On Fri, 2008-09-12 at 11:10 +0800, Liu Yu-B13201 wrote:
>
> > @@ -76,9 +76,9 @@ struct kvm_arch {
> >
> > struct kvm_vcpu_arch {
> > /* Unmodified copy of the guest's TLB. */
> > - struct tlbe guest_tlb[PPC44x_TLB_SIZE];
> > + struct kvmppc_44x_tlbe guest_tlb[PPC44x_TLB_SIZE];
> > /* TLB that's actually used when the guest is running. */
> > - struct tlbe shadow_tlb[PPC44x_TLB_SIZE];
> > + struct kvmppc_44x_tlbe shadow_tlb[PPC44x_TLB_SIZE];
> > /* Pages which are referenced in the shadow TLB. */
> > struct page *shadow_pages[PPC44x_TLB_SIZE];
> >
>
> I have been thinking about merge kvm_vcpu_arch of 44x and e500.
> But I'm not sure whether it's worth doing this.
We could do implement a scheme like to_vmx() in arch/x86/kvm/vmx.c . See
the thread around
http://article.gmane.org/gmane.comp.emulators.kvm.devel/8516/ for more
discussion.
This approach will get a little weird in emulate.c, since right now we
kvmppc_emulate_instruction() emulates all instructions for all cores. We
may need to split that up, for example:
emulate.c:
switch (opcode) {
case MFSPR:
switch (sprn) {
case SRR0: vcpu->arch.gpr[rt] = vcpu->arch.srr0;
...
default:
r = kvmppc_emulate_mfspr(vcpu, sprn, rt); // core hook
if (r)
return EMULATE_FAIL;
}
}
e500_emulate.c:
int kvmppc_emulate_mfspr(struct kvm_vcpu *vcpu, u32 sprn, unsigned gpr)
{
struct vcpu_e500 *vcpu_e500 = to_vcpu_e500(vcpu);
int r = 0;
switch (sprn) {
case MAS0: vcpu->arch.gpr[rt] = vcpu_e500->mas0;
...
default:
r = -1;
}
return r;
}
I will be traveling and won't have time to look at this next week, so if
this approach makes sense to you, do you want to send a patch?
--
Hollis Blanchard
IBM Linux Technology Center
--
To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
