PPC: Fix race in mtmsr paravirt implementation

The current implementation of mtmsr and mtmsrd are racy in that it does:

  * check (int_pending == 0)
  ---> host sets int_pending = 1 <---
  * write shared page
  * done

while instead we should check for int_pending after the shared page is written.

Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Avi Kivity <avi@redhat.com>
This commit is contained in:
Bharat Bhushan 2011-10-13 15:17:08 +05:30 committed by Avi Kivity
parent 95325e6b19
commit f9208427f7

View file

@ -167,6 +167,9 @@ maybe_stay_in_guest:
kvm_emulate_mtmsr_reg2:
ori r30, r0, 0
/* Put MSR into magic page because we don't call mtmsr */
STL64(r30, KVM_MAGIC_PAGE + KVM_MAGIC_MSR, 0)
/* Check if we have to fetch an interrupt */
lwz r31, (KVM_MAGIC_PAGE + KVM_MAGIC_INT)(0)
cmpwi r31, 0
@ -174,15 +177,10 @@ kvm_emulate_mtmsr_reg2:
/* Check if we may trigger an interrupt */
andi. r31, r30, MSR_EE
beq no_mtmsr
b do_mtmsr
bne do_mtmsr
no_mtmsr:
/* Put MSR into magic page because we don't call mtmsr */
STL64(r30, KVM_MAGIC_PAGE + KVM_MAGIC_MSR, 0)
SCRATCH_RESTORE
/* Go back to caller */