exception vs SIGALRM race on core2 CPUs (with fix!)

[stsp <stsp2@xxxxxxxxx>]

OK, I've finally found that this
fixes the race:

--- x86.c.old   2021-03-20 12:51:14.000000000 +0300
+++ x86.c       2021-06-26 02:28:37.082919492 +0300
@@ -9176,8 +9176,10 @@
                if (__xfer_to_guest_mode_work_pending()) {
                        srcu_read_unlock(&kvm->srcu, vcpu->srcu_idx);
                        r = xfer_to_guest_mode_handle_work(vcpu);
-                       if (r)
+                       if (r) {
+kvm_clear_exception_queue(vcpu);
                                return r;
+}
                        vcpu->srcu_idx = srcu_read_lock(&kvm->srcu);
                }
        }



This is where it returns to user
with the PF exception still pending.
So... any ideas?