File xsa262.patch of Package xen.11298
x86/HVM: guard against emulator driving ioreq state in weird ways
In the case where hvm_wait_for_io() calls wait_on_xen_event_channel(),
p->state ends up being read twice in succession: once to determine that
state != p->state, and then again at the top of the loop. This gives a
compromised emulator a chance to change the state back between the two
reads, potentially keeping Xen in a loop indefinitely.
Instead:
* Read p->state once in each of the wait_on_xen_event_channel() tests,
* re-use that value the next time around,
* and insist that the states continue to transition "forward" (with the
exception of the transition to STATE_IOREQ_NONE).
This is XSA-262.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: George Dunlap <george.dunlap@citrix.com>
--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -407,20 +407,31 @@ bool_t hvm_io_pending(struct vcpu *v)
static bool_t hvm_wait_for_io(struct hvm_ioreq_vcpu *sv, ioreq_t *p)
{
- unsigned int state;
+ unsigned int state = p->state, prev_state = STATE_IOREQ_NONE;
- /* NB. Optimised for common case (p->state == STATE_IOREQ_NONE). */
- while ( (state = p->state) != STATE_IOREQ_NONE )
+ smp_rmb();
+
+ while ( likely(state != STATE_IOREQ_NONE) )
{
- rmb();
- switch ( state )
+ if ( unlikely(state < prev_state) )
+ {
+ gdprintk(XENLOG_ERR, "Weird HVM ioreq state transition %u -> %u\n",
+ prev_state, state);
+ domain_crash(sv->vcpu->domain);
+ return 0; /* bail */
+ }
+
+ switch ( prev_state = state )
{
case STATE_IORESP_READY: /* IORESP_READY -> NONE */
hvm_io_assist(p);
- break;
+ return 1;
case STATE_IOREQ_READY: /* IOREQ_{READY,INPROCESS} -> IORESP_READY */
case STATE_IOREQ_INPROCESS:
- wait_on_xen_event_channel(sv->ioreq_evtchn, p->state != state);
+ wait_on_xen_event_channel(sv->ioreq_evtchn,
+ ({ state = p->state;
+ smp_rmb();
+ state != prev_state; }));
break;
default:
gdprintk(XENLOG_ERR, "Weird HVM iorequest state %u\n", state);