x86/HVM: avoid reading ioreq state more than once

Otherwise, especially when the compiler chooses to translate the
switch() to a jump table, unpredictable behavior (and in the jump table
case arbitrary code execution) can result.

This is XSA-166.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>

--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -460,7 +460,10 @@ static bool_t hvm_wait_for_io(struct hvm
 {
     while ( sv->pending )
     {
-        switch ( p->state )
+        unsigned int state = p->state;
+
+        rmb();
+        switch ( state )
         {
         case STATE_IOREQ_NONE:
             /*
@@ -471,18 +474,15 @@ static bool_t hvm_wait_for_io(struct hvm
             hvm_io_assist(sv, ~0ul);
             break;
         case STATE_IORESP_READY: /* IORESP_READY -> NONE */
-            rmb(); /* see IORESP_READY /then/ read contents of ioreq */
             p->state = STATE_IOREQ_NONE;
             hvm_io_assist(sv, p->data);
             break;
         case STATE_IOREQ_READY:  /* IOREQ_{READY,INPROCESS} -> IORESP_READY */
         case STATE_IOREQ_INPROCESS:
-            wait_on_xen_event_channel(sv->ioreq_evtchn,
-                                      (p->state != STATE_IOREQ_READY) &&
-                                      (p->state != STATE_IOREQ_INPROCESS));
+            wait_on_xen_event_channel(sv->ioreq_evtchn, p->state != state);
             break;
         default:
-            gdprintk(XENLOG_ERR, "Weird HVM iorequest state %d.\n", p->state);
+            gdprintk(XENLOG_ERR, "Weird HVM iorequest state %u\n", state);
             sv->pending = 0;
             domain_crash(sv->vcpu->domain);
             return 0; /* bail */