Messages in this thread | | | Date | Mon, 06 Jan 2014 18:41:25 +0000 | From | Dietmar Eggemann <> | Subject | Re: [RFC PATCH 8/8] sched: remove scheduler domain naming |
| |
On 20/12/13 14:08, Peter Zijlstra wrote: > On Fri, Dec 13, 2013 at 12:11:28PM +0000, dietmar.eggemann@arm.com wrote: >> From: Dietmar Eggemann <dietmar.eggemann@arm.com> >> >> In case the arch is allowed to define the conventional scheduler domain >> topology level (i.e. the one without SD_NUMA topology flag) layout, it is >> not feasible any more for the scheduler to name these levels. Therefore, >> this patch gets rid of of the sched_domain_topology_level structure >> member 'name' and the corresponding SD_INIT_NAME macro. It was only used >> when CONFIG_SCHED_DEBUG was set any way. > > Right, so for debug purposes it might be convenient to keep it; we could > simply put it in the topology array, something like: > > { cpu_smt_mask, SD_SHARE_CPU_POWER | SD_SHARE_PKG_RESOURCE, SD_NAME(smt) }, > > which would still allow us to make it go away on !debug, but does > provide us with a nice label to print for the debug topology prints.
I will incorporate this idea in my V2 patch set.
> > Alternatively we could do something like: > > #define SD_mask(name, flags) \ > { cpu_##name##_mask, (flags), .name = #name } > > to further reduce typing.
But the MC level cpu mask func ptr is called cpu_coregroup_mask.
-- Dietmar
>
| |