ch4/posix: using polling loop for recv and send queue

Trying do multiple polling for the recv and send queue to
improve the message latency. This helps the for multiple
incoming messages or deferred sends use cases. Two CVARs are
add to control the maximum number of loops in one progress call
If any of the queue is empty, POSIX will skip the rest iterations.

Author: Yanfei Guo

src/mpid/ch4/shm/posix/posix_init.c

@@ -47,6 +47,28 @@
       description : >-
         Controls topology-aware communication in POSIX.
 
+    - name        : MPIR_CVAR_CH4_SHM_POSIX_PROGRESS_RECV_LOOPS
+      category    : CH4
+      type        : int
+      default     : 1
+      class       : none
+      verbosity   : MPI_T_VERBOSITY_USER_BASIC
+      scope       : MPI_T_SCOPE_ALL_EQ
+      description : >-
+        Controls the maximum number of polling on recv queue
+        per progress.
+
+    - name        : MPIR_CVAR_CH4_SHM_POSIX_PROGRESS_SEND_LOOPS
+      category    : CH4
+      type        : int
+      default     : 1
+      class       : none
+      verbosity   : MPI_T_VERBOSITY_USER_BASIC
+      scope       : MPI_T_SCOPE_ALL_EQ
+      description : >-
+        Controls the maximum number of polling on send queue (for
+        deferred sends) per progress.
+
 === END_MPI_T_CVAR_INFO_BLOCK ===
 */
 

src/mpid/ch4/shm/posix/posix_progress.h

@@ -144,11 +144,25 @@ MPL_STATIC_INLINE_PREFIX int MPIDI_POSIX_progress(int vci, int *made_progress)
 
     MPIR_Assert(vci < MPIDI_POSIX_global.num_vcis);
 
-    mpi_errno = MPIDI_POSIX_progress_recv(vci, made_progress);
-    MPIR_ERR_CHECK(mpi_errno);
+    for (int i = 0; i < MPIR_CVAR_CH4_SHM_POSIX_PROGRESS_RECV_LOOPS; i++) {
+        int made_recv_progress = 0;
+        mpi_errno = MPIDI_POSIX_progress_recv(vci, &made_recv_progress);
+        MPIR_ERR_CHECK(mpi_errno);
+        if (!made_recv_progress)
+            break;
+
+        *made_progress |= made_recv_progress;
+    }
+
+    for (int i = 0; i < MPIR_CVAR_CH4_SHM_POSIX_PROGRESS_SEND_LOOPS; i++) {
+        int made_send_progress = 0;
+        mpi_errno = MPIDI_POSIX_progress_send(vci, &made_send_progress);
+        MPIR_ERR_CHECK(mpi_errno);
+        if (!made_send_progress)
+            break;
 
-    mpi_errno = MPIDI_POSIX_progress_send(vci, made_progress);
-    MPIR_ERR_CHECK(mpi_errno);
+        *made_progress |= made_send_progress;
+    }
 
   fn_exit:
     MPIR_FUNC_EXIT;