KVM: PPC: Implement 'skip instruction' mode

To fetch the last instruction we were interrupted on, we enable DR in early
exit code, where we are still in a very transitional phase between guest
and host state.

Most of the time this seemed to work, but another CPU can easily flush our
TLB and HTAB which makes us go in the Linux page fault handler which totally
breaks because we still use the guest's SLB entries.

To work around that, let's introduce a second KVM guest mode that defines
that whenever we get a trap, we don't call the Linux handler or go into
the KVM exit code, but just jump over the faulting instruction.

That way a potentially bad lwz doesn't trigger any faults and we can later
on interpret the invalid instruction we fetched as "fetch didn't work".

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Avi Kivity <avi@redhat.com>

1 files changed, 11 insertions, 5 deletions

diff --git a/arch/powerpc/kvm/book3s_64_slb.S b/arch/powerpc/kvm/book3s_64_slb.S
index 7188c11ed7d1..d07b88617b2c 100644
--- a/arch/powerpc/kvm/book3s_64_slb.S
+++ b/arch/powerpc/kvm/book3s_64_slb.S
@@ -212,10 +212,6 @@ kvmppc_handler_trampoline_exit:
         mfdar   r5
         mfdsisr r6
 
-        /* Unset guest state */
-        li      r9, 0
-        stb     r9, PACA_KVM_IN_GUEST(r13)
-
         /*
          * In order for us to easily get the last instruction,
          * we got the #vmexit at, we exploit the fact that the
@@ -233,18 +229,28 @@ kvmppc_handler_trampoline_exit:
 
 ld_last_inst:
         /* Save off the guest instruction we're at */
+
+        /* Set guest mode to 'jump over instruction' so if lwz faults
+         * we'll just continue at the next IP. */
+        li      r9, KVM_GUEST_MODE_SKIP
+        stb     r9, PACA_KVM_IN_GUEST(r13)
+
         /*    1) enable paging for data */
         mfmsr   r9
         ori     r11, r9, MSR_DR                 /* Enable paging for data */
         mtmsr   r11
         /*    2) fetch the instruction */
-        /* XXX implement PACA_KVM_IN_GUEST=2 path to safely jump over this */
+        li      r0, KVM_INST_FETCH_FAILED       /* In case lwz faults */
         lwz     r0, 0(r3)
         /*    3) disable paging again */
         mtmsr   r9
 
 no_ld_last_inst:
 
+        /* Unset guest mode */
+        li      r9, KVM_GUEST_MODE_NONE
+        stb     r9, PACA_KVM_IN_GUEST(r13)
+
         /* Restore bolted entries from the shadow and fix it along the way */
 
         /* We don't store anything in entry 0, so we don't need to take care of it */