Skip to content

Commit 9724096

Browse files
vegardaxboe
authored andcommitted
nbd: fix race in ioctl
Quentin ran into this bug: WARNING: CPU: 64 PID: 10085 at fs/sysfs/dir.c:31 sysfs_warn_dup+0x65/0x80 sysfs: cannot create duplicate filename '/devices/virtual/block/nbd3/pid' Modules linked in: nbd CPU: 64 PID: 10085 Comm: qemu-nbd Tainted: G D 4.6.0+ #7 0000000000000000 ffff8820330bba68 ffffffff814b8791 ffff8820330bbac8 0000000000000000 ffff8820330bbab8 ffffffff810d04ab ffff8820330bbaa8 0000001f00000296 0000000000017681 ffff8810380bf000 ffffffffa0001790 Call Trace: [<ffffffff814b8791>] dump_stack+0x4d/0x6c [<ffffffff810d04ab>] __warn+0xdb/0x100 [<ffffffff810d0574>] warn_slowpath_fmt+0x44/0x50 [<ffffffff81218c65>] sysfs_warn_dup+0x65/0x80 [<ffffffff81218a02>] sysfs_add_file_mode_ns+0x172/0x180 [<ffffffff81218a35>] sysfs_create_file_ns+0x25/0x30 [<ffffffff81594a76>] device_create_file+0x36/0x90 [<ffffffffa0000e8d>] __nbd_ioctl+0x32d/0x9b0 [nbd] [<ffffffff814cc8e8>] ? find_next_bit+0x18/0x20 [<ffffffff810f7c29>] ? select_idle_sibling+0xe9/0x120 [<ffffffff810f6cd7>] ? __enqueue_entity+0x67/0x70 [<ffffffff810f9bf0>] ? enqueue_task_fair+0x630/0xe20 [<ffffffff810efa76>] ? resched_curr+0x36/0x70 [<ffffffff810f0078>] ? check_preempt_curr+0x78/0x90 [<ffffffff810f00a2>] ? ttwu_do_wakeup+0x12/0x80 [<ffffffff810f01b1>] ? ttwu_do_activate.constprop.86+0x61/0x70 [<ffffffff810f0c15>] ? try_to_wake_up+0x185/0x2d0 [<ffffffff810f0d6d>] ? default_wake_function+0xd/0x10 [<ffffffff81105471>] ? autoremove_wake_function+0x11/0x40 [<ffffffffa0001577>] nbd_ioctl+0x67/0x94 [nbd] [<ffffffff814ac0fd>] blkdev_ioctl+0x14d/0x940 [<ffffffff811b0da2>] ? put_pipe_info+0x22/0x60 [<ffffffff811d96cc>] block_ioctl+0x3c/0x40 [<ffffffff811ba08d>] do_vfs_ioctl+0x8d/0x5e0 [<ffffffff811aa329>] ? ____fput+0x9/0x10 [<ffffffff810e9092>] ? task_work_run+0x72/0x90 [<ffffffff811ba627>] SyS_ioctl+0x47/0x80 [<ffffffff8185f5df>] entry_SYSCALL_64_fastpath+0x17/0x93 ---[ end trace 7899b295e4f850c8 ]--- It seems fairly obvious that device_create_file() is not being protected from being run concurrently on the same nbd. Quentin found the following relevant commits: 1a2ad21 nbd: add locking to nbd_ioctl 90b8f28 [PATCH] end of methods switch: remove the old ones d4430d6 [PATCH] beginning of methods conversion 08f8585 [PATCH] move block_device_operations to blkdev.h It would seem that the race was introduced in the process of moving nbd from BKL to unlocked ioctls. By setting nbd->task_recv while the mutex is held, we can prevent other processes from running concurrently (since nbd->task_recv is also checked while the mutex is held). Reported-and-tested-by: Quentin Casasnovas <[email protected]> Cc: Markus Pargmann <[email protected]> Cc: Paul Clements <[email protected]> Cc: Pavel Machek <[email protected]> Cc: Jens Axboe <[email protected]> Cc: Al Viro <[email protected]> Signed-off-by: Vegard Nossum <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
1 parent 77da160 commit 9724096

File tree

1 file changed

+4
-8
lines changed

1 file changed

+4
-8
lines changed

drivers/block/nbd.c

Lines changed: 4 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -451,14 +451,9 @@ static int nbd_thread_recv(struct nbd_device *nbd, struct block_device *bdev)
451451

452452
sk_set_memalloc(nbd->sock->sk);
453453

454-
nbd->task_recv = current;
455-
456454
ret = device_create_file(disk_to_dev(nbd->disk), &pid_attr);
457455
if (ret) {
458456
dev_err(disk_to_dev(nbd->disk), "device_create_file failed!\n");
459-
460-
nbd->task_recv = NULL;
461-
462457
return ret;
463458
}
464459

@@ -477,9 +472,6 @@ static int nbd_thread_recv(struct nbd_device *nbd, struct block_device *bdev)
477472
nbd_size_clear(nbd, bdev);
478473

479474
device_remove_file(disk_to_dev(nbd->disk), &pid_attr);
480-
481-
nbd->task_recv = NULL;
482-
483475
return ret;
484476
}
485477

@@ -788,6 +780,8 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
788780
if (!nbd->sock)
789781
return -EINVAL;
790782

783+
/* We have to claim the device under the lock */
784+
nbd->task_recv = current;
791785
mutex_unlock(&nbd->tx_lock);
792786

793787
nbd_parse_flags(nbd, bdev);
@@ -796,6 +790,7 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
796790
nbd_name(nbd));
797791
if (IS_ERR(thread)) {
798792
mutex_lock(&nbd->tx_lock);
793+
nbd->task_recv = NULL;
799794
return PTR_ERR(thread);
800795
}
801796

@@ -805,6 +800,7 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
805800
kthread_stop(thread);
806801

807802
mutex_lock(&nbd->tx_lock);
803+
nbd->task_recv = NULL;
808804

809805
sock_shutdown(nbd);
810806
nbd_clear_que(nbd);

0 commit comments

Comments
 (0)