On 10/17/2016 01:53 AM, David Gibson wrote: > On Fri, Oct 14, 2016 at 09:35:46AM +0200, Cédric Le Goater wrote: >> On 10/14/2016 07:32 AM, David Gibson wrote: >>> On Mon, Oct 03, 2016 at 09:24:46AM +0200, Cédric Le Goater wrote: >>>> From: Benjamin Herrenschmidt <b...@kernel.crashing.org> >>>> >>>> Instead of an array of fixed sized blocks, use a list, as we will need >>>> to have sources with variable number of interrupts. SPAPR only uses >>>> a single entry. Native will create more. If performance becomes an >>>> issue we can add some hashed lookup but for now this will do fine. >>>> >>>> Signed-off-by: Benjamin Herrenschmidt <b...@kernel.crashing.org> >>>> [ move the initialization of list to xics_common_initfn, >>>> restore xirr_owner after migration and move restoring to >>>> icp_post_load] >>>> Signed-off-by: Nikunj A Dadhania <nik...@linux.vnet.ibm.com> >>>> [ clg: removed the icp_post_load() changes from nikunj patchset v3: >>>> http://patchwork.ozlabs.org/patch/646008/ ] >>>> Signed-off-by: Cédric Le Goater <c...@kaod.org> >>> >>> I think this and 11/20 are good enough and sufficiently standalone >>> that I've merged them into ppc-for-2.8. >> >> These are a prereq for the patchset. > > Sorry, I wasn't clear. I realize they're prerequisites for the rest > of the patchset. What I meant is that they don't require the rest of > the patchset to apply themselves, and they make some sense even > without the rest of the series.
OK. I think there are a couple more cleanups we could do because the initialization is bit difficult to track. The way the ICPs are allocated is bit curious. I hope to have sometime for that after xics native is merged. I have made some migration tests on power8 tcg and kvm, we should be fine. Thanks, C.
