scsi: core: Avoid that SCSI device removal through sysfs triggers a deadlock
commit0ee223b2e1f67cb2de9c0e3247c510d846e74d63
authorBart Van Assche <bart.vanassche@wdc.com>
Thu, 2 Aug 2018 17:51:41 +0000 (2 10:51 -0700)
committerMartin K. Petersen <martin.petersen@oracle.com>
Thu, 2 Aug 2018 19:53:36 +0000 (2 15:53 -0400)
treec7c58775a87067aa3f63e36dfb6c6c565e01b2fe
parent2afc9166f79b8f6da5f347f48515215ceee4ae37
scsi: core: Avoid that SCSI device removal through sysfs triggers a deadlock

A long time ago the unfortunate decision was taken to add a self-deletion
attribute to the sysfs SCSI device directory. That decision was unfortunate
because self-deletion is really tricky. We can't drop that attribute
because widely used user space software depends on it, namely the
rescan-scsi-bus.sh script. Hence this patch that avoids that writing into
that attribute triggers a deadlock. See also commit 7973cbd9fbd9 ("[PATCH]
add sysfs attributes to scan and delete scsi_devices").

This patch avoids that self-removal triggers the following deadlock:

======================================================
WARNING: possible circular locking dependency detected
4.18.0-rc2-dbg+ #5 Not tainted
------------------------------------------------------
modprobe/6539 is trying to acquire lock:
000000008323c4cd (kn->count#202){++++}, at: kernfs_remove_by_name_ns+0x45/0x90

but task is already holding lock:
00000000a6ec2c69 (&shost->scan_mutex){+.+.}, at: scsi_remove_host+0x21/0x150 [scsi_mod]

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (&shost->scan_mutex){+.+.}:
       __mutex_lock+0xfe/0xc70
       mutex_lock_nested+0x1b/0x20
       scsi_remove_device+0x26/0x40 [scsi_mod]
       sdev_store_delete+0x27/0x30 [scsi_mod]
       dev_attr_store+0x3e/0x50
       sysfs_kf_write+0x87/0xa0
       kernfs_fop_write+0x190/0x230
       __vfs_write+0xd2/0x3b0
       vfs_write+0x101/0x270
       ksys_write+0xab/0x120
       __x64_sys_write+0x43/0x50
       do_syscall_64+0x77/0x230
       entry_SYSCALL_64_after_hwframe+0x49/0xbe

-> #0 (kn->count#202){++++}:
       lock_acquire+0xd2/0x260
       __kernfs_remove+0x424/0x4a0
       kernfs_remove_by_name_ns+0x45/0x90
       remove_files.isra.1+0x3a/0x90
       sysfs_remove_group+0x5c/0xc0
       sysfs_remove_groups+0x39/0x60
       device_remove_attrs+0x82/0xb0
       device_del+0x251/0x580
       __scsi_remove_device+0x19f/0x1d0 [scsi_mod]
       scsi_forget_host+0x37/0xb0 [scsi_mod]
       scsi_remove_host+0x9b/0x150 [scsi_mod]
       sdebug_driver_remove+0x4b/0x150 [scsi_debug]
       device_release_driver_internal+0x241/0x360
       device_release_driver+0x12/0x20
       bus_remove_device+0x1bc/0x290
       device_del+0x259/0x580
       device_unregister+0x1a/0x70
       sdebug_remove_adapter+0x8b/0xf0 [scsi_debug]
       scsi_debug_exit+0x76/0xe8 [scsi_debug]
       __x64_sys_delete_module+0x1c1/0x280
       do_syscall_64+0x77/0x230
       entry_SYSCALL_64_after_hwframe+0x49/0xbe

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&shost->scan_mutex);
                               lock(kn->count#202);
                               lock(&shost->scan_mutex);
  lock(kn->count#202);

 *** DEADLOCK ***

2 locks held by modprobe/6539:
 #0: 00000000efaf9298 (&dev->mutex){....}, at: device_release_driver_internal+0x68/0x360
 #1: 00000000a6ec2c69 (&shost->scan_mutex){+.+.}, at: scsi_remove_host+0x21/0x150 [scsi_mod]

stack backtrace:
CPU: 10 PID: 6539 Comm: modprobe Not tainted 4.18.0-rc2-dbg+ #5
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014
Call Trace:
 dump_stack+0xa4/0xf5
 print_circular_bug.isra.34+0x213/0x221
 __lock_acquire+0x1a7e/0x1b50
 lock_acquire+0xd2/0x260
 __kernfs_remove+0x424/0x4a0
 kernfs_remove_by_name_ns+0x45/0x90
 remove_files.isra.1+0x3a/0x90
 sysfs_remove_group+0x5c/0xc0
 sysfs_remove_groups+0x39/0x60
 device_remove_attrs+0x82/0xb0
 device_del+0x251/0x580
 __scsi_remove_device+0x19f/0x1d0 [scsi_mod]
 scsi_forget_host+0x37/0xb0 [scsi_mod]
 scsi_remove_host+0x9b/0x150 [scsi_mod]
 sdebug_driver_remove+0x4b/0x150 [scsi_debug]
 device_release_driver_internal+0x241/0x360
 device_release_driver+0x12/0x20
 bus_remove_device+0x1bc/0x290
 device_del+0x259/0x580
 device_unregister+0x1a/0x70
 sdebug_remove_adapter+0x8b/0xf0 [scsi_debug]
 scsi_debug_exit+0x76/0xe8 [scsi_debug]
 __x64_sys_delete_module+0x1c1/0x280
 do_syscall_64+0x77/0x230
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

See also https://www.mail-archive.com/linux-scsi@vger.kernel.org/msg54525.html.

Fixes: ac0ece9174ac ("scsi: use device_remove_file_self() instead of device_schedule_callback()")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
drivers/scsi/scsi_sysfs.c