LU-14391 lnet: optimize the Netlink packet size for routes

Currently Netlink by default sets its maximum packet size
to send back to user land to 64K. Some sites setup many
routes, above ~430, which exceed this limit. We can avoid
this limitation by calculate about the actually size of
the netlink packet and setting cb->min_dump_alloc. The
new max is then 4GB which should be plenty (27K of routes)

Test-Parameters: trivial testlist=sanity-lnet
Change-Id: Ica01f0cf290992a5d27b8ac2d09508d0a6e8151a
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/c/fs/lustre-release/+/54844
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Frank Sehr <fsehr@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

diff --git a/lnet/lnet/api-ni.c b/lnet/lnet/api-ni.c
index e2bb7c7..4c801e9 100644 (file)
--- a/lnet/lnet/api-ni.c
+++ b/lnet/lnet/api-ni.c
@@ -6789,6 +6789,21 @@ failed_alloc:
        return rc;
 }
 
+/* Size of the message send by lnet_genl_send_scalar_list().
+ * Length is from genlmsg_len() on the msg created.
+ */
+#define ROUTER_MSG_MIN_SIZE            284
+/* For 'value' packet it contains
+ *     net             LNET_NIDSTR_SIZE
+ *     gateway         LNET_NIDSTR_SIZE
+ *     hop             u32
+ *     priority        u32
+ *     health sensit.. u32
+ *     state           "down" largest string (5)
+ *     type            "single-hop" largest string (10)
+ */
+#define ROUTER_MSG_VALUES_SIZE         (LNET_NIDSTR_SIZE * 2 + 27)
+
 /* LNet route ->start() handler for GET requests */
 static int lnet_route_show_start(struct netlink_callback *cb)
 {
@@ -6796,6 +6811,7 @@ static int lnet_route_show_start(struct netlink_callback *cb)
 #ifdef HAVE_NL_PARSE_WITH_EXT_ACK
        struct netlink_ext_ack *extack = NULL;
 #endif
+       unsigned long len = ROUTER_MSG_MIN_SIZE;
        struct lnet_genl_route_list *rlist;
        int msg_len = genlmsg_len(gnlh);
        int rc = 0;
@@ -6936,6 +6952,14 @@ static int lnet_route_show_start(struct netlink_callback *cb)
                        }
                }
        }
+
+       len += ROUTER_MSG_VALUES_SIZE * rlist->lgrl_count;
+       if (len > BIT(sizeof(cb->min_dump_alloc) << 3)) {
+               NL_SET_ERR_MSG(extack, "Netlink msg is too large");
+               rc = -EMSGSIZE;
+       } else {
+               cb->min_dump_alloc = len;
+       }
 report_err:
        mutex_unlock(&the_lnet.ln_api_mutex);
 

diff --git a/lustre/tests/sanity-lnet.sh b/lustre/tests/sanity-lnet.sh
index af84e49..d35e310 100755 (executable)
--- a/lustre/tests/sanity-lnet.sh
+++ b/lustre/tests/sanity-lnet.sh
@@ -1708,6 +1708,21 @@ test_110() {
 }
 run_test 110 "Add NI using a specific TCP / IP address"
 
+test_111() {
+       [[ $(uname -r | grep "3.10") ]] &&
+               skip "Unsupported on RHEL7"
+
+       reinit_dlc || return $?
+       add_net "${NETTYPE}" "${INTERFACES[0]}"
+
+       for index in {2..500}
+       do
+               do_lnetctl route add --net ${NETTYPE}${index} --gateway ${GW_NID}
+       done
+       do_lnetctl route show || return $?
+}
+run_test 111 "Test many routes"
+
 test_200() {
        [[ ${NETTYPE} == tcp* ]] ||
                skip "Need tcp NETTYPE"