Messages in this thread | | | Date | Tue, 23 Nov 2010 21:13:33 -0600 (CST) | From | Christoph Lameter <> | Subject | Re: [thiscpuops upgrade 10/10] Lockless (and preemptless) fastpaths for slub |
| |
On Wed, 24 Nov 2010, Eric Dumazet wrote:
> Hmm, this only works for power of two NR_CPUS, or else one cpu 'tid' > could wrap on another cpu tid.
Hmm... That essentially like using a bitfield then.
> I suggest using 4096 (or roundup_pow_of_two(CONFIG_NR_CPUS))
There could be machines with 64k cpus soon. So I guess roundupp2 it will have to be.
| |