On Tue, Sep 20, 2016 at 10:04:58PM +0530, Nikunj A Dadhania wrote: > PowerPC failed to handle broadcast TLB flush operations. Executing > instructions that are defined architecturally as synchronizing global TLB > should have a global effect. > > * tlbie on BookS > * tlbivax on BookE > * H_CALLs (H_REMOVE, H_BULK_REMOVE and H_PROTECT) in case of pseries, > since they contain a tlbie on a real hypervisor. > > The implementation provides a single point that can be used in MTTCG for > doing async-flushes. > > The patchset introduces bit-flags in CPUPPCState::tlb_need_flush: > > TLB_NEED_LOCAL_FLUSH (0x1) - Flush local tlb > TLB_NEED_GLOBAL_FLUSH (0x2) - Flush tlb on other cpus. > > Note: Tested patches on commit ebc231d(Merge remote-tracking branch > 'remotes/awilliam/tags/vfio-fixes-20160915.0' into staging), as ppc64 > system emulation is broken after that.
Applied to ppc-for-2.8. I did make a couple of small tweaks to
patches 2 & 3.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature
