mptcp: fix possible stall on recvmsg()

recvmsg() can enter an infinite loop if the caller provides the
MSG_WAITALL, the data present in the receive queue is not sufficient to
fulfill the request, and no more data is received by the peer.

When the above happens, mptcp_wait_data() will always return with
no wait, as the MPTCP_DATA_READY flag checked by such function is
set and never cleared in such code path.

Leveraging the above syzbot was able to trigger an RCU stall:

rcu: INFO: rcu_preempt self-detected stall on CPU
rcu:    0-...!: (10499 ticks this GP) idle=0af/1/0x4000000000000000 softirq=10678/10678 fqs=1
        (t=10500 jiffies g=13089 q=109)
rcu: rcu_preempt kthread starved for 10497 jiffies! g13089 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x0 ->cpu=1
rcu:    Unless rcu_preempt kthread gets sufficient CPU time, OOM is now expected behavior.
rcu: RCU grace-period kthread stack dump:
task:rcu_preempt     state:R  running task     stack:28696 pid:   14 ppid:     2 flags:0x00004000
Call Trace:
 context_switch kernel/sched/core.c:4955 [inline]
 __schedule+0x940/0x26f0 kernel/sched/core.c:6236
 schedule+0xd3/0x270 kernel/sched/core.c:6315
 schedule_timeout+0x14a/0x2a0 kernel/time/timer.c:1881
 rcu_gp_fqs_loop+0x186/0x810 kernel/rcu/tree.c:1955
 rcu_gp_kthread+0x1de/0x320 kernel/rcu/tree.c:2128
 kthread+0x405/0x4f0 kernel/kthread.c:327
 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:295
rcu: Stack dump where RCU GP kthread last ran:
Sending NMI from CPU 0 to CPUs 1:
NMI backtrace for cpu 1
CPU: 1 PID: 8510 Comm: syz-executor827 Not tainted 5.15.0-rc2-next-20210920-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:bytes_is_nonzero mm/kasan/generic.c:84 [inline]
RIP: 0010:memory_is_nonzero mm/kasan/generic.c:102 [inline]
RIP: 0010:memory_is_poisoned_n mm/kasan/generic.c:128 [inline]
RIP: 0010:memory_is_poisoned mm/kasan/generic.c:159 [inline]
RIP: 0010:check_region_inline mm/kasan/generic.c:180 [inline]
RIP: 0010:kasan_check_range+0xc8/0x180 mm/kasan/generic.c:189
Code: 38 00 74 ed 48 8d 50 08 eb 09 48 83 c0 01 48 39 d0 74 7a 80 38 00 74 f2 48 89 c2 b8 01 00 00 00 48 85 d2 75 56 5b 5d 41 5c c3 <48> 85 d2 74 5e 48 01 ea eb 09 48 83 c0 01 48 39 d0 74 50 80 38 00
RSP: 0018:ffffc9000cd676c8 EFLAGS: 00000283
RAX: ffffed100e9a110e RBX: ffffed100e9a110f RCX: ffffffff88ea062a
RDX: 0000000000000001 RSI: 0000000000000008 RDI: ffff888074d08870
RBP: ffffed100e9a110e R08: 0000000000000001 R09: ffff888074d08877
R10: ffffed100e9a110e R11: 0000000000000000 R12: ffff888074d08000
R13: ffff888074d08000 R14: ffff888074d08088 R15: ffff888074d08000
FS:  0000555556d8e300(0000) GS:ffff8880b9d00000(0000) knlGS:0000000000000000
S:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000020000180 CR3: 0000000068909000 CR4: 00000000001506e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 instrument_atomic_read_write include/linux/instrumented.h:101 [inline]
 test_and_clear_bit include/asm-generic/bitops/instrumented-atomic.h:83 [inline]
 mptcp_release_cb+0x14a/0x210 net/mptcp/protocol.c:3016
 release_sock+0xb4/0x1b0 net/core/sock.c:3204
 mptcp_wait_data net/mptcp/protocol.c:1770 [inline]
 mptcp_recvmsg+0xfd1/0x27b0 net/mptcp/protocol.c:2080
 inet6_recvmsg+0x11b/0x5e0 net/ipv6/af_inet6.c:659
 sock_recvmsg_nosec net/socket.c:944 [inline]
 ____sys_recvmsg+0x527/0x600 net/socket.c:2626
 ___sys_recvmsg+0x127/0x200 net/socket.c:2670
 do_recvmmsg+0x24d/0x6d0 net/socket.c:2764
 __sys_recvmmsg net/socket.c:2843 [inline]
 __do_sys_recvmmsg net/socket.c:2866 [inline]
 __se_sys_recvmmsg net/socket.c:2859 [inline]
 __x64_sys_recvmmsg+0x20b/0x260 net/socket.c:2859
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x7fc200d2dc39
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 41 15 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffc5758e5a8 EFLAGS: 00000246 ORIG_RAX: 000000000000012b
RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007fc200d2dc39
RDX: 0000000000000002 RSI: 00000000200017c0 RDI: 0000000000000003
RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000f0b5ff
R10: 0000000000000100 R11: 0000000000000246 R12: 0000000000000003
R13: 00007ffc5758e5d0 R14: 00007ffc5758e5c0 R15: 0000000000000003

Fix the issue by replacing the MPTCP_DATA_READY bit with direct
inspection of the msk receive queue.

Reported-and-tested-by: syzbot+3360da629681aa0d22fe@syzkaller.appspotmail.com
Fixes: 7a6a6cbc3e59 ("mptcp: recvmsg() can drain data from multiple subflow")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

0 files changed, 0 insertions, 0 deletions