rcu: increment quiescent state counter in ksoftirqd()
authorEric Dumazet <dada1@cosmosbay.com>
Wed, 4 Mar 2009 20:11:56 +0000 (12:11 -0800)
committerIngo Molnar <mingo@elte.hu>
Wed, 4 Mar 2009 21:08:45 +0000 (22:08 +0100)
commit64ca5ab913f1594ef316556e65f5eae63ff50cee
tree44cecee89609fbec9e4d7ed853b0eaf5c3782fc3
parent559595a985e106d2fa9f0c79b7f5805453fed593
rcu: increment quiescent state counter in ksoftirqd()

If a machine is flooded by network frames, a cpu can loop
100% of its time inside ksoftirqd() without calling schedule().
This can delay RCU grace period to insane values.

Adding rcu_qsctr_inc() call in ksoftirqd() solves this problem.

Paul: "This regression was a result of the recent change from
"schedule()" to "cond_resched()", which got rid of that quiescent
state in the common case where a reschedule is not needed".

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
kernel/softirq.c