All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: Greg KH <gregkh@linuxfoundation.org>,
	torvalds@linux-foundation.org, akpm@linux-foundation.org,
	alan@lxorguk.ukuu.org.uk,
	Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>,
	Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Subject: [ 03/44] nilfs2: fix deadlock issue between chcp and thaw ioctls
Date: Mon, 13 Aug 2012 15:02:10 -0700	[thread overview]
Message-ID: <20120813220142.436840284@linuxfoundation.org> (raw)
In-Reply-To: <20120813220142.113186818@linuxfoundation.org>

From: Greg KH <gregkh@linuxfoundation.org>

3.0-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

commit 572d8b3945a31bee7c40d21556803e4807fd9141 upstream.

An fs-thaw ioctl causes deadlock with a chcp or mkcp -s command:

 chcp            D ffff88013870f3d0     0  1325   1324 0x00000004
 ...
 Call Trace:
   nilfs_transaction_begin+0x11c/0x1a0 [nilfs2]
   wake_up_bit+0x20/0x20
   copy_from_user+0x18/0x30 [nilfs2]
   nilfs_ioctl_change_cpmode+0x7d/0xcf [nilfs2]
   nilfs_ioctl+0x252/0x61a [nilfs2]
   do_page_fault+0x311/0x34c
   get_unmapped_area+0x132/0x14e
   do_vfs_ioctl+0x44b/0x490
   __set_task_blocked+0x5a/0x61
   vm_mmap_pgoff+0x76/0x87
   __set_current_blocked+0x30/0x4a
   sys_ioctl+0x4b/0x6f
   system_call_fastpath+0x16/0x1b
 thaw            D ffff88013870d890     0  1352   1351 0x00000004
 ...
 Call Trace:
   rwsem_down_failed_common+0xdb/0x10f
   call_rwsem_down_write_failed+0x13/0x20
   down_write+0x25/0x27
   thaw_super+0x13/0x9e
   do_vfs_ioctl+0x1f5/0x490
   vm_mmap_pgoff+0x76/0x87
   sys_ioctl+0x4b/0x6f
   filp_close+0x64/0x6c
   system_call_fastpath+0x16/0x1b

where the thaw ioctl deadlocked at thaw_super() when called while chcp was
waiting at nilfs_transaction_begin() called from
nilfs_ioctl_change_cpmode().  This deadlock is 100% reproducible.

This is because nilfs_ioctl_change_cpmode() first locks sb->s_umount in
read mode and then waits for unfreezing in nilfs_transaction_begin(),
whereas thaw_super() locks sb->s_umount in write mode.  The locking of
sb->s_umount here was intended to make snapshot mounts and the downgrade
of snapshots to checkpoints exclusive.

This fixes the deadlock issue by replacing the sb->s_umount usage in
nilfs_ioctl_change_cpmode() with a dedicated mutex which protects snapshot
mounts.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Cc: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Tested-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 fs/nilfs2/ioctl.c     |    4 ++--
 fs/nilfs2/super.c     |    3 +++
 fs/nilfs2/the_nilfs.c |    1 +
 fs/nilfs2/the_nilfs.h |    2 ++
 4 files changed, 8 insertions(+), 2 deletions(-)

--- a/fs/nilfs2/ioctl.c
+++ b/fs/nilfs2/ioctl.c
@@ -182,7 +182,7 @@ static int nilfs_ioctl_change_cpmode(str
 	if (copy_from_user(&cpmode, argp, sizeof(cpmode)))
 		goto out;
 
-	down_read(&inode->i_sb->s_umount);
+	mutex_lock(&nilfs->ns_snapshot_mount_mutex);
 
 	nilfs_transaction_begin(inode->i_sb, &ti, 0);
 	ret = nilfs_cpfile_change_cpmode(
@@ -192,7 +192,7 @@ static int nilfs_ioctl_change_cpmode(str
 	else
 		nilfs_transaction_commit(inode->i_sb); /* never fails */
 
-	up_read(&inode->i_sb->s_umount);
+	mutex_unlock(&nilfs->ns_snapshot_mount_mutex);
 out:
 	mnt_drop_write(filp->f_path.mnt);
 	return ret;
--- a/fs/nilfs2/super.c
+++ b/fs/nilfs2/super.c
@@ -951,6 +951,8 @@ static int nilfs_attach_snapshot(struct
 	struct nilfs_root *root;
 	int ret;
 
+	mutex_lock(&nilfs->ns_snapshot_mount_mutex);
+
 	down_read(&nilfs->ns_segctor_sem);
 	ret = nilfs_cpfile_is_snapshot(nilfs->ns_cpfile, cno);
 	up_read(&nilfs->ns_segctor_sem);
@@ -975,6 +977,7 @@ static int nilfs_attach_snapshot(struct
 	ret = nilfs_get_root_dentry(s, root, root_dentry);
 	nilfs_put_root(root);
  out:
+	mutex_unlock(&nilfs->ns_snapshot_mount_mutex);
 	return ret;
 }
 
--- a/fs/nilfs2/the_nilfs.c
+++ b/fs/nilfs2/the_nilfs.c
@@ -76,6 +76,7 @@ struct the_nilfs *alloc_nilfs(struct blo
 	nilfs->ns_bdev = bdev;
 	atomic_set(&nilfs->ns_ndirtyblks, 0);
 	init_rwsem(&nilfs->ns_sem);
+	mutex_init(&nilfs->ns_snapshot_mount_mutex);
 	INIT_LIST_HEAD(&nilfs->ns_dirty_files);
 	INIT_LIST_HEAD(&nilfs->ns_gc_inodes);
 	spin_lock_init(&nilfs->ns_inode_lock);
--- a/fs/nilfs2/the_nilfs.h
+++ b/fs/nilfs2/the_nilfs.h
@@ -47,6 +47,7 @@ enum {
  * @ns_flags: flags
  * @ns_bdev: block device
  * @ns_sem: semaphore for shared states
+ * @ns_snapshot_mount_mutex: mutex to protect snapshot mounts
  * @ns_sbh: buffer heads of on-disk super blocks
  * @ns_sbp: pointers to super block data
  * @ns_sbwtime: previous write time of super block
@@ -99,6 +100,7 @@ struct the_nilfs {
 
 	struct block_device    *ns_bdev;
 	struct rw_semaphore	ns_sem;
+	struct mutex		ns_snapshot_mount_mutex;
 
 	/*
 	 * used for



  parent reply	other threads:[~2012-08-13 22:34 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-08-13 22:02 [ 00/44] 3.0.41-stable review Greg Kroah-Hartman
2012-08-13 22:02 ` [ 01/44] [IA64] Redefine ATOMIC_INIT and ATOMIC64_INIT to drop the casts Greg Kroah-Hartman
2012-08-13 22:02 ` [ 02/44] SUNRPC: return negative value in case rpcbind client creation error Greg Kroah-Hartman
2012-08-13 22:02 ` Greg Kroah-Hartman [this message]
2012-08-13 22:02 ` [ 04/44] pcdp: use early_ioremap/early_iounmap to access pcdp table Greg Kroah-Hartman
2012-08-13 22:02 ` [ 05/44] mm: fix wrong argument of migrate_huge_pages() in soft_offline_huge_page() Greg Kroah-Hartman
2012-08-13 22:02 ` [ 06/44] ARM: 7467/1: mutex: use generic xchg-based implementation for ARMv6+ Greg Kroah-Hartman
2012-08-15 14:02   ` Ben Hutchings
2012-08-13 22:02 ` [ 07/44] ARM: 7477/1: vfp: Always save VFP state in vfp_pm_suspend on UP Greg Kroah-Hartman
2012-08-14 20:01   ` Herton Ronaldo Krzesinski
2012-08-15 14:05     ` Greg Kroah-Hartman
2012-08-15 14:50       ` Herton Ronaldo Krzesinski
2012-08-13 22:02 ` [ 08/44] ARM: 7478/1: errata: extend workaround for erratum #720789 Greg Kroah-Hartman
2012-08-13 22:02 ` [ 09/44] ARM: 7479/1: mm: avoid NULL dereference when flushing gate_vma with VIVT caches Greg Kroah-Hartman
2012-08-13 22:02 ` [ 10/44] ALSA: hda - remove quirk for Dell Vostro 1015 Greg Kroah-Hartman
2012-08-14  5:17   ` David Henningsson
2012-08-14  5:43     ` Takashi Iwai
2012-08-15 14:03     ` Greg Kroah-Hartman
2012-08-13 22:02 ` [ 11/44] mm: mmu_notifier: fix freed page still mapped in secondary MMU Greg Kroah-Hartman
2012-08-13 22:02 ` [ 12/44] mac80211: cancel mesh path timer Greg Kroah-Hartman
2012-08-13 22:02 ` [ 13/44] x86, nops: Missing break resulting in incorrect selection on Intel Greg Kroah-Hartman
2012-08-13 22:02 ` [ 14/44] random: Add support for architectural random hooks Greg Kroah-Hartman
2012-08-13 22:02 ` [ 15/44] fix typo/thinko in get_random_bytes() Greg Kroah-Hartman
2012-08-13 22:02 ` [ 16/44] random: Use arch_get_random_int instead of cycle counter if avail Greg Kroah-Hartman
2012-08-13 22:02 ` [ 17/44] random: Use arch-specific RNG to initialize the entropy store Greg Kroah-Hartman
2012-08-13 22:02 ` [ 18/44] random: Adjust the number of loops when initializing Greg Kroah-Hartman
2012-08-13 22:02 ` [ 19/44] drivers/char/random.c: fix boot id uniqueness race Greg Kroah-Hartman
2012-08-13 22:02 ` [ 20/44] random: make add_interrupt_randomness() do something sane Greg Kroah-Hartman
2012-08-13 22:02 ` [ 21/44] random: use lockless techniques in the interrupt path Greg Kroah-Hartman
2012-08-13 22:02 ` [ 22/44] random: create add_device_randomness() interface Greg Kroah-Hartman
2012-08-13 22:02 ` [ 23/44] usb: feed USB device information to the /dev/random driver Greg Kroah-Hartman
2012-08-13 22:02 ` [ 24/44] net: feed /dev/random with the MAC address when registering a device Greg Kroah-Hartman
2012-08-13 22:02 ` [ 25/44] random: use the arch-specific rng in xfer_secondary_pool Greg Kroah-Hartman
2012-08-13 22:02 ` [ 26/44] random: add new get_random_bytes_arch() function Greg Kroah-Hartman
2012-08-13 22:02 ` [ 27/44] random: add tracepoints for easier debugging and verification Greg Kroah-Hartman
2012-08-13 22:02 ` [ 28/44] MAINTAINERS: Theodore Tso is taking over the random driver Greg Kroah-Hartman
2012-08-13 22:02 ` [ 29/44] rtc: wm831x: Feed the write counter into device_add_randomness() Greg Kroah-Hartman
2012-08-13 22:02 ` [ 30/44] mfd: wm831x: Feed the device UUID " Greg Kroah-Hartman
2012-08-13 22:02 ` [ 31/44] random: remove rand_initialize_irq() Greg Kroah-Hartman
2012-08-13 22:02 ` [ 32/44] random: Add comment to random_initialize() Greg Kroah-Hartman
2012-08-13 22:02 ` [ 33/44] dmi: Feed DMI table to /dev/random driver Greg Kroah-Hartman
2012-08-13 22:02 ` [ 34/44] random: mix in architectural randomness in extract_buf() Greg Kroah-Hartman
2012-08-13 22:02 ` [ 35/44] x86, microcode: microcode_core.c simple_strtoul cleanup Greg Kroah-Hartman
2012-08-13 22:02 ` [ 36/44] x86, microcode: Sanitize per-cpu microcode reloading interface Greg Kroah-Hartman
2012-08-15  0:26   ` Henrique de Moraes Holschuh
2012-08-15 14:06     ` Greg Kroah-Hartman
2012-08-15 16:30       ` Henrique de Moraes Holschuh
2012-08-15 18:26         ` Greg Kroah-Hartman
2012-08-13 22:02 ` [ 37/44] mm: hugetlbfs: close race during teardown of hugetlbfs shared page tables Greg Kroah-Hartman
2012-08-13 22:02 ` [ 38/44] ARM: mxs: Remove MMAP_MIN_ADDR setting from mxs_defconfig Greg Kroah-Hartman
2012-08-13 22:02 ` [ 39/44] ARM: pxa: remove irq_to_gpio from ezx-pcap driver Greg Kroah-Hartman
2012-08-13 22:02 ` [ 40/44] cfg80211: process pending events when unregistering net device Greg Kroah-Hartman
2012-08-13 22:02 ` [ 41/44] cfg80211: fix interface combinations check for ADHOC(IBSS) Greg Kroah-Hartman
2012-08-13 22:02 ` [ 42/44] e1000e: NIC goes up and immediately goes down Greg Kroah-Hartman
2012-08-13 22:02 ` [ 43/44] Input: wacom - Bamboo One 1024 pressure fix Greg Kroah-Hartman
2012-08-13 22:02 ` [ 44/44] rt61pci: fix NULL pointer dereference in config_lna_gain Greg Kroah-Hartman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120813220142.436840284@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=akpm@linux-foundation.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=fernando@oss.ntt.co.jp \
    --cc=konishi.ryusuke@lab.ntt.co.jp \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.