it depends if the table is empty or not, and its size.

Some RT tasks want to to be wakeup every ms for example.

With a 8MB hash table (one million slots on x86_64), the full scan is longer 
than 1 ms, even if no IRQ came and stole the cpu...