We no longer pass in a bitmap of rq_flag_bits bits to __btrfs_map_block.
It will always be a REQ_OP, or the btrfs specific REQ_GET_READ_MIRRORS,
so this drops the bit tests.
Signed-off-by: NMike Christie <mchristi@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NJens Axboe <axboe@fb.com>

This should be the easier cases to convert btrfs to
bio_set_op_attrs/bio_op.
They are mostly just cut and replace type of changes.
Signed-off-by: NMike Christie <mchristi@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NJens Axboe <axboe@fb.com>

This has callers of submit_bio/submit_bio_wait set the bio->bi_rw
instead of passing it in. This makes that use the same as
generic_make_request and how we set the other bio fields.
Signed-off-by: NMike Christie <mchristi@redhat.com>

Fixed up fs/ext4/crypto.c
Signed-off-by: NJens Axboe <axboe@fb.com>

While iterating and copying extents from the source device, the device
replace code keeps adjusting a left cursor that is used to make sure that
once we finish processing a device extent, any future writes to extents
from the corresponding block group will get into both the source and
target devices. This left cursor is also used for resuming the device
replace operation at mount time.

However using this left cursor to decide whether writes go into both
devices or only the source device is not enough to guarantee we don't
miss copying extents into the target device. There are two cases where
the current approach fails. The first one is related to when there are
holes in the device and they get allocated for new block groups while
the device replace operation is iterating the device extents (more on
this explained below). The second one is that when that loop over the
device extents finishes, we start dellaloc, wait for all ordered extents
and then commit the current transaction, we might have got new block
groups allocated that are now using a device extent that has an offset
greater then or equals to the value of the left cursor, in which case
writes to extents belonging to these new block groups will get issued
only to the source device.

For the first case where the current approach of using a left cursor
fails, consider the source device currently has the following layout:

  [ extent bg A ] [ hole, unallocated space ] [extent bg B ]
  3Gb             4Gb                         5Gb

While we are iterating the device extents from the source device using
the commit root of the device tree, the following happens:

        CPU 1                                            CPU 2

                      <we are at transaction N>

  scrub_enumerate_chunks()
    --> searches the device tree for
        extents belonging to the source
        device using the device tree's
        commit root
    --> 1st iteration finds extent belonging to
        block group A

        --> sets block group A to RO mode
            (btrfs_inc_block_group_ro)

        --> sets cursor left to found_key.offset
            which is 3Gb

        --> scrub_chunk() starts
            copies all allocated extents from
            block group's A stripe at source
            device into target device

                                                           btrfs_alloc_chunk()
                                                             --> allocates device extent
                                                                 in the range [4Gb, 5Gb[
                                                                 from the source device for
                                                                 a new block group C

                                                           extent allocated from block
                                                           group C for a direct IO,
                                                           buffered write or btree node/leaf

                                                           extent is written to, perhaps
                                                           in response to a writepages()
                                                           call from the VM or directly
                                                           through direct IO

                                                           the write is made only against
                                                           the source device and not against
                                                           the target device because the
                                                           extent's offset is in the interval
                                                           [4Gb, 5Gb[ which is larger then
                                                           the value of cursor_left (3Gb)

        --> scrub_chunks() finishes

        --> updates left cursor from 3Gb to
            4Gb

        --> btrfs_dec_block_group_ro() sets
            block group A back to RW mode

                             <we are still at transaction N>

    --> 2nd iteration finds extent belonging to
        block group B - it did not find the new
        extent in the range [4Gb, 5Gb[ for block
        group C because we are using the device
        tree's commit root or even because the
        block group's items are not all yet
        inserted in the respective btrees, that is,
        the block group is still attached to some
        transaction handle's new_bgs list and
        btrfs_create_pending_block_groups() was
        not called yet against that transaction
        handle, so the device extent items were
        not yet inserted into the devices tree

                             <we are still at transaction N>

        --> so we end not copying anything from the newly
            allocated device extent from the source device
            to the target device

So fix this by making __btrfs_map_block() always redirect writes to the
target device as well, independently of the left cursor's value. With
this change the left cursor is now used only for the purpose of tracking
progress and allow a mount operation to resume a device replace.
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Reviewed-by: NJosef Bacik <jbacik@fb.com>

When it's finishing, the device replace code iterates all extent maps
representing block group and for each one that has a stripe that refers
to the source device, it replaces its device with the target device.
However when it replaces the source device with the target device it,
the target device still has an ID of 0ULL (BTRFS_DEV_REPLACE_DEVID),
only after its ID is changed to match the one from the source device.
This leads to races with the chunk removal code that can temporarly see
a device with an ID of 0ULL and then attempt to use that ID to remove
items from the device tree and fail, causing a transaction abort:

[ 9238.594364] BTRFS info (device sdf): dev_replace from /dev/sdf (devid 3) to /dev/sde finished
[ 9238.594377] ------------[ cut here ]------------
[ 9238.594402] WARNING: CPU: 14 PID: 21566 at fs/btrfs/volumes.c:2771 btrfs_remove_chunk+0x2e5/0x793 [btrfs]
[ 9238.594403] BTRFS: Transaction aborted (error 1)
[ 9238.594416] Modules linked in: btrfs crc32c_generic acpi_cpufreq xor tpm_tis tpm raid6_pq ppdev parport_pc processor psmouse parport i2c_piix4 evdev sg i2c_core se
rio_raw pcspkr button loop autofs4 ext4 crc16 jbd2 mbcache sr_mod cdrom sd_mod ata_generic virtio_scsi ata_piix virtio_pci libata virtio_ring virtio e1000 scsi_mod fl
oppy [last unloaded: btrfs]
[ 9238.594418] CPU: 14 PID: 21566 Comm: btrfs-cleaner Not tainted 4.6.0-rc7-btrfs-next-29+ #1
[ 9238.594419] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014
[ 9238.594421]  0000000000000000 ffff88017f1dbc60 ffffffff8126b42c ffff88017f1dbcb0
[ 9238.594422]  0000000000000000 ffff88017f1dbca0 ffffffff81052b14 00000ad37f1dbd18
[ 9238.594423]  0000000000000001 ffff88018068a558 ffff88005c4b9c00 ffff880233f60db0
[ 9238.594424] Call Trace:
[ 9238.594428]  [<ffffffff8126b42c>] dump_stack+0x67/0x90
[ 9238.594430]  [<ffffffff81052b14>] __warn+0xc2/0xdd
[ 9238.594432]  [<ffffffff81052b7a>] warn_slowpath_fmt+0x4b/0x53
[ 9238.594434]  [<ffffffff8116c311>] ? kmem_cache_free+0x128/0x188
[ 9238.594450]  [<ffffffffa04d43f5>] btrfs_remove_chunk+0x2e5/0x793 [btrfs]
[ 9238.594452]  [<ffffffff8108e456>] ? arch_local_irq_save+0x9/0xc
[ 9238.594464]  [<ffffffffa04a26fa>] btrfs_delete_unused_bgs+0x317/0x382 [btrfs]
[ 9238.594476]  [<ffffffffa04a961d>] cleaner_kthread+0x1ad/0x1c7 [btrfs]
[ 9238.594489]  [<ffffffffa04a9470>] ? btree_invalidatepage+0x8e/0x8e [btrfs]
[ 9238.594490]  [<ffffffff8106f403>] kthread+0xd4/0xdc
[ 9238.594494]  [<ffffffff8149e242>] ret_from_fork+0x22/0x40
[ 9238.594495]  [<ffffffff8106f32f>] ? kthread_stop+0x286/0x286
[ 9238.594496] ---[ end trace 183efbe50275f059 ]---

The sequence of steps leading to this is like the following:

              CPU 1                                           CPU 2

 btrfs_dev_replace_finishing()

   at this point
   dev_replace->tgtdev->devid ==
   BTRFS_DEV_REPLACE_DEVID (0ULL)

   ...

   btrfs_start_transaction()
   btrfs_commit_transaction()

                                                     btrfs_delete_unused_bgs()
                                                       btrfs_remove_chunk()

                                                         looks up for the extent map
                                                         corresponding to the chunk

                                                         lock_chunks() (chunk_mutex)
                                                         check_system_chunk()
                                                         unlock_chunks() (chunk_mutex)

   locks fs_info->chunk_mutex

   btrfs_dev_replace_update_device_in_mapping_tree()
     --> iterates fs_info->mapping_tree and
         replaces the device in every extent
         map's map->stripes[] with
         dev_replace->tgtdev, which still has
         an id of 0ULL (BTRFS_DEV_REPLACE_DEVID)

                                                         iterates over all stripes from
                                                         the extent map

                                                           --> calls btrfs_free_dev_extent()
                                                               passing it the target device
                                                               that still has an ID of 0ULL

                                                           --> btrfs_free_dev_extent() fails
                                                             --> aborts current transaction

   finishes setting up the target device,
   namely it sets tgtdev->devid to the value
   of srcdev->devid (which is necessarily > 0)

   frees the srcdev

   unlocks fs_info->chunk_mutex

So fix this by taking the device list mutex while processing the stripes
for the chunk's extent map. This is similar to the race between device
replace and block group creation that was fixed by commit 50460e37
("Btrfs: fix race when finishing dev replace leading to transaction abort").
Signed-off-by: NFilipe Manana <fdmanana@suse.com>
Reviewed-by: NJosef Bacik <jbacik@fb.com>

Signed-off-by: NNicholas D Steeves <nsteeves@gmail.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

While reading sys_chunk_array in superblock, btrfs creates a temporary
extent buffer.  Since we don't use it after finishing reading
 sys_chunk_array, we don't need to keep it in memory.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Let's gather the UUID related functions under one hood.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: NMatt Fleming <matt@codeblueprint.co.uk>
Cc: Dmitry Kasatkin <dmitry.kasatkin@gmail.com>
Cc: Mimi Zohar <zohar@linux.vnet.ibm.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Move scratch super outside of the chunk lock to avoid below
lockdep warning. The better place to scratch super is in
the function btrfs_rm_dev_replace_free_srcdev() just before
free_device, which is outside of the chunk lock as well.

To reproduce:
  (fresh boot)
  mkfs.btrfs -f -draid5 -mraid5 /dev/sdc /dev/sdd /dev/sde
  mount /dev/sdc /btrfs
  dd if=/dev/zero of=/btrfs/tf1 bs=4096 count=100
  (get devmgt from https://github.com/asj/devmgt.git)
  devmgt detach /dev/sde
  dd if=/dev/zero of=/btrfs/tf1 bs=4096 count=100
  sync
  btrfs replace start -Brf 3 /dev/sdf /btrfs <--
  devmgt attach host7

======================================================
[ INFO: possible circular locking dependency detected ]
4.6.0-rc2asj+ #1 Not tainted
---------------------------------------------------

btrfs/2174 is trying to acquire lock:
(sb_writers){.+.+.+}, at:
[<ffffffff812449b4>] __sb_start_write+0xb4/0xf0

but task is already holding lock:
(&fs_info->chunk_mutex){+.+.+.}, at:
[<ffffffffa05c5f55>] btrfs_dev_replace_finishing+0x145/0x980 [btrfs]

which lock already depends on the new lock.

Chain exists of:
sb_writers --> &fs_devs->device_list_mutex --> &fs_info->chunk_mutex
Possible unsafe locking scenario:
CPU0				CPU1
----				----
lock(&fs_info->chunk_mutex);
				lock(&fs_devs->device_list_mutex);
				lock(&fs_info->chunk_mutex);
lock(sb_writers);

*** DEADLOCK ***

-> #0 (sb_writers){.+.+.+}:
[<ffffffff810e6415>] __lock_acquire+0x1bc5/0x1ee0
[<ffffffff810e707e>] lock_acquire+0xbe/0x210
[<ffffffff810df49a>] percpu_down_read+0x4a/0xa0
[<ffffffff812449b4>] __sb_start_write+0xb4/0xf0
[<ffffffff81265534>] mnt_want_write+0x24/0x50
[<ffffffff812508a2>] path_openat+0x952/0x1190
[<ffffffff81252451>] do_filp_open+0x91/0x100
[<ffffffff8123f5cc>] file_open_name+0xfc/0x140
[<ffffffff8123f643>] filp_open+0x33/0x60
[<ffffffffa0572bb6>] update_dev_time+0x16/0x40 [btrfs]
[<ffffffffa057f60d>] btrfs_scratch_superblocks+0x5d/0xb0 [btrfs]
[<ffffffffa057f70e>] btrfs_rm_dev_replace_remove_srcdev+0xae/0xd0 [btrfs]
[<ffffffffa05c62c5>] btrfs_dev_replace_finishing+0x4b5/0x980 [btrfs]
[<ffffffffa05c6ae8>] btrfs_dev_replace_start+0x358/0x530 [btrfs]
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

btrfs_map_block can go horribly wrong in the face of fs corruption, lets agree
to not be assholes and panic at any possible chance things are all fucked up.
Signed-off-by: NJosef Bacik <jbacik@fb.com>
[ removed type casts ]
Signed-off-by: NDavid Sterba <dsterba@suse.com>

The struct 'map_lookup' uses type int for @stripe_len, while
btrfs_chunk_stripe_len() can return a u64 value, and it may end up with
@stripe_len being undefined value and it can lead to 'divide error' in
 __btrfs_map_block().

This changes 'map_lookup' to use type u64 for stripe_len, also right now
we only use BTRFS_STRIPE_LEN for stripe_len, so this adds a valid checker for
BTRFS_STRIPE_LEN.
Reported-by: NVegard Nossum <vegard.nossum@oracle.com>
Reported-by: NQuentin Casasnovas <quentin.casasnovas@oracle.com>
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Reviewed-by: NDavid Sterba <dsterba@suse.com>
[ folded division fix to scrub_raid56_parity ]
Signed-off-by: NDavid Sterba <dsterba@suse.com>

When the replace target fails, the target device will be taken
out of fs device list, scratch + update_dev_time and freed. However
we could do the scratch  + update_dev_time and free part after the
device has been taken out of device list, so that we don't have to
hold the device_list_mutex and uuid_mutex locks.

Reported issue:

[ 5375.718845] ======================================================
[ 5375.718846] [ INFO: possible circular locking dependency detected ]
[ 5375.718849] 4.4.5-scst31x-debug-11+ #40 Not tainted
[ 5375.718849] -------------------------------------------------------
[ 5375.718851] btrfs-health/4662 is trying to acquire lock:
[ 5375.718861]  (sb_writers){.+.+.+}, at: [<ffffffff812214f7>] __sb_start_write+0xb7/0xf0
[ 5375.718862]
[ 5375.718862] but task is already holding lock:
[ 5375.718907]  (&fs_devs->device_list_mutex){+.+.+.}, at: [<ffffffffa028263c>] btrfs_destroy_dev_replace_tgtdev+0x3c/0x150 [btrfs]
[ 5375.718907]
[ 5375.718907] which lock already depends on the new lock.
[ 5375.718907]
[ 5375.718908]
[ 5375.718908] the existing dependency chain (in reverse order) is:
[ 5375.718911]
[ 5375.718911] -> #3 (&fs_devs->device_list_mutex){+.+.+.}:
[ 5375.718917]        [<ffffffff810da4be>] lock_acquire+0xce/0x1e0
[ 5375.718921]        [<ffffffff81633949>] mutex_lock_nested+0x69/0x3c0
[ 5375.718940]        [<ffffffffa0219bf6>] btrfs_show_devname+0x36/0x210 [btrfs]
[ 5375.718945]        [<ffffffff81267079>] show_vfsmnt+0x49/0x150
[ 5375.718948]        [<ffffffff81240b07>] m_show+0x17/0x20
[ 5375.718951]        [<ffffffff81246868>] seq_read+0x2d8/0x3b0
[ 5375.718955]        [<ffffffff8121df28>] __vfs_read+0x28/0xd0
[ 5375.718959]        [<ffffffff8121e806>] vfs_read+0x86/0x130
[ 5375.718962]        [<ffffffff8121f4c9>] SyS_read+0x49/0xa0
[ 5375.718966]        [<ffffffff81637976>] entry_SYSCALL_64_fastpath+0x16/0x7a
[ 5375.718968]
[ 5375.718968] -> #2 (namespace_sem){+++++.}:
[ 5375.718971]        [<ffffffff810da4be>] lock_acquire+0xce/0x1e0
[ 5375.718974]        [<ffffffff81635199>] down_write+0x49/0x80
[ 5375.718977]        [<ffffffff81243593>] lock_mount+0x43/0x1c0
[ 5375.718979]        [<ffffffff81243c13>] do_add_mount+0x23/0xd0
[ 5375.718982]        [<ffffffff81244afb>] do_mount+0x27b/0xe30
[ 5375.718985]        [<ffffffff812459dc>] SyS_mount+0x8c/0xd0
[ 5375.718988]        [<ffffffff81637976>] entry_SYSCALL_64_fastpath+0x16/0x7a
[ 5375.718991]
[ 5375.718991] -> #1 (&sb->s_type->i_mutex_key#5){+.+.+.}:
[ 5375.718994]        [<ffffffff810da4be>] lock_acquire+0xce/0x1e0
[ 5375.718996]        [<ffffffff81633949>] mutex_lock_nested+0x69/0x3c0
[ 5375.719001]        [<ffffffff8122d608>] path_openat+0x468/0x1360
[ 5375.719004]        [<ffffffff8122f86e>] do_filp_open+0x7e/0xe0
[ 5375.719007]        [<ffffffff8121da7b>] do_sys_open+0x12b/0x210
[ 5375.719010]        [<ffffffff8121db7e>] SyS_open+0x1e/0x20
[ 5375.719013]        [<ffffffff81637976>] entry_SYSCALL_64_fastpath+0x16/0x7a
[ 5375.719015]
[ 5375.719015] -> #0 (sb_writers){.+.+.+}:
[ 5375.719018]        [<ffffffff810d97ca>] __lock_acquire+0x17ba/0x1ae0
[ 5375.719021]        [<ffffffff810da4be>] lock_acquire+0xce/0x1e0
[ 5375.719026]        [<ffffffff810d3bef>] percpu_down_read+0x4f/0xa0
[ 5375.719028]        [<ffffffff812214f7>] __sb_start_write+0xb7/0xf0
[ 5375.719031]        [<ffffffff81242eb4>] mnt_want_write+0x24/0x50
[ 5375.719035]        [<ffffffff8122ded2>] path_openat+0xd32/0x1360
[ 5375.719037]        [<ffffffff8122f86e>] do_filp_open+0x7e/0xe0
[ 5375.719040]        [<ffffffff8121d8a4>] file_open_name+0xe4/0x130
[ 5375.719043]        [<ffffffff8121d923>] filp_open+0x33/0x60
[ 5375.719073]        [<ffffffffa02776a6>] update_dev_time+0x16/0x40 [btrfs]
[ 5375.719099]        [<ffffffffa02825be>] btrfs_scratch_superblocks+0x4e/0x90 [btrfs]
[ 5375.719123]        [<ffffffffa0282665>] btrfs_destroy_dev_replace_tgtdev+0x65/0x150 [btrfs]
[ 5375.719150]        [<ffffffffa02c6c80>] btrfs_dev_replace_finishing+0x6b0/0x990 [btrfs]
[ 5375.719175]        [<ffffffffa02c729e>] btrfs_dev_replace_start+0x33e/0x540 [btrfs]
[ 5375.719199]        [<ffffffffa02c7f58>] btrfs_auto_replace_start+0xf8/0x140 [btrfs]
[ 5375.719222]        [<ffffffffa02464e6>] health_kthread+0x246/0x490 [btrfs]
[ 5375.719225]        [<ffffffff810a70df>] kthread+0xef/0x110
[ 5375.719229]        [<ffffffff81637d2f>] ret_from_fork+0x3f/0x70
[ 5375.719230]
[ 5375.719230] other info that might help us debug this:
[ 5375.719230]
[ 5375.719233] Chain exists of:
[ 5375.719233]   sb_writers --> namespace_sem --> &fs_devs->device_list_mutex
[ 5375.719233]
[ 5375.719234]  Possible unsafe locking scenario:
[ 5375.719234]
[ 5375.719234]        CPU0                    CPU1
[ 5375.719235]        ----                    ----
[ 5375.719236]   lock(&fs_devs->device_list_mutex);
[ 5375.719238]                                lock(namespace_sem);
[ 5375.719239]                                lock(&fs_devs->device_list_mutex);
[ 5375.719241]   lock(sb_writers);
[ 5375.719241]
[ 5375.719241]  *** DEADLOCK ***
[ 5375.719241]
[ 5375.719243] 4 locks held by btrfs-health/4662:
[ 5375.719266]  #0:  (&fs_info->health_mutex){+.+.+.}, at: [<ffffffffa0246303>] health_kthread+0x63/0x490 [btrfs]
[ 5375.719293]  #1:  (&fs_info->dev_replace.lock_finishing_cancel_unmount){+.+.+.}, at: [<ffffffffa02c6611>] btrfs_dev_replace_finishing+0x41/0x990 [btrfs]
[ 5375.719319]  #2:  (uuid_mutex){+.+.+.}, at: [<ffffffffa0282620>] btrfs_destroy_dev_replace_tgtdev+0x20/0x150 [btrfs]
[ 5375.719343]  #3:  (&fs_devs->device_list_mutex){+.+.+.}, at: [<ffffffffa028263c>] btrfs_destroy_dev_replace_tgtdev+0x3c/0x150 [btrfs]
[ 5375.719343]
[ 5375.719343] stack backtrace:
[ 5375.719347] CPU: 2 PID: 4662 Comm: btrfs-health Not tainted 4.4.5-scst31x-debug-11+ #40
[ 5375.719348] Hardware name: Supermicro SYS-6018R-WTRT/X10DRW-iT, BIOS 1.0c 01/07/2015
[ 5375.719352]  0000000000000000 ffff880856f73880 ffffffff813529e3 ffffffff826182a0
[ 5375.719354]  ffffffff8260c090 ffff880856f738c0 ffffffff810d667c ffff880856f73930
[ 5375.719357]  ffff880861f32b40 ffff880861f32b68 0000000000000003 0000000000000004
[ 5375.719357] Call Trace:
[ 5375.719363]  [<ffffffff813529e3>] dump_stack+0x85/0xc2
[ 5375.719366]  [<ffffffff810d667c>] print_circular_bug+0x1ec/0x260
[ 5375.719369]  [<ffffffff810d97ca>] __lock_acquire+0x17ba/0x1ae0
[ 5375.719373]  [<ffffffff810f606d>] ? debug_lockdep_rcu_enabled+0x1d/0x20
[ 5375.719376]  [<ffffffff810da4be>] lock_acquire+0xce/0x1e0
[ 5375.719378]  [<ffffffff812214f7>] ? __sb_start_write+0xb7/0xf0
[ 5375.719383]  [<ffffffff810d3bef>] percpu_down_read+0x4f/0xa0
[ 5375.719385]  [<ffffffff812214f7>] ? __sb_start_write+0xb7/0xf0
[ 5375.719387]  [<ffffffff812214f7>] __sb_start_write+0xb7/0xf0
[ 5375.719389]  [<ffffffff81242eb4>] mnt_want_write+0x24/0x50
[ 5375.719393]  [<ffffffff8122ded2>] path_openat+0xd32/0x1360
[ 5375.719415]  [<ffffffffa02462a0>] ? btrfs_congested_fn+0x180/0x180 [btrfs]
[ 5375.719418]  [<ffffffff810f606d>] ? debug_lockdep_rcu_enabled+0x1d/0x20
[ 5375.719420]  [<ffffffff8122f86e>] do_filp_open+0x7e/0xe0
[ 5375.719423]  [<ffffffff810f615d>] ? rcu_read_lock_sched_held+0x6d/0x80
[ 5375.719426]  [<ffffffff81201a9b>] ? kmem_cache_alloc+0x26b/0x5d0
[ 5375.719430]  [<ffffffff8122e7d4>] ? getname_kernel+0x34/0x120
[ 5375.719433]  [<ffffffff8121d8a4>] file_open_name+0xe4/0x130
[ 5375.719436]  [<ffffffff8121d923>] filp_open+0x33/0x60
[ 5375.719462]  [<ffffffffa02776a6>] update_dev_time+0x16/0x40 [btrfs]
[ 5375.719485]  [<ffffffffa02825be>] btrfs_scratch_superblocks+0x4e/0x90 [btrfs]
[ 5375.719506]  [<ffffffffa0282665>] btrfs_destroy_dev_replace_tgtdev+0x65/0x150 [btrfs]
[ 5375.719530]  [<ffffffffa02c6c80>] btrfs_dev_replace_finishing+0x6b0/0x990 [btrfs]
[ 5375.719554]  [<ffffffffa02c6b23>] ? btrfs_dev_replace_finishing+0x553/0x990 [btrfs]
[ 5375.719576]  [<ffffffffa02c729e>] btrfs_dev_replace_start+0x33e/0x540 [btrfs]
[ 5375.719598]  [<ffffffffa02c7f58>] btrfs_auto_replace_start+0xf8/0x140 [btrfs]
[ 5375.719621]  [<ffffffffa02464e6>] health_kthread+0x246/0x490 [btrfs]
[ 5375.719641]  [<ffffffffa02463d8>] ? health_kthread+0x138/0x490 [btrfs]
[ 5375.719661]  [<ffffffffa02462a0>] ? btrfs_congested_fn+0x180/0x180 [btrfs]
[ 5375.719663]  [<ffffffff810a70df>] kthread+0xef/0x110
[ 5375.719666]  [<ffffffff810a6ff0>] ? kthread_create_on_node+0x200/0x200
[ 5375.719669]  [<ffffffff81637d2f>] ret_from_fork+0x3f/0x70
[ 5375.719672]  [<ffffffff810a6ff0>] ? kthread_create_on_node+0x200/0x200
[ 5375.719697] ------------[ cut here ]------------
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Reported-by: NYauhen Kharuzhy <yauhen.kharuzhy@zavadatar.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Currently, we don't allow the user to try and rebalance to a dup profile
on a multi-device filesystem.  In most cases, this is a perfectly sensible
restriction as raid1 uses the same amount of space and provides better
protection.

However, when reshaping a multi-device filesystem down to a single device
filesystem, this requires the user to convert metadata and system chunks
to single profile before deleting devices, and then convert again to dup,
which leaves a period of time where metadata integrity is reduced.

This patch removes the single-device-only restriction from converting to
dup profile to remove this potential data integrity reduction.
Signed-off-by: NAustin S. Hemmelgarn <ahferroin7@gmail.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Creates helper fucntion as needed by the device delete
and replace operations. Also now it checks if the next
device being assigned is an active device.
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Yauhen reported in the ML that s_bdev is null at mount, and
s_bdev gets updated to some device when missing device is
replaced, as because bdev is null for missing device, things
gets matched up. Fix this by checking if s_bdev is set. I
didn't want to completely remove updating s_bdev because
the future multi device support at vfs layer may need it.
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Reported-by: NYauhen Kharuzhy <yauhen.kharuzhy@zavadatar.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

For clarity how we are going to find the device, let's call it a device
specifier, devspec for short. Also rename the arguments that are a
leftover from previous function purpose.
Reviewed-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

We should avoid duplicating the device constraints, let's use the
btrfs_raid_array in btrfs_check_raid_min_devices.
Reviewed-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Reviewed-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Before this patch, btrfs_check_raid_min_devices would do an off-by-one
check of the constraints and not the miminmum check, as its name
suggests. This is not a problem if the only caller is device remove, but
would be confusing for others.

Add an argument with the exact number and let the caller(s) decide if
this needs any adjustments, like when device replace is running.
Reviewed-by: NAnand Jain <anand.jain@oracle.com>
Tested-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Underscores are for special functions, use the full prefix for better
stacktrace recognition.
Reviewed-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Optimize check for stale device to only be checked when there is device
added or changed. If there is no update to the device, there is no need
to call btrfs_free_stale_device().
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

This introduces new ioctl BTRFS_IOC_RM_DEV_V2, which uses enhanced struct
btrfs_ioctl_vol_args_v2 to carry devid as an user argument.

The patch won't delete the old ioctl interface and so kernel remains
backward compatible with user land progs.

Test case/script:
echo "0 $(blockdev --getsz /dev/sdf) linear /dev/sdf 0" | dmsetup create bad_disk
mkfs.btrfs -f -d raid1 -m raid1 /dev/sdd /dev/sde /dev/mapper/bad_disk
mount /dev/sdd /btrfs
dmsetup suspend bad_disk
echo "0 $(blockdev --getsz /dev/sdf) error /dev/sdf 0" | dmsetup load bad_disk
dmsetup resume bad_disk
echo "bad disk failed. now deleting/replacing"
btrfs dev del  3  /btrfs
echo $?
btrfs fi show /btrfs
umount /btrfs
btrfs-show-super /dev/sdd | egrep num_device
dmsetup remove bad_disk
wipefs -a /dev/sdf
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Reported-by: NMartin <m_btrfs@ml1.co.uk>
[ adjust messages, s/disk/device/ ]
Signed-off-by: NDavid Sterba <dsterba@suse.com>

With the previous patches now the btrfs_scratch_superblocks() is ready to
be used in btrfs_rm_device() so use it.
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
[ use GFP_KERNEL ]
Signed-off-by: NDavid Sterba <dsterba@suse.com>

The operation of device replace and device delete follows same steps upto
some depth with in btrfs kernel, however they don't share codes. This
enhancement will help replace and delete to share codes.
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

btrfs_rm_device() has a section of the code which can be replaced
btrfs_find_device_by_user_input()
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

The patch renames btrfs_dev_replace_find_srcdev() to
btrfs_find_device_by_user_input() and moves it to volumes.c, so that
delete device can use it.
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

__check_raid_min_device() which was pealed from btrfs_rm_device()
maintianed its original code to show the block move. This patch cleans up
__check_raid_min_device().
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

move a section of btrfs_rm_device() code to check for min number of the
devices into the function __check_raid_min_devices()
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

A part of code from btrfs_scan_one_device() is moved to a new function
btrfs_read_disk_super(), so that former function looks cleaner. (In this
process it also moves the code which ensures null terminating label). So
this creates easy opportunity to merge various duplicate codes on read
disk super. Earlier attempt to merge duplicate codes highlighted that
there were some issues for which there are duplicate codes (to read disk
super), however it was not clear what was the issue. So until we figure
that out, its better to keep them in a separate functions.
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
[ use GFP_KERNEL, PAGE_CACHE_ removal related fixups ]
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Now we force to create empty block group to keep data profile alive,
however, in the below example, we eventually get an empty block group
while we're trying to get more space for other types (metadata/system),

- Before,
block group "A": size=2G, used=1.2G
block group "B": size=2G, used=512M

- After "btrfs balance start -dusage=50 mount_point",
block group "A": size=2G, used=(1.2+0.5)G
block group "C": size=2G, used=0

Since there is no data in block group C, it won't be deleted
automatically and we have to get the unused 2G until the next mount.

Balance itself just moves data and doesn't remove data, so it's safe
to not create such a empty block group if we already have data
 allocated in other block groups.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

btrfs_std_error() handles errors, puts FS into readonly mode
(as of now). So its good idea to rename it to btrfs_handle_fs_error().
Signed-off-by: NAnand Jain <anand.jain@oracle.com>
Reviewed-by: NDavid Sterba <dsterba@suse.com>
[ edit changelog ]
Signed-off-by: NDavid Sterba <dsterba@suse.com>

PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time
ago with promise that one day it will be possible to implement page
cache with bigger chunks than PAGE_SIZE.

This promise never materialized.  And unlikely will.

We have many places where PAGE_CACHE_SIZE assumed to be equal to
PAGE_SIZE.  And it's constant source of confusion on whether
PAGE_CACHE_* or PAGE_* constant should be used in a particular case,
especially on the border between fs and mm.

Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much
breakage to be doable.

Let's stop pretending that pages in page cache are special.  They are
not.

The changes are pretty straight-forward:

 - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>;

 - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>;

 - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN};

 - page_cache_get() -> get_page();

 - page_cache_release() -> put_page();

This patch contains automated changes generated with coccinelle using
script below.  For some reason, coccinelle doesn't patch header files.
I've called spatch for them manually.

The only adjustment after coccinelle is revert of changes to
PAGE_CAHCE_ALIGN definition: we are going to drop it later.

There are few places in the code where coccinelle didn't reach.  I'll
fix them manually in a separate patch.  Comments and documentation also
will be addressed with the separate patch.

virtual patch

@@
expression E;
@@
- E << (PAGE_CACHE_SHIFT - PAGE_SHIFT)
+ E

@@
expression E;
@@
- E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT)
+ E

@@
@@
- PAGE_CACHE_SHIFT
+ PAGE_SHIFT

@@
@@
- PAGE_CACHE_SIZE
+ PAGE_SIZE

@@
@@
- PAGE_CACHE_MASK
+ PAGE_MASK

@@
expression E;
@@
- PAGE_CACHE_ALIGN(E)
+ PAGE_ALIGN(E)

@@
expression E;
@@
- page_cache_get(E)
+ get_page(E)

@@
expression E;
@@
- page_cache_release(E)
+ put_page(E)
Signed-off-by: NKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: NMichal Hocko <mhocko@suse.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NAdam Buchbinder <adam.buchbinder@gmail.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Xfstests btrfs/011 complains about a deadlock warning,

[ 1226.649039] =========================================================
[ 1226.649039] [ INFO: possible irq lock inversion dependency detected ]
[ 1226.649039] 4.1.0+ #270 Not tainted
[ 1226.649039] ---------------------------------------------------------
[ 1226.652955] kswapd0/46 just changed the state of lock:
[ 1226.652955]  (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0
[ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past:
[ 1226.652955]  (&fs_info->dev_replace.lock){+.+.+.}

and interrupts could create inverse lock ordering between them.

[ 1226.652955]
other info that might help us debug this:
[ 1226.652955] Chain exists of:
  &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock

[ 1226.652955]  Possible interrupt unsafe locking scenario:

[ 1226.652955]        CPU0                    CPU1
[ 1226.652955]        ----                    ----
[ 1226.652955]   lock(&fs_info->dev_replace.lock);
[ 1226.652955]                                local_irq_disable();
[ 1226.652955]                                lock(&delayed_node->mutex);
[ 1226.652955]                                lock(&found->groups_sem);
[ 1226.652955]   <Interrupt>
[ 1226.652955]     lock(&delayed_node->mutex);
[ 1226.652955]
 *** DEADLOCK ***

Commit 084b6e7c ("btrfs: Fix a lockdep warning when running xfstest.") tried
to fix a similar one that has the exactly same warning, but with that, we still
run to this.

The above lock chain comes from
btrfs_commit_transaction
  ->btrfs_run_delayed_items
    ...
    ->__btrfs_update_delayed_inode
      ...
      ->__btrfs_cow_block
         ...
         ->find_free_extent
            ->cache_block_group
              ->load_free_space_cache
                ->btrfs_readpages
                  ->submit_one_bio
                    ...
                    ->__btrfs_map_block
                      ->btrfs_dev_replace_lock

However, with high memory pressure, tasks which hold dev_replace.lock can
be interrupted by kswapd and then kswapd is intended to release memory occupied
by superblock, inodes and dentries, where we may call evict_inode, and it comes
to

[ 1226.652955]  [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0
[ 1226.652955]  [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30
[ 1226.652955]  [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700

delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads
to a ABBA deadlock.

To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but
things are simpler here since we only needs read's spinlock to blocking lock.

With this, btrfs/011 no more produces warnings in dmesg.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Signed-off-by: NDavid Sterba <dsterba@suse.com>

No visible change.
Signed-off-by: NDavid Sterba <dsterba@suse.com>

We can safely use GFP_KERNEL in the functions called from the ioctl
handlers.
Signed-off-by: NDavid Sterba <dsterba@suse.com>

This reverts commit 14e46e04.

This ends up doing sysfs operations from deep in balance (where we
should be GFP_NOFS) and under heavy balance load, we're making races
against sysfs internals.

Revert it for now while we figure things out.
Signed-off-by: NChris Mason <clm@fb.com>

The files under /sys/fs/UUID/features get out of sync with the actual
incompat bits set for the filesystem if they change after mount (eg. the
LZO compression).

Synchronize the feature bits with the sysfs files representing them
right after we set/clear them.
Signed-off-by: NDavid Sterba <dsterba@suse.com>

Old code used bbio->raid_map to determine whether in raid56
write/recover operation, because we didn't't have bbio->map_type.

Now we have direct way for this condition, rid of using
the function-relative data, and make the code more readable.
Signed-off-by: NZhao Lei <zhaolei@cn.fujitsu.com>
Signed-off-by: NChris Mason <clm@fb.com>