linux-toradex.git/fs/btrfs/ioctl.c, branch v4.4.3

Btrfs: fix page reading in extent_same ioctl leading to csum errors

2016-02-25T20:01:15+00:00

commit 313140023026ae542ad76e7e268c56a1eaa2c28e upstream.

In the extent_same ioctl, we were grabbing the pages (locked) and
attempting to read them without bothering about any concurrent IO
against them. That is, we were not checking for any ongoing ordered
extents nor waiting for them to complete, which leads to a race where
the extent_same() code gets a checksum verification error when it
reads the pages, producing a message like the following in dmesg
and making the operation fail to user space with -ENOMEM:

[18990.161265] BTRFS warning (device sdc): csum failed ino 259 off 495616 csum 685204116 expected csum 1515870868

Fix this by using btrfs_readpage() for reading the pages instead of
extent_read_full_page_nolock(), which waits for any concurrent ordered
extents to complete and locks the io range. Also do better error handling
and don't treat all failures as -ENOMEM, as that's clearly misleasing,
becoming identical to the checks and operation of prepare_uptodate_page().

The use of extent_read_full_page_nolock() was required before
commit f441460202cb ("btrfs: fix deadlock with extent-same and readpage"),
as we had the range locked in an inode's io tree before attempting to
read the pages.

Fixes: f441460202cb ("btrfs: fix deadlock with extent-same and readpage")
Signed-off-by: Filipe Manana 
Signed-off-by: Greg Kroah-Hartman

Btrfs: fix invalid page accesses in extent_same (dedup) ioctl

2016-02-25T20:01:15+00:00

commit e0bd70c67bf996b360f706b6c643000f2e384681 upstream.

In the extent_same ioctl we are getting the pages for the source and
target ranges and unlocking them immediately after, which is incorrect
because later we attempt to map them (with kmap_atomic) and access their
contents at btrfs_cmp_data(). When we do such access the pages might have
been relocated or removed from memory, which leads to an invalid memory
access. This issue is detected on a kernel with CONFIG_DEBUG_PAGEALLOC=y
which produces a trace like the following:

186736.677437] general protection fault: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
[186736.680382] Modules linked in: btrfs dm_flakey dm_mod ppdev xor raid6_pq sha256_generic hmac drbg ansi_cprng acpi_cpufreq evdev sg aesni_intel aes_x86_64
parport_pc ablk_helper tpm_tis psmouse parport i2c_piix4 tpm cryptd i2c_core lrw processor button serio_raw pcspkr gf128mul glue_helper loop autofs4 ext4
crc16 mbcache jbd2 sd_mod sr_mod cdrom ata_generic virtio_scsi ata_piix libata virtio_pci virtio_ring crc32c_intel scsi_mod e1000 virtio floppy [last
unloaded: btrfs]
[186736.681319] CPU: 13 PID: 10222 Comm: duperemove Tainted: G        W       4.4.0-rc6-btrfs-next-18+ #1
[186736.681319] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014
[186736.681319] task: ffff880132600400 ti: ffff880362284000 task.ti: ffff880362284000
[186736.681319] RIP: 0010:[]  [] memcmp+0xb/0x22
[186736.681319] RSP: 0018:ffff880362287d70  EFLAGS: 00010287
[186736.681319] RAX: 000002c002468acf RBX: 0000000012345678 RCX: 0000000000000000
[186736.681319] RDX: 0000000000001000 RSI: 0005d129c5cf9000 RDI: 0005d129c5cf9000
[186736.681319] RBP: ffff880362287d70 R08: 0000000000000000 R09: 0000000000001000
[186736.681319] R10: ffff880000000000 R11: 0000000000000476 R12: 0000000000001000
[186736.681319] R13: ffff8802f91d4c88 R14: ffff8801f2a77830 R15: ffff880352e83e40
[186736.681319] FS:  00007f27b37fe700(0000) GS:ffff88043dda0000(0000) knlGS:0000000000000000
[186736.681319] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[186736.681319] CR2: 00007f27a406a000 CR3: 0000000217421000 CR4: 00000000001406e0
[186736.681319] Stack:
[186736.681319]  ffff880362287ea0 ffffffffa048d0bd 000000000009f000 0000000000001000
[186736.681319]  0100000000000000 ffff8801f2a77850 ffff8802f91d49b0 ffff880132600400
[186736.681319]  00000000000004f8 ffff8801c1efbe41 0000000000000000 0000000000000038
[186736.681319] Call Trace:
[186736.681319]  [] btrfs_ioctl+0x24cb/0x2731 [btrfs]
[186736.681319]  [] ? arch_local_irq_save+0x9/0xc
[186736.681319]  [] ? rcu_read_unlock+0x3e/0x5d
[186736.681319]  [] do_vfs_ioctl+0x42b/0x4ea
[186736.681319]  [] ? __fget_light+0x62/0x71
[186736.681319]  [] SyS_ioctl+0x57/0x79
[186736.681319]  [] entry_SYSCALL_64_fastpath+0x12/0x6f
[186736.681319] Code: 0a 3c 6e 74 0d 3c 79 74 04 3c 59 75 0c c6 06 01 eb 03 c6 06 00 31 c0 eb 05 b8 ea ff ff ff 5d c3 55 31 c9 48 89 e5 48 39 d1 74 13 <0f> b6
04 0f 44 0f b6 04 0e 48 ff c1 44 29 c0 74 ea eb 02 31 c0

(gdb) list *(btrfs_ioctl+0x24cb)
0x5e0e1 is in btrfs_ioctl (fs/btrfs/ioctl.c:2972).
2967                    dst_addr = kmap_atomic(dst_page);
2968
2969                    flush_dcache_page(src_page);
2970                    flush_dcache_page(dst_page);
2971
2972                    if (memcmp(addr, dst_addr, cmp_len))
2973                            ret = BTRFS_SAME_DATA_DIFFERS;
2974
2975                    kunmap_atomic(addr);
2976                    kunmap_atomic(dst_addr);

So fix this by making sure we keep the pages locked and respect the same
locking order as everywhere else: get and lock the pages first and then
lock the range in the inode's io tree (like for example at
__btrfs_buffered_write() and extent_readpages()). If an ordered extent
is found after locking the range in the io tree, unlock the range,
unlock the pages, wait for the ordered extent to complete and repeat the
entire locking process until no overlapping ordered extents are found.

Signed-off-by: Filipe Manana 
Signed-off-by: Greg Kroah-Hartman

Merge branch 'for-linus-4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

2015-11-07T01:17:13+00:00

Pull btrfs updates from Chris Mason:
 "We have a lot of subvolume quota improvements in here, along with big
  piles of cleanups from Dave Sterba and Anand Jain and others.

  Josef pitched in a batch of allocator fixes based on production use
  here at FB.  We found that mount -o ssd_spread greatly improved our
  performance on hardware raid5/6, but it exposed some CPU bottlenecks
  in the allocator.  These patches make a huge difference"

* 'for-linus-4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (100 commits)
  Btrfs: fix hole punching when using the no-holes feature
  Btrfs: find_free_extent: Do not erroneously skip LOOP_CACHING_WAIT state
  btrfs: Fix a data space underflow warning
  btrfs: qgroup: Fix a rebase bug which will cause qgroup double free
  btrfs: qgroup: Fix a race in delayed_ref which leads to abort trans
  btrfs: clear PF_NOFREEZE in cleaner_kthread()
  btrfs: qgroup: Don't copy extent buffer to do qgroup rescan
  btrfs: add balance filters limits, stripes and usage to supported mask
  btrfs: extend balance filter usage to take minimum and maximum
  btrfs: add balance filter for stripes
  btrfs: extend balance filter limit to take minimum and maximum
  btrfs: fix use after free iterating extrefs
  btrfs: check unsupported filters in balance arguments
  Btrfs: fix regression running delayed references when using qgroups
  Btrfs: fix regression when running delayed references
  Btrfs: don't do extra bitmap search in one bit case
  Btrfs: keep track of largest extent in bitmaps
  Btrfs: don't keep trying to build clusters if we are fragmented
  Btrfs: cut down on loops through the allocator
  Btrfs: don't continue setting up space cache when enospc
  ...

btrfs: check unsupported filters in balance arguments

2015-10-27T02:38:26+00:00

We don't verify that all the balance filter arguments supplemented by
the flags are actually known to the kernel. Thus we let it silently pass
and do nothing.

At the moment this means only the 'limit' filter, but we're going to add
a few more soon so it's better to have that fixed. Also in older stable
kernels so that it works with newer userspace tools.

Cc: stable@vger.kernel.org # 3.16+
Signed-off-by: David Sterba 
Signed-off-by: Chris Mason

Btrfs: fix regression running delayed references when using qgroups

2015-10-25T19:53:26+00:00

In the kernel 4.2 merge window we had a big changes to the implementation
of delayed references and qgroups which made the no_quota field of delayed
references not used anymore. More specifically the no_quota field is not
used anymore as of:

  commit 0ed4792af0e8 ("btrfs: qgroup: Switch to new extent-oriented qgroup mechanism.")

Leaving the no_quota field actually prevents delayed references from
getting merged, which in turn cause the following BUG_ON(), at
fs/btrfs/extent-tree.c, to be hit when qgroups are enabled:

  static int run_delayed_tree_ref(...)
  {
     (...)
     BUG_ON(node->ref_mod != 1);
     (...)
  }

This happens on a scenario like the following:

  1) Ref1 bytenr X, action = BTRFS_ADD_DELAYED_REF, no_quota = 1, added.

  2) Ref2 bytenr X, action = BTRFS_DROP_DELAYED_REF, no_quota = 0, added.
     It's not merged with Ref1 because Ref1->no_quota != Ref2->no_quota.

  3) Ref3 bytenr X, action = BTRFS_ADD_DELAYED_REF, no_quota = 1, added.
     It's not merged with the reference at the tail of the list of refs
     for bytenr X because the reference at the tail, Ref2 is incompatible
     due to Ref2->no_quota != Ref3->no_quota.

  4) Ref4 bytenr X, action = BTRFS_DROP_DELAYED_REF, no_quota = 0, added.
     It's not merged with the reference at the tail of the list of refs
     for bytenr X because the reference at the tail, Ref3 is incompatible
     due to Ref3->no_quota != Ref4->no_quota.

  5) We run delayed references, trigger merging of delayed references,
     through __btrfs_run_delayed_refs() -> btrfs_merge_delayed_refs().

  6) Ref1 and Ref3 are merged as Ref1->no_quota = Ref3->no_quota and
     all other conditions are satisfied too. So Ref1 gets a ref_mod
     value of 2.

  7) Ref2 and Ref4 are merged as Ref2->no_quota = Ref4->no_quota and
     all other conditions are satisfied too. So Ref2 gets a ref_mod
     value of 2.

  8) Ref1 and Ref2 aren't merged, because they have different values
     for their no_quota field.

  9) Delayed reference Ref1 is picked for running (select_delayed_ref()
     always prefers references with an action == BTRFS_ADD_DELAYED_REF).
     So run_delayed_tree_ref() is called for Ref1 which triggers the
     BUG_ON because Ref1->red_mod != 1 (equals 2).

So fix this by removing the no_quota field, as it's not used anymore as
of commit 0ed4792af0e8 ("btrfs: qgroup: Switch to new extent-oriented
qgroup mechanism.").

The use of no_quota was also buggy in at least two places:

1) At delayed-refs.c:btrfs_add_delayed_tree_ref() - we were setting
   no_quota to 0 instead of 1 when the following condition was true:
   is_fstree(ref_root) || !fs_info->quota_enabled

2) At extent-tree.c:__btrfs_inc_extent_ref() - we were attempting to
   reset a node's no_quota when the condition "!is_fstree(root_objectid)
   || !root->fs_info->quota_enabled" was true but we did it only in
   an unused local stack variable, that is, we never reset the no_quota
   value in the node itself.

This fixes the remainder of problems several people have been having when
running delayed references, mostly while a balance is running in parallel,
on a 4.2+ kernel.

Very special thanks to Stéphane Lesimple for helping debugging this issue
and testing this fix on his multi terabyte filesystem (which took more
than one day to balance alone, plus fsck, etc).

Also, this fixes deadlock issue when using the clone ioctl with qgroups
enabled, as reported by Elias Probst in the mailing list. The deadlock
happens because after calling btrfs_insert_empty_item we have our path
holding a write lock on a leaf of the fs/subvol tree and then before
releasing the path we called check_ref() which did backref walking, when
qgroups are enabled, and tried to read lock the same leaf. The trace for
this case is the following:

  INFO: task systemd-nspawn:6095 blocked for more than 120 seconds.
  (...)
  Call Trace:
    [] schedule+0x74/0x83
    [] btrfs_tree_read_lock+0xc0/0xea
    [] ? wait_woken+0x74/0x74
    [] btrfs_search_old_slot+0x51a/0x810
    [] btrfs_next_old_leaf+0xdf/0x3ce
    [] ? ulist_add_merge+0x1b/0x127
    [] __resolve_indirect_refs+0x62a/0x667
    [] ? btrfs_clear_lock_blocking_rw+0x78/0xbe
    [] find_parent_nodes+0xaf3/0xfc6
    [] __btrfs_find_all_roots+0x92/0xf0
    [] btrfs_find_all_roots+0x45/0x65
    [] ? btrfs_get_tree_mod_seq+0x2b/0x88
    [] check_ref+0x64/0xc4
    [] btrfs_clone+0x66e/0xb5d
    [] btrfs_ioctl_clone+0x48f/0x5bb
    [] ? native_sched_clock+0x28/0x77
    [] btrfs_ioctl+0xabc/0x25cb
  (...)

The problem goes away by eleminating check_ref(), which no longer is
needed as its purpose was to get a value for the no_quota field of
a delayed reference (this patch removes the no_quota field as mentioned
earlier).

Reported-by: Stéphane Lesimple 
Tested-by: Stéphane Lesimple 
Reported-by: Elias Probst 
Reported-by: Peter Becker 
Reported-by: Malte Schröder 
Reported-by: Derek Dongray 
Reported-by: Erkki Seppala 
Cc: stable@vger.kernel.org  # 4.2+
Signed-off-by: Filipe Manana 
Reviewed-by: Qu Wenruo

btrfs: qgroup: Cleanup old inaccurate facilities

2015-10-22T01:41:06+00:00

Cleanup the old facilities which use old btrfs_qgroup_reserve() function
call, replace them with the newer version, and remove the "__" prefix in
them.

Also, make btrfs_qgroup_reserve/free() functions private, as they are
now only used inside qgroup codes.

Now, the whole btrfs qgroup is swithed to use the new reserve facilities.

Signed-off-by: Qu Wenruo 
Signed-off-by: Chris Mason

btrfs: extent-tree: Switch to new delalloc space reserve and release

2015-10-22T01:41:05+00:00

Use new __btrfs_delalloc_reserve_space() and
__btrfs_delalloc_release_space() to reserve and release space for
delalloc.

Signed-off-by: Qu Wenruo 
Signed-off-by: Chris Mason

Merge branch 'integration-4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/fdmanana/linux into for-linus-4.4

2015-10-22T01:23:59+00:00

Merge branch 'cleanups/for-4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux into for-linus-4.4

2015-10-22T01:21:40+00:00

btrfs: fix possible leak in btrfs_ioctl_balance()

2015-10-22T01:10:02+00:00

Commit 8eb934591f8b ("btrfs: check unsupported filters in balance
arguments") adds a jump to exit label out_bargs in case the argument
check fails. At this point in addition to the bargs memory, the
memory for struct btrfs_balance_control has already been allocated.
Ownership of bctl is passed to btrfs_balance() in the good case,
thus the memory is not freed due to the introduced jump. Make sure
that the memory gets freed in any case as necessary. Detected by
Coverity CID 1328378.

Signed-off-by: Christian Engelmayer 
Reviewed-by: David Sterba 
Signed-off-by: Chris Mason