Skip to content

L
linux
Switch branch/tag
  1. 17 Jul, 2014 14 commits
  3. 16 Jul, 2014 6 commits
    • Thomas Gleixner's avatar
      rtmutex: Plug slow unlock race · 0c57bb13
      Thomas Gleixner authored 
      commit 27e35715 upstream.

When the rtmutex fast path is enabled the slow unlock function can
create the following situation:

spin_lock(foo->m->wait_lock);
foo->m->owner = NULL;
	    			rt_mutex_lock(foo->m); <-- fast path
				free = atomic_dec_and_test(foo->refcnt);
				rt_mutex_unlock(foo->m); <-- fast path
				if (free)
				   kfree(foo);

spin_unlock(foo->m->wait_lock); <--- Use after free.

Plug the race by changing the slow unlock to the following scheme:

     while (!rt_mutex_has_waiters(m)) {
     	    /* Clear the waiters bit in m->owner */
	    clear_rt_mutex_waiters(m);
      	    owner = rt_mutex_owner(m);
      	    spin_unlock(m->wait_lock);
      	    if (cmpxchg(m->owner, owner, 0) == owner)
      	       return;
      	    spin_lock(m->wait_lock);
     }

So in case of a new waiter incoming while the owner tries the slow
path unlock we have two situations:

 unlock(wait_lock);
					lock(wait_lock);
 cmpxchg(p, owner, 0) == owner
 	    	   			mark_rt_mutex_waiters(lock);
	 				acquire(lock);

Or:

 unlock(wait_lock);
					lock(wait_lock);
	 				mark_rt_mutex_waiters(lock);
 cmpxchg(p, owner, 0) != owner
					enqueue_waiter();
					unlock(wait_lock);
 lock(wait_lock);
 wakeup_next waiter();
 unlock(wait_lock);
					lock(wait_lock);
					acquire(lock);

If the fast path is disabled, then the simple

   m->owner = NULL;
   unlock(m->wait_lock);

is sufficient as all access to m->owner is serialized via
m->wait_lock;

Also document and clarify the wakeup_next_waiter function as suggested
by Oleg Nesterov.
Reported-by: default avatarSteven Rostedt <rostedt@goodmis.org>
Signed-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Reviewed-by: default avatarSteven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140611183852.937945560@linutronix.deSigned-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Signed-off-by: default avatarMike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      0c57bb13
    • Thomas Gleixner's avatar
      rtmutex: Handle deadlock detection smarter · a2e64fcd
      Thomas Gleixner authored 
      commit 3d5c9340 upstream.

Even in the case when deadlock detection is not requested by the
caller, we can detect deadlocks. Right now the code stops the lock
chain walk and keeps the waiter enqueued, even on itself. Silly not to
yell when such a scenario is detected and to keep the waiter enqueued.

Return -EDEADLK unconditionally and handle it at the call sites.

The futex calls return -EDEADLK. The non futex ones dequeue the
waiter, throw a warning and put the task into a schedule loop.

Tagged for stable as it makes the code more robust.
Signed-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Brad Mouring <bmouring@ni.com>
Link: http://lkml.kernel.org/r/20140605152801.836501969@linutronix.deSigned-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Signed-off-by: default avatarMike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      a2e64fcd
    • Thomas Gleixner's avatar
      rtmutex: Detect changes in the pi lock chain · 7ae4100d
      Thomas Gleixner authored 
      commit 82084984 upstream.

When we walk the lock chain, we drop all locks after each step. So the
lock chain can change under us before we reacquire the locks. That's
harmless in principle as we just follow the wrong lock path. But it
can lead to a false positive in the dead lock detection logic:

T0 holds L0
T0 blocks on L1 held by T1
T1 blocks on L2 held by T2
T2 blocks on L3 held by T3
T4 blocks on L4 held by T4

Now we walk the chain

lock T1 -> lock L2 -> adjust L2 -> unlock T1 ->
     lock T2 ->  adjust T2 ->  drop locks

T2 times out and blocks on L0

Now we continue:

lock T2 -> lock L0 -> deadlock detected, but it's not a deadlock at all.

Brad tried to work around that in the deadlock detection logic itself,
but the more I looked at it the less I liked it, because it's crystal
ball magic after the fact.

We actually can detect a chain change very simple:

lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 ->

     next_lock = T2->pi_blocked_on->lock;

drop locks

T2 times out and blocks on L0

Now we continue:

lock T2 ->

     if (next_lock != T2->pi_blocked_on->lock)
     	   return;

So if we detect that T2 is now blocked on a different lock we stop the
chain walk. That's also correct in the following scenario:

lock T1 -> lock L2 -> adjust L2 -> unlock T1 -> lock T2 -> adjust T2 ->

     next_lock = T2->pi_blocked_on->lock;

drop locks

T3 times out and drops L3
T2 acquires L3 and blocks on L4 now

Now we continue:

lock T2 ->

     if (next_lock != T2->pi_blocked_on->lock)
     	   return;

We don't have to follow up the chain at that point, because T2
propagated our priority up to T4 already.

[ Folded a cleanup patch from peterz ]
Signed-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Reported-by: default avatarBrad Mouring <bmouring@ni.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140605152801.930031935@linutronix.deSigned-off-by: default avatarMike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      7ae4100d
    • Thomas Gleixner's avatar
      rtmutex: Fix deadlock detector for real · 9e30d699
      Thomas Gleixner authored 
      commit 397335f0 upstream.

The current deadlock detection logic does not work reliably due to the
following early exit path:

	/*
	 * Drop out, when the task has no waiters. Note,
	 * top_waiter can be NULL, when we are in the deboosting
	 * mode!
	 */
	if (top_waiter && (!task_has_pi_waiters(task) ||
			   top_waiter != task_top_pi_waiter(task)))
		goto out_unlock_pi;

So this not only exits when the task has no waiters, it also exits
unconditionally when the current waiter is not the top priority waiter
of the task.

So in a nested locking scenario, it might abort the lock chain walk
and therefor miss a potential deadlock.

Simple fix: Continue the chain walk, when deadlock detection is
enabled.

We also avoid the whole enqueue, if we detect the deadlock right away
(A-A). It's an optimization, but also prevents that another waiter who
comes in after the detection and before the task has undone the damage
observes the situation and detects the deadlock and returns
-EDEADLOCK, which is wrong as the other task is not in a deadlock
situation.
Signed-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: default avatarSteven Rostedt <rostedt@goodmis.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Link: http://lkml.kernel.org/r/20140522031949.725272460@linutronix.deSigned-off-by: default avatarThomas Gleixner <tglx@linutronix.de>
Signed-off-by: default avatarMike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      9e30d699
    • Florian Westphal's avatar
      netfilter: nf_nat: fix oops on netns removal · 26e4e17c
      Florian Westphal authored 
      [ Upstream commit 945b2b2d ]

Quoting Samu Kallio:

 Basically what's happening is, during netns cleanup,
 nf_nat_net_exit gets called before ipv4_net_exit. As I understand
 it, nf_nat_net_exit is supposed to kill any conntrack entries which
 have NAT context (through nf_ct_iterate_cleanup), but for some
 reason this doesn't happen (perhaps something else is still holding
 refs to those entries?).

 When ipv4_net_exit is called, conntrack entries (including those
 with NAT context) are cleaned up, but the
 nat_bysource hashtable is long gone - freed in nf_nat_net_exit. The
 bug happens when attempting to free a conntrack entry whose NAT hash
 'prev' field points to a slot in the freed hash table (head for that
 bin).

We ignore conntracks with null nat bindings.  But this is wrong,
as these are in bysource hash table as well.

Restore nat-cleaning for the netns-is-being-removed case.

bug:
https://bugzilla.kernel.org/show_bug.cgi?id=65191

Cc: <stable@vger.kernel.org> # 3.15.x
Cc: <stable@vger.kernel.org> # 3.14.x
Cc: <stable@vger.kernel.org> # 3.12.x
Cc: <stable@vger.kernel.org> # 3.10.x
Fixes: c2d421e1 ('netfilter: nf_nat: fix race when unloading protocol modules')
Reported-by: default avatarSamu Kallio <samu.kallio@aberdeencloud.com>
Debugged-by: default avatarSamu Kallio <samu.kallio@aberdeencloud.com>
Signed-off-by: default avatarFlorian Westphal <fw@strlen.de>
Tested-by: default avatarSamu Kallio <samu.kallio@aberdeencloud.com>
Signed-off-by: default avatarPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      26e4e17c
    • Julian Anastasov's avatar
      ipvs: stop tot_stats estimator only under CONFIG_SYSCTL · f730d2ab
      Julian Anastasov authored 
      [ Upstream commit 9802d21e ]

The tot_stats estimator is started only when CONFIG_SYSCTL
is defined. But it is stopped without checking CONFIG_SYSCTL.
Fix the crash by moving ip_vs_stop_estimator into
ip_vs_control_net_cleanup_sysctl.

The change is needed after commit 14e40546
("IPVS: Add __ip_vs_control_{init,cleanup}_sysctl()") from 2.6.39.

Cc: <stable@vger.kernel.org> # 3.15.x
Cc: <stable@vger.kernel.org> # 3.14.x
Cc: <stable@vger.kernel.org> # 3.12.x
Cc: <stable@vger.kernel.org> # 3.10.x
Cc: <stable@vger.kernel.org> # 3.2.x
Reported-by: default avatarJet Chen <jet.chen@intel.com>
Tested-by: default avatarJet Chen <jet.chen@intel.com>
Signed-off-by: default avatarJulian Anastasov <ja@ssi.bg>
Sgned-off-by: default avatarSimon Horman <horms@verge.net.au>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      f730d2ab
  5. 14 Jul, 2014 1 commit
  7. 04 Jul, 2014 10 commits
    • Jiri Slaby's avatar
      Linux 3.12.24 · 8097be3b
      Jiri Slaby authored
      8097be3b
    • Jie Liu's avatar
      xfs: don't perform discard if the given range length is less than block size · 0af3f136
      Jie Liu authored 
      commit f9fd0135 upstream.

For discard operation, we should return EINVAL if the given range length
is less than a block size, otherwise it will go through the file system
to discard data blocks as the end range might be evaluated to -1, e.g,
/xfs7: 9811378176 bytes were trimmed

This issue can be triggered via xfstests/generic/288.

Also, it seems to get the request queue pointer via bdev_get_queue()
instead of the hard code pointer dereference is not a bad thing.
Signed-off-by: default avatarJie Liu <jeff.liu@oracle.com>
Reviewed-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      0af3f136
    • Dave Chinner's avatar
      xfs: xfs_remove deadlocks due to inverted AGF vs AGI lock ordering · f7009499
      Dave Chinner authored 
      commit 27320369 upstream.

Removing an inode from the namespace involves removing the directory
entry and dropping the link count on the inode. Removing the
directory entry can result in locking an AGF (directory blocks were
freed) and removing a link count can result in placing the inode on
an unlinked list which results in locking an AGI.

The big problem here is that we have an ordering constraint on AGF
and AGI locking - inode allocation locks the AGI, then can allocate
a new extent for new inodes, locking the AGF after the AGI.
Similarly, freeing the inode removes the inode from the unlinked
list, requiring that we lock the AGI first, and then freeing the
inode can result in an inode chunk being freed and hence freeing
disk space requiring that we lock an AGF.

Hence the ordering that is imposed by other parts of the code is AGI
before AGF. This means we cannot remove the directory entry before
we drop the inode reference count and put it on the unlinked list as
this results in a lock order of AGF then AGI, and this can deadlock
against inode allocation and freeing. Therefore we must drop the
link counts before we remove the directory entry.

This is still safe from a transactional point of view - it is not
until we get to xfs_bmap_finish() that we have the possibility of
multiple transactions in this operation. Hence as long as we remove
the directory entry and drop the link count in the first transaction
of the remove operation, there are no transactional constraints on
the ordering here.

Change the ordering of the operations in the xfs_remove() function
to align the ordering of AGI and AGF locking to match that of the
rest of the code.
Signed-off-by: default avatarDave Chinner <dchinner@redhat.com>
Reviewed-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      f7009499
    • Jie Liu's avatar
      xfs: fix the extent count when allocating an new indirection array entry · 67820ad0
      Jie Liu authored 
      commit bb86d21c upstream.

At xfs_iext_add(), if extent(s) are being appended to the last page in
the indirection array and the new extent(s) don't fit in the page, the
number of extents(erp->er_extcount) in a new allocated entry should be
the minimum value between count and XFS_LINEAR_EXTS, instead of count.

For now, there is no existing test case can demonstrates a problem with
the er_extcount being set incorrectly here, but it obviously like a bug.
Signed-off-by: default avatarJie Liu <jeff.liu@oracle.com>
Reviewed-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      67820ad0
    • Geyslan G. Bem's avatar
      xfs: fix possible NULL dereference in xlog_verify_iclog · 9af76725
      Geyslan G. Bem authored 
      commit 643f7c4e upstream.

In xlog_verify_iclog a debug check of the incore log buffers prints an
error if icptr is null and then goes on to dereference the pointer
regardless.  Convert this to an assert so that the intention is clear.
This was reported by Coverty.
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Reviewed-by: default avatarEric Sandeen <sandeen@redhat.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      9af76725
    • Dave Chinner's avatar
      xfs: prevent stack overflows from page cache allocation · 71024660
      Dave Chinner authored 
      commit ad22c7a0 upstream.

Page cache allocation doesn't always go through ->begin_write and
hence we don't always get the opportunity to set the allocation
context to GFP_NOFS. Failing to do this means we open up the direct
relcaim stack to recurse into the filesystem and consume a
significant amount of stack.

On RHEL6.4 kernels we are seeing ra_submit() and
generic_file_splice_read() from an nfsd context recursing into the
filesystem via the inode cache shrinker and evicting inodes. This is
causing truncation to be run (e.g EOF block freeing) and causing
bmap btree block merges and free space btree block splits to occur.
These btree manipulations are occurring with the call chain already
30 functions deep and hence there is not enough stack space to
complete such operations.

To avoid these specific overruns, we need to prevent the page cache
allocation from recursing via direct reclaim. We can do that because
the allocation functions take the allocation context from that which
is stored in the mapping for the inode. We don't set that right now,
so the default is GFP_HIGHUSER_MOVABLE, which is effectively a
GFP_KERNEL context. We need it to be the equivalent of GFP_NOFS, so
when we initialise an inode, set the mapping gfp mask appropriately.

This makes the use of AOP_FLAG_NOFS redundant from other parts of
the XFS IO path, so get rid of it.
Signed-off-by: default avatarDave Chinner <dchinner@redhat.com>
Reviewed-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      71024660
    • Eric Sandeen's avatar
      xfs: don't break from growfs ag update loop on error · e465c957
      Eric Sandeen authored 
      commit 59e5a0e8 upstream.

When xfs_growfs_data_private() is updating backup superblocks,
it bails out on the first error encountered, whether reading or
writing:

* If we get an error writing out the alternate superblocks,
* just issue a warning and continue.  The real work is
* already done and committed.

This can cause a problem later during repair, because repair
looks at all superblocks, and picks the most prevalent one
as correct.  If we bail out early in the backup superblock
loop, we can end up with more "bad" matching superblocks than
good, and a post-growfs repair may revert the filesystem to
the old geometry.

With the combination of superblock verifiers and old bugs,
we're more likely to encounter read errors due to verification.

And perhaps even worse, we don't even properly write any of the
newly-added superblocks in the new AGs.

Even with this change, growfs will still say:

  xfs_growfs: XFS_IOC_FSGROWFSDATA xfsctl failed: Structure needs cleaning
  data blocks changed from 319815680 to 335216640

which might be confusing to the user, but it at least communicates
that something has gone wrong, and dmesg will probably highlight
the need for an xfs_repair.

And this is still best-effort; if verifiers fail on more than
half the backup supers, they may still "win" - but that's probably
best left to repair to more gracefully handle by doing its own
strict verification as part of the backup super "voting."
Signed-off-by: default avatarEric Sandeen <sandeen@redhat.com>
Acked-by: default avatarDave Chinner <david@fromorbit.com>
Reviewed-by: default avatarMark Tinguely <tinguely@sgi.com>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      e465c957
    • Eric Sandeen's avatar
      xfs: don't emit corruption noise on fs probes · 22a16208
      Eric Sandeen authored 
      commit 31625f28 upstream.

If we get EWRONGFS due to probing of non-xfs filesystems,
there's no need to issue the scary corruption error and backtrace.
Signed-off-by: default avatarEric Sandeen <sandeen@redhat.com>
Reviewed-by: default avatarMark Tinguely <tinguely@sgi.com>
Reviewed-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      22a16208
    • Dave Chinner's avatar
      xfs: prevent deadlock trying to cover an active log · 192fedc5
      Dave Chinner authored 
      commit 2c6e24ce upstream.

Recent analysis of a deadlocked XFS filesystem from a kernel
crash dump indicated that the filesystem was stuck waiting for log
space. The short story of the hang on the RHEL6 kernel is this:

	- the tail of the log is pinned by an inode
	- the inode has been pushed by the xfsaild
	- the inode has been flushed to it's backing buffer and is
	  currently flush locked and hence waiting for backing
	  buffer IO to complete and remove it from the AIL
	- the backing buffer is marked for write - it is on the
	  delayed write queue
	- the inode buffer has been modified directly and logged
	  recently due to unlinked inode list modification
	- the backing buffer is pinned in memory as it is in the
	  active CIL context.
	- the xfsbufd won't start buffer writeback because it is
	  pinned
	- xfssyncd won't force the log because it sees the log as
	  needing to be covered and hence wants to issue a dummy
	  transaction to move the log covering state machine along.

Hence there is no trigger to force the CIL to the log and hence
unpin the inode buffer and therefore complete the inode IO, remove
it from the AIL and hence move the tail of the log along, allowing
transactions to start again.

Mainline kernels also have the same deadlock, though the signature
is slightly different - the inode buffer never reaches the delayed
write lists because xfs_buf_item_push() sees that it is pinned and
hence never adds it to the delayed write list that the xfsaild
flushes.

There are two possible solutions here. The first is to simply force
the log before trying to cover the log and so ensure that the CIL is
emptied before we try to reserve space for the dummy transaction in
the xfs_log_worker(). While this might work most of the time, it is
still racy and is no guarantee that we don't get stuck in
xfs_trans_reserve waiting for log space to come free. Hence it's not
the best way to solve the problem.

The second solution is to modify xfs_log_need_covered() to be aware
of the CIL. We only should be attempting to cover the log if there
is no current activity in the log - covering the log is the process
of ensuring that the head and tail in the log on disk are identical
(i.e. the log is clean and at idle). Hence, by definition, if there
are items in the CIL then the log is not at idle and so we don't
need to attempt to cover it.

When we don't need to cover the log because it is active or idle, we
issue a log force from xfs_log_worker() - if the log is idle, then
this does nothing.  However, if the log is active due to there being
items in the CIL, it will force the items in the CIL to the log and
unpin them.

In the case of the above deadlock scenario, instead of
xfs_log_worker() getting stuck in xfs_trans_reserve() attempting to
cover the log, it will instead force the log, thereby unpinning the
inode buffer, allowing IO to be issued and complete and hence
removing the inode that was pinning the tail of the log from the
AIL. At that point, everything will start moving along again. i.e.
the xfs_log_worker turns back into a watchdog that can alleviate
deadlocks based around pinned items that prevent the tail of the log
from being moved...
Signed-off-by: default avatarDave Chinner <dchinner@redhat.com>
Reviewed-by: default avatarEric Sandeen <sandeen@redhat.com>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      192fedc5
    • Jie Liu's avatar
      xfs: fix the wrong new_size/rnew_size at xfs_iext_realloc_direct() · 5b093812
      Jie Liu authored 
      commit 17ec81c1 upstream.

At xfs_iext_realloc_direct(), the new_size is changed by adding
if_bytes if originally the extent records are stored at the inline
extent buffer, and we have to switch from it to a direct extent
list for those new allocated extents, this is wrong. e.g,

Create a file with three extents which was showing as following,

xfs_io -f -c "truncate 100m" /xfs/testme

for i in $(seq 0 5 10); do
	offset=$(($i * $((1 << 20))))
	xfs_io -c "pwrite $offset 1m" /xfs/testme
done

Inline
------
irec:	if_bytes	bytes_diff	new_size
1st	0		16		16
2nd	16		16		32

Switching
---------						rnew_size
3rd	32		16		48 + 32 = 80	roundup=128

In this case, the desired value of new_size should be 48, and then
it will be roundup to 64 and be assigned to rnew_size.

However, this issue has been covered by resetting the if_bytes to
the new_size which is calculated at the begnning of xfs_iext_add()
before leaving out this function, and in turn make the rnew_size
correctly again. Hence, this can not be detected via xfstestes.

This patch fix above problem and revise the new_size comments at
xfs_iext_realloc_direct() to make it more readable.  Also, fix the
comments while switching from the inline extent buffer to a direct
extent list to reflect this change.
Signed-off-by: default avatarJie Liu <jeff.liu@oracle.com>
Reviewed-by: default avatarDave Chinner <dchinner@redhat.com>
Signed-off-by: default avatarBen Myers <bpm@sgi.com>
Signed-off-by: default avatarJiri Slaby <jslaby@suse.cz>
      5b093812
  9. 02 Jul, 2014 9 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7