Commit c4fef01b authored by Yunsheng Lin's avatar Yunsheng Lin Committed by David S. Miller

net: sched: implement TCQ_F_CAN_BYPASS for lockless qdisc

Currently pfifo_fast has both TCQ_F_CAN_BYPASS and TCQ_F_NOLOCK
flag set, but queue discipline by-pass does not work for lockless
qdisc because skb is always enqueued to qdisc even when the qdisc
is empty, see __dev_xmit_skb().

This patch calls sch_direct_xmit() to transmit the skb directly
to the driver for empty lockless qdisc, which aviod enqueuing
and dequeuing operation.

As qdisc->empty is not reliable to indicate a empty qdisc because
there is a time window between enqueuing and setting qdisc->empty.
So we use the MISSED state added in commit a90c57f2 ("net:
sched: fix packet stuck problem for lockless qdisc"), which
indicate there is lock contention, suggesting that it is better
not to do the qdisc bypass in order to avoid packet out of order
problem.

In order to make MISSED state reliable to indicate a empty qdisc,
we need to ensure that testing and clearing of MISSED state is
within the protection of qdisc->seqlock, only setting MISSED state
can be done without the protection of qdisc->seqlock. A MISSED
state testing is added without the protection of qdisc->seqlock to
aviod doing unnecessary spin_trylock() for contention case.

As the enqueuing is not within the protection of qdisc->seqlock,
there is still a potential data race as mentioned by Jakub [1]:

      thread1               thread2             thread3
qdisc_run_begin() # true
                        qdisc_run_begin(q)
                             set(MISSED)
pfifo_fast_dequeue
  clear(MISSED)
  # recheck the queue
qdisc_run_end()
                            enqueue skb1
                                             qdisc empty # true
                                          qdisc_run_begin() # true
                                          sch_direct_xmit() # skb2
                         qdisc_run_begin()
                            set(MISSED)

When above happens, skb1 enqueued by thread2 is transmited after
skb2 is transmited by thread3 because MISSED state setting and
enqueuing is not under the qdisc->seqlock. If qdisc bypass is
disabled, skb1 has better chance to be transmited quicker than
skb2.

This patch does not take care of the above data race, because we
view this as similar as below:
Even at the same time CPU1 and CPU2 write the skb to two socket
which both heading to the same qdisc, there is no guarantee that
which skb will hit the qdisc first, because there is a lot of
factor like interrupt/softirq/cache miss/scheduling afffecting
that.

There are below cases that need special handling:
1. When MISSED state is cleared before another round of dequeuing
   in pfifo_fast_dequeue(), and __qdisc_run() might not be able to
   dequeue all skb in one round and call __netif_schedule(), which
   might result in a non-empty qdisc without MISSED set. In order
   to avoid this, the MISSED state is set for lockless qdisc and
   __netif_schedule() will be called at the end of qdisc_run_end.

2. The MISSED state also need to be set for lockless qdisc instead
   of calling __netif_schedule() directly when requeuing a skb for
   a similar reason.

3. For netdev queue stopped case, the MISSED case need clearing
   while the netdev queue is stopped, otherwise there may be
   unnecessary __netif_schedule() calling. So a new DRAINING state
   is added to indicate this case, which also indicate a non-empty
   qdisc.

4. As there is already netif_xmit_frozen_or_stopped() checking in
   dequeue_skb() and sch_direct_xmit(), which are both within the
   protection of qdisc->seqlock, but the same checking in
   __dev_xmit_skb() is without the protection, which might cause
   empty indication of a lockless qdisc to be not reliable. So
   remove the checking in __dev_xmit_skb(), and the checking in
   the protection of qdisc->seqlock seems enough to avoid the cpu
   consumption problem for netdev queue stopped case.

1. https://lkml.org/lkml/2021/5/29/215Acked-by: default avatarJakub Kicinski <kuba@kernel.org>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com> # flexcan
Signed-off-by: default avatarYunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent dd25296a
...@@ -37,8 +37,15 @@ enum qdisc_state_t { ...@@ -37,8 +37,15 @@ enum qdisc_state_t {
__QDISC_STATE_SCHED, __QDISC_STATE_SCHED,
__QDISC_STATE_DEACTIVATED, __QDISC_STATE_DEACTIVATED,
__QDISC_STATE_MISSED, __QDISC_STATE_MISSED,
__QDISC_STATE_DRAINING,
}; };
#define QDISC_STATE_MISSED BIT(__QDISC_STATE_MISSED)
#define QDISC_STATE_DRAINING BIT(__QDISC_STATE_DRAINING)
#define QDISC_STATE_NON_EMPTY (QDISC_STATE_MISSED | \
QDISC_STATE_DRAINING)
struct qdisc_size_table { struct qdisc_size_table {
struct rcu_head rcu; struct rcu_head rcu;
struct list_head list; struct list_head list;
...@@ -145,6 +152,11 @@ static inline bool qdisc_is_running(struct Qdisc *qdisc) ...@@ -145,6 +152,11 @@ static inline bool qdisc_is_running(struct Qdisc *qdisc)
return (raw_read_seqcount(&qdisc->running) & 1) ? true : false; return (raw_read_seqcount(&qdisc->running) & 1) ? true : false;
} }
static inline bool nolock_qdisc_is_empty(const struct Qdisc *qdisc)
{
return !(READ_ONCE(qdisc->state) & QDISC_STATE_NON_EMPTY);
}
static inline bool qdisc_is_percpu_stats(const struct Qdisc *q) static inline bool qdisc_is_percpu_stats(const struct Qdisc *q)
{ {
return q->flags & TCQ_F_CPUSTATS; return q->flags & TCQ_F_CPUSTATS;
...@@ -206,10 +218,8 @@ static inline void qdisc_run_end(struct Qdisc *qdisc) ...@@ -206,10 +218,8 @@ static inline void qdisc_run_end(struct Qdisc *qdisc)
spin_unlock(&qdisc->seqlock); spin_unlock(&qdisc->seqlock);
if (unlikely(test_bit(__QDISC_STATE_MISSED, if (unlikely(test_bit(__QDISC_STATE_MISSED,
&qdisc->state))) { &qdisc->state)))
clear_bit(__QDISC_STATE_MISSED, &qdisc->state);
__netif_schedule(qdisc); __netif_schedule(qdisc);
}
} else { } else {
write_seqcount_end(&qdisc->running); write_seqcount_end(&qdisc->running);
} }
......
...@@ -3852,10 +3852,33 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, ...@@ -3852,10 +3852,33 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
qdisc_calculate_pkt_len(skb, q); qdisc_calculate_pkt_len(skb, q);
if (q->flags & TCQ_F_NOLOCK) { if (q->flags & TCQ_F_NOLOCK) {
if (q->flags & TCQ_F_CAN_BYPASS && nolock_qdisc_is_empty(q) &&
qdisc_run_begin(q)) {
/* Retest nolock_qdisc_is_empty() within the protection
* of q->seqlock to protect from racing with requeuing.
*/
if (unlikely(!nolock_qdisc_is_empty(q))) {
rc = q->enqueue(skb, q, &to_free) &
NET_XMIT_MASK;
__qdisc_run(q);
qdisc_run_end(q);
goto no_lock_out;
}
qdisc_bstats_cpu_update(q, skb);
if (sch_direct_xmit(skb, q, dev, txq, NULL, true) &&
!nolock_qdisc_is_empty(q))
__qdisc_run(q);
qdisc_run_end(q);
return NET_XMIT_SUCCESS;
}
rc = q->enqueue(skb, q, &to_free) & NET_XMIT_MASK; rc = q->enqueue(skb, q, &to_free) & NET_XMIT_MASK;
if (likely(!netif_xmit_frozen_or_stopped(txq))) qdisc_run(q);
qdisc_run(q);
no_lock_out:
if (unlikely(to_free)) if (unlikely(to_free))
kfree_skb_list(to_free); kfree_skb_list(to_free);
return rc; return rc;
......
...@@ -52,6 +52,8 @@ static void qdisc_maybe_clear_missed(struct Qdisc *q, ...@@ -52,6 +52,8 @@ static void qdisc_maybe_clear_missed(struct Qdisc *q,
*/ */
if (!netif_xmit_frozen_or_stopped(txq)) if (!netif_xmit_frozen_or_stopped(txq))
set_bit(__QDISC_STATE_MISSED, &q->state); set_bit(__QDISC_STATE_MISSED, &q->state);
else
set_bit(__QDISC_STATE_DRAINING, &q->state);
} }
/* Main transmission queue. */ /* Main transmission queue. */
...@@ -164,9 +166,13 @@ static inline void dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q) ...@@ -164,9 +166,13 @@ static inline void dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q)
skb = next; skb = next;
} }
if (lock)
if (lock) {
spin_unlock(lock); spin_unlock(lock);
__netif_schedule(q); set_bit(__QDISC_STATE_MISSED, &q->state);
} else {
__netif_schedule(q);
}
} }
static void try_bulk_dequeue_skb(struct Qdisc *q, static void try_bulk_dequeue_skb(struct Qdisc *q,
...@@ -409,7 +415,11 @@ void __qdisc_run(struct Qdisc *q) ...@@ -409,7 +415,11 @@ void __qdisc_run(struct Qdisc *q)
while (qdisc_restart(q, &packets)) { while (qdisc_restart(q, &packets)) {
quota -= packets; quota -= packets;
if (quota <= 0) { if (quota <= 0) {
__netif_schedule(q); if (q->flags & TCQ_F_NOLOCK)
set_bit(__QDISC_STATE_MISSED, &q->state);
else
__netif_schedule(q);
break; break;
} }
} }
...@@ -698,13 +708,14 @@ static struct sk_buff *pfifo_fast_dequeue(struct Qdisc *qdisc) ...@@ -698,13 +708,14 @@ static struct sk_buff *pfifo_fast_dequeue(struct Qdisc *qdisc)
if (likely(skb)) { if (likely(skb)) {
qdisc_update_stats_at_dequeue(qdisc, skb); qdisc_update_stats_at_dequeue(qdisc, skb);
} else if (need_retry && } else if (need_retry &&
test_bit(__QDISC_STATE_MISSED, &qdisc->state)) { READ_ONCE(qdisc->state) & QDISC_STATE_NON_EMPTY) {
/* Delay clearing the STATE_MISSED here to reduce /* Delay clearing the STATE_MISSED here to reduce
* the overhead of the second spin_trylock() in * the overhead of the second spin_trylock() in
* qdisc_run_begin() and __netif_schedule() calling * qdisc_run_begin() and __netif_schedule() calling
* in qdisc_run_end(). * in qdisc_run_end().
*/ */
clear_bit(__QDISC_STATE_MISSED, &qdisc->state); clear_bit(__QDISC_STATE_MISSED, &qdisc->state);
clear_bit(__QDISC_STATE_DRAINING, &qdisc->state);
/* Make sure dequeuing happens after clearing /* Make sure dequeuing happens after clearing
* STATE_MISSED. * STATE_MISSED.
...@@ -1222,6 +1233,7 @@ static void dev_reset_queue(struct net_device *dev, ...@@ -1222,6 +1233,7 @@ static void dev_reset_queue(struct net_device *dev,
spin_unlock_bh(qdisc_lock(qdisc)); spin_unlock_bh(qdisc_lock(qdisc));
if (nolock) { if (nolock) {
clear_bit(__QDISC_STATE_MISSED, &qdisc->state); clear_bit(__QDISC_STATE_MISSED, &qdisc->state);
clear_bit(__QDISC_STATE_DRAINING, &qdisc->state);
spin_unlock_bh(&qdisc->seqlock); spin_unlock_bh(&qdisc->seqlock);
} }
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment