IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
No longer used as of commit 5846a3c268 ("btrfs: qgroup: Fix a race in
delayed_ref which leads to abort trans").
Signed-off-by: Filipe Manana <fdmanana@suse.com>
With commit 56f23fdbb6 ("Btrfs: fix file/data loss caused by fsync after
rename and new inode") we got simple fix for a functional issue when the
following sequence of actions is done:
at transaction N
create file A at directory D
at transaction N + M (where M >= 1)
move/rename existing file A from directory D to directory E
create a new file named A at directory D
fsync the new file
power fail
The solution was to simply detect such scenario and fallback to a full
transaction commit when we detect it. However this turned out to had a
significant impact on throughput (and a bit on latency too) for benchmarks
using the dbench tool, which simulates real workloads from smbd (Samba)
servers. For example on a test vm (with a debug kernel):
Unpatched:
Throughput 19.1572 MB/sec 32 clients 32 procs max_latency=1005.229 ms
Patched:
Throughput 23.7015 MB/sec 32 clients 32 procs max_latency=809.206 ms
The patched results (this patch is applied) are similar to the results of
a kernel with the commit 56f23fdbb6 ("Btrfs: fix file/data loss caused
by fsync after rename and new inode") reverted.
This change avoids the fallback to a transaction commit and instead makes
sure all the names of the conflicting inode (the one that had a name in a
past transaction that matches the name of the new file in the same parent
directory) are logged so that at log replay time we don't lose neither the
new file nor the old file, and the old file gets the name it was renamed
to.
This also ends up avoiding a full transaction commit for a similar case
that involves an unlink instead of a rename of the old file:
at transaction N
create file A at directory D
at transaction N + M (where M >= 1)
remove file A
create a new file named A at directory D
fsync the new file
power fail
Signed-off-by: Filipe Manana <fdmanana@suse.com>
When we attempt to read an inode from disk, we end up always returning an
-ESTALE error to the caller regardless of the actual failure reason, which
can be an out of memory problem (when allocating a path), some error found
when reading from the fs/subvolume btree (like a genuine IO error) or the
inode does not exists. So lets start returning the real error code to the
callers so that they don't treat all -ESTALE errors as meaning that the
inode does not exists (such as during orphan cleanup). This will also be
needed for a subsequent patch in the same series dealing with a special
fsync case.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
When doing an incremental send, if we find a new/modified/deleted extent,
reference or xattr without having previously processed the corresponding
inode item we end up exexuting a BUG_ON(). This is because whenever an
extent, xattr or reference is added, modified or deleted, we always expect
to have the corresponding inode item updated. However there are situations
where this will not happen due to transient -ENOMEM or -ENOSPC errors when
doing delayed inode updates.
For example, when punching holes we can succeed in deleting and modifying
(shrinking) extents but later fail to do the delayed inode update. So after
such failure we close our transaction handle and right after a snapshot of
the fs/subvol tree can be made and used later for a send operation. The
same thing can happen during truncate, link, unlink, and xattr related
operations.
So instead of executing a BUG_ON, make send return an -EIO error and print
an informative error message do dmesg/syslog.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
The caller of send_utimes() is supposed to be sure that the inode number
it passes to this function does actually exists in the send snapshot.
However due to logic/algorithm bugs (such as the one fixed by the patch
titled "Btrfs: send, fix invalid leaf accesses due to incorrect utimes
operations"), this might not be the case and when that happens it makes
send_utimes() access use an unrelated leaf item as the target inode item
or access beyond a leaf's boundaries (when the leaf is full and
path->slots[0] matches the number of items in the leaf).
So if the call to btrfs_search_slot() done by send_utimes() does not find
the inode item, just make sure send_utimes() returns -ENOENT and does not
silently accesses unrelated leaf items or does invalid leaf accesses, also
allowing us to easialy and deterministically catch such algorithmic/logic
bugs.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
During an incremental send, if we have delayed rename operations for inodes
that were children of directories which were removed in the send snapshot,
we can end up accessing incorrect items in a leaf or accessing beyond the
last item of the leaf due to issuing utimes operations for the removed
inodes. Consider the following example:
Parent snapshot:
. (ino 256)
|--- a/ (ino 257)
| |--- c/ (ino 262)
|
|--- b/ (ino 258)
| |--- d/ (ino 263)
|
|--- del/ (ino 261)
|--- x/ (ino 259)
|--- y/ (ino 260)
Send snapshot:
. (ino 256)
|--- a/ (ino 257)
|
|--- b/ (ino 258)
|
|--- c/ (ino 262)
| |--- y/ (ino 260)
|
|--- d/ (ino 263)
|--- x/ (ino 259)
1) When processing inodes 259 and 260, we end up delaying their rename
operations because their parents, inodes 263 and 262 respectively, were
not yet processed and therefore not yet renamed;
2) When processing inode 262, its rename operation is issued and right
after the rename operation for inode 260 is issued. However right after
issuing the rename operation for inode 260, at send.c:apply_dir_move(),
we issue utimes operations for all current and past parents of inode
260. This means we try to send a utimes operation for its old parent,
inode 261 (deleted in the send snapshot), which does not cause any
immediate and deterministic failure, because when the target inode is
not found in the send snapshot, the send.c:send_utimes() function
ignores it and uses the leaf region pointed to by path->slots[0],
which can be any unrelated item (belonging to other inode) or it can
be a region outside the leaf boundaries, if the leaf is full and
path->slots[0] matches the number of items in the leaf. So we end
up either successfully sending a utimes operation, which is fine
and irrelevant because the old parent (inode 261) will end up being
deleted later, or we end up doing an invalid memory access tha
crashes the kernel.
So fix this by making apply_dir_move() issue utimes operations only for
parents that still exist in the send snapshot. In a separate patch we
will make send_utimes() return an error (-ENOENT) if the given inode
does not exists in the send snapshot.
Signed-off-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
[Rewrote change log to be more detailed and better organized]
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Under certain situations, when doing an incremental send, we can end up
not freeing orphan_dir_info structures as soon as they are no longer
needed. Instead we end up freeing them only after finishing the send
stream, which causes a warning to be emitted:
[282735.229200] ------------[ cut here ]------------
[282735.229968] WARNING: CPU: 9 PID: 10588 at fs/btrfs/send.c:6298 btrfs_ioctl_send+0xe2f/0xe51 [btrfs]
[282735.231282] Modules linked in: btrfs crc32c_generic xor raid6_pq acpi_cpufreq tpm_tis ppdev tpm parport_pc psmouse parport sg pcspkr i2c_piix4 i2c_core evdev processor serio_raw button loop autofs4 ext4 crc16 jbd2 mbcache sr_mod cdrom sd_mod ata_generic virtio_scsi ata_piix libata virtio_pci virtio_ring virtio e1000 scsi_mod floppy [last unloaded: btrfs]
[282735.237130] CPU: 9 PID: 10588 Comm: btrfs Tainted: G W 4.6.0-rc7-btrfs-next-31+ #1
[282735.239309] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014
[282735.240160] 0000000000000000 ffff880224273ca8 ffffffff8126b42c 0000000000000000
[282735.240160] 0000000000000000 ffff880224273ce8 ffffffff81052b14 0000189a24273ac8
[282735.240160] ffff8802210c9800 0000000000000000 0000000000000001 0000000000000000
[282735.240160] Call Trace:
[282735.240160] [<ffffffff8126b42c>] dump_stack+0x67/0x90
[282735.240160] [<ffffffff81052b14>] __warn+0xc2/0xdd
[282735.240160] [<ffffffff81052beb>] warn_slowpath_null+0x1d/0x1f
[282735.240160] [<ffffffffa03c99d5>] btrfs_ioctl_send+0xe2f/0xe51 [btrfs]
[282735.240160] [<ffffffffa0398358>] btrfs_ioctl+0x14f/0x1f81 [btrfs]
[282735.240160] [<ffffffff8108e456>] ? arch_local_irq_save+0x9/0xc
[282735.240160] [<ffffffff8118da05>] vfs_ioctl+0x18/0x34
[282735.240160] [<ffffffff8118e00c>] do_vfs_ioctl+0x550/0x5be
[282735.240160] [<ffffffff81196f0c>] ? __fget+0x6b/0x77
[282735.240160] [<ffffffff81196fa1>] ? __fget_light+0x62/0x71
[282735.240160] [<ffffffff8118e0d1>] SyS_ioctl+0x57/0x79
[282735.240160] [<ffffffff8149e025>] entry_SYSCALL_64_fastpath+0x18/0xa8
[282735.240160] [<ffffffff81100c6b>] ? time_hardirqs_off+0x9/0x14
[282735.240160] [<ffffffff8108e87d>] ? trace_hardirqs_off_caller+0x1f/0xaa
[282735.256343] ---[ end trace a4539270c8056f93 ]---
Consider the following example:
Parent snapshot:
. (ino 256)
|--- a/ (ino 257)
| |--- c/ (ino 260)
|
|--- del/ (ino 259)
|--- tmp/ (ino 258)
|--- x/ (ino 261)
|--- y/ (ino 262)
Send snapshot:
. (ino 256)
|--- a/ (ino 257)
| |--- x/ (ino 261)
| |--- y/ (ino 262)
|
|--- c/ (ino 260)
|--- tmp/ (ino 258)
1) When processing inode 258, we end up delaying its rename operation
because it has an ancestor (in the send snapshot) that has a higher
inode number (inode 260) which was also renamed in the send snapshot,
therefore we delay the rename of inode 258 so that it happens after
inode 260 is renamed;
2) When processing inode 259, we end up delaying its deletion (rmdir
operation) because it has a child inode (258) that has its rename
operation delayed. At this point we allocate an orphan_dir_info
structure and tag inode 258 so that we later attempt to see if we
can delete (rmdir) inode 259 once inode 258 is renamed;
3) When we process inode 260, after renaming it we finally do the rename
operation for inode 258. Once we issue the rename operation for inode
258 we notice that this inode was tagged so that we attempt to see
if at this point we can delete (rmdir) inode 259. But at this point
we can not still delete inode 259 because it has 2 children, inodes
261 and 262, that were not yet processed and therefore not yet
moved (renamed) away from inode 259. We end up not freeing the
orphan_dir_info structure allocated in step 2;
4) We process inodes 261 and 262, and once we move/rename inode 262
we issue the rmdir operation for inode 260;
5) We finish the send stream and notice that red black tree that
contains orphan_dir_info structures is not empty, so we emit
a warning and then free any orphan_dir_structures left.
So fix this by freeing an orphan_dir_info structure once we try to
apply a pending rename operation if we can not delete yet the tagged
directory.
A test case for fstests follows soon.
Signed-off-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
[Modified changelog to be more detailed and easier to understand]
Under certain situations, an incremental send operation can contain
a rmdir operation that will make the receiving end fail when attempting
to execute it, because the target directory is not yet empty.
Consider the following example:
Parent snapshot:
. (ino 256)
|--- a/ (ino 257)
| |--- c/ (ino 260)
|
|--- del/ (ino 259)
|--- tmp/ (ino 258)
|--- x/ (ino 261)
Send snapshot:
. (ino 256)
|--- a/ (ino 257)
| |--- x/ (ino 261)
|
|--- c/ (ino 260)
|--- tmp/ (ino 258)
1) When processing inode 258, we delay its rename operation because inode
260 is its new parent in the send snapshot and it was not yet renamed
(since 260 > 258, that is, beyond the current progress);
2) When processing inode 259, we realize we can not yet send an rmdir
operation (against inode 259) because inode 258 was still not yet
renamed/moved away from inode 259. Therefore we update data structures
so that after inode 258 is renamed, we try again to see if we can
finally send an rmdir operation for inode 259;
3) When we process inode 260, we send a rename operation for it followed
by a rename operation for inode 258. Once we send the rename operation
for inode 258 we then check if we can finally issue an rmdir for its
previous parent, inode 259, by calling the can_rmdir() function with
a value of sctx->cur_ino + 1 (260 + 1 = 261) for its "progress"
argument. This makes can_rmdir() return true (value 1) because even
though there's still a child inode of inode 259 that was not yet
renamed/moved, which is inode 261, the given value of progress (261)
is not lower then 261 (that is, not lower than the inode number of
some child of inode 259). So we end up sending a rmdir operation for
inode 259 before its child inode 261 is processed and renamed.
So fix this by passing the correct progress value to the call to
can_rmdir() from within apply_dir_move() (where we issue delayed rename
operations), which should match stcx->cur_ino (the number of the inode
currently being processed) and not sctx->cur_ino + 1.
A test case for fstests follows soon.
Signed-off-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
[Rewrote change log to be more detailed, clear and well formatted]
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Example scenario:
Parent snapshot:
. (ino 277)
|---- tmp/ (ino 278)
|---- pre/ (ino 280)
| |---- wait_dir/ (ino 281)
|
|---- desc/ (ino 282)
|---- ance/ (ino 283)
| |---- below_ance/ (ino 279)
|
|---- other_dir/ (ino 284)
Send snapshot:
. (ino 277)
|---- tmp/ (ino 278)
|---- other_dir/ (ino 284)
|---- below_ance/ (ino 279)
| |---- pre/ (ino 280)
|
|---- wait_dir/ (ino 281)
|---- desc/ (ino 282)
|---- ance/ (ino 283)
While computing the send stream the following steps happen:
1) While processing inode 279 we end up delaying its rename operation
because its new parent in the send snapshot, inode 284, was not
yet processed and therefore not yet renamed;
2) Later when processing inode 280 we end up renaming it immediately to
"ance/below_once/pre" and not delay its rename operation because its
new parent (inode 279 in the send snapshot) has its rename operation
delayed and inode 280 is not an encestor of inode 279 (its parent in
the send snapshot) in the parent snapshot;
3) When processing inode 281 we end up delaying its rename operation
because its new parent in the send snapshot, inode 284, was not yet
processed and therefore not yet renamed;
4) When processing inode 282 we do not delay its rename operation because
its parent in the send snapshot, inode 281, already has its own rename
operation delayed and our current inode (282) is not an ancestor of
inode 281 in the parent snapshot. Therefore inode 282 is renamed to
"ance/below_ance/pre/wait_dir";
5) When processing inode 283 we realize that we can rename it because one
of its ancestors in the send snapshot, inode 281, has its rename
operation delayed and inode 283 is not an ancestor of inode 281 in the
parent snapshot. So a rename operation to rename inode 283 to
"ance/below_ance/pre/wait_dir/desc/ance" is issued. This path is
invalid due to a missing path building loop that was undetected by
the incremental send implementation, as inode 283 ends up getting
included twice in the path (once with its path in the parent snapshot).
Therefore its rename operation must wait before the ancestor inode 284
is renamed.
Fix this by not terminating the rename dependency checks when we find an
ancestor, in the send snapshot, that has its rename operation delayed. So
that we continue doing the same checks if the current inode is not an
ancestor, in the parent snapshot, of an ancestor in the send snapshot we
are processing in the loop.
The problem and reproducer were reported by Robbie Ko, as part of a patch
titled "Btrfs: incremental send, avoid ancestor rename to descendant".
However the fix was unnecessarily complicated and can be addressed with
much less code and effort.
Reported-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
The function path_loop() can return a negative integer, signaling an
error, 0 if there's no path loop and 1 if there's a path loop. We were
treating any non zero values as meaning that a path loop exists. Fix
this by explicitly checking for errors and gracefully return them to
user space.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
When doing an incremental send we can end up not moving directories that
have the same name. This happens when the same parent directory has
different child directories with the same name in the parent and send
snapshots.
For example, consider the following scenario:
Parent snapshot:
. (ino 256)
|---- d/ (ino 257)
| |--- p1/ (ino 258)
|
|---- p1/ (ino 259)
Send snapshot:
. (ino 256)
|--- d/ (ino 257)
|--- p1/ (ino 259)
|--- p1/ (ino 258)
The directory named "d" (inode 257) has in both snapshots an entry with
the name "p1" but it refers to different inodes in both snapshots (inode
258 in the parent snapshot and inode 259 in the send snapshot). When
attempting to move inode 258, the operation is delayed because its new
parent, inode 259, was not yet moved/renamed (as the stream is currently
processing inode 258). Then when processing inode 259, we also end up
delaying its move/rename operation so that it happens after inode 258 is
moved/renamed. This decision to delay the move/rename rename operation
of inode 259 is due to the fact that the new parent inode (257) still
has inode 258 as its child, which has the same name has inode 259. So
we end up with inode 258 move/rename operation waiting for inode's 259
move/rename operation, which in turn it waiting for inode's 258
move/rename. This results in ending the send stream without issuing
move/rename operations for inodes 258 and 259 and generating the
following warnings in syslog/dmesg:
[148402.979747] ------------[ cut here ]------------
[148402.980588] WARNING: CPU: 14 PID: 4117 at fs/btrfs/send.c:6177 btrfs_ioctl_send+0xe03/0xe51 [btrfs]
[148402.981928] Modules linked in: btrfs crc32c_generic xor raid6_pq acpi_cpufreq tpm_tis ppdev tpm parport_pc psmouse parport sg pcspkr i2c_piix4 i2c_core evdev processor serio_raw button loop autofs4 ext4 crc16 jbd2 mbcache sr_mod cdrom sd_mod ata_generic virtio_scsi ata_piix libata virtio_pci virtio_ring virtio e1000 scsi_mod floppy [last unloaded: btrfs]
[148402.986999] CPU: 14 PID: 4117 Comm: btrfs Tainted: G W 4.6.0-rc7-btrfs-next-31+ #1
[148402.988136] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014
[148402.988136] 0000000000000000 ffff88022139fca8 ffffffff8126b42c 0000000000000000
[148402.988136] 0000000000000000 ffff88022139fce8 ffffffff81052b14 000018212139fac8
[148402.988136] ffff88022b0db400 0000000000000000 0000000000000001 0000000000000000
[148402.988136] Call Trace:
[148402.988136] [<ffffffff8126b42c>] dump_stack+0x67/0x90
[148402.988136] [<ffffffff81052b14>] __warn+0xc2/0xdd
[148402.988136] [<ffffffff81052beb>] warn_slowpath_null+0x1d/0x1f
[148402.988136] [<ffffffffa04bc831>] btrfs_ioctl_send+0xe03/0xe51 [btrfs]
[148402.988136] [<ffffffffa048b358>] btrfs_ioctl+0x14f/0x1f81 [btrfs]
[148402.988136] [<ffffffff8108e456>] ? arch_local_irq_save+0x9/0xc
[148402.988136] [<ffffffff8108eb51>] ? __lock_is_held+0x3c/0x57
[148402.988136] [<ffffffff8118da05>] vfs_ioctl+0x18/0x34
[148402.988136] [<ffffffff8118e00c>] do_vfs_ioctl+0x550/0x5be
[148402.988136] [<ffffffff81196f0c>] ? __fget+0x6b/0x77
[148402.988136] [<ffffffff81196fa1>] ? __fget_light+0x62/0x71
[148402.988136] [<ffffffff8118e0d1>] SyS_ioctl+0x57/0x79
[148402.988136] [<ffffffff8149e025>] entry_SYSCALL_64_fastpath+0x18/0xa8
[148402.988136] [<ffffffff8108e89d>] ? trace_hardirqs_off_caller+0x3f/0xaa
[148403.011373] ---[ end trace a4539270c8056f8b ]---
[148403.012296] ------------[ cut here ]------------
[148403.013071] WARNING: CPU: 14 PID: 4117 at fs/btrfs/send.c:6194 btrfs_ioctl_send+0xe19/0xe51 [btrfs]
[148403.014447] Modules linked in: btrfs crc32c_generic xor raid6_pq acpi_cpufreq tpm_tis ppdev tpm parport_pc psmouse parport sg pcspkr i2c_piix4 i2c_core evdev processor serio_raw button loop autofs4 ext4 crc16 jbd2 mbcache sr_mod cdrom sd_mod ata_generic virtio_scsi ata_piix libata virtio_pci virtio_ring virtio e1000 scsi_mod floppy [last unloaded: btrfs]
[148403.019708] CPU: 14 PID: 4117 Comm: btrfs Tainted: G W 4.6.0-rc7-btrfs-next-31+ #1
[148403.020104] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014
[148403.020104] 0000000000000000 ffff88022139fca8 ffffffff8126b42c 0000000000000000
[148403.020104] 0000000000000000 ffff88022139fce8 ffffffff81052b14 000018322139fac8
[148403.020104] ffff88022b0db400 0000000000000000 0000000000000001 0000000000000000
[148403.020104] Call Trace:
[148403.020104] [<ffffffff8126b42c>] dump_stack+0x67/0x90
[148403.020104] [<ffffffff81052b14>] __warn+0xc2/0xdd
[148403.020104] [<ffffffff81052beb>] warn_slowpath_null+0x1d/0x1f
[148403.020104] [<ffffffffa04bc847>] btrfs_ioctl_send+0xe19/0xe51 [btrfs]
[148403.020104] [<ffffffffa048b358>] btrfs_ioctl+0x14f/0x1f81 [btrfs]
[148403.020104] [<ffffffff8108e456>] ? arch_local_irq_save+0x9/0xc
[148403.020104] [<ffffffff8108eb51>] ? __lock_is_held+0x3c/0x57
[148403.020104] [<ffffffff8118da05>] vfs_ioctl+0x18/0x34
[148403.020104] [<ffffffff8118e00c>] do_vfs_ioctl+0x550/0x5be
[148403.020104] [<ffffffff81196f0c>] ? __fget+0x6b/0x77
[148403.020104] [<ffffffff81196fa1>] ? __fget_light+0x62/0x71
[148403.020104] [<ffffffff8118e0d1>] SyS_ioctl+0x57/0x79
[148403.020104] [<ffffffff8149e025>] entry_SYSCALL_64_fastpath+0x18/0xa8
[148403.020104] [<ffffffff8108e89d>] ? trace_hardirqs_off_caller+0x3f/0xaa
[148403.038981] ---[ end trace a4539270c8056f8c ]---
There's another issue caused by similar (but more complex) changes in the
directory hierarchy that makes move/rename operations fail, described with
the following example:
Parent snapshot:
.
|---- a/ (ino 262)
| |---- c/ (ino 268)
|
|---- d/ (ino 263)
|---- ance/ (ino 267)
|---- e/ (ino 264)
|---- f/ (ino 265)
|---- ance/ (ino 266)
Send snapshot:
.
|---- a/ (ino 262)
|---- c/ (ino 268)
| |---- ance/ (ino 267)
|
|---- d/ (ino 263)
| |---- ance/ (ino 266)
|
|---- f/ (ino 265)
|---- e/ (ino 264)
When the inode 265 is processed, the path for inode 267 is computed, which
at that time corresponds to "d/ance", and it's stored in the names cache.
Later on when processing inode 266, we end up orphanizing (renaming to a
name matching the pattern o<ino>-<gen>-<seq>) inode 267 because it has
the same name as inode 266 and it's currently a child of the new parent
directory (inode 263) for inode 266. After the orphanization and while we
are still processing inode 266, a rename operation for inode 266 is
generated. However the source path for that rename operation is incorrect
because it ends up using the old, pre-orphanization, name of inode 267.
The no longer valid name for inode 267 was previously cached when
processing inode 265 and it remains usable and considered valid until
the inode currently being processed has a number greater than 267.
This resulted in the receiving side failing with the following error:
ERROR: rename d/ance/ance -> d/ance failed: No such file or directory
So fix these issues by detecting such circular dependencies for rename
operations and by clearing the cached name of an inode once the inode
is orphanized.
A test case for fstests will follow soon.
Signed-off-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
[Rewrote change log to be more detailed and organized, and improved
comments]
Signed-off-by: Filipe Manana <fdmanana@suse.com>
When we start an fsync we start ordered extents for all delalloc ranges.
However before attempting to log the inode, we only wait for those ordered
extents if we are not doing a full sync (bit BTRFS_INODE_NEEDS_FULL_SYNC
is set in the inode's flags). This means that if an ordered extent
completes with an IO error before we check if we can skip logging the
inode, we will not catch and report the IO error to user space. This is
because on an IO error, when the ordered extent completes we do not
update the inode, so if the inode was not previously updated by the
current transaction we end up not logging it through calls to fsync and
therefore not check its mapping flags for the presence of IO errors.
Fix this by checking for errors in the flags of the inode's mapping when
we notice we can skip logging the inode.
This caused sporadic failures in the test generic/331 (which explicitly
tests for IO errors during an fsync call).
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Commit 56244ef151 was almost but not quite enough to fix the
reservation math after btrfs_copy_from_user returned partial copies.
Some users are still seeing warnings in btrfs_destroy_inode, and with a
long enough test run I'm able to trigger them as well.
This patch fixes the accounting math again, bringing it much closer to
the way it was before the sectorsize conversion Chandan did. The
problem is accounting for the offset into the page/sector when we do a
partial copy. This one just uses the dirty_sectors variable which
should already be updated properly.
Signed-off-by: Chris Mason <clm@fb.com>
cc: stable@vger.kernel.org # v4.6+
The new enospc code makes it possible to deadlock if we don't use
FLUSH_LIMIT during reservations inside a transaction. This enforces
the correct flush type to avoid both deadlocks and assertions
Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
We used to allow you to set FLUSH_ALL and then just wouldn't do things like
commit transactions or wait on ordered extents if we noticed you were in a
transaction. However now that all the flushing for FLUSH_ALL is asynchronous
we've lost the ability to tell, and we could end up deadlocking. So instead use
FLUSH_LIMIT in reserve_metadata_bytes in relocation and then return -EAGAIN if
we error out to preserve the previous behavior. I've also added an ASSERT() to
catch anybody else who tries to do this. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Since we set the reloc control before we've reserved our space for relocation we
could race with a root being dirtied and not actually have space to do our init
reloc root. So once we've allocated it and set it up go ahead and make our
reservation before setting the relocate control, that way anybody who tries to
do the reloc root init has space to use. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is the case all the time anyway except for relocation which could be doing
a reloc root for a non ref counted root, in which case we'd end up with some
random block rsv rather than the one we have our reservation in. If there isn't
enough space in the block rsv we are trying to steal from we'll BUG() because we
expect there to be space for the orphan to make its reservation. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Traditionally we've calculated the global block rsv by guessing how much of the
metadata used amount was the extent tree, and then taking the data size and
figuring out how large the csum tree would have to be to hold that much data.
This is imprecise and falls down on MIXED file systems as we can't trust the
data used amount. This resulted in failures for xfstests generic/333 because it
creates lots of clones, which explodes out the extent tree. Our global reserve
calculations were woefully inaccurate in this case which meant we got into a
situation where we did not have enough reserved to do our work.
We know we only use the global block rsv for the extent, csum, and root trees,
so just get the bytes used for these trees and use that as the basis of our
global reserve. Since these are not reference counted trees the bytes_used
value will be accurate. This fixed the transaction aborts seen with
generic/333. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Instead of doing fs_info->fs_root in need_async_flush, which may not be set
during recovery when mounting, just pass the root itself in, which makes more
sense as thats what btrfs_calc_reclaim_metadata_size takes.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reported-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We do this check when we start the async reclaimer thread, might as well check
before we kick it off to save us some cycles. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We were doing trace_btrfs_release_reserved_extent() in pin_down_extent which
isn't quite right because we will go through and free that extent later when we
unpin, so it messes up apps that are accounting for the reservation space. We
were also unconditionally doing it in __btrfs_free_reserved_extent(), when we
only actually free the reservation instead of pinning the extent. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We want to track when we're triggering flushing from our reservation code and
what flushing is being done when we start flushing. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We can sometimes drop the reservation we had for our inode, so we need to remove
that amount from to_reserve so that our tracepoint reports a valid amount of
space.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Pinned extents are an important metric to keep track of for enospc.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Our enospc flushing sucks. It is born from a time where we were early
enospc'ing constantly because multiple threads would race in for the same
reservation and randomly starve other ones out. So I came up with this solution
to block any other reservations from happening while one guy tried to flush
stuff to satisfy his reservation. This gives us pretty good correctness, but
completely crap latency.
The solution I've come up with is ticketed reservations. Basically we try to
make our reservation, and if we can't we put a ticket on a list in order and
kick off an async flusher thread. This async flusher thread does the same old
flushing we always did, just asynchronously. As space is freed and added back
to the space_info it checks and sees if we have any tickets that need
satisfying, and adds space to the tickets and wakes up anything we've satisfied.
Once the flusher thread stops making progress it wakes up all the current
tickets and tells them to take a hike.
There is a priority list for things that can't flush, since the async flusher
could do anything we need to avoid deadlocks. These guys get priority for
having their reservation made, and will still do manual flushing themselves in
case the async flusher isn't running.
This patch gives us significantly better latencies. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
I'm writing a tool to visualize the enospc system inside btrfs, I need this
tracepoint in order to keep track of the block groups in the system. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
These were hidden behind enospc_debug, which isn't helpful as they indicate
actual bugs, unlike the rest of the enospc_debug stuff which is really debug
information. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We reserve space for the inode update when we first reserve space for writing to
a file. However there are lots of ways that we can use this reservation and not
have it for subsequent ordered extents. Previously we'd fall through and try to
reserve metadata bytes for this, then we'd just steal the full reservation from
the delalloc_block_rsv, and if that didn't have enough space we'd steal the full
reservation from the global reserve. The problem with this is we can easily
just return ENOSPC and fallback to updating the inode item directly. In the
worst case (assuming 4k nodesize) we'd steal 64kib from the global reserve if we
fall all the way through, however if we just fallback and update the inode
directly we'd only steal 4k * BTRFS_PATH_MAX in the worst case which is 32kib.
We would have also just added the extent item for the inode so we likely will
have already cow'ed down most of the way to the leaf containing the inode item,
so we are more often than not only need one or two nodesize's worth of
reservations. Given the reservation for the extent itself is also a worst case
we will likely already have space to cover the inode update.
This change will make us behave better in the theoretical worst case, and much
better in the case that we don't have our reservation and cannot reserve more
metadata. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There are a few races in the metadata reservation stuff. First we add the bytes
to the block_rsv well after we've set the bit on the inode saying that we have
space for it and after we've reserved the bytes. So use the normal
btrfs_block_rsv_add helper for this case. Secondly we can flush delalloc
extents when we try to reserve space for our write, which means that we could
have used up the space for the inode and we wouldn't know because we only check
before the reservation. So instead make sure we are always reserving space for
the inode update, and then if we don't need it release those bytes afterward.
Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
So btrfs_block_rsv_migrate just unconditionally calls block_rsv_migrate_bytes.
Not only this but it unconditionally changes the size of the block_rsv. This
isn't a bug strictly speaking, but it makes truncate block rsv's look funny
because every time we migrate bytes over its size grows, even though we only
want it to be a specific size. So collapse this into one function that takes an
update_size argument and make truncate and evict not update the size for
consistency sake. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For some reason we're adding bytes_readonly to the space info after we update
the space info with the block group info. This creates a tiny race where we
could over-reserve space because we haven't yet taken out the bytes_readonly
bit. Since we already know this information at the time we call
update_space_info, just pass it along so it can be updated all at once. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Pull btrfs fixes part 2 from Chris Mason:
"This has one patch from Omar to bring iterate_shared back to btrfs.
We have a tree of work we queue up for directory items and it doesn't
lend itself well to shared access. While we're cleaning it up, Omar
has changed things to use an exclusive lock when there are delayed
items"
* 'for-linus-4.7-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
Btrfs: fix ->iterate_shared() by upgrading i_rwsem for delayed nodes
Pull btrfs fixes from Chris Mason:
"I have a two part pull this time because one of the patches Dave
Sterba collected needed to be against v4.7-rc2 or higher (we used
rc4). I try to make my for-linus-xx branch testable on top of the
last major so we can hand fixes to people on the list more easily, so
I've split this pull in two.
This first part has some fixes and two performance improvements that
we've been testing for some time.
Josef's two performance fixes are most notable. The transid tracking
patch makes a big improvement on pretty much every workload"
* 'for-linus-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
Btrfs: Force stripesize to the value of sectorsize
btrfs: fix disk_i_size update bug when fallocate() fails
Btrfs: fix error handling in map_private_extent_buffer
Btrfs: fix error return code in btrfs_init_test_fs()
Btrfs: don't do nocow check unless we have to
btrfs: fix deadlock in delayed_ref_async_start
Btrfs: track transid for delayed ref flushing
Commit fe742fd4f9 ("Revert "btrfs: switch to ->iterate_shared()"")
backed out the conversion to ->iterate_shared() for Btrfs because the
delayed inode handling in btrfs_real_readdir() is racy. However, we can
still do readdir in parallel if there are no delayed nodes.
This is a temporary fix which upgrades the shared inode lock to an
exclusive lock only when we have delayed items until we come up with a
more complete solution. While we're here, rename the
btrfs_{get,put}_delayed_items functions to make it very clear that
they're just for readdir.
Tested with xfstests and by doing a parallel kernel build:
while make tinyconfig && make -j4 && git clean dqfx; do
:
done
along with a bunch of parallel finds in another shell:
while true; do
for ((i=0; i<4; i++)); do
find . >/dev/null &
done
wait
done
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Btrfs code currently assumes stripesize to be same as
sectorsize. However Btrfs-progs (until commit
df05c7ed455f519e6e15e46196392e4757257305) has been setting
btrfs_super_block->stripesize to a value of 4096.
This commit makes sure that the value of btrfs_super_block->stripesize
is a power of 2. Later, it unconditionally sets btrfs_root->stripesize
to sectorsize.
Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
When doing truncate operation, btrfs_setsize() will first call
truncate_setsize() to set new inode->i_size, but if later
btrfs_truncate() fails, btrfs_setsize() will call
"i_size_write(inode, BTRFS_I(inode)->disk_i_size)" to reset the
inmemory inode size, now bug occurs. It's because for truncate
case btrfs_ordered_update_i_size() directly uses inode->i_size
to update BTRFS_I(inode)->disk_i_size, indeed we should use the
"offset" argument to update disk_i_size. Here is the call graph:
==>btrfs_truncate()
====>btrfs_truncate_inode_items()
======>btrfs_ordered_update_i_size(inode, last_size, NULL);
Here btrfs_ordered_update_i_size()'s offset argument is last_size.
And below test case can reveal this bug:
dd if=/dev/zero of=fs.img bs=$((1024*1024)) count=100
dev=$(losetup --show -f fs.img)
mkdir -p /mnt/mntpoint
mkfs.btrfs -f $dev
mount $dev /mnt/mntpoint
cd /mnt/mntpoint
echo "workdir is: /mnt/mntpoint"
blocksize=$((128 * 1024))
dd if=/dev/zero of=testfile bs=$blocksize count=1
sync
count=$((17*1024*1024*1024/blocksize))
echo "file size is:" $((count*blocksize))
for ((i = 1; i <= $count; i++)); do
i=$((i + 1))
dst_offset=$((blocksize * i))
xfs_io -f -c "reflink testfile 0 $dst_offset $blocksize"\
testfile > /dev/null
done
sync
truncate --size 0 testfile
ls -l testfile
du -sh testfile
exit
In this case, truncate operation will fail for enospc reason and
"du -sh testfile" returns value greater than 0, but testfile's
size is 0, we need to reflect correct inode->i_size.
Signed-off-by: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
map_private_extent_buffer() can return -EINVAL in two different cases,
1. when the requested contents span two pages if nodesize is larger
than pagesize,
2. when it detects something insane.
The 2nd one used to be only a WARN_ON(1), and we decided to return a error
to callers, but we didn't fix up all its callers, which will be
addressed by this patch.
Without this, btrfs may end up with 'general protection', ie.
reading invalid memory.
Reported-by: Vegard Nossum <vegard.nossum@oracle.com>
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Fix to return a negative error code from the kern_mount() error handling
case instead of 0(ret is set to 0 by register_filesystem), as done
elsewhere in this function.
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Before we write into prealloc/nocow space we have to make sure that there are no
references to the extents we are writing into, which means checking the extent
tree and csum tree in the case of nocow. So we don't want to do the nocow dance
unless we can't reserve data space, since it's a serious drag on performance.
With the following sequence
fallocate -l10737418240 /mnt/btrfs-test/file
cp --reflink /mnt/btrfs-test/file /mnt/btrfs-test/link
fio --name=randwrite --rw=randwrite --bs=4k --filename=/mnt/btrfs-test/file \
--end_fsync=1
we get the worst case scenario where we have to fall back on to doing the check
anyway.
Without this patch
lat (usec): min=5, max=111598, avg=27.65, stdev=124.51
write: io=10240MB, bw=126876KB/s, iops=31718, runt= 82646msec
With this patch
lat (usec): min=3, max=91210, avg=14.09, stdev=110.62
write: io=10240MB, bw=212753KB/s, iops=53188, runt= 49286msec
We get twice the throughput, half of the runtime, and half of the average
latency. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
[ PAGE_CACHE_ removal related fixups ]
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
"Btrfs: track transid for delayed ref flushing" was deadlocking on
btrfs_attach_transaction because its not safe to call from the async
delayed ref start code. This commit brings back btrfs_join_transaction
instead and checks for a blocked commit.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
Using the offwakecputime bpf script I noticed most of our time was spent waiting
on the delayed ref throttling. This is what is supposed to happen, but
sometimes the transaction can commit and then we're waiting for throttling that
doesn't matter anymore. So change this stuff to be a little smarter by tracking
the transid we were in when we initiated the throttling. If the transaction we
get is different then we can just bail out. This resulted in a 50% speedup in
my fs_mark test, and reduced the amount of time spent throttling by 60 seconds
over the entire run (which is about 30 minutes). Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
Pull btrfs fixes from Chris Mason:
"The most user visible change here is a fix for our recent superblock
validation checks that were causing problems on non-4k pagesized
systems"
* 'for-linus-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
Btrfs: btrfs_check_super_valid: Allow 4096 as stripesize
btrfs: remove build fixup for qgroup_account_snapshot
btrfs: use new error message helper in qgroup_account_snapshot
btrfs: avoid blocking open_ctree from cleaner_kthread
Btrfs: don't BUG_ON() in btrfs_orphan_add
btrfs: account for non-CoW'd blocks in btrfs_abort_transaction
Btrfs: check if extent buffer is aligned to sectorsize
btrfs: Use correct format specifier
Older btrfs-progs/mkfs.btrfs sets 4096 as the stripesize. Hence
restricting stripesize to be equal to sectorsize would cause super block
validation to return an error on architectures where PAGE_SIZE is not
equal to 4096.
Hence as a workaround, this commit allows stripesize to be set to 4096
bytes.
Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Introduced in 2c1984f244 ("btrfs: build fixup for
qgroup_account_snapshot") as temporary bisectability build fixup.
Signed-off-by: David Sterba <dsterba@suse.com>
This fixes a problem introduced in commit 2f3165ecf1
"btrfs: don't force mounts to wait for cleaner_kthread to delete one or more subvolumes".
open_ctree eventually calls btrfs_replay_log which in turn calls
btrfs_commit_super which tries to lock the cleaner_mutex, causing a
recursive mutex deadlock during mount.
Instead of playing whack-a-mole trying to keep up with all the
functions that may want to lock cleaner_mutex, put all the cleaner_mutex
lockers back where they were, and attack the problem more directly:
keep cleaner_kthread asleep until the filesystem is mounted.
When filesystems are mounted read-only and later remounted read-write,
open_ctree did not set fs_info->open and neither does anything else.
Set this flag in btrfs_remount so that neither btrfs_delete_unused_bgs
nor cleaner_kthread get confused by the common case of "/" filesystem
read-only mount followed by read-write remount.
Signed-off-by: Zygo Blaxell <ce3g8jdj@umail.furryterror.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is just a screwup for developers, so change it to an ASSERT() so developers
notice when things go wrong and deal with the error appropriately if ASSERT()
isn't enabled. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
The test for !trans->blocks_used in btrfs_abort_transaction is
insufficient to determine whether it's safe to drop the transaction
handle on the floor. btrfs_cow_block, informed by should_cow_block,
can return blocks that have already been CoW'd in the current
transaction. trans->blocks_used is only incremented for new block
allocations. If an operation overlaps the blocks in the current
transaction entirely and must abort the transaction, we'll happily
let it clean up the trans handle even though it may have modified
the blocks and will commit an incomplete operation.
In the long-term, I'd like to do closer tracking of when the fs
is actually modified so we can still recover as gracefully as possible,
but that approach will need some discussion. In the short term,
since this is the only code using trans->blocks_used, let's just
switch it to a bool indicating whether any blocks were used and set
it when should_cow_block returns false.
Cc: stable@vger.kernel.org # 3.4+
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Thanks to fuzz testing, we can pass an invalid bytenr to extent buffer
via alloc_extent_buffer(). An unaligned eb can have more pages than it
should have, which ends up extent buffer's leak or some corrupted content
in extent buffer.
This adds a warning to let us quickly know what was happening.
Now that alloc_extent_buffer() no more returns NULL, this changes its
caller and callers of its caller to match with the new error
handling.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Component mirror_num of struct btrfsic_block is defined
as unsigned int. Use %u as format specifier.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
