| Message ID | 20240603093625.4055-2-fw@strlen.de (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | net: tcp: un-pin tw timer | expand |
On Mon, Jun 3, 2024 at 11:37 AM 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 > vs. reschedule and timer firing: timer_del_sync() is replaced with > timer_shutdown(). > > This means that tw_timer may still be running on another CPU. However, as > the timer owns a reference on tw sk that is only put at the end this > should be fine. > > 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> > --- > include/net/inet_timewait_sock.h | 6 ++- > net/dccp/minisocks.c | 3 +- > net/ipv4/inet_timewait_sock.c | 74 ++++++++++++++++++++++++++------ > net/ipv4/tcp_minisocks.c | 3 +- > 4 files changed, 68 insertions(+), 18 deletions(-) > > 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..2b1d836df64e 100644 > --- a/net/ipv4/inet_timewait_sock.c > +++ b/net/ipv4/inet_timewait_sock.c > @@ -44,14 +44,14 @@ void inet_twsk_bind_unhash(struct inet_timewait_sock *tw, > __sock_put((struct sock *)tw); > } > > -/* Must be called with locally disabled BHs. */ > -static void inet_twsk_kill(struct inet_timewait_sock *tw) > +static void __inet_twsk_kill(struct inet_timewait_sock *tw, spinlock_t *lock) > +__releases(lock) > { > struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; > - spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > struct inet_bind_hashbucket *bhead, *bhead2; > > - spin_lock(lock); > + lockdep_assert_held(lock); > + > sk_nulls_del_node_init_rcu((struct sock *)tw); > spin_unlock(lock); > > @@ -71,6 +71,16 @@ static void inet_twsk_kill(struct inet_timewait_sock *tw) > inet_twsk_put(tw); > } > > +/* Must be called with locally disabled BHs. */ > +static void inet_twsk_kill(struct inet_timewait_sock *tw) > +{ > + struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; > + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > + > + spin_lock(lock); > + __inet_twsk_kill(tw, lock); > +} > + > void inet_twsk_free(struct inet_timewait_sock *tw) > { > struct module *owner = tw->tw_prot->owner; > @@ -96,9 +106,13 @@ 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 +143,33 @@ 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) > { > @@ -217,8 +238,37 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); > */ > void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) > { > - if (del_timer_sync(&tw->tw_timer)) > - inet_twsk_kill(tw); > + 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). > + * > + * With plain timer_shutdown_sync() and without grabbing the ehash > + * lock, we can 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); > + if (timer_shutdown(&tw->tw_timer)) { > + /* releases @lock */ > + __inet_twsk_kill(tw, lock); > + } else { If we do not have a sync variant here, I think that inet_twsk_purge() could return while ongoing timers are alive. tcp_sk_exit_batch() would then possibly hit : WARN_ON_ONCE(!refcount_dec_and_test(&net->ipv4.tcp_death_row.tw_refcount)); The alive timer are releasing tw->tw_dr->tw_refcount at the end of inet_twsk_kill() Presumably, adding some udelay(200) in inet_twsk_kill() could help a syzkaller instance to trigger a bug. > + spin_unlock(lock); > + } > + > inet_twsk_put(tw); > } > EXPORT_SYMBOL(inet_twsk_deschedule_put); > 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 > -- > 2.44.2 >
The patch header should be "net: tcp/dccp ....". On 2024/6/3 17:36, Florian Westphal 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 > vs. reschedule and timer firing: timer_del_sync() is replaced with > timer_shutdown(). > > This means that tw_timer may still be running on another CPU. However, as > the timer owns a reference on tw sk that is only put at the end this > should be fine. > > 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> > --- > include/net/inet_timewait_sock.h | 6 ++- > net/dccp/minisocks.c | 3 +- > net/ipv4/inet_timewait_sock.c | 74 ++++++++++++++++++++++++++------ > net/ipv4/tcp_minisocks.c | 3 +- > 4 files changed, 68 insertions(+), 18 deletions(-) > > 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..2b1d836df64e 100644 > --- a/net/ipv4/inet_timewait_sock.c > +++ b/net/ipv4/inet_timewait_sock.c > @@ -44,14 +44,14 @@ void inet_twsk_bind_unhash(struct inet_timewait_sock *tw, > __sock_put((struct sock *)tw); > } > > -/* Must be called with locally disabled BHs. */ > -static void inet_twsk_kill(struct inet_timewait_sock *tw) > +static void __inet_twsk_kill(struct inet_timewait_sock *tw, spinlock_t *lock) > +__releases(lock) > { > struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; > - spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > struct inet_bind_hashbucket *bhead, *bhead2; > > - spin_lock(lock); > + lockdep_assert_held(lock); > + > sk_nulls_del_node_init_rcu((struct sock *)tw); > spin_unlock(lock); > > @@ -71,6 +71,16 @@ static void inet_twsk_kill(struct inet_timewait_sock *tw) > inet_twsk_put(tw); > } > > +/* Must be called with locally disabled BHs. */ > +static void inet_twsk_kill(struct inet_timewait_sock *tw) > +{ > + struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; > + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > + > + spin_lock(lock); > + __inet_twsk_kill(tw, lock); > +} > + > void inet_twsk_free(struct inet_timewait_sock *tw) > { > struct module *owner = tw->tw_prot->owner; > @@ -96,9 +106,13 @@ 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 +143,33 @@ 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) > { > @@ -217,8 +238,37 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); > */ > void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) > { > - if (del_timer_sync(&tw->tw_timer)) > - inet_twsk_kill(tw); > + 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). > + * > + * With plain timer_shutdown_sync() and without grabbing the ehash > + * lock, we can 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); > + if (timer_shutdown(&tw->tw_timer)) { > + /* releases @lock */ > + __inet_twsk_kill(tw, lock); > + } else { > + spin_unlock(lock); > + } > + > inet_twsk_put(tw); > } > EXPORT_SYMBOL(inet_twsk_deschedule_put); > 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
Eric Dumazet <edumazet@google.com> wrote: > On Mon, Jun 3, 2024 at 11:37 AM Florian Westphal <fw@strlen.de> wrote: > > + spin_lock(lock); > > + if (timer_shutdown(&tw->tw_timer)) { > > + /* releases @lock */ > > + __inet_twsk_kill(tw, lock); > > + } else { > > If we do not have a sync variant here, I think that inet_twsk_purge() > could return while ongoing timers are alive. Yes. We can't use sync variant, it would deadlock on ehash spinlock. > tcp_sk_exit_batch() would then possibly hit : > > WARN_ON_ONCE(!refcount_dec_and_test(&net->ipv4.tcp_death_row.tw_refcount)); > > The alive timer are releasing tw->tw_dr->tw_refcount at the end of > inet_twsk_kill() Theoretically the tw socket can be unlinked from the tw hash already (inet_twsk_purge won't encounter it), but timer is still running. Only solution I see is to schedule() in tcp_sk_exit_batch() until tw_refcount has dropped to the expected value, i.e. something like static void tcp_wait_for_tw_timers(struct net *n) { while (refcount_read(&n->ipv4.tcp_death_row.tw_refcount) > 1)) schedule(); } Any better idea? I started to sketch a patch that keeps PINNED as-is but schedules almost all of the actual work to a work item. Idea was that it would lower RT latencies to acceptable level but it got so ugly that I did not follow this path. I could resurrect this if you think its worth a try.
On Mon, Jun 3, 2024 at 1:21 PM Florian Westphal <fw@strlen.de> wrote: > > Eric Dumazet <edumazet@google.com> wrote: > > On Mon, Jun 3, 2024 at 11:37 AM Florian Westphal <fw@strlen.de> wrote: > > > + spin_lock(lock); > > > + if (timer_shutdown(&tw->tw_timer)) { > > > + /* releases @lock */ > > > + __inet_twsk_kill(tw, lock); > > > + } else { > > > > If we do not have a sync variant here, I think that inet_twsk_purge() > > could return while ongoing timers are alive. > > Yes. > > We can't use sync variant, it would deadlock on ehash spinlock. > > > tcp_sk_exit_batch() would then possibly hit : > > > > WARN_ON_ONCE(!refcount_dec_and_test(&net->ipv4.tcp_death_row.tw_refcount)); > > > > The alive timer are releasing tw->tw_dr->tw_refcount at the end of > > inet_twsk_kill() > > Theoretically the tw socket can be unlinked from the tw hash already > (inet_twsk_purge won't encounter it), but timer is still running. > > Only solution I see is to schedule() in tcp_sk_exit_batch() until > tw_refcount has dropped to the expected value, i.e. something like > > static void tcp_wait_for_tw_timers(struct net *n) > { > while (refcount_read(&n->ipv4.tcp_death_row.tw_refcount) > 1)) > schedule(); > } > > Any better idea? Maybe usleep_range(500, 1000) > > I started to sketch a patch that keeps PINNED as-is but schedules almost > all of the actual work to a work item. > > Idea was that it would lower RT latencies to acceptable level but it got > so ugly that I did not follow this path. > > I could resurrect this if you think its worth a try. I would rather avoid a work queue.
On 2024-06-03 13:21:52 [+0200], Florian Westphal wrote: > Theoretically the tw socket can be unlinked from the tw hash already > (inet_twsk_purge won't encounter it), but timer is still running. > > Only solution I see is to schedule() in tcp_sk_exit_batch() until > tw_refcount has dropped to the expected value, i.e. something like > > static void tcp_wait_for_tw_timers(struct net *n) > { > while (refcount_read(&n->ipv4.tcp_death_row.tw_refcount) > 1)) > schedule(); > } > > Any better idea? I need to read the thread but from the context, couldn't you - grab a refcount - timer_shutdown_sync() on it ? Sebastian
Florian Westphal <fw@strlen.de> wrote: > Eric Dumazet <edumazet@google.com> wrote: > > On Mon, Jun 3, 2024 at 11:37 AM Florian Westphal <fw@strlen.de> wrote: > > > + spin_lock(lock); > > > + if (timer_shutdown(&tw->tw_timer)) { > > > + /* releases @lock */ > > > + __inet_twsk_kill(tw, lock); > > > + } else { > > > > If we do not have a sync variant here, I think that inet_twsk_purge() > > could return while ongoing timers are alive. > > Yes. > > We can't use sync variant, it would deadlock on ehash spinlock. > > > tcp_sk_exit_batch() would then possibly hit : > > > > WARN_ON_ONCE(!refcount_dec_and_test(&net->ipv4.tcp_death_row.tw_refcount)); > > > > The alive timer are releasing tw->tw_dr->tw_refcount at the end of > > inet_twsk_kill() > > Theoretically the tw socket can be unlinked from the tw hash already > (inet_twsk_purge won't encounter it), but timer is still running. > > Only solution I see is to schedule() in tcp_sk_exit_batch() until > tw_refcount has dropped to the expected value, i.e. something like > > static void tcp_wait_for_tw_timers(struct net *n) > { > while (refcount_read(&n->ipv4.tcp_death_row.tw_refcount) > 1)) > schedule(); > } > > Any better idea? Actually, I think we can solve this in a much simpler way. Instead of replacing: void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) { if (del_timer_sync(&tw->tw_timer)) inet_twsk_kill(tw); inet_twsk_put(tw); } With: spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); spin_lock(lock); if (timer_shutdown(&tw->tw_timer)) { (Which gets us into the tcp_sk_exit_batch trouble Eric points out), we can simply add "empty" ehash lock unlock pair before calling del_timer_sync(): void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) { + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); + spin_lock(lock) + spin_unlock(lock) if (del_timer_sync(&tw->tw_timer)) inet_twsk_kill(tw); inet_twsk_put(tw); } Rationale: inet_twsk_deschedule_put() cannot be called before hashdance_schedule calls refcount_set(&tw->tw_refcnt, 3). Before this any refcount_inc_not_zero fails so we never get into deschedule_put. Hashdance_schedule holds the ehash lock when it sets the tw refcount. The lock is released only after the timer is up and running. When inet_twsk_deschedule_put() is called, and hashdance_schedule is not yet done, the spinlock/unlock pair will guarantee that the timer is up after the spin_unlock. I think this is much better than the schedule loop waiting for tw_dr refcount to drop, it mainly needs a comment to explain what this is doing. Thoughts?
On Mon, Jun 3, 2024 at 3:32 PM Florian Westphal <fw@strlen.de> wrote: > > Florian Westphal <fw@strlen.de> wrote: > > Eric Dumazet <edumazet@google.com> wrote: > > > On Mon, Jun 3, 2024 at 11:37 AM Florian Westphal <fw@strlen.de> wrote: > > > > + spin_lock(lock); > > > > + if (timer_shutdown(&tw->tw_timer)) { > > > > + /* releases @lock */ > > > > + __inet_twsk_kill(tw, lock); > > > > + } else { > > > > > > If we do not have a sync variant here, I think that inet_twsk_purge() > > > could return while ongoing timers are alive. > > > > Yes. > > > > We can't use sync variant, it would deadlock on ehash spinlock. > > > > > tcp_sk_exit_batch() would then possibly hit : > > > > > > WARN_ON_ONCE(!refcount_dec_and_test(&net->ipv4.tcp_death_row.tw_refcount)); > > > > > > The alive timer are releasing tw->tw_dr->tw_refcount at the end of > > > inet_twsk_kill() > > > > Theoretically the tw socket can be unlinked from the tw hash already > > (inet_twsk_purge won't encounter it), but timer is still running. > > > > Only solution I see is to schedule() in tcp_sk_exit_batch() until > > tw_refcount has dropped to the expected value, i.e. something like > > > > static void tcp_wait_for_tw_timers(struct net *n) > > { > > while (refcount_read(&n->ipv4.tcp_death_row.tw_refcount) > 1)) > > schedule(); > > } > > > > Any better idea? > > Actually, I think we can solve this in a much simpler way. > > Instead of replacing: > > void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) > { > if (del_timer_sync(&tw->tw_timer)) > inet_twsk_kill(tw); > inet_twsk_put(tw); > } > > With: > spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > spin_lock(lock); > if (timer_shutdown(&tw->tw_timer)) { > > (Which gets us into the tcp_sk_exit_batch trouble Eric points out), > we can simply add "empty" ehash lock unlock pair before calling > del_timer_sync(): > > void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) > { > + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); > + spin_lock(lock) > + spin_unlock(lock) > > if (del_timer_sync(&tw->tw_timer)) Sounds good, maybe use timer_shutdown_sync() here ? > inet_twsk_kill(tw); > inet_twsk_put(tw); > } > > Rationale: > inet_twsk_deschedule_put() cannot be called before hashdance_schedule > calls refcount_set(&tw->tw_refcnt, 3). > > Before this any refcount_inc_not_zero fails so we never get into > deschedule_put. > > Hashdance_schedule holds the ehash lock when it sets the tw refcount. > The lock is released only after the timer is up and running. > > When inet_twsk_deschedule_put() is called, and hashdance_schedule > is not yet done, the spinlock/unlock pair will guarantee that > the timer is up after the spin_unlock. > > I think this is much better than the schedule loop waiting for tw_dr > refcount to drop, it mainly needs a comment to explain what this is > doing. > > Thoughts?
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..2b1d836df64e 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -44,14 +44,14 @@ void inet_twsk_bind_unhash(struct inet_timewait_sock *tw, __sock_put((struct sock *)tw); } -/* Must be called with locally disabled BHs. */ -static void inet_twsk_kill(struct inet_timewait_sock *tw) +static void __inet_twsk_kill(struct inet_timewait_sock *tw, spinlock_t *lock) +__releases(lock) { struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; - spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); struct inet_bind_hashbucket *bhead, *bhead2; - spin_lock(lock); + lockdep_assert_held(lock); + sk_nulls_del_node_init_rcu((struct sock *)tw); spin_unlock(lock); @@ -71,6 +71,16 @@ static void inet_twsk_kill(struct inet_timewait_sock *tw) inet_twsk_put(tw); } +/* Must be called with locally disabled BHs. */ +static void inet_twsk_kill(struct inet_timewait_sock *tw) +{ + struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); + + spin_lock(lock); + __inet_twsk_kill(tw, lock); +} + void inet_twsk_free(struct inet_timewait_sock *tw) { struct module *owner = tw->tw_prot->owner; @@ -96,9 +106,13 @@ 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 +143,33 @@ 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) { @@ -217,8 +238,37 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); */ void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) { - if (del_timer_sync(&tw->tw_timer)) - inet_twsk_kill(tw); + 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). + * + * With plain timer_shutdown_sync() and without grabbing the ehash + * lock, we can 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); + if (timer_shutdown(&tw->tw_timer)) { + /* releases @lock */ + __inet_twsk_kill(tw, lock); + } else { + spin_unlock(lock); + } + inet_twsk_put(tw); } EXPORT_SYMBOL(inet_twsk_deschedule_put); 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