From 3b82f227ce96a3debc13357e71493f5df72486cc Mon Sep 17 00:00:00 2001 From: Richard Cochran Date: Tue, 30 Aug 2016 22:44:55 +0200 Subject: [PATCH] clock: Monitor the link status using a RT netlink socket. Poll for link up/down events. When a link goes down, the port becomes faulty until the link goes up again. We keep the fault timer from the existing fault detection, but a downed link prevents clear the fault. The new state machine is depicted in this ascii art diagram: +--------+ Fault +---------+ | |------------>| | | UP | | FAULT | | |<------------| | +--------+ Timeout +---------+ A | / | | / Link-Up | | Link-Down / | | / | V / +--------+ / Link-Down | | / | DOWN |<--------/ | | +--------+ If the fault timer occurs in the DOWN state, we simply ignore it. After all, without the link the port is useless. There is one case where the new code changes the existing behavior. If the link quickly does down and then up again while another fault (and its timer) are active, then we will enter the UP state without waiting for the fault timer expiration. However, this behavior is acceptable because when a link goes up, you are starting with a clean slate. Signed-off-by: Richard Cochran --- clock.c | 62 +++++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 52 insertions(+), 10 deletions(-) diff --git a/clock.c b/clock.c index a0124d2..c716f01 100644 --- a/clock.c +++ b/clock.c @@ -39,6 +39,7 @@ #include "servo.h" #include "stats.h" #include "print.h" +#include "rtnl.h" #include "sk.h" #include "tlv.h" #include "tsproc.h" @@ -271,6 +272,9 @@ void clock_destroy(struct clock *c) LIST_FOREACH_SAFE(p, &c->ports, list, tmp) { clock_remove_port(c, p); } + if (c->pollfd[0].fd >= 0) { + rtnl_close(c->pollfd[0].fd); + } port_close(c->uds_port); free(c->pollfd); hash_destroy(c->index2port, NULL); @@ -320,6 +324,25 @@ static void clock_freq_est_reset(struct clock *c) c->fest.count = 0; } +static void clock_link_status(void *ctx, int index, int linkup) +{ + struct clock *c = ctx; + struct port *p; + char key[16]; + + snprintf(key, sizeof(key), "%d", index); + p = hash_lookup(c->index2port, key); + if (!p) { + return; + } + port_link_status_set(p, linkup); + if (linkup) { + port_dispatch(p, EV_FAULT_CLEARED, 0); + } else { + port_dispatch(p, EV_FAULT_DETECTED, 0); + } +} + static void clock_management_send_error(struct port *p, struct ptp_message *msg, int error_id) { @@ -1096,13 +1119,16 @@ struct clock *clock_create(enum clock_type type, struct config *config, LIST_INIT(&c->ports); c->last_port_number = 0; - /* - * Create the UDS interface. - */ if (clock_resize_pollfd(c, 0)) { pr_err("failed to allocate pollfd"); return NULL; } + + /* Open a RT netlink socket. */ + c->pollfd[0].fd = rtnl_open(); + c->pollfd[0].events = POLLIN|POLLPRI; + + /* Create the UDS interface. */ c->uds_port = port_open(phc_index, timestamping, 0, udsif, c); if (!c->uds_port) { pr_err("failed to open the UDS port"); @@ -1129,7 +1155,9 @@ struct clock *clock_create(enum clock_type type, struct config *config, port_dispatch(p, EV_INITIALIZE, 0); } port_dispatch(c->uds_port, EV_INITIALIZE, 0); - + if (c->pollfd[0].fd >= 0) { + rtnl_link_query(c->pollfd[0].fd); + } return c; } @@ -1194,9 +1222,12 @@ static int clock_resize_pollfd(struct clock *c, int new_nports) { struct pollfd *new_pollfd; - /* Need to allocate one extra block of fds for uds */ + /* + * Need to allocate one descriptor for RT netlink and one + * whole extra block of fds for UDS. + */ new_pollfd = realloc(c->pollfd, - (new_nports + 1) * N_CLOCK_PFD * + (1 + (new_nports + 1) * N_CLOCK_PFD) * sizeof(struct pollfd)); if (!new_pollfd) return -1; @@ -1221,7 +1252,7 @@ static void clock_fill_pollfd(struct pollfd *dest, struct port *p) static void clock_check_pollfd(struct clock *c) { struct port *p; - struct pollfd *dest = c->pollfd; + struct pollfd *dest = c->pollfd + 1; if (c->pollfd_valid) return; @@ -1437,7 +1468,7 @@ int clock_poll(struct clock *c) struct port *p; clock_check_pollfd(c); - cnt = poll(c->pollfd, (c->nports + 1) * N_CLOCK_PFD, -1); + cnt = poll(c->pollfd, 1 + (c->nports + 1) * N_CLOCK_PFD, -1); if (cnt < 0) { if (EINTR == errno) { return 0; @@ -1449,7 +1480,13 @@ int clock_poll(struct clock *c) return 0; } + /* Check the RT netlink. */ cur = c->pollfd; + if (cur->revents & (POLLIN|POLLPRI)) { + rtnl_link_status(cur->fd, clock_link_status, c); + } + + cur++; LIST_FOREACH(p, &c->ports, list) { /* Let the ports handle their events. */ for (i = err = 0; i < N_POLLFD && !err; i++) { @@ -1468,10 +1505,15 @@ int clock_poll(struct clock *c) } } - /* Check the fault timer. */ + /* + * When the fault timer expires we clear the fault, + * but only if the link is up. + */ if (cur[N_POLLFD].revents & (POLLIN|POLLPRI)) { clock_fault_timeout(p, 0); - port_dispatch(p, EV_FAULT_CLEARED, 0); + if (port_link_status_get(p)) { + port_dispatch(p, EV_FAULT_CLEARED, 0); + } } cur += N_CLOCK_PFD;