Fix race condition on reporting successful per-proc cfg reload

Set the initial per-proc reload status (to SENT) before actually sending the IPC job, to avoid any race condition (in updating the status) with the IPC callback (ran in the target process).

Fixes #2584
Many thanks to @etamme on providing the troubleshooting logs

(cherry picked from commit e85072c)

cfg_reload.c

@@ -274,10 +274,16 @@ static inline void send_cmd_to_all_procs(ipc_rpc_f *rpc)
 	for( i=1 ; i<counted_max_processes ; i++) {
 		if ( (pt[i].flags&(OSS_PROC_NO_IPC|OSS_PROC_NEEDS_SCRIPT))==
 		OSS_PROC_NEEDS_SCRIPT ) {
-			if (ipc_send_rpc( i, rpc, (void*)(long)srr_ctx->seq_no)<0)
-				srr_ctx->proc_status[i] = RELOAD_FAILED;
-			else
-				srr_ctx->proc_status[i] = RELOAD_SENT;
+			/* set the status before sending, to avoid any race condition
+			 * with running the callback function */
+			srr_ctx->proc_status[i] = RELOAD_SENT;
+			if (i==process_no) {
+				/* run line the cmd for the proc itself */
+				rpc( process_no, (void*)(long)srr_ctx->seq_no);
+			} else {
+				if (ipc_send_rpc( i, rpc, (void*)(long)srr_ctx->seq_no)<0)
+					srr_ctx->proc_status[i] = RELOAD_FAILED;
+			}
 		}
 	}
 }