| Message ID | 20240604140903.31939-2-fw@strlen.de (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | net: tcp: un-pin tw timer | expand |
On Tue, Jun 4, 2024 at 4:11 PM Florian Westphal <fw@strlen.de> wrote: > > From: Valentin Schneider <vschneid@redhat.com> > > The TCP timewait timer is proving to be problematic for setups where > scheduler CPU isolation is achieved at runtime via cpusets (as opposed to > statically via isolcpus=domains). > > What happens there is a CPU goes through tcp_time_wait(), arming the > time_wait timer, then gets isolated. TCP_TIMEWAIT_LEN later, the timer > fires, causing interference for the now-isolated CPU. This is conceptually > similar to the issue described in commit e02b93124855 ("workqueue: Unbind > kworkers before sending them to exit()") > > Move inet_twsk_schedule() to within inet_twsk_hashdance(), with the ehash > lock held. Expand the lock's critical section from inet_twsk_kill() to > inet_twsk_deschedule_put(), serializing the scheduling vs descheduling of > the timer. IOW, this prevents the following race: > > tcp_time_wait() > inet_twsk_hashdance() > inet_twsk_deschedule_put() > del_timer_sync() > inet_twsk_schedule() > > Thanks to Paolo Abeni for suggesting to leverage the ehash lock. > > This also restores a comment from commit ec94c2696f0b ("tcp/dccp: avoid > one atomic operation for timewait hashdance") as inet_twsk_hashdance() had > a "Step 1" and "Step 3" comment, but the "Step 2" had gone missing. > > inet_twsk_deschedule_put() now acquires the ehash spinlock to synchronize > with inet_twsk_hashdance_schedule(). > > To ease possible regression search, actual un-pin is done in next patch. > > Link: https://lore.kernel.org/all/ZPhpfMjSiHVjQkTk@localhost.localdomain/ > Signed-off-by: Valentin Schneider <vschneid@redhat.com> > Signed-off-by: Florian Westphal <fw@strlen.de> Reviewed-by: Eric Dumazet <edumazet@google.com> Thanks.
On 2024-06-04 16:08:47 [+0200], Florian Westphal wrote: > --- a/include/net/inet_timewait_sock.h > +++ b/include/net/inet_timewait_sock.h > @@ -93,8 +93,10 @@ struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, > struct inet_timewait_death_row *dr, > const int state); > > -void inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, > - struct inet_hashinfo *hashinfo); > +void inet_twsk_hashdance_schedule(struct inet_timewait_sock *tw, > + struct sock *sk, > + struct inet_hashinfo *hashinfo, > + int timeo); Since this is gone now, could please also update the reference to in tcp_twsk_unique(). Sebastian
On 04/06/24 16:08, Florian Westphal wrote: > @@ -217,7 +228,34 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); > */ > void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) > { > - if (del_timer_sync(&tw->tw_timer)) > + struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; > + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > + > + /* inet_twsk_purge() walks over all sockets, including tw ones, > + * and removes them via inet_twsk_deschedule_put() after a > + * refcount_inc_not_zero(). > + * > + * inet_twsk_hashdance_schedule() must (re)init the refcount before > + * arming the timer, i.e. inet_twsk_purge can obtain a reference to > + * a twsk that did not yet schedule the timer. > + * > + * The ehash lock synchronizes these two: > + * After acquiring the lock, the timer is always scheduled (else > + * timer_shutdown returns false), because hashdance_schedule releases > + * the ehash lock only after completing the timer initialization. > + * > + * Without grabbing the ehash lock, we get: > + * 1) cpu x sets twsk refcount to 3 > + * 2) cpu y bumps refcount to 4 > + * 3) cpu y calls inet_twsk_deschedule_put() and shuts timer down > + * 4) cpu x tries to start timer, but mod_timer is a noop post-shutdown > + * -> timer refcount is never decremented. > + */ > + spin_lock(lock); > + /* Makes sure hashdance_schedule() has completed */ > + spin_unlock(lock); Ah, this is clever! Thanks a ton for finding a way to get this to work!
diff --git a/include/net/inet_timewait_sock.h b/include/net/inet_timewait_sock.h index 2a536eea9424..5b43d220243d 100644 --- a/include/net/inet_timewait_sock.h +++ b/include/net/inet_timewait_sock.h @@ -93,8 +93,10 @@ struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, struct inet_timewait_death_row *dr, const int state); -void inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, - struct inet_hashinfo *hashinfo); +void inet_twsk_hashdance_schedule(struct inet_timewait_sock *tw, + struct sock *sk, + struct inet_hashinfo *hashinfo, + int timeo); void __inet_twsk_schedule(struct inet_timewait_sock *tw, int timeo, bool rearm); diff --git a/net/dccp/minisocks.c b/net/dccp/minisocks.c index 251a57cf5822..deb52d7d31b4 100644 --- a/net/dccp/minisocks.c +++ b/net/dccp/minisocks.c @@ -59,11 +59,10 @@ void dccp_time_wait(struct sock *sk, int state, int timeo) * we complete the initialization. */ local_bh_disable(); - inet_twsk_schedule(tw, timeo); /* Linkage updates. * Note that access to tw after this point is illegal. */ - inet_twsk_hashdance(tw, sk, &dccp_hashinfo); + inet_twsk_hashdance_schedule(tw, sk, &dccp_hashinfo, timeo); local_bh_enable(); } else { /* Sorry, if we're out of memory, just CLOSE this diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index e28075f0006e..628d33a41ce5 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -96,9 +96,13 @@ static void inet_twsk_add_node_rcu(struct inet_timewait_sock *tw, * Enter the time wait state. This is called with locally disabled BH. * Essentially we whip up a timewait bucket, copy the relevant info into it * from the SK, and mess with hash chains and list linkage. + * + * The caller must not access @tw anymore after this function returns. */ -void inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, - struct inet_hashinfo *hashinfo) +void inet_twsk_hashdance_schedule(struct inet_timewait_sock *tw, + struct sock *sk, + struct inet_hashinfo *hashinfo, + int timeo) { const struct inet_sock *inet = inet_sk(sk); const struct inet_connection_sock *icsk = inet_csk(sk); @@ -129,26 +133,33 @@ void inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, spin_lock(lock); + /* Step 2: Hash TW into tcp ehash chain */ inet_twsk_add_node_rcu(tw, &ehead->chain); /* Step 3: Remove SK from hash chain */ if (__sk_nulls_del_node_init_rcu(sk)) sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1); - spin_unlock(lock); + /* Ensure above writes are committed into memory before updating the + * refcount. + * Provides ordering vs later refcount_inc(). + */ + smp_wmb(); /* tw_refcnt is set to 3 because we have : * - one reference for bhash chain. * - one reference for ehash chain. * - one reference for timer. - * We can use atomic_set() because prior spin_lock()/spin_unlock() - * committed into memory all tw fields. * Also note that after this point, we lost our implicit reference * so we are not allowed to use tw anymore. */ refcount_set(&tw->tw_refcnt, 3); + + inet_twsk_schedule(tw, timeo); + + spin_unlock(lock); } -EXPORT_SYMBOL_GPL(inet_twsk_hashdance); +EXPORT_SYMBOL_GPL(inet_twsk_hashdance_schedule); static void tw_timer_handler(struct timer_list *t) { @@ -217,7 +228,34 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); */ void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) { - if (del_timer_sync(&tw->tw_timer)) + struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); + + /* inet_twsk_purge() walks over all sockets, including tw ones, + * and removes them via inet_twsk_deschedule_put() after a + * refcount_inc_not_zero(). + * + * inet_twsk_hashdance_schedule() must (re)init the refcount before + * arming the timer, i.e. inet_twsk_purge can obtain a reference to + * a twsk that did not yet schedule the timer. + * + * The ehash lock synchronizes these two: + * After acquiring the lock, the timer is always scheduled (else + * timer_shutdown returns false), because hashdance_schedule releases + * the ehash lock only after completing the timer initialization. + * + * Without grabbing the ehash lock, we get: + * 1) cpu x sets twsk refcount to 3 + * 2) cpu y bumps refcount to 4 + * 3) cpu y calls inet_twsk_deschedule_put() and shuts timer down + * 4) cpu x tries to start timer, but mod_timer is a noop post-shutdown + * -> timer refcount is never decremented. + */ + spin_lock(lock); + /* Makes sure hashdance_schedule() has completed */ + spin_unlock(lock); + + if (timer_shutdown_sync(&tw->tw_timer)) inet_twsk_kill(tw); inet_twsk_put(tw); } diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 538c06f95918..47de6f3efc85 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -344,11 +344,10 @@ void tcp_time_wait(struct sock *sk, int state, int timeo) * we complete the initialization. */ local_bh_disable(); - inet_twsk_schedule(tw, timeo); /* Linkage updates. * Note that access to tw after this point is illegal. */ - inet_twsk_hashdance(tw, sk, net->ipv4.tcp_death_row.hashinfo); + inet_twsk_hashdance_schedule(tw, sk, net->ipv4.tcp_death_row.hashinfo, timeo); local_bh_enable(); } else { /* Sorry, if we're out of memory, just CLOSE this