linux-toradex.git/fs, branch v4.9.121

proc: Fix proc_sys_prune_dcache to hold a sb reference

2018-08-15T16:14:43+00:00

commit 2fd1d2c4ceb2248a727696962cf3370dc9f5a0a4 upstream.

Andrei Vagin writes:
FYI: This bug has been reproduced on 4.11.7
> BUG: Dentry ffff895a3dd01240{i=4e7c09a,n=lo}  still in use (1) [unmount of proc proc]
> ------------[ cut here ]------------
> WARNING: CPU: 1 PID: 13588 at fs/dcache.c:1445 umount_check+0x6e/0x80
> CPU: 1 PID: 13588 Comm: kworker/1:1 Not tainted 4.11.7-200.fc25.x86_64 #1
> Hardware name: CompuLab sbc-flt1/fitlet, BIOS SBCFLT_0.08.04 06/27/2015
> Workqueue: events proc_cleanup_work
> Call Trace:
>  dump_stack+0x63/0x86
>  __warn+0xcb/0xf0
>  warn_slowpath_null+0x1d/0x20
>  umount_check+0x6e/0x80
>  d_walk+0xc6/0x270
>  ? dentry_free+0x80/0x80
>  do_one_tree+0x26/0x40
>  shrink_dcache_for_umount+0x2d/0x90
>  generic_shutdown_super+0x1f/0xf0
>  kill_anon_super+0x12/0x20
>  proc_kill_sb+0x40/0x50
>  deactivate_locked_super+0x43/0x70
>  deactivate_super+0x5a/0x60
>  cleanup_mnt+0x3f/0x90
>  mntput_no_expire+0x13b/0x190
>  kern_unmount+0x3e/0x50
>  pid_ns_release_proc+0x15/0x20
>  proc_cleanup_work+0x15/0x20
>  process_one_work+0x197/0x450
>  worker_thread+0x4e/0x4a0
>  kthread+0x109/0x140
>  ? process_one_work+0x450/0x450
>  ? kthread_park+0x90/0x90
>  ret_from_fork+0x2c/0x40
> ---[ end trace e1c109611e5d0b41 ]---
> VFS: Busy inodes after unmount of proc. Self-destruct in 5 seconds.  Have a nice day...
> BUG: unable to handle kernel NULL pointer dereference at           (null)
> IP: _raw_spin_lock+0xc/0x30
> PGD 0

Fix this by taking a reference to the super block in proc_sys_prune_dcache.

The superblock reference is the core of the fix however the sysctl_inodes
list is converted to a hlist so that hlist_del_init_rcu may be used.  This
allows proc_sys_prune_dache to remove inodes the sysctl_inodes list, while
not causing problems for proc_sys_evict_inode when if it later choses to
remove the inode from the sysctl_inodes list.  Removing inodes from the
sysctl_inodes list allows proc_sys_prune_dcache to have a progress
guarantee, while still being able to drop all locks.  The fact that
head->unregistering is set in start_unregistering ensures that no more
inodes will be added to the the sysctl_inodes list.

Previously the code did a dance where it delayed calling iput until the
next entry in the list was being considered to ensure the inode remained on
the sysctl_inodes list until the next entry was walked to.  The structure
of the loop in this patch does not need that so is much easier to
understand and maintain.

Cc: stable@vger.kernel.org
Reported-by: Andrei Vagin 
Tested-by: Andrei Vagin 
Fixes: ace0c791e6c3 ("proc/sysctl: Don't grab i_lock under sysctl_lock.")
Fixes: d6cffbbe9a7e ("proc/sysctl: prune stale dentries during unregistering")
Signed-off-by: "Eric W. Biederman" 
Signed-off-by: Greg Kroah-Hartman

proc/sysctl: Don't grab i_lock under sysctl_lock.

2018-08-15T16:14:43+00:00

commit ace0c791e6c3cf5ef37cad2df69f0d90ccc40ffb upstream.

Konstantin Khlebnikov  writes:
> This patch has locking problem. I've got lockdep splat under LTP.
>
> [ 6633.115456] ======================================================
> [ 6633.115502] [ INFO: possible circular locking dependency detected ]
> [ 6633.115553] 4.9.10-debug+ #9 Tainted: G             L
> [ 6633.115584] -------------------------------------------------------
> [ 6633.115627] ksm02/284980 is trying to acquire lock:
> [ 6633.115659]  (&sb->s_type->i_lock_key#4){+.+...}, at: [] igrab+0x1e/0x80
> [ 6633.115834] but task is already holding lock:
> [ 6633.115882]  (sysctl_lock){+.+...}, at: [] unregister_sysctl_table+0x6b/0x110
> [ 6633.116026] which lock already depends on the new lock.
> [ 6633.116026]
> [ 6633.116080]
> [ 6633.116080] the existing dependency chain (in reverse order) is:
> [ 6633.116117]
> -> #2 (sysctl_lock){+.+...}:
> -> #1 (&(&dentry->d_lockref.lock)->rlock){+.+...}:
> -> #0 (&sb->s_type->i_lock_key#4){+.+...}:
>
> d_lock nests inside i_lock
> sysctl_lock nests inside d_lock in d_compare
>
> This patch adds i_lock nesting inside sysctl_lock.

Al Viro  replied:
> Once ->unregistering is set, you can drop sysctl_lock just fine.  So I'd
> try something like this - use rcu_read_lock() in proc_sys_prune_dcache(),
> drop sysctl_lock() before it and regain after.  Make sure that no inodes
> are added to the list ones ->unregistering has been set and use RCU list
> primitives for modifying the inode list, with sysctl_lock still used to
> serialize its modifications.
>
> Freeing struct inode is RCU-delayed (see proc_destroy_inode()), so doing
> igrab() is safe there.  Since we don't drop inode reference until after we'd
> passed beyond it in the list, list_for_each_entry_rcu() should be fine.

I agree with Al Viro's analsysis of the situtation.

Fixes: d6cffbbe9a7e ("proc/sysctl: prune stale dentries during unregistering")
Reported-by: Konstantin Khlebnikov 
Tested-by: Konstantin Khlebnikov 
Suggested-by: Al Viro 
Signed-off-by: "Eric W. Biederman" 
Signed-off-by: Greg Kroah-Hartman

proc/sysctl: prune stale dentries during unregistering

2018-08-15T16:14:43+00:00

commit d6cffbbe9a7e51eb705182965a189457c17ba8a3 upstream.

Currently unregistering sysctl table does not prune its dentries.
Stale dentries could slowdown sysctl operations significantly.

For example, command:

 # for i in {1..100000} ; do unshare -n -- sysctl -a &> /dev/null ; done
 creates a millions of stale denties around sysctls of loopback interface:

 # sysctl fs.dentry-state
 fs.dentry-state = 25812579  24724135        45      0       0       0

 All of them have matching names thus lookup have to scan though whole
 hash chain and call d_compare (proc_sys_compare) which checks them
 under system-wide spinlock (sysctl_lock).

 # time sysctl -a > /dev/null
 real    1m12.806s
 user    0m0.016s
 sys     1m12.400s

Currently only memory reclaimer could remove this garbage.
But without significant memory pressure this never happens.

This patch collects sysctl inodes into list on sysctl table header and
prunes all their dentries once that table unregisters.

Konstantin Khlebnikov  writes:
> On 10.02.2017 10:47, Al Viro wrote:
>> how about >> the matching stats *after* that patch?
>
> dcache size doesn't grow endlessly, so stats are fine
>
> # sysctl fs.dentry-state
> fs.dentry-state = 92712	58376	45	0	0	0
>
> # time sysctl -a &>/dev/null
>
> real	0m0.013s
> user	0m0.004s
> sys	0m0.008s

Signed-off-by: Konstantin Khlebnikov 
Suggested-by: Al Viro 
Signed-off-by: Eric W. Biederman 
Signed-off-by: Greg Kroah-Hartman

fix __legitimize_mnt()/mntput() race

2018-08-15T16:14:43+00:00

commit 119e1ef80ecfe0d1deb6378d4ab41f5b71519de1 upstream.

__legitimize_mnt() has two problems - one is that in case of success
the check of mount_lock is not ordered wrt preceding increment of
refcount, making it possible to have successful __legitimize_mnt()
on one CPU just before the otherwise final mntpu() on another,
with __legitimize_mnt() not seeing mntput() taking the lock and
mntput() not seeing the increment done by __legitimize_mnt().
Solved by a pair of barriers.

Another is that failure of __legitimize_mnt() on the second
read_seqretry() leaves us with reference that'll need to be
dropped by caller; however, if that races with final mntput()
we can end up with caller dropping rcu_read_lock() and doing
mntput() to release that reference - with the first mntput()
having freed the damn thing just as rcu_read_lock() had been
dropped.  Solution: in "do mntput() yourself" failure case
grab mount_lock, check if MNT_DOOMED has been set by racing
final mntput() that has missed our increment and if it has -
undo the increment and treat that as "failure, caller doesn't
need to drop anything" case.

It's not easy to hit - the final mntput() has to come right
after the first read_seqretry() in __legitimize_mnt() *and*
manage to miss the increment done by __legitimize_mnt() before
the second read_seqretry() in there.  The things that are almost
impossible to hit on bare hardware are not impossible on SMP
KVM, though...

Reported-by: Oleg Nesterov 
Fixes: 48a066e72d97 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro 
Signed-off-by: Greg Kroah-Hartman

fix mntput/mntput race

2018-08-15T16:14:42+00:00

commit 9ea0a46ca2c318fcc449c1e6b62a7230a17888f1 upstream.

mntput_no_expire() does the calculation of total refcount under mount_lock;
unfortunately, the decrement (as well as all increments) are done outside
of it, leading to false positives in the "are we dropping the last reference"
test.  Consider the following situation:
	* mnt is a lazy-umounted mount, kept alive by two opened files.  One
of those files gets closed.  Total refcount of mnt is 2.  On CPU 42
mntput(mnt) (called from __fput()) drops one reference, decrementing component
	* After it has looked at component #0, the process on CPU 0 does
mntget(), incrementing component #0, gets preempted and gets to run again -
on CPU 69.  There it does mntput(), which drops the reference (component #69)
and proceeds to spin on mount_lock.
	* On CPU 42 our first mntput() finishes counting.  It observes the
decrement of component #69, but not the increment of component #0.  As the
result, the total it gets is not 1 as it should've been - it's 0.  At which
point we decide that vfsmount needs to be killed and proceed to free it and
shut the filesystem down.  However, there's still another opened file
on that filesystem, with reference to (now freed) vfsmount, etc. and we are
screwed.

It's not a wide race, but it can be reproduced with artificial slowdown of
the mnt_get_count() loop, and it should be easier to hit on SMP KVM setups.

Fix consists of moving the refcount decrement under mount_lock; the tricky
part is that we want (and can) keep the fast case (i.e. mount that still
has non-NULL ->mnt_ns) entirely out of mount_lock.  All places that zero
mnt->mnt_ns are dropping some reference to mnt and they call synchronize_rcu()
before that mntput().  IOW, if mntput() observes (under rcu_read_lock())
a non-NULL ->mnt_ns, it is guaranteed that there is another reference yet to
be dropped.

Reported-by: Jann Horn 
Tested-by: Jann Horn 
Fixes: 48a066e72d97 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro 
Signed-off-by: Greg Kroah-Hartman

make sure that __dentry_kill() always invalidates d_seq, unhashed or not

2018-08-15T16:14:42+00:00

commit 4c0d7cd5c8416b1ef41534d19163cb07ffaa03ab upstream.

RCU pathwalk relies upon the assumption that anything that changes
->d_inode of a dentry will invalidate its ->d_seq.  That's almost
true - the one exception is that the final dput() of already unhashed
dentry does *not* touch ->d_seq at all.  Unhashing does, though,
so for anything we'd found by RCU dcache lookup we are fine.
Unfortunately, we can *start* with an unhashed dentry or jump into
it.

We could try and be careful in the (few) places where that could
happen.  Or we could just make the final dput() invalidate the damn
thing, unhashed or not.  The latter is much simpler and easier to
backport, so let's do it that way.

Reported-by: "Dae R. Jeong" 
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro 
Signed-off-by: Greg Kroah-Hartman

root dentries need RCU-delayed freeing

2018-08-15T16:14:42+00:00

commit 90bad5e05bcdb0308cfa3d3a60f5c0b9c8e2efb3 upstream.

Since mountpoint crossing can happen without leaving lazy mode,
root dentries do need the same protection against having their
memory freed without RCU delay as everything else in the tree.

It's partially hidden by RCU delay between detaching from the
mount tree and dropping the vfsmount reference, but the starting
point of pathwalk can be on an already detached mount, in which
case umount-caused RCU delay has already passed by the time the
lazy pathwalk grabs rcu_read_lock().  If the starting point
happens to be at the root of that vfsmount *and* that vfsmount
covers the entire filesystem, we get trouble.

Fixes: 48a066e72d97 ("RCU'd vsfmounts")
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro 
Signed-off-by: Greg Kroah-Hartman

ext4: fix check to prevent initializing reserved inodes

2018-08-15T16:14:41+00:00

commit 5012284700775a4e6e3fbe7eac4c543c4874b559 upstream.

Commit 8844618d8aa7: "ext4: only look at the bg_flags field if it is
valid" will complain if block group zero does not have the
EXT4_BG_INODE_ZEROED flag set.  Unfortunately, this is not correct,
since a freshly created file system has this flag cleared.  It gets
almost immediately after the file system is mounted read-write --- but
the following somewhat unlikely sequence will end up triggering a
false positive report of a corrupted file system:

   mkfs.ext4 /dev/vdc
   mount -o ro /dev/vdc /vdc
   mount -o remount,rw /dev/vdc

Instead, when initializing the inode table for block group zero, test
to make sure that itable_unused count is not too large, since that is
the case that will result in some or all of the reserved inodes
getting cleared.

This fixes the failures reported by Eric Whiteney when running
generic/230 and generic/231 in the the nojournal test case.

Fixes: 8844618d8aa7 ("ext4: only look at the bg_flags field if it is valid")
Reported-by: Eric Whitney 
Signed-off-by: Theodore Ts'o 
Signed-off-by: Greg Kroah-Hartman

jfs: Fix inconsistency between memory allocation and ea_buf->max_size

2018-08-09T10:18:00+00:00

commit 92d34134193e5b129dc24f8d79cb9196626e8d7a upstream.

The code is assuming the buffer is max_size length, but we weren't
allocating enough space for it.

Signed-off-by: Shankara Pailoor 
Signed-off-by: Dave Kleikamp 
Cc: Guenter Roeck 
Signed-off-by: Greg Kroah-Hartman

Btrfs: fix file data corruption after cloning a range and fsync

2018-08-09T10:18:00+00:00

commit bd3599a0e142cd73edd3b6801068ac3f48ac771a upstream.

When we clone a range into a file we can end up dropping existing
extent maps (or trimming them) and replacing them with new ones if the
range to be cloned overlaps with a range in the destination inode.
When that happens we add the new extent maps to the list of modified
extents in the inode's extent map tree, so that a "fast" fsync (the flag
BTRFS_INODE_NEEDS_FULL_SYNC not set in the inode) will see the extent maps
and log corresponding extent items. However, at the end of range cloning
operation we do truncate all the pages in the affected range (in order to
ensure future reads will not get stale data). Sometimes this truncation
will release the corresponding extent maps besides the pages from the page
cache. If this happens, then a "fast" fsync operation will miss logging
some extent items, because it relies exclusively on the extent maps being
present in the inode's extent tree, leading to data loss/corruption if
the fsync ends up using the same transaction used by the clone operation
(that transaction was not committed in the meanwhile). An extent map is
released through the callback btrfs_invalidatepage(), which gets called by
truncate_inode_pages_range(), and it calls __btrfs_releasepage(). The
later ends up calling try_release_extent_mapping() which will release the
extent map if some conditions are met, like the file size being greater
than 16Mb, gfp flags allow blocking and the range not being locked (which
is the case during the clone operation) nor being the extent map flagged
as pinned (also the case for cloning).

The following example, turned into a test for fstests, reproduces the
issue:

  $ mkfs.btrfs -f /dev/sdb
  $ mount /dev/sdb /mnt

  $ xfs_io -f -c "pwrite -S 0x18 9000K 6908K" /mnt/foo
  $ xfs_io -f -c "pwrite -S 0x20 2572K 156K" /mnt/bar

  $ xfs_io -c "fsync" /mnt/bar
  # reflink destination offset corresponds to the size of file bar,
  # 2728Kb minus 4Kb.
  $ xfs_io -c ""reflink ${SCRATCH_MNT}/foo 0 2724K 15908K" /mnt/bar
  $ xfs_io -c "fsync" /mnt/bar

  $ md5sum /mnt/bar
  95a95813a8c2abc9aa75a6c2914a077e  /mnt/bar

  

  $ mount /dev/sdb /mnt
  $ md5sum /mnt/bar
  207fd8d0b161be8a84b945f0df8d5f8d  /mnt/bar
  # digest should be 95a95813a8c2abc9aa75a6c2914a077e like before the
  # power failure

In the above example, the destination offset of the clone operation
corresponds to the size of the "bar" file minus 4Kb. So during the clone
operation, the extent map covering the range from 2572Kb to 2728Kb gets
trimmed so that it ends at offset 2724Kb, and a new extent map covering
the range from 2724Kb to 11724Kb is created. So at the end of the clone
operation when we ask to truncate the pages in the range from 2724Kb to
2724Kb + 15908Kb, the page invalidation callback ends up removing the new
extent map (through try_release_extent_mapping()) when the page at offset
2724Kb is passed to that callback.

Fix this by setting the bit BTRFS_INODE_NEEDS_FULL_SYNC whenever an extent
map is removed at try_release_extent_mapping(), forcing the next fsync to
search for modified extents in the fs/subvolume tree instead of relying on
the presence of extent maps in memory. This way we can continue doing a
"fast" fsync if the destination range of a clone operation does not
overlap with an existing range or if any of the criteria necessary to
remove an extent map at try_release_extent_mapping() is not met (file
size not bigger then 16Mb or gfp flags do not allow blocking).

CC: stable@vger.kernel.org # 3.16+
Signed-off-by: Filipe Manana 
Signed-off-by: David Sterba 
Signed-off-by: Sudip Mukherjee 
Signed-off-by: Greg Kroah-Hartman