On 03/21/2011 08:48 PM, Michael Tokarev wrote:
21.03.2011 20:12, Eric Dumazet wrote: > Le lundi 21 mars 2011 Ã 19:02 +0200, Avi Kivity a Ãcrit : > >> Any ideas on how to fix it? We could pre-allocate IDs and batch them in >> per-cpu caches, but it seems like a lot of work. >> > > Hmm, I dont know what syscalls kvm do, but even a timer_gettime() has to > lock this idr_lock. I wonder why my testcase - 160 kvm guests on a 2-core AMD AthlonII CPU - shows no contention of that sort, not at all... Yes sure, all these (linux) guests are idling at login: prompt, but I'd expect at least some gettimeofday() and some load average, -- everything stays close to zero.
For this workload spinlock contention only happens on large machines; it (mostly) doesn't matter how many guests you have on a small host.
-- I have a truly marvellous patch that fixes the bug which this signature is too narrow to contain. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
