[alert,fpv] Make AlertCheck1_A actually true for async alerts

This assertion wasn't true because a skew on e.g. the ack signal from
the alert receiver to the alert sender can completely lock up
communication between them.

It seems that Jasper didn't manage to find the loop it would need for
a counterexample with this sort of liveness proof, so it just hung.
With the extra condition (which, in hindsight, is definitely
required), the proof goes through nicely.

Signed-off-by: Rupert Swarbrick <[email protected]>

Author: rswarbrick

hw/ip/prim/fpv/vip/prim_alert_rxtx_async_assert_fpv.sv

@@ -196,11 +196,13 @@ module prim_alert_rxtx_async_assert_fpv
   //
   // We want to disable the assertion if there is a reset or the alert complex is still
   // initialising. We also want to disable it if there has been an injected error: the alerts that
-  // should come out in that situation rely on the ping mechanism.
+  // should come out in that situation rely on the ping mechanism. This injected error will either
+  // be detected with error_setreg_q or seen_skew_q.
   //
   // Finally, we disable it if the sender clears its alert (because ack was dropped again at the end
   // of an alert handshake).
-  AlertCheck1_A: assert property (disable iff (!rst_ni || init_pending || error_setreg_q ||
+  AlertCheck1_A: assert property (disable iff (!rst_ni || init_pending ||
+                                               error_setreg_q || seen_skew_q ||
                                                i_prim_alert_sender.alert_clr)
                                   alert_req_i || alert_test_i |=> s_eventually alert_o);