job_submit/lua - add cpus_per_task field

bug 511

NEWS

@@ -7,6 +7,7 @@ documents those changes that are of interest to users and admins.
     more than one numeric range in brackets (e.g. rack[0-3]_blade[0-63]").
  -- Add notification if using proctrack/cgroup and task/cgroup when oom hits.
  -- Corrections to advanced reservation logic with overlapping jobs.
+ -- job_submit/lua - add cpus_per_task field to those available.
 
 * Changes in Slurm 2.6.4
 ========================

src/plugins/job_submit/lua/job_submit_lua.c

@@ -375,6 +375,8 @@ static int _get_job_req_field (lua_State *L)
 		lua_pushnumber (L, job_desc->contiguous);
 	} else if (!strcmp(name, "cores_per_socket")) {
 		lua_pushnumber (L, job_desc->cores_per_socket);
+	} else if (!strcmp(name, "cpus_per_task")) {
+		lua_pushnumber (L, job_desc->cpus_per_task);
 	} else if (!strcmp(name, "default_account")) {
 		lua_pushstring (L, _get_default_account(job_desc->user_id));
 	} else if (!strcmp(name, "dependency")) {
@@ -499,6 +501,8 @@ static int _set_job_req_field (lua_State *L)
 		job_desc->contiguous = luaL_checknumber(L, 3);
 	} else if (!strcmp(name, "cores_per_socket")) {
 		job_desc->cores_per_socket = luaL_checknumber(L, 3);
+	} else if (!strcmp(name, "cpus_per_task")) {
+		job_desc->cpus_per_task = luaL_checknumber(L, 3);
 	} else if (!strcmp(name, "dependency")) {
 		value_str = luaL_checkstring(L, 3);
 		xfree(job_desc->dependency);