Messages in this thread | | | Subject | Re: [PATCH 2/2]x86: spread tlb flush vector between nodes | From | Shaohua Li <> | Date | Wed, 20 Oct 2010 16:44:22 +0800 |
| |
On Wed, 2010-10-20 at 15:30 +0800, Andi Kleen wrote: > Hi Shaohua, > > > + if (nr_online_nodes > NUM_INVALIDATE_TLB_VECTORS) > > + nr_node_vecs = 1; > > + else > > + nr_node_vecs = NUM_INVALIDATE_TLB_VECTORS/nr_online_nodes; > > Does this build without CONFIG_NUMA? AFAIK nr_online_nodes is only > defined for a numa kernel. yes it's ok without CONFIG_NUMA
> > + > > +static int tlb_cpuhp_notify(struct notifier_block *n, > > + unsigned long action, void *hcpu) > > +{ > > + switch (action & 0xf) { > > + case CPU_ONLINE: > > + case CPU_DEAD: > > + calculate_tlb_offset(); > > I still think the notifier is overkill and a static mapping at boot time > would be fine. this hasn't much overhead, so please keep it. I'm afraid static mapping calculation for hotplug cpu will introduce more complex.
| |