On Sun, Apr 28, 2019 at 09:45:15PM +1000, Nicholas Piggin wrote: > This is the KVM update to the new idle code. A few improvements: > > - Idle sleepers now always return to caller rather than branch out > to KVM first. > - This allows optimisations like very fast return to caller when no > state has been lost. > - KVM no longer requires nap_state_lost because it controls NVGPR > save/restore itself on the way in and out. > - The heavy idle wakeup KVM request check can be moved out of the > normal host idle code and into the not-performance-critical offline > code. > - KVM nap code now returns from where it is called, which makes the > flow a bit easier to follow.
One question below... > diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S > b/arch/powerpc/kvm/book3s_hv_rmhandlers.S > index 58d0f1ba845d..f66191d8f841 100644 > --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S > +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S ... > @@ -2656,6 +2662,9 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300) > > lis r3, LPCR_PECEDP@h /* Do wake on privileged doorbell */ > > + /* Go back to host stack */ > + ld r1, HSTATE_HOST_R1(r13) At this point we are in kvmppc_h_cede, which we branched to from hcall_try_real_mode, which came from the guest exit path, where we have already loaded r1 from HSTATE_HOST_R1(r13). So if there is a path to get here with r1 not already set to HSTATE_HOST_R1(r13), then I missed it - please point it out to me. Otherwise this statement seems superfluous. Paul.
