On Tue, 2019-03-19 at 17:53 +1100, Alexey Kardashevskiy wrote: > > On 19/03/2019 15:04, Suraj Jitindar Singh wrote: > > Implement a virtual mode handler for the H_CALL H_PAGE_INIT which > > can be > > used to zero or copy a guest page. The page is defined to be 4k and > > must > > be 4k aligned. > > > > The in-kernel handler halves the time to handle this H_CALL > > compared to > > handling it in userspace for a radix guest. > > > > Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@xxxxxxxxx> > > --- > > arch/powerpc/kvm/book3s_hv.c | 39 > > +++++++++++++++++++++++++++++++++++++++ > > 1 file changed, 39 insertions(+) > > > > diff --git a/arch/powerpc/kvm/book3s_hv.c > > b/arch/powerpc/kvm/book3s_hv.c > > index 7179ea783f4f..fa29cc4900c2 100644 > > --- a/arch/powerpc/kvm/book3s_hv.c > > +++ b/arch/powerpc/kvm/book3s_hv.c > > @@ -883,6 +883,39 @@ static int kvmppc_copy_guest(struct kvm *kvm, > > gpa_t to, gpa_t from, > > return 0; > > } > > > > +static int kvmppc_h_page_init(struct kvm_vcpu *vcpu, unsigned long > > flags, > > > > > + unsigned long dest, unsigned long > > src) > > +{ > > + int ret = H_FUNCTION; > > + u64 pg_sz = 1UL << 12; /* 4K page size */ > > + u64 pg_mask = pg_sz - 1; > > > Ditch pg_sz and just use SZ_4K? The page size is not going to change > here ever. Yep, will do. Didn't realise that was a thing > > > > + > > + /* Check for invalid flags (H_PAGE_SET_LOANED covers all > > CMO flags) */ > > + if (flags & ~(H_ICACHE_INVALIDATE | H_ICACHE_SYNCHRONIZE | > > + H_ZERO_PAGE | H_COPY_PAGE | > > H_PAGE_SET_LOANED)) > > + return H_PARAMETER; > > + > > + /* dest (and src if copy_page flag set) must be page > > aligned */ > > + if ((dest & pg_mask) || ((flags & H_COPY_PAGE) && (src & > > pg_mask))) > > + return H_PARAMETER; > > + > > + /* zero and/or copy the page as determined by the flags */ > > + if (flags & H_ZERO_PAGE) { > > + ret = kvm_clear_guest(vcpu->kvm, dest, pg_sz); > > + if (ret < 0) > > + return H_PARAMETER; > > + } > > + if (flags & H_COPY_PAGE) { > > > If H_COPY_PAGE is set, is there any good reason to do kvm_clear_guest > first even if H_ZERO_PAGE is set? It should not happen in practice > and > it is no harm but also no sense. True, if both are set then we can get away with only doing copy since it will have the same effect. > > > + ret = kvmppc_copy_guest(vcpu->kvm, dest, src, > > pg_sz); > > + if (ret < 0) > > + return H_PARAMETER; > > + } > > > else if (src) > return H_PARAMETER; I'm not sure that's technically an error so I'm going to leave it out. > > > > > + > > + /* We can ignore the remaining flags */ > > + > > + return H_SUCCESS; > > +} > > + > > static int kvm_arch_vcpu_yield_to(struct kvm_vcpu *target) > > { > > struct kvmppc_vcore *vcore = target->arch.vcore; > > @@ -1083,6 +1116,11 @@ int kvmppc_pseries_do_hcall(struct kvm_vcpu > > *vcpu) > > if (nesting_enabled(vcpu->kvm)) > > ret = > > kvmhv_copy_tofrom_guest_nested(vcpu); > > break; > > + case H_PAGE_INIT: > > + ret = kvmppc_h_page_init(vcpu, > > kvmppc_get_gpr(vcpu, 4), > > + kvmppc_get_gpr(vcpu, 5), > > + kvmppc_get_gpr(vcpu, 6)); > > + break; > > default: > > return RESUME_HOST; > > } > > @@ -1127,6 +1165,7 @@ static int kvmppc_hcall_impl_hv(unsigned long > > cmd) > > case H_IPOLL: > > case H_XIRR_X: > > #endif > > + case H_PAGE_INIT: > > return 1; > > } > > > > > >