Branch HEAD

. Show_error command
. warning if server group has userland test nodes
. lst depends on LND_LIBS

diff --git a/lnet/include/lnet/lnetst.h b/lnet/include/lnet/lnetst.h
index b5e69cb..ea5cd49 100644 (file)
--- a/lnet/include/lnet/lnetst.h
+++ b/lnet/include/lnet/lnetst.h
@@ -396,6 +396,7 @@ typedef struct {
                                                                lstio_bulk_param_t,
                                                                lstio_ping_param_t,
                                                                ... more */
+        int                    *lstio_tes_retp;         /* OUT: private returned value */
         struct list_head       *lstio_tes_resultp;      /* OUT: list head of result buffer */
 } lstio_test_args_t;
 

diff --git a/lnet/selftest/conctl.c b/lnet/selftest/conctl.c
index 7cb1378..a6fc031 100644 (file)
--- a/lnet/selftest/conctl.c
+++ b/lnet/selftest/conctl.c
@@ -685,9 +685,11 @@ int lst_test_add_ioctl(lstio_test_args_t *args)
         char           *srcgrp = NULL;
         char           *dstgrp = NULL;
         void           *param = NULL;
+        int             ret = 0;
         int             rc = -ENOMEM;
 
         if (args->lstio_tes_resultp == NULL ||
+            args->lstio_tes_retp == NULL ||
             args->lstio_tes_bat_name == NULL || /* no specified batch */
             args->lstio_tes_bat_nmlen <= 0 ||
             args->lstio_tes_bat_nmlen > LST_NAME_SIZE ||
@@ -743,7 +745,11 @@ int lst_test_add_ioctl(lstio_test_args_t *args)
                             args->lstio_tes_concur,
                             args->lstio_tes_dist, args->lstio_tes_span,
                             srcgrp, dstgrp, param, args->lstio_tes_param_len,
-                            args->lstio_tes_resultp);
+                            &ret, args->lstio_tes_resultp);
+
+        if (ret != 0)
+                rc = (copy_to_user(args->lstio_tes_retp, &ret, sizeof(ret))) ?
+                     -EFAULT : 0;
 out:
         if (name != NULL)
                 LIBCFS_FREE(name, args->lstio_tes_bat_nmlen + 1);

diff --git a/lnet/selftest/console.c b/lnet/selftest/console.c
index f474c0b..87ef348 100644 (file)
--- a/lnet/selftest/console.c
+++ b/lnet/selftest/console.c
@@ -1204,7 +1204,7 @@ again:
 int
 lstcon_test_add(char *name, int type, int loop, int concur,
                 int dist, int span, char *src_name, char * dst_name,
-                void *param, int paramlen, struct list_head *result_up)
+                void *param, int paramlen, int *retp, struct list_head *result_up)
                 
 {
         lstcon_group_t  *src_grp = NULL;
@@ -1236,6 +1236,9 @@ lstcon_test_add(char *name, int type, int loop, int concur,
                 goto out;
         }
 
+        if (dst_grp->grp_userland)
+                *retp = 1;
+
         LIBCFS_ALLOC(test, offsetof(lstcon_test_t, tes_param[paramlen]));
         if (!test) {
                 CERROR("Can't allocate test descriptor\n");
@@ -1849,6 +1852,9 @@ lstcon_acceptor_handle (srpc_server_rpc_t *rpc)
         ndl->ndl_node->nd_state   = LST_NODE_ACTIVE;
         ndl->ndl_node->nd_timeout = console_session.ses_timeout;
 
+        if (grp->grp_userland == 0)
+                grp->grp_userland = 1;
+
         strcpy(jrep->join_session, console_session.ses_name);
         jrep->join_timeout = console_session.ses_timeout;
         jrep->join_status  = 0;

diff --git a/lnet/selftest/console.h b/lnet/selftest/console.h
index cc72930..deb02e4 100644 (file)
--- a/lnet/selftest/console.h
+++ b/lnet/selftest/console.h
@@ -38,6 +38,7 @@ typedef struct {
 typedef struct {
         struct list_head        grp_link;       /* chain on global group list */
         int                     grp_ref;        /* reference count */
+        int                     grp_userland;   /* has userland nodes */
         int                     grp_nnode;      /* # of nodes */
         char                    grp_name[LST_NAME_SIZE]; /* group name */
 
@@ -184,7 +185,7 @@ extern int lstcon_nodes_stat(int count, lnet_process_id_t *ids_up,
                              int timeout, struct list_head *result_up);
 extern int lstcon_test_add(char *name, int type, int loop, int concur,
                            int dist, int span, char *src_name, char * dst_name,
-                           void *param, int paramlen, struct list_head *result_up);
+                           void *param, int paramlen, int *retp, struct list_head *result_up);
 #endif
 
 #endif  

diff --git a/lnet/utils/Makefile.am b/lnet/utils/Makefile.am
index 86e7a70..71d49b5 100644 (file)
--- a/lnet/utils/Makefile.am
+++ b/lnet/utils/Makefile.am
@@ -68,7 +68,7 @@ endif
 
 if LIBLUSTRE
 LIB_SELFTEST = $(top_builddir)/lnet/libcfs/libcfs.a $(top_builddir)/lnet/lnet/liblnet.a $(top_builddir)/lnet/selftest/libselftest.a
-liblst.a : $(LIB_SELFTEST)
+liblst.a : $(LIB_SELFTEST) $(LND_LIBS)
        sh $(srcdir)/genlib.sh "$(LIBS)" "$(LND_LIBS)" "$(PTHREAD_LIBS)"
 
 lstclient_SOURCES = lstclient.c

diff --git a/lnet/utils/lst.c b/lnet/utils/lst.c
index dee1359..b5221a3 100644 (file)
--- a/lnet/utils/lst.c
+++ b/lnet/utils/lst.c
@@ -1896,6 +1896,132 @@ out:
 }
 
 int
+jt_lst_show_error(int argc, char **argv)
+{
+        struct list_head      head;
+        lst_stat_req_param_t *srp;
+        lstcon_rpc_ent_t     *ent;
+        sfw_counters_t       *sfwk;
+        srpc_counters_t      *srpc;
+        lnet_counters_t      *lnet;
+        lnet_process_id_t    *idsp   = NULL;
+        char                 *name   = NULL;
+        int                   optidx = 0;
+        int                   count  = 0;
+        int                   type   = 0;
+        int                   timeout = 5;
+        int                   ecount = 0;
+        int                   rc;
+        int                   c;
+
+
+        static struct option show_error_opts[] =
+        {
+                {"group"  , required_argument, 0, 'g' },
+                {"nodes"  , required_argument, 0, 'n' },
+                {0,         0,                 0,  0  }
+        };
+
+        if (session_key == 0) {
+                fprintf(stderr,
+                        "Can't find env LST_SESSION or value is not valid\n");
+                return -1;
+        }
+
+        while (1) {
+                c = getopt_long(argc, argv, "g:n:", show_error_opts, &optidx);
+
+                if (c == -1)
+                        break;
+        
+                switch (c) {
+                case 'g':
+                        type = LST_OPC_GROUP;
+                        name = optarg;
+                        break;
+                case 'n': 
+                        type = LST_OPC_NODES;
+                        name = optarg;
+                        break;
+                default:
+                        lst_print_usage(argv[0]);
+                        return -1;
+                }
+        }
+
+        if (optind != argc || type == 0) {
+                lst_print_usage(argv[0]);
+                return -1;
+        }
+
+        if (name == NULL) {
+                fprintf(stderr, "Missing name of target (group | nodes)\n");
+                return -1;
+        }
+
+        rc = lst_get_node_count(type, name, &count, &idsp);
+        if (rc < 0) {
+                fprintf(stderr, "Failed to get count of nodes from %s: %s\n",
+                        name, strerror(errno));
+                return -1;
+        }
+        
+        CFS_INIT_LIST_HEAD(&head);
+
+        rc = lst_alloc_rpcent(&head, count, sizeof(sfw_counters_t) +
+                                            sizeof(srpc_counters_t) +
+                                            sizeof(lnet_counters_t));
+        if (rc != 0) {
+                fprintf(stderr, "Out of memory\n");
+                goto out;
+        }
+
+        rc = lst_stat_ioctl(name, count, idsp, timeout, &head);
+        if (rc == -1) {
+                lst_print_error(name, "Failed to show errors of %s: %s\n",
+                                name, strerror(errno));
+                goto out;
+        }
+
+        list_for_each_entry(ent, &head, rpe_link) {
+                if (ent->rpe_rpc_errno != 0) {
+                        ecount ++;
+                        fprintf(stderr, "RPC failure, can't show error on %s\n",
+                                libcfs_id2str(ent->rpe_peer));
+                        continue;
+                }
+
+                if (ent->rpe_fwk_errno != 0) {
+                        ecount ++;
+                        fprintf(stderr, "Framework failure, can't show error on %s\n",
+                                libcfs_id2str(ent->rpe_peer));
+                        continue;
+                }
+
+                sfwk = (sfw_counters_t *)&ent->rpe_payload[0];
+                srpc = (srpc_counters_t *)((char *)sfwk + sizeof(*sfwk));
+                lnet = (lnet_counters_t *)((char *)srpc + sizeof(*srpc));
+
+                if (srpc->errors == 0 &&
+                    sfwk->brw_errors == 0 && sfwk->ping_errors == 0)
+                        continue;
+
+                ecount ++;
+                fprintf(stderr, "[%s]: %d RPC errors, %d brw errors, %d ping errors\n",
+                        libcfs_id2str(ent->rpe_peer), srpc->errors, 
+                        sfwk->brw_errors, sfwk->ping_errors);
+        }
+
+        fprintf(stdout, "Total %d errors in %s\n", ecount, name);
+out:
+        lst_free_rpcent(&head);
+        if (idsp != NULL)
+                free(idsp);
+
+        return 0;
+}
+
+int
 lst_add_batch_ioctl (char *name)
 {
         lstio_batch_add_args_t  args = {
@@ -2737,7 +2863,7 @@ lst_get_test_param(char *test, int argc, char **argv, void **param, int *plen)
 int
 lst_add_test_ioctl(char *batch, int type, int loop, int concur,
                    int dist, int span, char *sgrp, char *dgrp,
-                   void *param, int plen, struct list_head *resultp)
+                   void *param, int plen, int *retp, struct list_head *resultp)
 {
         lstio_test_args_t args = {
                 .lstio_tes_key          = session_key,
@@ -2754,6 +2880,7 @@ lst_add_test_ioctl(char *batch, int type, int loop, int concur,
                 .lstio_tes_dgrp_name    = dgrp,
                 .lstio_tes_param_len    = plen,
                 .lstio_tes_param        = param,
+                .lstio_tes_retp         = retp,
                 .lstio_tes_resultp      = resultp,
         };
 
@@ -2778,6 +2905,7 @@ jt_lst_add_test(int argc, char **argv)
         int               plen   = 0;
         int               fcount = 0;
         int               tcount = 0;
+        int               ret    = 0;
         int               type;
         int               rc;
         int               c;
@@ -2888,10 +3016,16 @@ jt_lst_add_test(int argc, char **argv)
         }
 
         rc = lst_add_test_ioctl(batch, type, loop, concur,
-                                dist, span, from, to, param, plen, &head);
+                                dist, span, from, to, param, plen, &ret, &head);
 
         if (rc == 0) {
                 fprintf(stdout, "Test was added successfully\n");
+                if (ret != 0) {
+                        fprintf(stdout, "Server group contains userland test "
+                                "nodes, old version of tcplnd can't accept "
+                                "connection request\n");
+                }
+
                 goto out;
         }
 
@@ -2931,6 +3065,8 @@ static command_t lst_cmdlist[] = {
         {"stat",                jt_lst_stat,            NULL,
          "Usage: lst stat [--bw] [--rate] [--read] [--write] [--max] [--min] [--avg] "
          " [--timeout #] [--delay #] GROUP [GROUP]"                                     },
+        {"show_error",          jt_lst_show_error,      NULL,
+         "Usage: lst show_error [--group NAME] | [--nodes IDS]"                         },
         {"add_batch",           jt_lst_add_batch,       NULL,
          "Usage: lst add_batch NAME"                                                    },
         {"run",                 jt_lst_start_batch,     NULL,