kernel/linux.git/net/mptcp/protocol.c, branch linux-6.0.y

mptcp: fix lockdep false positive

2023-01-12T11:00:30+00:00

[ Upstream commit fec3adfd754ccc99a7230e8ab9f105b65fb07bcc ] MattB reported a lockdep splat in the mptcp listener code cleanup: WARNING: possible circular locking dependency detected packetdrill/14278 is trying to acquire lock: ffff888017d868f0 ((work_completion)(&msk->work)){+.+.}-{0:0}, at: __flush_work (kernel/workqueue.c:3069) but task is already holding lock: ffff888017d84130 (sk_lock-AF_INET){+.+.}-{0:0}, at: mptcp_close (net/mptcp/protocol.c:2973) which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #1 (sk_lock-AF_INET){+.+.}-{0:0}: __lock_acquire (kernel/locking/lockdep.c:5055) lock_acquire (kernel/locking/lockdep.c:466) lock_sock_nested (net/core/sock.c:3463) mptcp_worker (net/mptcp/protocol.c:2614) process_one_work (kernel/workqueue.c:2294) worker_thread (include/linux/list.h:292) kthread (kernel/kthread.c:376) ret_from_fork (arch/x86/entry/entry_64.S:312) -> #0 ((work_completion)(&msk->work)){+.+.}-{0:0}: check_prev_add (kernel/locking/lockdep.c:3098) validate_chain (kernel/locking/lockdep.c:3217) __lock_acquire (kernel/locking/lockdep.c:5055) lock_acquire (kernel/locking/lockdep.c:466) __flush_work (kernel/workqueue.c:3070) __cancel_work_timer (kernel/workqueue.c:3160) mptcp_cancel_work (net/mptcp/protocol.c:2758) mptcp_subflow_queue_clean (net/mptcp/subflow.c:1817) __mptcp_close_ssk (net/mptcp/protocol.c:2363) mptcp_destroy_common (net/mptcp/protocol.c:3170) mptcp_destroy (include/net/sock.h:1495) __mptcp_destroy_sock (net/mptcp/protocol.c:2886) __mptcp_close (net/mptcp/protocol.c:2959) mptcp_close (net/mptcp/protocol.c:2974) inet_release (net/ipv4/af_inet.c:432) __sock_release (net/socket.c:651) sock_close (net/socket.c:1367) __fput (fs/file_table.c:320) task_work_run (kernel/task_work.c:181 (discriminator 1)) exit_to_user_mode_prepare (include/linux/resume_user_mode.h:49) syscall_exit_to_user_mode (kernel/entry/common.c:130) do_syscall_64 (arch/x86/entry/common.c:87) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:120) other info that might help us debug this: Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(sk_lock-AF_INET); lock((work_completion)(&msk->work)); lock(sk_lock-AF_INET); lock((work_completion)(&msk->work)); *** DEADLOCK *** The report is actually a false positive, since the only existing lock nesting is the msk socket lock acquired by the mptcp work. cancel_work_sync() is invoked without the relevant socket lock being held, but under a different (the msk listener) socket lock. We could silence the splat adding a per workqueue dynamic lockdep key, but that looks overkill. Instead just tell lockdep the msk socket lock is not held around cancel_work_sync(). Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/322 Fixes: 30e51b923e43 ("mptcp: fix unreleased socket in accept queue") Reported-by: Matthieu Baerts Reviewed-by: Mat Martineau Signed-off-by: Paolo Abeni Signed-off-by: Mat Martineau Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

mptcp: don't orphan ssk in mptcp_close()

2022-12-08T10:30:17+00:00

[ Upstream commit fe94800184f22d4778628f1321dce5acb7513d84 ] All of the subflows of a msk will be orphaned in mptcp_close(), which means the subflows are in DEAD state. After then, DATA_FIN will be sent, and the other side will response with a DATA_ACK for this DATA_FIN. However, if the other side still has pending data, the data that received on these subflows will not be passed to the msk, as they are DEAD and subflow_data_ready() will not be called in tcp_data_ready(). Therefore, these data can't be acked, and they will be retransmitted again and again, until timeout. Fix this by setting ssk->sk_socket and ssk->sk_wq to 'NULL', instead of orphaning the subflows in __mptcp_close(), as Paolo suggested. Fixes: e16163b6e2b7 ("mptcp: refactor shutdown and close") Reviewed-by: Biao Jiang Reviewed-by: Mengen Sun Signed-off-by: Menglong Dong Reviewed-by: Paolo Abeni Signed-off-by: Matthieu Baerts Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

mptcp: set msk local address earlier

2022-11-03T15:00:31+00:00

[ Upstream commit e72e4032637f4646554794ac28a3abecc6c2416d ] The mptcp_pm_nl_get_local_id() code assumes that the msk local address is available at that point. For passive sockets, we initialize such address at accept() time. Depending on the running configuration and the user-space timing, a passive MPJ subflow can join the msk socket before accept() completes. In such case, the PM assigns a wrong local id to the MPJ subflow and later PM netlink operations will end-up touching the wrong/unexpected subflow. All the above causes sporadic self-tests failures, especially when the host is heavy loaded. Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/308 Fixes: 01cacb00b35c ("mptcp: add netlink-based PM") Fixes: d045b9eb95a9 ("mptcp: introduce implicit endpoints") Reviewed-by: Mat Martineau Signed-off-by: Paolo Abeni Signed-off-by: Mat Martineau Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

mptcp: fix unreleased socket in accept queue

2022-09-29T02:05:21+00:00

The mptcp socket and its subflow sockets in accept queue can't be released after the process exit. While the release of a mptcp socket in listening state, the corresponding tcp socket will be released too. Meanwhile, the tcp socket in the unaccept queue will be released too. However, only init subflow is in the unaccept queue, and the joined subflow is not in the unaccept queue, which makes the joined subflow won't be released, and therefore the corresponding unaccepted mptcp socket will not be released to. This can be reproduced easily with following steps: 1. create 2 namespace and veth: $ ip netns add mptcp-client $ ip netns add mptcp-server $ sysctl -w net.ipv4.conf.all.rp_filter=0 $ ip netns exec mptcp-client sysctl -w net.mptcp.enabled=1 $ ip netns exec mptcp-server sysctl -w net.mptcp.enabled=1 $ ip link add red-client netns mptcp-client type veth peer red-server \ netns mptcp-server $ ip -n mptcp-server address add 10.0.0.1/24 dev red-server $ ip -n mptcp-server address add 192.168.0.1/24 dev red-server $ ip -n mptcp-client address add 10.0.0.2/24 dev red-client $ ip -n mptcp-client address add 192.168.0.2/24 dev red-client $ ip -n mptcp-server link set red-server up $ ip -n mptcp-client link set red-client up 2. configure the endpoint and limit for client and server: $ ip -n mptcp-server mptcp endpoint flush $ ip -n mptcp-server mptcp limits set subflow 2 add_addr_accepted 2 $ ip -n mptcp-client mptcp endpoint flush $ ip -n mptcp-client mptcp limits set subflow 2 add_addr_accepted 2 $ ip -n mptcp-client mptcp endpoint add 192.168.0.2 dev red-client id \ 1 subflow 3. listen and accept on a port, such as 9999. The nc command we used here is modified, which makes it use mptcp protocol by default. $ ip netns exec mptcp-server nc -l -k -p 9999 4. open another *two* terminal and use each of them to connect to the server with the following command: $ ip netns exec mptcp-client nc 10.0.0.1 9999 Input something after connect to trigger the connection of the second subflow. So that there are two established mptcp connections, with the second one still unaccepted. 5. exit all the nc command, and check the tcp socket in server namespace. And you will find that there is one tcp socket in CLOSE_WAIT state and can't release forever. Fix this by closing all of the unaccepted mptcp socket in mptcp_subflow_queue_clean() with __mptcp_close(). Now, we can ensure that all unaccepted mptcp sockets will be cleaned by __mptcp_close() before they are released, so mptcp_sock_destruct(), which is used to clean the unaccepted mptcp socket, is not needed anymore. The selftests for mptcp is ran for this commit, and no new failures. Fixes: f296234c98a8 ("mptcp: Add handling of incoming MP_JOIN requests") Fixes: 6aeed9045071 ("mptcp: fix race on unaccepted mptcp sockets") Cc: stable@vger.kernel.org Reviewed-by: Jiang Biao Reviewed-by: Mengen Sun Acked-by: Paolo Abeni Signed-off-by: Menglong Dong Signed-off-by: Mat Martineau Signed-off-by: Jakub Kicinski

mptcp: factor out __mptcp_close() without socket lock

2022-09-29T02:05:21+00:00

Factor out __mptcp_close() from mptcp_close(). The caller of __mptcp_close() should hold the socket lock, and cancel mptcp work when __mptcp_close() returns true. This function will be used in the next commit. Fixes: f296234c98a8 ("mptcp: Add handling of incoming MP_JOIN requests") Fixes: 6aeed9045071 ("mptcp: fix race on unaccepted mptcp sockets") Cc: stable@vger.kernel.org Reviewed-by: Jiang Biao Reviewed-by: Mengen Sun Acked-by: Paolo Abeni Signed-off-by: Menglong Dong Signed-off-by: Mat Martineau Signed-off-by: Jakub Kicinski

mptcp: fix fwd memory accounting on coalesce

2022-09-13T08:18:44+00:00

The intel bot reported a memory accounting related splat: [ 240.473094] ------------[ cut here ]------------ [ 240.478507] page_counter underflow: -4294828518 nr_pages=4294967290 [ 240.485500] WARNING: CPU: 2 PID: 14986 at mm/page_counter.c:56 page_counter_cancel+0x96/0xc0 [ 240.570849] CPU: 2 PID: 14986 Comm: mptcp_connect Tainted: G S 5.19.0-rc4-00739-gd24141fe7b48 #1 [ 240.581637] Hardware name: HP HP Z240 SFF Workstation/802E, BIOS N51 Ver. 01.63 10/05/2017 [ 240.590600] RIP: 0010:page_counter_cancel+0x96/0xc0 [ 240.596179] Code: 00 00 00 45 31 c0 48 89 ef 5d 4c 89 c6 41 5c e9 40 fd ff ff 4c 89 e2 48 c7 c7 20 73 39 84 c6 05 d5 b1 52 04 01 e8 e7 95 f3 01 <0f> 0b eb a9 48 89 ef e8 1e 25 fc ff eb c3 66 66 2e 0f 1f 84 00 00 [ 240.615639] RSP: 0018:ffffc9000496f7c8 EFLAGS: 00010082 [ 240.621569] RAX: 0000000000000000 RBX: ffff88819c9c0120 RCX: 0000000000000000 [ 240.629404] RDX: 0000000000000027 RSI: 0000000000000004 RDI: fffff5200092deeb [ 240.637239] RBP: ffff88819c9c0120 R08: 0000000000000001 R09: ffff888366527a2b [ 240.645069] R10: ffffed106cca4f45 R11: 0000000000000001 R12: 00000000fffffffa [ 240.652903] R13: ffff888366536118 R14: 00000000fffffffa R15: ffff88819c9c0000 [ 240.660738] FS: 00007f3786e72540(0000) GS:ffff888366500000(0000) knlGS:0000000000000000 [ 240.669529] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 240.675974] CR2: 00007f966b346000 CR3: 0000000168cea002 CR4: 00000000003706e0 [ 240.683807] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 240.691641] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 240.699468] Call Trace: [ 240.702613] [ 240.705413] page_counter_uncharge+0x29/0x80 [ 240.710389] drain_stock+0xd0/0x180 [ 240.714585] refill_stock+0x278/0x580 [ 240.718951] __sk_mem_reduce_allocated+0x222/0x5c0 [ 240.729248] __mptcp_update_rmem+0x235/0x2c0 [ 240.734228] __mptcp_move_skbs+0x194/0x6c0 [ 240.749764] mptcp_recvmsg+0xdfa/0x1340 [ 240.763153] inet_recvmsg+0x37f/0x500 [ 240.782109] sock_read_iter+0x24a/0x380 [ 240.805353] new_sync_read+0x420/0x540 [ 240.838552] vfs_read+0x37f/0x4c0 [ 240.842582] ksys_read+0x170/0x200 [ 240.864039] do_syscall_64+0x5c/0x80 [ 240.872770] entry_SYSCALL_64_after_hwframe+0x46/0xb0 [ 240.878526] RIP: 0033:0x7f3786d9ae8e [ 240.882805] Code: c0 e9 b6 fe ff ff 50 48 8d 3d 6e 18 0a 00 e8 89 e8 01 00 66 0f 1f 84 00 00 00 00 00 64 8b 04 25 18 00 00 00 85 c0 75 14 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 66 0f 1f 84 00 00 00 00 00 48 83 ec 28 [ 240.902259] RSP: 002b:00007fff7be81e08 EFLAGS: 00000246 ORIG_RAX: 0000000000000000 [ 240.910533] RAX: ffffffffffffffda RBX: 0000000000002000 RCX: 00007f3786d9ae8e [ 240.918368] RDX: 0000000000002000 RSI: 00007fff7be87ec0 RDI: 0000000000000005 [ 240.926206] RBP: 0000000000000005 R08: 00007f3786e6a230 R09: 00007f3786e6a240 [ 240.934046] R10: fffffffffffff288 R11: 0000000000000246 R12: 0000000000002000 [ 240.941884] R13: 00007fff7be87ec0 R14: 00007fff7be87ec0 R15: 0000000000002000 [ 240.949741] [ 240.952632] irq event stamp: 27367 [ 240.956735] hardirqs last enabled at (27366): [] mem_cgroup_uncharge_skmem+0x6a/0x80 [ 240.966848] hardirqs last disabled at (27367): [] refill_stock+0x282/0x580 [ 240.976017] softirqs last enabled at (27360): [] mptcp_recvmsg+0xaf/0x1340 [ 240.985273] softirqs last disabled at (27364): [] __mptcp_move_skbs+0x18c/0x6c0 [ 240.994872] ---[ end trace 0000000000000000 ]--- After commit d24141fe7b48 ("mptcp: drop SK_RECLAIM_* macros"), if rmem_fwd_alloc become negative, mptcp_rmem_uncharge() can try to reclaim a negative amount of pages, since the expression: reclaimable >= PAGE_SIZE will evaluate to true for any negative value of the int 'reclaimable': 'PAGE_SIZE' is an unsigned long and the negative integer will be promoted to a (very large) unsigned long value. Still after the mentioned commit, kfree_skb_partial() in mptcp_try_coalesce() will reclaim most of just released fwd memory, so that following charging of the skb delta size will lead to negative fwd memory values. At that point a racing recvmsg() can trigger the splat. Address the issue switching the order of the memory accounting operations. The fwd memory can still transiently reach negative values, but that will happen in an atomic scope and no code path could touch/use such value. Reported-by: kernel test robot Fixes: d24141fe7b48 ("mptcp: drop SK_RECLAIM_* macros") Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts Signed-off-by: Matthieu Baerts Link: https://lore.kernel.org/r/20220906180404.1255873-1-matthieu.baerts@tessares.net Signed-off-by: Paolo Abeni

net: Fix data-races around sysctl_max_skb_frags.

2022-08-24T12:46:58+00:00

While reading sysctl_max_skb_frags, it can be changed concurrently. Thus, we need to add READ_ONCE() to its readers. Fixes: 5f74f82ea34c ("net:Add sysctl_max_skb_frags") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller

mptcp: do not queue data on closed subflows

2022-08-05T07:51:28+00:00

Dipanjan reported a syzbot splat at close time: WARNING: CPU: 1 PID: 10818 at net/ipv4/af_inet.c:153 inet_sock_destruct+0x6d0/0x8e0 net/ipv4/af_inet.c:153 Modules linked in: uio_ivshmem(OE) uio(E) CPU: 1 PID: 10818 Comm: kworker/1:16 Tainted: G OE 5.19.0-rc6-g2eae0556bb9d #2 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 Workqueue: events mptcp_worker RIP: 0010:inet_sock_destruct+0x6d0/0x8e0 net/ipv4/af_inet.c:153 Code: 21 02 00 00 41 8b 9c 24 28 02 00 00 e9 07 ff ff ff e8 34 4d 91 f9 89 ee 4c 89 e7 e8 4a 47 60 ff e9 a6 fc ff ff e8 20 4d 91 f9 <0f> 0b e9 84 fe ff ff e8 14 4d 91 f9 0f 0b e9 d4 fd ff ff e8 08 4d RSP: 0018:ffffc9001b35fa78 EFLAGS: 00010246 RAX: 0000000000000000 RBX: 00000000002879d0 RCX: ffff8881326f3b00 RDX: 0000000000000000 RSI: ffff8881326f3b00 RDI: 0000000000000002 RBP: ffff888179662674 R08: ffffffff87e983a0 R09: 0000000000000000 R10: 0000000000000005 R11: 00000000000004ea R12: ffff888179662400 R13: ffff888179662428 R14: 0000000000000001 R15: ffff88817e38e258 FS: 0000000000000000(0000) GS:ffff8881f5f00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000020007bc0 CR3: 0000000179592000 CR4: 0000000000150ee0 Call Trace: __sk_destruct+0x4f/0x8e0 net/core/sock.c:2067 sk_destruct+0xbd/0xe0 net/core/sock.c:2112 __sk_free+0xef/0x3d0 net/core/sock.c:2123 sk_free+0x78/0xa0 net/core/sock.c:2134 sock_put include/net/sock.h:1927 [inline] __mptcp_close_ssk+0x50f/0x780 net/mptcp/protocol.c:2351 __mptcp_destroy_sock+0x332/0x760 net/mptcp/protocol.c:2828 mptcp_worker+0x5d2/0xc90 net/mptcp/protocol.c:2586 process_one_work+0x9cc/0x1650 kernel/workqueue.c:2289 worker_thread+0x623/0x1070 kernel/workqueue.c:2436 kthread+0x2e9/0x3a0 kernel/kthread.c:376 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:302 The root cause of the problem is that an mptcp-level (re)transmit can race with mptcp_close() and the packet scheduler checks the subflow state before acquiring the socket lock: we can try to (re)transmit on an already closed ssk. Fix the issue checking again the subflow socket status under the subflow socket lock protection. Additionally add the missing check for the fallback-to-tcp case. Fixes: d5f49190def6 ("mptcp: allow picking different xmit subflows") Reported-by: Dipanjan Das Reviewed-by: Mat Martineau Signed-off-by: Paolo Abeni Signed-off-by: Mat Martineau Signed-off-by: David S. Miller

mptcp: move subflow cleanup in mptcp_destroy_common()

2022-08-05T07:51:28+00:00

If the mptcp socket creation fails due to a CGROUP_INET_SOCK_CREATE eBPF program, the MPTCP protocol ends-up leaking all the subflows: the related cleanup happens in __mptcp_destroy_sock() that is not invoked in such code path. Address the issue moving the subflow sockets cleanup in the mptcp_destroy_common() helper, which is invoked in every msk cleanup path. Additionally get rid of the intermediate list_splice_init step, which is an unneeded relic from the past. The issue is present since before the reported root cause commit, but any attempt to backport the fix before that hash will require a complete rewrite. Fixes: e16163b6e2 ("mptcp: refactor shutdown and close") Reported-by: Nguyen Dinh Phi Reviewed-by: Mat Martineau Co-developed-by: Nguyen Dinh Phi Signed-off-by: Nguyen Dinh Phi Signed-off-by: Paolo Abeni Signed-off-by: Mat Martineau Signed-off-by: David S. Miller

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2022-07-29T01:21:16+00:00

No conflicts. Signed-off-by: Jakub Kicinski