From patchwork Mon Apr 15 11:34:35 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Valentin Schneider X-Patchwork-Id: 13629834 X-Patchwork-Delegate: kuba@kernel.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id F02CD627FC for ; Mon, 15 Apr 2024 11:35:04 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.133.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1713180906; cv=none; b=nr5Cypank/gzU7dgDSmYx0eWS03VatLrJjJzIbZTp6w8UgsP0RidLk9qSR0J4T96Uwv+5P9F4LZcs2ehoYcbq8oPfQUtInR5eNffz+oZfkXklndR3zEZ7erUekLXsZvWrvyn552N5fMqGLDldn/AeqJjNQVv1O8QyUF++zekUDI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1713180906; c=relaxed/simple; bh=tYQjaVztYSLksCUGTNtr0f1PS0mrayvCCU6iPNxYEsw=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=X52T2+3GM/exzUgRMJjh7CMtuXfc2Ds0rfCj62vc5N3oKUAnmb8BuJ0+kjFccsn6pQfK/Fn/broKv9nmAZdp6r+s5uI8XK6Z1ic7/JBGe9xq0dLWt4J1ZZ+nr0JDDaHCnfBLdXCQW4ll9KhYTF1V9GFp3RqmArEfqw6zAUlFy7A= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=SiBdg508; arc=none smtp.client-ip=170.10.133.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="SiBdg508" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1713180904; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=7f1OoUgw4H495Xl3fEpds6BjBxuOUZU0DulXURwGAsY=; b=SiBdg508gikV7E5EQMhoXYIismF0rVRs+GGPZm+FBWIesRlVNgI/INdh6sMdkrqgH7L8u/ TMSpk0/gFffz+UI06qcAvuRrx2nL2chCspMXIYP1yhkbNeYp4antHBQC/qk04fRjbfOTRR 1vNMQU/dXWTokoQ+IsPvKc8NRInpC5E= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-636--qhKjR8DOYursTvlK__Fjg-1; Mon, 15 Apr 2024 07:34:59 -0400 X-MC-Unique: -qhKjR8DOYursTvlK__Fjg-1 Received: from smtp.corp.redhat.com (int-mx07.intmail.prod.int.rdu2.redhat.com [10.11.54.7]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 301CD1049BC9; Mon, 15 Apr 2024 11:34:59 +0000 (UTC) Received: from vschneid-thinkpadt14sgen2i.remote.csb (unknown [10.39.193.251]) by smtp.corp.redhat.com (Postfix) with ESMTPS id D6D991C06889; Mon, 15 Apr 2024 11:34:56 +0000 (UTC) From: Valentin Schneider To: dccp@vger.kernel.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, linux-rt-users@vger.kernel.org Cc: "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , mleitner@redhat.com, David Ahern , Juri Lelli , Tomas Glozar , Sebastian Andrzej Siewior , Thomas Gleixner Subject: [PATCH v5 1/2] SQUASH: tcp/dcpp: Convert timewait timer into a delayed_work Date: Mon, 15 Apr 2024 13:34:35 +0200 Message-ID: <20240415113436.3261042-2-vschneid@redhat.com> In-Reply-To: <20240415113436.3261042-1-vschneid@redhat.com> References: <20240415113436.3261042-1-vschneid@redhat.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.4.1 on 10.11.54.7 X-Patchwork-Delegate: bpf@iogearbox.net !NOTE! this changes behaviour, as the tw_timer is TIMER_PINNED and delayed works currently cannot behave as such. TIMER_PINNED timers are enqueued onto the local base, whereas delayed_works have their timer enqueued on the global base AND also check for HK_TYPE_TIMER isolation. The split between this commit and the next is mainly there for ease of reviewing. This commit should be squashed with the next one. Signed-off-by: Valentin Schneider --- include/net/inet_timewait_sock.h | 2 +- net/ipv4/inet_diag.c | 2 +- net/ipv4/inet_timewait_sock.c | 16 +++++++++------- net/ipv4/tcp_ipv4.c | 2 +- net/ipv6/tcp_ipv6.c | 2 +- .../testing/selftests/bpf/progs/bpf_iter_tcp4.c | 2 +- .../testing/selftests/bpf/progs/bpf_iter_tcp6.c | 2 +- 7 files changed, 15 insertions(+), 13 deletions(-) diff --git a/include/net/inet_timewait_sock.h b/include/net/inet_timewait_sock.h index f28da08a37b4e..c4d64f1f8d415 100644 --- a/include/net/inet_timewait_sock.h +++ b/include/net/inet_timewait_sock.h @@ -72,7 +72,7 @@ struct inet_timewait_sock { tw_tos : 8; u32 tw_txhash; u32 tw_priority; - struct timer_list tw_timer; + struct delayed_work tw_expiry_work; struct inet_bind_bucket *tw_tb; struct inet_bind2_bucket *tw_tb2; }; diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c index 7adace541fe29..ab11b688f1eeb 100644 --- a/net/ipv4/inet_diag.c +++ b/net/ipv4/inet_diag.c @@ -444,7 +444,7 @@ static int inet_twsk_diag_fill(struct sock *sk, r->idiag_state = tw->tw_substate; r->idiag_timer = 3; - tmo = tw->tw_timer.expires - jiffies; + tmo = tw->tw_expiry_work.timer.expires - jiffies; r->idiag_expires = jiffies_delta_to_msecs(tmo); r->idiag_rqueue = 0; r->idiag_wqueue = 0; diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index e8de45d34d56a..7a2dcbaa1a61e 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -150,11 +150,13 @@ void inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, } EXPORT_SYMBOL_GPL(inet_twsk_hashdance); -static void tw_timer_handler(struct timer_list *t) +static void tw_expiry_workfn(struct work_struct *work) { - struct inet_timewait_sock *tw = from_timer(tw, t, tw_timer); - + struct inet_timewait_sock *tw = container_of( + work, struct inet_timewait_sock, tw_expiry_work.work); + local_bh_disable(); inet_twsk_kill(tw); + local_bh_enable(); } struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, @@ -192,7 +194,7 @@ struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, tw->tw_prot = sk->sk_prot_creator; atomic64_set(&tw->tw_cookie, atomic64_read(&sk->sk_cookie)); twsk_net_set(tw, sock_net(sk)); - timer_setup(&tw->tw_timer, tw_timer_handler, TIMER_PINNED); + INIT_DELAYED_WORK(&tw->tw_expiry_work, tw_expiry_workfn); /* * Because we use RCU lookups, we should not set tw_refcnt * to a non null value before everything is setup for this @@ -217,7 +219,7 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); */ void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) { - if (del_timer_sync(&tw->tw_timer)) + if (cancel_delayed_work_sync(&tw->tw_expiry_work)) inet_twsk_kill(tw); inet_twsk_put(tw); } @@ -255,10 +257,10 @@ void __inet_twsk_schedule(struct inet_timewait_sock *tw, int timeo, bool rearm) __NET_INC_STATS(twsk_net(tw), kill ? LINUX_MIB_TIMEWAITKILLED : LINUX_MIB_TIMEWAITED); - BUG_ON(mod_timer(&tw->tw_timer, jiffies + timeo)); + BUG_ON(!queue_delayed_work(system_unbound_wq, &tw->tw_expiry_work, timeo)); refcount_inc(&tw->tw_dr->tw_refcount); } else { - mod_timer_pending(&tw->tw_timer, jiffies + timeo); + mod_delayed_work(system_unbound_wq, &tw->tw_expiry_work, timeo); } } EXPORT_SYMBOL_GPL(__inet_twsk_schedule); diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index a22ee58387518..4b106f017a81f 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -2920,7 +2920,7 @@ static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i) static void get_timewait4_sock(const struct inet_timewait_sock *tw, struct seq_file *f, int i) { - long delta = tw->tw_timer.expires - jiffies; + long delta = tw->tw_expiry_work.timer.expires - jiffies; __be32 dest, src; __u16 destp, srcp; diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 3f4cba49e9ee6..58cd12fdc91a5 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -2242,7 +2242,7 @@ static void get_tcp6_sock(struct seq_file *seq, struct sock *sp, int i) static void get_timewait6_sock(struct seq_file *seq, struct inet_timewait_sock *tw, int i) { - long delta = tw->tw_timer.expires - jiffies; + long delta = tw->tw_expiry_work.timer.expires - jiffies; const struct in6_addr *dest, *src; __u16 destp, srcp; diff --git a/tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c b/tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c index 92267abb462fc..a429f4eb9939c 100644 --- a/tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c +++ b/tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c @@ -152,7 +152,7 @@ static int dump_tw_sock(struct seq_file *seq, struct tcp_timewait_sock *ttw, __be32 dest, src; long delta; - delta = tw->tw_timer.expires - bpf_jiffies64(); + delta = tw->tw_expiry_work.timer.expires - bpf_jiffies64(); dest = tw->tw_daddr; src = tw->tw_rcv_saddr; destp = bpf_ntohs(tw->tw_dport); diff --git a/tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c b/tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c index 943f7bba180e7..795bb34c95f72 100644 --- a/tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c +++ b/tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c @@ -157,7 +157,7 @@ static int dump_tw_sock(struct seq_file *seq, struct tcp_timewait_sock *ttw, __u16 destp, srcp; long delta; - delta = tw->tw_timer.expires - bpf_jiffies64(); + delta = tw->tw_expiry_work.timer.expires - bpf_jiffies64(); dest = &tw->tw_v6_daddr; src = &tw->tw_v6_rcv_saddr; destp = bpf_ntohs(tw->tw_dport); From patchwork Mon Apr 15 11:34:36 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Valentin Schneider X-Patchwork-Id: 13629835 X-Patchwork-Delegate: kuba@kernel.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 31AA060ED1 for ; Mon, 15 Apr 2024 11:35:07 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1713180908; cv=none; b=XwMtOUdBd3g6dAcYPYN3xfhVPfqL1VOByQq9HpdIplx4iU9y+QpNCdCXlBgsaid/qbXJ3p9uGqZJLV9dHpRXWF7n2Vh8dJxW2UviMgiHlrP4LOcwwV/ay/tjh/vMtCkEABtTUnxRizTujC/eYcecc7S5fzo06hd5nPOoETLiB08= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1713180908; c=relaxed/simple; bh=st9rueJJE/E+kECJZK5RAvKXy5xFFjU4dzJnO6U1R/8=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=sQEdoecw3Q4LXq13gdLOObg6pTQXS93q1aSYZ8mpXU8Z/5CZc48JF6X+nLuH82/DM6sfhFXcDb2PC49tH4DitHT0Yu5fhknIN5kbck7bkm7XfZxoKm/edqOXr88rrlYiWlvRemDNVfcrAP3ff9N4c5A4p24ScJ4wyw9aS430PwA= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=a/V7tALp; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="a/V7tALp" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1713180906; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=CqhmKrYfV//Q+UEBRlZmTN84xreN6GieAMgq8Wr615k=; b=a/V7tALpQPGDVm7dg7G/O+zY1RfhYcDTg5D17Hv7Ex5WWIuNJ8drElUanGrwTctaFCKeJc EjnihBmw1qjNUc0w+J7KV71L/akWTPDfscbqqsH93+M2tl8qKEWPs6TnrS43rvOjKpvIUj cjAvHuMZkwvb9n4IwF4S6rjE3NAfNzg= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-245-aqh76rrpN5WZWRiKHjz2GQ-1; Mon, 15 Apr 2024 07:35:03 -0400 X-MC-Unique: aqh76rrpN5WZWRiKHjz2GQ-1 Received: from smtp.corp.redhat.com (int-mx07.intmail.prod.int.rdu2.redhat.com [10.11.54.7]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 5FAE68007BA; Mon, 15 Apr 2024 11:35:02 +0000 (UTC) Received: from vschneid-thinkpadt14sgen2i.remote.csb (unknown [10.39.193.251]) by smtp.corp.redhat.com (Postfix) with ESMTPS id 6B81E1C06701; Mon, 15 Apr 2024 11:34:59 +0000 (UTC) From: Valentin Schneider To: dccp@vger.kernel.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, linux-rt-users@vger.kernel.org Cc: "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , mleitner@redhat.com, David Ahern , Juri Lelli , Tomas Glozar , Sebastian Andrzej Siewior , Thomas Gleixner Subject: [PATCH v5 2/2] tcp/dcpp: Un-pin tw_timer Date: Mon, 15 Apr 2024 13:34:36 +0200 Message-ID: <20240415113436.3261042-3-vschneid@redhat.com> In-Reply-To: <20240415113436.3261042-1-vschneid@redhat.com> References: <20240415113436.3261042-1-vschneid@redhat.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.4.1 on 10.11.54.7 X-Patchwork-Delegate: kuba@kernel.org Context ======= 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 e02b93124855 ("workqueue: Unbind kworkers before sending them to exit()") Changes ======= 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() Lastly, to prevent the ehash lock from being taken both in timer context and in timer deletion context (which leads to deadlocks), move the inet_twsk_kill() from timer expiry context to workqueue context by changing the tw_timer to a delayed_work. Thanks to Paolo for suggesting to leverage the ehash lock. Note on partial reverts ======================= This partially reverts ec94c2696f0b ("tcp/dccp: avoid one atomic operation for timewait hashdance") This also reinstores a comment from 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. Link: https://lore.kernel.org/all/ZPhpfMjSiHVjQkTk@localhost.localdomain/ Signed-off-by: Valentin Schneider --- include/net/inet_timewait_sock.h | 6 ++-- net/dccp/minisocks.c | 9 +++-- net/ipv4/inet_timewait_sock.c | 59 +++++++++++++++++++++++--------- net/ipv4/tcp_minisocks.c | 9 +++-- 4 files changed, 55 insertions(+), 28 deletions(-) diff --git a/include/net/inet_timewait_sock.h b/include/net/inet_timewait_sock.h index c4d64f1f8d415..93a50eed41872 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 64d805b27adde..dd0b09553f142 100644 --- a/net/dccp/minisocks.c +++ b/net/dccp/minisocks.c @@ -58,11 +58,10 @@ 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); + /* Linkage updates. */ + inet_twsk_hashdance_schedule(tw, sk, &dccp_hashinfo, timeo); + /* Access to tw after this point is illegal. */ + inet_twsk_put(tw); 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 7a2dcbaa1a61e..165b9cdf27e1a 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; @@ -97,8 +107,10 @@ static void inet_twsk_add_node_rcu(struct inet_timewait_sock *tw, * Essentially we whip up a timewait bucket, copy the relevant info into it * from the SK, and mess with hash chains and list linkage. */ -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 +141,33 @@ 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); - /* tw_refcnt is set to 3 because we have : + /* Ensure above writes are committed into memory before updating the + * refcount. + * Provides ordering vs later refcount_inc(). + */ + smp_wmb(); + /* tw_refcnt is set to 4 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. + * - one reference for ourself (our caller will release it). */ - refcount_set(&tw->tw_refcnt, 3); + refcount_set(&tw->tw_refcnt, 4); + + inet_twsk_schedule(tw, timeo); + + spin_unlock(lock); + } -EXPORT_SYMBOL_GPL(inet_twsk_hashdance); +EXPORT_SYMBOL_GPL(inet_twsk_hashdance_schedule); static void tw_expiry_workfn(struct work_struct *work) { @@ -219,8 +238,16 @@ EXPORT_SYMBOL_GPL(inet_twsk_alloc); */ void inet_twsk_deschedule_put(struct inet_timewait_sock *tw) { - if (cancel_delayed_work_sync(&tw->tw_expiry_work)) - inet_twsk_kill(tw); + struct inet_hashinfo *hashinfo = tw->tw_dr->hashinfo; + spinlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash); + + spin_lock(lock); + if (cancel_delayed_work_sync(&tw->tw_expiry_work)) { + /* 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 f0761f060a837..77e2515409e40 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -343,11 +343,10 @@ 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); + /* Linkage updates. */ + inet_twsk_hashdance_schedule(tw, sk, net->ipv4.tcp_death_row.hashinfo, timeo); + /* Access to tw after this point is illegal. */ + inet_twsk_put(tw); local_bh_enable(); } else { /* Sorry, if we're out of memory, just CLOSE this