CSIT-844: fix binary api rx pthread heap push/pop

We need to push/pop the rx pthread's heap without affecting other
thread(s).

Search clib_per_cpu_mheaps, locate an unused slot. Duplicate the main
thread heap pointer in that slot, and set __os_thread_index
appropriately.

Miscellaneous cleanups. Print exec_inband results as a vector, instead
of as a format string. Don't bail out of vpp_api_test with results
pending, e.g. at the end of a vpp_api_test script. Even though vpp
will eventuallly garbage-collect them, We don't want to leave
allocated reply messages lurking in the api message allocation
rings...

Change-Id: I0e8a25d1ff0d3700249dc330d079db16c2fcbc55
Signed-off-by: Dave Barach <dave@barachs.net>

1 files changed, 13 insertions, 0 deletions

diff --git a/src/vat/main.c b/src/vat/main.c
index e2c9b70889a..aa990a31d93 100644
--- a/src/vat/main.c
+++ b/src/vat/main.c
@@ -308,6 +308,7 @@ main (int argc, char **argv)
   u8 *heap;
   mheap_t *h;
   int i;
+  f64 timeout;
 
   clib_mem_init (0, 128 << 20);
 
@@ -432,6 +433,18 @@ main (int argc, char **argv)
       fclose (vam->ifp);
     }
 
+  /*
+   * Particularly when running a script, don't be in a hurry to leave.
+   * A reply message queued to this process will end up constipating
+   * the allocation rings.
+   */
+  timeout = vat_time_now (vam) + 2.0;
+  while (vam->result_ready == 0 && vat_time_now (vam) < timeout)
+    ;
+
+  if (vat_time_now (vam) > timeout)
+    clib_warning ("BUG: message reply spin-wait timeout");
+
   vl_client_disconnect_from_vlib ();
   exit (0);
 }