Hi Liu, sorry, but there was a US holiday yesterday. :)
On Sunday 25 May 2008 22:32:13 Liu Yu wrote:
> When guest invalidate a large tlb map,
> there may be more than one corresponding shadow tlb maps
> that need to be invalidated.
>
> Use eaddr and eend to find these shadow tlb maps.
Hmm, good catch.
Did you discover this through inspection, or do you have a guest that triggers
this bug? I don't believe there is large page support for e500 Linux
userspace (and the kernel mappings don't ever need invalidation)?
> diff --git a/arch/powerpc/kvm/44x_tlb.c b/arch/powerpc/kvm/44x_tlb.c
> index 75dff7c..04abd83 100644
> --- a/arch/powerpc/kvm/44x_tlb.c
> +++ b/arch/powerpc/kvm/44x_tlb.c
> @@ -177,7 +177,8 @@ void kvmppc_mmu_map(struct kvm_vcpu *vcpu, u64 gvaddr,
gfn_t gfn, u64 asid,
> vcpu->arch.msr & MSR_PR);
> }
>
> -void kvmppc_mmu_invalidate(struct kvm_vcpu *vcpu, u64 eaddr, u64 asid)
> +void kvmppc_mmu_invalidate(struct kvm_vcpu *vcpu, gva_t eaddr,
> + gva_t eend, u32 asid)
> {
> unsigned int pid = asid & 0xff;
> int i;
> @@ -191,7 +192,7 @@ void kvmppc_mmu_invalidate(struct kvm_vcpu *vcpu, u64
eaddr, u64 asid)
> if (!get_tlb_v(stlbe))
> continue;
>
> - if (eaddr < get_tlb_eaddr(stlbe))
> + if (eend < get_tlb_eaddr(stlbe))
> continue;
>
> if (eaddr > get_tlb_end(stlbe))
> diff --git a/arch/powerpc/kvm/emulate.c b/arch/powerpc/kvm/emulate.c
> index 0000974..8c605d0 100644
> --- a/arch/powerpc/kvm/emulate.c
> +++ b/arch/powerpc/kvm/emulate.c
> @@ -137,7 +137,7 @@ static int kvmppc_emul_tlbwe(struct kvm_vcpu *vcpu, u32
inst)
> if (tlbe->word0 & PPC44x_TLB_VALID) {
> eaddr = get_tlb_eaddr(tlbe);
> asid = (tlbe->word0 & PPC44x_TLB_TS) | tlbe->tid;
> - kvmppc_mmu_invalidate(vcpu, eaddr, asid);
> + kvmppc_mmu_invalidate(vcpu, eaddr, get_tlb_end(tlbe), asid);
> }
>
> switch (ws) {
Looks good, but do you have a testcase to verify with?
--
Hollis Blanchard
IBM Linux Technology Center
--
To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
