Hi,
I was fuzzing linux 3.18 (plus a minor, unrelated patch in the memory management subsystem) using trinity. I got the following lock-dep splat, which seems to originate from the autofs4 code:
[ 8890.416041] ======================================================
[ 8890.416041] [ INFO: possible circular locking dependency detected ]
[ 8890.416041] 3.18.0+ #25 Tainted: G W
[ 8890.416041] -------------------------------------------------------
[ 8890.416041] bash/4644 is trying to acquire lock:
[ 8890.416041] (&sbi->pipe_mutex){+.+.+.}, at: [<ffffffff812a25fb>] autofs4_notify_daemon+0x1cb/0x340
[ 8890.416041]
but task is already holding lock:
[ 8890.416041] (&sig->cred_guard_mutex){+.+.+.}, at: [<ffffffff8121f1e4>] prepare_bprm_creds+0x34/0x90
[ 8890.416041]
which lock already depends on the new lock.
[ 8890.416041]
the existing dependency chain (in reverse order) is:
[ 8890.416041]
-> #2 (&sig->cred_guard_mutex){+.+.+.}:
[ 8890.416041] [<ffffffff810cc5b6>] validate_chain.isra.36+0x486/0x930
[ 8890.416041] [<ffffffff810ced48>] __lock_acquire+0x378/0xbe0
[ 8890.416041] [<ffffffff810cf677>] lock_acquire+0xc7/0x160
[ 8890.416041] [<ffffffff816d15e1>] mutex_lock_interruptible_nested+0x81/0x5c0
[ 8890.416041] [<ffffffff8128628d>] proc_pid_attr_write+0xfd/0x150
[ 8890.416041] [<ffffffff81216e48>] __kernel_write+0x68/0x160
[ 8890.416041] [<ffffffff81249445>] write_pipe_buf+0x85/0xb0
[ 8890.416041] [<ffffffff81248f5e>] __splice_from_pipe+0x11e/0x190
[ 8890.416041] [<ffffffff8124afde>] splice_from_pipe+0x5e/0x90
[ 8890.416041] [<ffffffff8124b04d>] default_file_splice_write+0x1d/0x30
[ 8890.416041] [<ffffffff8124b51a>] SyS_splice+0x25a/0x800
[ 8890.416041] [<ffffffff816d65e9>] system_call_fastpath+0x12/0x17
[ 8890.416041]
-> #1 (&pipe->mutex/1){+.+.+.}:
[ 8890.416041] [<ffffffff810cc5b6>] validate_chain.isra.36+0x486/0x930
[ 8890.416041] [<ffffffff810ced48>] __lock_acquire+0x378/0xbe0
[ 8890.416041] [<ffffffff810cf677>] lock_acquire+0xc7/0x160
[ 8890.416041] [<ffffffff816d20b1>] mutex_lock_nested+0x81/0x510
[ 8890.416041] [<ffffffff8122022f>] pipe_write+0x3f/0x490
[ 8890.416041] [<ffffffff81216b3e>] new_sync_write+0x8e/0xd0
[ 8890.416041] [<ffffffff812a2613>] autofs4_notify_daemon+0x1e3/0x340
[ 8890.416041] [<ffffffff812a2f65>] autofs4_wait+0x7f5/0xa20
[ 8890.416041] [<ffffffff812a0ce0>] autofs4_mount_wait+0x50/0xf0
[ 8890.416041] [<ffffffff812a1609>] autofs4_d_automount+0xe9/0x240
[ 8890.416041] [<ffffffff81221f93>] follow_managed+0x143/0x310
[ 8890.416041] [<ffffffff816c96e2>] lookup_slow+0x7d/0xad
[ 8890.416041] [<ffffffff81226732>] link_path_walk+0x792/0x860
[ 8890.416041] [<ffffffff8122740b>] path_openat+0xab/0x640
[ 8890.416041] [<ffffffff81229a69>] do_filp_open+0x49/0xc0
[ 8890.416041] [<ffffffff81216397>] do_sys_open+0x137/0x240
[ 8890.416041] [<ffffffff812164be>] SyS_open+0x1e/0x20
[ 8890.416041] [<ffffffff816d65e9>] system_call_fastpath+0x12/0x17
[ 8890.416041]
-> #0 (&sbi->pipe_mutex){+.+.+.}:
[ 8890.416041] [<ffffffff810cc10f>] check_prevs_add+0x9af/0x9d0
[ 8890.416041] [<ffffffff810cc5b6>] validate_chain.isra.36+0x486/0x930
[ 8890.416041] [<ffffffff810ced48>] __lock_acquire+0x378/0xbe0
[ 8890.416041] [<ffffffff810cf677>] lock_acquire+0xc7/0x160
[ 8890.416041] [<ffffffff816d20b1>] mutex_lock_nested+0x81/0x510
[ 8890.416041] [<ffffffff812a25fb>] autofs4_notify_daemon+0x1cb/0x340
[ 8890.416041] [<ffffffff812a2f65>] autofs4_wait+0x7f5/0xa20
[ 8890.416041] [<ffffffff812a0ce0>] autofs4_mount_wait+0x50/0xf0
[ 8890.416041] [<ffffffff812a1609>] autofs4_d_automount+0xe9/0x240
[ 8890.416041] [<ffffffff81221f93>] follow_managed+0x143/0x310
[ 8890.416041] [<ffffffff816c96e2>] lookup_slow+0x7d/0xad
[ 8890.416041] [<ffffffff812266a7>] link_path_walk+0x707/0x860
[ 8890.416041] [<ffffffff8122740b>] path_openat+0xab/0x640
[ 8890.416041] [<ffffffff81229a69>] do_filp_open+0x49/0xc0
[ 8890.416041] [<ffffffff8121e546>] do_open_exec+0x26/0xf0
[ 8890.416041] [<ffffffff8121f458>] do_execve_common.isra.24+0x218/0x710
[ 8890.416041] [<ffffffff8121fb99>] SyS_execve+0x29/0x30
[ 8890.416041] [<ffffffff816d6bf9>] stub_execve+0x69/0xa0
[ 8890.416041]
other info that might help us debug this:
[ 8890.416041] Chain exists of:
&sbi->pipe_mutex --> &pipe->mutex/1 --> &sig->cred_guard_mutex
[ 8890.416041] Possible unsafe locking scenario:
[ 8890.416041] CPU0 CPU1
[ 8890.416041] ---- ----
[ 8890.416041] lock(&sig->cred_guard_mutex);
[ 8890.416041] lock(&pipe->mutex/1);
[ 8890.416041] lock(&sig->cred_guard_mutex);
[ 8890.416041] lock(&sbi->pipe_mutex);
[ 8890.416041]
*** DEADLOCK ***
[ 8890.416041] 1 lock held by bash/4644:
[ 8890.416041] #0: (&sig->cred_guard_mutex){+.+.+.}, at: [<ffffffff8121f1e4>] prepare_bprm_creds+0x34/0x90
[ 8890.416041]
stack backtrace:
[ 8890.416041] CPU: 1 PID: 4644 Comm: bash Tainted: G W 3.18.0+ #25
[ 8890.416041] Hardware name: Red Hat KVM, BIOS Bochs 01/01/2007
[ 8890.416041] 0000000000000000 0000000043c40f29 ffff88004dc775b8 ffffffff816cbade
[ 8890.416041] 0000000000000000 ffffffff829fc330 ffff88004dc77608 ffffffff816c4bf0
[ 8890.416041] ffff88000d401b70 ffff88004dc77648 ffff88004dc77608 0000000000000000
[ 8890.416041] Call Trace:
[ 8890.416041] [<ffffffff816cbade>] dump_stack+0x4e/0x68
[ 8890.416041] [<ffffffff816c4bf0>] print_circular_bug+0x1fb/0x20c
[ 8890.416041] [<ffffffff810cc10f>] check_prevs_add+0x9af/0x9d0
[ 8890.416041] [<ffffffff8105f4d7>] ? kvm_clock_read+0x27/0x40
[ 8890.416041] [<ffffffff810cc5b6>] validate_chain.isra.36+0x486/0x930
[ 8890.416041] [<ffffffff810ced48>] __lock_acquire+0x378/0xbe0
[ 8890.416041] [<ffffffff810b02c6>] ? local_clock+0x16/0x30
[ 8890.416041] [<ffffffff810cf677>] lock_acquire+0xc7/0x160
[ 8890.416041] [<ffffffff812a25fb>] ? autofs4_notify_daemon+0x1cb/0x340
[ 8890.416041] [<ffffffff816d20b1>] mutex_lock_nested+0x81/0x510
[ 8890.416041] [<ffffffff812a25fb>] ? autofs4_notify_daemon+0x1cb/0x340
[ 8890.416041] [<ffffffff812a25fb>] ? autofs4_notify_daemon+0x1cb/0x340
[ 8890.416041] [<ffffffff810cd11d>] ? trace_hardirqs_on+0xd/0x10
[ 8890.416041] [<ffffffff812a25fb>] autofs4_notify_daemon+0x1cb/0x340
[ 8890.416041] [<ffffffff812a2f65>] autofs4_wait+0x7f5/0xa20
[ 8890.416041] [<ffffffff812a28fb>] ? autofs4_wait+0x18b/0xa20
[ 8890.416041] [<ffffffff812a15ff>] ? autofs4_d_automount+0xdf/0x240
[ 8890.416041] [<ffffffff812a0ce0>] autofs4_mount_wait+0x50/0xf0
[ 8890.416041] [<ffffffff812a1609>] autofs4_d_automount+0xe9/0x240
[ 8890.416041] [<ffffffff81221f93>] follow_managed+0x143/0x310
[ 8890.416041] [<ffffffff816c96e2>] lookup_slow+0x7d/0xad
[ 8890.416041] [<ffffffff812266a7>] link_path_walk+0x707/0x860
[ 8890.416041] [<ffffffff812242e5>] ? path_init+0x4b5/0x700
[ 8890.416041] [<ffffffff8122740b>] path_openat+0xab/0x640
[ 8890.416041] [<ffffffff8105f4d7>] ? kvm_clock_read+0x27/0x40
[ 8890.416041] [<ffffffff81020009>] ? sched_clock+0x9/0x10
[ 8890.416041] [<ffffffff81020009>] ? sched_clock+0x9/0x10
[ 8890.416041] [<ffffffff810affc5>] ? sched_clock_local+0x25/0x90
[ 8890.416041] [<ffffffff81229a69>] do_filp_open+0x49/0xc0
[ 8890.416041] [<ffffffff8121f43c>] ? do_execve_common.isra.24+0x1fc/0x710
[ 8890.416041] [<ffffffff8121f3f2>] ? do_execve_common.isra.24+0x1b2/0x710
[ 8890.416041] [<ffffffff8121e546>] do_open_exec+0x26/0xf0
[ 8890.416041] [<ffffffff8121f458>] do_execve_common.isra.24+0x218/0x710
[ 8890.416041] [<ffffffff8121f377>] ? do_execve_common.isra.24+0x137/0x710
[ 8890.416041] [<ffffffff8121fb99>] SyS_execve+0x29/0x30
[ 8890.416041] [<ffffffff816d6bf9>] stub_execve+0x69/0xa0
Thanks,
--Shachar
--
To unsubscribe from this list: send the line "unsubscribe autofs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
