aboutsummaryrefslogtreecommitdiff
path: root/net/rxrpc/peer_event.c
diff options
context:
space:
mode:
authorDavid Howells <dhowells@redhat.com>2018-08-08 11:30:02 +0100
committerDavid S. Miller <davem@davemloft.net>2018-08-08 19:10:26 -0700
commit330bdcfadceea5e9a1526d731711e163f9a90975 (patch)
treebbe3790891058a942b1cc833888f7f9bec6e1a9a /net/rxrpc/peer_event.c
parentf39cc1c7f32dd93d19fb14dbf973fd2b5c0f7103 (diff)
download96b-common-330bdcfadceea5e9a1526d731711e163f9a90975.tar.gz
rxrpc: Fix the keepalive generator [ver #2]
AF_RXRPC has a keepalive message generator that generates a message for a peer ~20s after the last transmission to that peer to keep firewall ports open. The implementation is incorrect in the following ways: (1) It mixes up ktime_t and time64_t types. (2) It uses ktime_get_real(), the output of which may jump forward or backward due to adjustments to the time of day. (3) If the current time jumps forward too much or jumps backwards, the generator function will crank the base of the time ring round one slot at a time (ie. a 1s period) until it catches up, spewing out VERSION packets as it goes. Fix the problem by: (1) Only using time64_t. There's no need for sub-second resolution. (2) Use ktime_get_seconds() rather than ktime_get_real() so that time isn't perceived to go backwards. (3) Simplifying rxrpc_peer_keepalive_worker() by splitting it into two parts: (a) The "worker" function that manages the buckets and the timer. (b) The "dispatch" function that takes the pending peers and potentially transmits a keepalive packet before putting them back in the ring into the slot appropriate to the revised last-Tx time. (4) Taking everything that's pending out of the ring and splicing it into a temporary collector list for processing. In the case that there's been a significant jump forward, the ring gets entirely emptied and then the time base can be warped forward before the peers are processed. The warping can't happen if the ring isn't empty because the slot a peer is in is keepalive-time dependent, relative to the base time. (5) Limit the number of iterations of the bucket array when scanning it. (6) Set the timer to skip any empty slots as there's no point waking up if there's nothing to do yet. This can be triggered by an incoming call from a server after a reboot with AF_RXRPC and AFS built into the kernel causing a peer record to be set up before userspace is started. The system clock is then adjusted by userspace, thereby potentially causing the keepalive generator to have a meltdown - which leads to a message like: watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [kworker/0:1:23] ... Workqueue: krxrpcd rxrpc_peer_keepalive_worker EIP: lock_acquire+0x69/0x80 ... Call Trace: ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? _raw_spin_lock_bh+0x29/0x60 ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? __lock_acquire+0x3d3/0x870 ? process_one_work+0x110/0x340 ? process_one_work+0x166/0x340 ? process_one_work+0x110/0x340 ? worker_thread+0x39/0x3c0 ? kthread+0xdb/0x110 ? cancel_delayed_work+0x90/0x90 ? kthread_stop+0x70/0x70 ? ret_from_fork+0x19/0x24 Fixes: ace45bec6d77 ("rxrpc: Fix firewall route keepalive") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/rxrpc/peer_event.c')
-rw-r--r--net/rxrpc/peer_event.c156
1 files changed, 88 insertions, 68 deletions
diff --git a/net/rxrpc/peer_event.c b/net/rxrpc/peer_event.c
index 0ed8b651cec2..4f9da2f51c69 100644
--- a/net/rxrpc/peer_event.c
+++ b/net/rxrpc/peer_event.c
@@ -350,97 +350,117 @@ void rxrpc_peer_add_rtt(struct rxrpc_call *call, enum rxrpc_rtt_rx_trace why,
}
/*
- * Perform keep-alive pings with VERSION packets to keep any NAT alive.
+ * Perform keep-alive pings.
*/
-void rxrpc_peer_keepalive_worker(struct work_struct *work)
+static void rxrpc_peer_keepalive_dispatch(struct rxrpc_net *rxnet,
+ struct list_head *collector,
+ time64_t base,
+ u8 cursor)
{
- struct rxrpc_net *rxnet =
- container_of(work, struct rxrpc_net, peer_keepalive_work);
struct rxrpc_peer *peer;
- unsigned long delay;
- ktime_t base, now = ktime_get_real();
- s64 diff;
- u8 cursor, slot;
+ const u8 mask = ARRAY_SIZE(rxnet->peer_keepalive) - 1;
+ time64_t keepalive_at;
+ int slot;
- base = rxnet->peer_keepalive_base;
- cursor = rxnet->peer_keepalive_cursor;
+ spin_lock_bh(&rxnet->peer_hash_lock);
- _enter("%u,%lld", cursor, ktime_sub(now, base));
+ while (!list_empty(collector)) {
+ peer = list_entry(collector->next,
+ struct rxrpc_peer, keepalive_link);
-next_bucket:
- diff = ktime_to_ns(ktime_sub(now, base));
- if (diff < 0)
- goto resched;
+ list_del_init(&peer->keepalive_link);
+ if (!rxrpc_get_peer_maybe(peer))
+ continue;
- _debug("at %u", cursor);
- spin_lock_bh(&rxnet->peer_hash_lock);
-next_peer:
- if (!rxnet->live) {
spin_unlock_bh(&rxnet->peer_hash_lock);
- goto out;
- }
- /* Everything in the bucket at the cursor is processed this second; the
- * bucket at cursor + 1 goes now + 1s and so on...
- */
- if (hlist_empty(&rxnet->peer_keepalive[cursor])) {
- if (hlist_empty(&rxnet->peer_keepalive_new)) {
- spin_unlock_bh(&rxnet->peer_hash_lock);
- goto emptied_bucket;
+ keepalive_at = peer->last_tx_at + RXRPC_KEEPALIVE_TIME;
+ slot = keepalive_at - base;
+ _debug("%02x peer %u t=%d {%pISp}",
+ cursor, peer->debug_id, slot, &peer->srx.transport);
+
+ if (keepalive_at <= base ||
+ keepalive_at > base + RXRPC_KEEPALIVE_TIME) {
+ rxrpc_send_keepalive(peer);
+ slot = RXRPC_KEEPALIVE_TIME;
}
- hlist_move_list(&rxnet->peer_keepalive_new,
- &rxnet->peer_keepalive[cursor]);
+ /* A transmission to this peer occurred since last we examined
+ * it so put it into the appropriate future bucket.
+ */
+ slot += cursor;
+ slot &= mask;
+ spin_lock_bh(&rxnet->peer_hash_lock);
+ list_add_tail(&peer->keepalive_link,
+ &rxnet->peer_keepalive[slot & mask]);
+ rxrpc_put_peer(peer);
}
- peer = hlist_entry(rxnet->peer_keepalive[cursor].first,
- struct rxrpc_peer, keepalive_link);
- hlist_del_init(&peer->keepalive_link);
- if (!rxrpc_get_peer_maybe(peer))
- goto next_peer;
-
spin_unlock_bh(&rxnet->peer_hash_lock);
+}
- _debug("peer %u {%pISp}", peer->debug_id, &peer->srx.transport);
+/*
+ * Perform keep-alive pings with VERSION packets to keep any NAT alive.
+ */
+void rxrpc_peer_keepalive_worker(struct work_struct *work)
+{
+ struct rxrpc_net *rxnet =
+ container_of(work, struct rxrpc_net, peer_keepalive_work);
+ const u8 mask = ARRAY_SIZE(rxnet->peer_keepalive) - 1;
+ time64_t base, now, delay;
+ u8 cursor, stop;
+ LIST_HEAD(collector);
-recalc:
- diff = ktime_divns(ktime_sub(peer->last_tx_at, base), NSEC_PER_SEC);
- if (diff < -30 || diff > 30)
- goto send; /* LSW of 64-bit time probably wrapped on 32-bit */
- diff += RXRPC_KEEPALIVE_TIME - 1;
- if (diff < 0)
- goto send;
+ now = ktime_get_seconds();
+ base = rxnet->peer_keepalive_base;
+ cursor = rxnet->peer_keepalive_cursor;
+ _enter("%lld,%u", base - now, cursor);
- slot = (diff > RXRPC_KEEPALIVE_TIME - 1) ? RXRPC_KEEPALIVE_TIME - 1 : diff;
- if (slot == 0)
- goto send;
+ if (!rxnet->live)
+ return;
- /* A transmission to this peer occurred since last we examined it so
- * put it into the appropriate future bucket.
+ /* Remove to a temporary list all the peers that are currently lodged
+ * in expired buckets plus all new peers.
+ *
+ * Everything in the bucket at the cursor is processed this
+ * second; the bucket at cursor + 1 goes at now + 1s and so
+ * on...
*/
- slot = (slot + cursor) % ARRAY_SIZE(rxnet->peer_keepalive);
spin_lock_bh(&rxnet->peer_hash_lock);
- hlist_add_head(&peer->keepalive_link, &rxnet->peer_keepalive[slot]);
- rxrpc_put_peer(peer);
- goto next_peer;
-
-send:
- rxrpc_send_keepalive(peer);
- now = ktime_get_real();
- goto recalc;
+ list_splice_init(&rxnet->peer_keepalive_new, &collector);
+
+ stop = cursor + ARRAY_SIZE(rxnet->peer_keepalive);
+ while (base <= now && (s8)(cursor - stop) < 0) {
+ list_splice_tail_init(&rxnet->peer_keepalive[cursor & mask],
+ &collector);
+ base++;
+ cursor++;
+ }
-emptied_bucket:
- cursor++;
- if (cursor >= ARRAY_SIZE(rxnet->peer_keepalive))
- cursor = 0;
- base = ktime_add_ns(base, NSEC_PER_SEC);
- goto next_bucket;
+ base = now;
+ spin_unlock_bh(&rxnet->peer_hash_lock);
-resched:
rxnet->peer_keepalive_base = base;
rxnet->peer_keepalive_cursor = cursor;
- delay = nsecs_to_jiffies(-diff) + 1;
- timer_reduce(&rxnet->peer_keepalive_timer, jiffies + delay);
-out:
+ rxrpc_peer_keepalive_dispatch(rxnet, &collector, base, cursor);
+ ASSERT(list_empty(&collector));
+
+ /* Schedule the timer for the next occupied timeslot. */
+ cursor = rxnet->peer_keepalive_cursor;
+ stop = cursor + RXRPC_KEEPALIVE_TIME - 1;
+ for (; (s8)(cursor - stop) < 0; cursor++) {
+ if (!list_empty(&rxnet->peer_keepalive[cursor & mask]))
+ break;
+ base++;
+ }
+
+ now = ktime_get_seconds();
+ delay = base - now;
+ if (delay < 1)
+ delay = 1;
+ delay *= HZ;
+ if (rxnet->live)
+ timer_reduce(&rxnet->peer_keepalive_timer, jiffies + delay);
+
_leave("");
}