Messages in this thread |  | | | Date | Fri, 22 Aug 2008 12:52:26 -0700 | | From | "Paul E. McKenney" <> | | Subject | Re: [PATCH 2/2] smp_call_function: use rwlocks on queues rather than rcu |
| |
On Fri, Aug 22, 2008 at 01:36:37PM -0500, Christoph Lameter wrote: > Paul E. McKenney wrote: > > >>> So on your these large boxes, read-only cachelines are preferentially > >>> ejected from the cache, so that one should write to per-CPU data > >>> occasionally to keep it resident? Or is the issue the long RCU grace > >>> periods which allow the structure being freed to age out of all relevant > >>> caches? (My guess would be the second.) > >> The issue are the RCU grace period that are generally long enough to make the > >> cacheline fall out of all caches. > > > > Would it make sense to push the freed-by-RCU memory further up the > > hierarchy, so that such memory is not mistaken for recently freed > > hot-in-cache memory? > > That would mean passing a gfp flag like __GFP_COLD on free from RCU? Or how > would that work at the higher levels?
I was indeed thinking in terms of the free from RCU being specially marked.
Thanx, Paul
|  |