Skip to contentSkip to conversation

KVM: PPC: Book3S: PR: Don't clobber our exit handler id

PATCH New Alexander Graf 1 participant 1 message Dec 2, 2013

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/4 KVM: PPC: Book3S: PR: Don't clobber our exit handler id

We call a C helper to save all svcpu fields into our vcpu. The C ABI states that r12 is considered volatile. However, we keep our exit handler id in r12 currently.

So we need to save it away into a non-volatile register instead that definitely does get preserved across the C call.

This bug usually didn't hit anyone yet since gcc is smart enough to generate code that doesn't even need r12 which means it stayed identical throughout the call by sheer luck. But we can't rely on that.

---

v1 -> v2:

- save/restore r12 in memory instead of abusing r14

---
 arch/powerpc/kvm/book3s_interrupts.S | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)
arch/powerpc/kvm/book3s_interrupts.S +10 -3
index f4dd041..5e7cb32 100644
@@ -132,9 +132,17 @@ kvm_start_lightweight:
 	 *
 	 */
 
+	PPC_LL	r3, GPR4(r1)		/* vcpu pointer */
+
+	/*
+	 * kvmppc_copy_from_svcpu can clobber volatile registers, save
+	 * the exit handler id to the vcpu and restore it from there later.
+	 */
+	stw	r12, VCPU_TRAP(r3)
+
 	/* Transfer reg values from shadow vcpu back to vcpu struct */
 	/* On 64-bit, interrupts are still off at this point */
-	PPC_LL	r3, GPR4(r1)		/* vcpu pointer */
+
 	GET_SHADOW_VCPU(r4)
 	bl	FUNC(kvmppc_copy_from_svcpu)
 	nop
@@ -151,7 +159,6 @@ kvm_start_lightweight:
 	 */
 	ld	r3, PACA_SPRG3(r13)
 	mtspr	SPRN_SPRG3, r3
-
 #endif /* CONFIG_PPC_BOOK3S_64 */
 
 	/* R7 = vcpu */
@@ -177,7 +184,7 @@ kvm_start_lightweight:
 	PPC_STL	r31, VCPU_GPR(R31)(r7)
 
 	/* Pass the exit number as 3rd argument to kvmppc_handle_exit */
-	mr	r5, r12
+	lwz	r5, VCPU_TRAP(r7)
 
 	/* Restore r3 (kvm_run) and r4 (vcpu) */
 	REST_2GPRS(3, r1)
Signature · 1 line
1.8.1.4