This patch provides a band aid to improve the 'out of the box'
behaviour of btrfs for disks that are detected as being an ssd. In a
general purpose mixed workload scenario, the current ssd mode causes
overallocation of available raw disk space for data, while leaving
behind increasing amounts of unused fragmented free space. This
situation leads to early ENOSPC problems which are harming user
experience and adoption of btrfs as a general purpose filesystem.
This patch modifies the data extent allocation behaviour of the ssd mode
to make it behave identical to nossd mode. The metadata behaviour and
additional ssd_spread option stay untouched so far.
Recommendations for future development are to reconsider the current
oversimplified nossd / ssd distinction and the broken detection
mechanism based on the rotational attribute in sysfs and provide
experienced users with a more flexible way to choose allocator behaviour
for data and metadata, optimized for certain use cases, while keeping
sane 'out of the box' default settings. The internals of the current
btrfs code have more potential than what currently gets exposed to the
user to choose from.
The SSD story...
In the first year of btrfs development, around early 2008, btrfs
gained a mount option which enables specific functionality for
filesystems on solid state devices. The first occurance of this
functionality is in commit e18e4809, labeled "Add mount -o ssd, which
includes optimizations for seek free storage".
The effect on allocating free space for doing (data) writes is to
'cluster' writes together, writing them out in contiguous space, as
opposed to a 'tetris' way of putting all separate writes into any free
space fragment that fits (which is what the -o nossd behaviour does).
A somewhat simplified explanation of what happens is that, when for
example, the 'cluster' size is set to 2MiB, when we do some writes, the
data allocator will search for a free space block that is 2MiB big, and
put the writes in there. The ssd mode itself might allow a 2MiB cluster
to be composed of multiple free space extents with some existing data in
between, while the additional ssd_spread mount option kills off this
option and requires fully free space.
The idea behind this is (commit 536ac8ae): "The [...] clusters make it
more likely a given IO will completely overwrite the ssd block, so it
doesn't have to do an internal rwm cycle."; ssd block meaning nand erase
block. So, effectively this means applying a "locality based algorithm"
and trying to outsmart the actual ssd.
Since then, various changes have been made to the involved code, but the
basic idea is still present, and gets activated whenever the ssd mount
option is active. This also happens by default, when the rotational flag
as seen at /sys/block/<device>/queue/rotational is set to 0.
However, there's a number of problems with this approach.
First, what the optimization is trying to do is outsmart the ssd by
assuming there is a relation between the physical address space of the
block device as seen by btrfs and the actual physical storage of the
ssd, and then adjusting data placement. However, since the introduction
of the Flash Translation Layer (FTL) which is a part of the internal
controller of an ssd, these attempts are futile. The use of good quality
FTL in consumer ssd products might have been limited in 2008, but this
situation has changed drastically soon after that time. Today, even the
flash memory in your automatic cat feeding machine or your grandma's
wheelchair has a full featured one.
Second, the behaviour as described above results in the filesystem being
filled up with badly fragmented free space extents because of relatively
small pieces of space that are freed up by deletes, but not selected
again as part of a 'cluster'. Since the algorithm prefers allocating a
new chunk over going back to tetris mode, the end result is a filesystem
in which all raw space is allocated, but which is composed of
underutilized chunks with a 'shotgun blast' pattern of fragmented free
space. Usually, the next problematic thing that happens is the
filesystem wanting to allocate new space for metadata, which causes the
filesystem to fail in spectacular ways.
Third, the default mount options you get for an ssd ('ssd' mode enabled,
'discard' not enabled), in combination with spreading out writes over
the full address space and ignoring freed up space leads to worst case
behaviour in providing information to the ssd itself, since it will
never learn that all the free space left behind is actually free. There
are two ways to let an ssd know previously written data does not have to
be preserved, which are sending explicit signals using discard or
fstrim, or by simply overwriting the space with new data. The worst
case behaviour is the btrfs ssd_spread mount option in combination with
not having discard enabled. It has a side effect of minimizing the reuse
of free space previously written in.
Fourth, the rotational flag in /sys/ does not reliably indicate if the
device is a locally attached ssd. For example, iSCSI or NBD displays as
non-rotational, while a loop device on an ssd shows up as rotational.
The combination of the second and third problem effectively means that
despite all the good intentions, the btrfs ssd mode reliably causes the
ssd hardware and the filesystem structures and performance to be choked
to death. The clickbait version of the title of this story would have
been "Btrfs ssd optimizations considered harmful for ssds".
The current nossd 'tetris' mode (even still without discard) allows a
pattern of overwriting much more previously used space, causing many
more implicit discards to happen because of the overwrite information
the ssd gets. The actual location in the physical address space, as seen
from the point of view of btrfs is irrelevant, because the actual writes
to the low level flash are reordered anyway thanks to the FTL.
Changes made in the code
1. Make ssd mode data allocation identical to tetris mode, like nossd.
2. Adjust and clean up filesystem mount messages so that we can easily
identify if a kernel has this patch applied or not, when providing
support to end users. Also, make better use of the *_and_info helpers to
only trigger messages on actual state changes.
Backporting notes
Notes for whoever wants to backport this patch to their 4.9 LTS kernel:
* First apply commit 951e7966 "btrfs: drop the nossd flag when
remounting with -o ssd", or fixup the differences manually.
* The rest of the conflicts are because of the fs_info refactoring. So,
for example, instead of using fs_info, it's root->fs_info in
extent-tree.c
Signed-off-by: Hans van Kranenburg <hans.van.kranenburg@mendix.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Although this bio has no data attached, it will reach this condition
(bio->bi_opf & REQ_PREFLUSH) and then update the flush_gen of dev_state
in __btrfsic_submit_bio. So we should still submit it through integrity
checker. Otherwise, the integrity checker will throw the following warning
when I mount a newly created btrfs filesystem.
[10264.755497] btrfs: attempt to write superblock which references block M @29523968 (sdb1/1111654400/0) which is not flushed out of disk's write cache (block flush_gen=1, dev->flush_gen=0)!
[10264.755498] btrfs: attempt to write superblock which references block M @29523968 (sdb1/37912576/0) which is not flushed out of disk's write cache (block flush_gen=1, dev->flush_gen=0)!
Signed-off-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When doing an incremental send it's possible that the computed send stream
contains clone operations that will fail on the receiver if the receiver
has compression enabled and the clone operations target a sector sized
extent that starts at a zero file offset, is not compressed on the source
filesystem but ends up being compressed and inlined at the destination
filesystem.
Example scenario:
$ mkfs.btrfs -f /dev/sdb
$ mount -o compress /dev/sdb /mnt
# By doing a direct IO write, the data is not compressed.
$ xfs_io -f -d -c "pwrite -S 0xab 0 4K" /mnt/foobar
$ btrfs subvolume snapshot -r /mnt /mnt/mysnap1
$ xfs_io -c "reflink /mnt/foobar 0 8K 4K" /mnt/foobar
$ btrfs subvolume snapshot -r /mnt /mnt/mysnap2
$ btrfs send -f /tmp/1.snap /mnt/mysnap1
$ btrfs send -f /tmp/2.snap -p /mnt/mysnap1 /mnt/mysnap2
$ umount /mnt
$ mkfs.btrfs -f /dev/sdc
$ mount -o compress /dev/sdc /mnt
$ btrfs receive -f /tmp/1.snap /mnt
$ btrfs receive -f /tmp/2.snap /mnt
ERROR: failed to clone extents to foobar
Operation not supported
The same could be achieved by mounting the source filesystem without
compression and doing a buffered IO write instead of a direct IO one,
and mounting the destination filesystem with compression enabled.
So fix this by issuing regular write operations in the send stream
instead of clone operations when the source offset is zero and the
range has a length matching the sector size.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There is a corner case that slips through the checkers in functions
reading extent buffer, ie.
if (start < eb->len) and (start + len > eb->len),
then
a) map_private_extent_buffer() returns immediately because
it's thinking the range spans across two pages,
b) and the checkers in read_extent_buffer(), WARN_ON(start > eb->len)
and WARN_ON(start + len > eb->start + eb->len), both are OK in this
corner case, but it'd actually try to access the eb->pages out of
bounds because of (start + len > eb->len).
The case is found by switching extent inline ref type from shared data
ref to non-shared data ref, which is a kind of metadata corruption.
It'd use the wrong helper to access the eb,
eg. btrfs_extent_data_ref_root(eb, ref) is used but the %ref passing
here is "struct btrfs_shared_data_ref". And if the extent item
happens to be the first item in the eb, then offset/length will get
over eb->len which ends up an invalid memory access.
This is adding proper checks in order to avoid invalid memory access,
ie. 'general protection fault', before it's too late.
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The buffer passed to btrfs_ioctl_tree_search* functions have to be at least
sizeof(struct btrfs_ioctl_search_header). If this is not the case then the
ioctl should return -EOVERFLOW and set the uarg->buf_size to the minimum
required size. Currently btrfs_ioctl_tree_search_v2 would return an -EOVERFLOW
error with ->buf_size being set to the value passed by user space. Fix this by
removing the size check and relying on search_ioctl, which already includes it
and correctly sets buf_size.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently the code checks whether we should do data checksumming in
btrfs_submit_direct and the boolean result of this check is passed to
btrfs_submit_direct_hook, in turn passing it to __btrfs_submit_dio_bio which
actually consumes it. The last function actually has all the necessary context
to figure out whether to skip the check or not, so let's move the check closer
to where it's being consumed. No functional changes.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Chris Mason <clm@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If the range being cleared was not marked for defrag and we are not
about to clear the range from the defrag status, we don't need to
lock and unlock the inode.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Chris Mason <clm@fb.com>
Reviewed-by: Wang Shilong <wangshilong1991@gmail.com>
Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The error return variable ret is initialized to zero and then is
checked to see if it is non-zero in the if-block that follows it.
It is therefore impossible for ret to be non-zero after the if-block
hence the check is redundant and can be removed.
Detected by CoverityScan, CID#1021040 ("Logically dead code")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The internal free space tree management routines are always exposed for
testing purposes. Make them dependent on SANITY_TESTS being on so that
they are exposed only when they really have to.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This variable was added in 1abe9b8a13 ("Btrfs: add initial tracepointi
support for btrfs"), yet it never really got used, only assigned to. So
let's remove it.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have a WARN_ON(!var) inside an if branch which is executed (among
others) only when var is true.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We aren't using this define, so removing it.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Though BTRFS_FSID_SIZE and BTRFS_UUID_SIZE are of the same size, we
should use the matching constant for the fsid buffer.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Our dir_context->pos is supposed to hold the next position we're
supposed to look. If we successfully insert a delayed dir index we
could end up with a duplicate entry because we don't increase ctx->pos
after doing the dir_emit.
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>
Readdir does dir_emit while under the btree lock. dir_emit can trigger
the page fault which means we can deadlock. Fix this by allocating a
buffer on opening a directory and copying the readdir into this buffer
and doing dir_emit from outside of the tree lock.
Thread A
readdir <holding tree lock>
dir_emit
<page fault>
down_read(mmap_sem)
Thread B
mmap write
down_write(mmap_sem)
page_mkwrite
wait_ordered_extents
Process C
finish_ordered_extent
insert_reserved_file_extent
try to lock leaf <hang>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ copy the deadlock scenario to changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
Currently should_alloc_chunk uses ->total_bytes - ->bytes_readonly to
signify the total amount of bytes in this space info. However, given
Jeff's patch which adds bytes_pinned and bytes_may_use to the calculation
of num_allocated it becomes a lot more clear to just eliminate num_bytes
altogether and add the bytes_readonly to the amount of used space. That
way we don't change the results of the following statements. In the
process also start using btrfs_space_info_used.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In a heavy write scenario, we can end up with a large number of pinned bytes.
This can translate into (very) premature ENOSPC because pinned bytes
must be accounted for when allowing a reservation but aren't accounted for
when deciding whether to create a new chunk.
This patch adds the accounting to should_alloc_chunk so that we can
create the chunk.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is a minimal patch intended to be backported to older kernels.
We're going to extend the string specifying the compression method and
this would fail on kernels before that change (the string is compared
exactly).
Relax the string matching only to the prefix, ie. ignoring anything that
goes after "zlib" or "lzo", regardless of th format extension we decide
to use. This applies to the mount options and properties.
That way, patched old kernels could be booted on systems already
utilizing the new compression spec.
Applicable since commit 63541927c8, v3.14.
Signed-off-by: David Sterba <dsterba@suse.com>
Currently, the BTRFS_INODE_NOCOMPRESS will prevent any compression on a
given file, except when the mount is force-compress. As users have
reported on IRC, this will also prevent compression when requested by
defrag (btrfs fi defrag -c file).
The nocompress flag is set automatically by filesystem when the ratios
are bad and the user would have to manually drop the bit in order to
make defrag -c work. This is not good from the usability perspective.
This patch will raise priority for the defrag -c over nocompress, ie.
any file with NOCOMPRESS bit set will get defragmented. The bit will
remain untouched.
Alternate option was to also drop the nocompress bit and keep the
decision logic as is, but I think this is not the right solution.
Signed-off-by: David Sterba <dsterba@suse.com>
Add new value for compression to distinguish between defrag and
property. Previously, a single variable was used and this caused clashes
when the per-file 'compression' was set and a defrag -c was called.
The property-compression is loaded when the file is open, defrag will
overwrite the same variable and reset to 0 (ie. NONE) at when the file
defragmentaion is finished. That's considered a usability bug.
Now we won't touch the property value, use the defrag-compression. The
precedence of defrag is higher than for property (and whole-filesystem).
Signed-off-by: David Sterba <dsterba@suse.com>
This is preparatory for separating inode compression requested by defrag
and set via properties. This will fix a usability bug when defrag will
reset compression type to NONE. If the file has compression set via
property, it will not apply anymore (until next mount or reset through
command line).
We're going to fix that by adding another variable just for the defrag
call and won't touch the property. The defrag will have higher priority
when deciding whether to compress the data.
Signed-off-by: David Sterba <dsterba@suse.com>
Add skeleton code for compresison heuristics. Now it iterates over all
the pages, but in the end always says "yes, compress please", ie it does
not change the current behaviour.
In the future we're going to add various heuristics to analyze the data.
This patch can be used as a baseline for measuring if the effectivness
and performance.
Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ enhanced changelog, modified comments ]
Signed-off-by: David Sterba <dsterba@suse.com>
Correctly account for IO when waiting for a submitted bio in scrub. This
only for the accounting purposes and should not change other behaviour.
Signed-off-by: David Sterba <dsterba@suse.com>
Correctly account for IO when waiting for a submitted DIO read, the case
when we're retrying. This only for the accounting purposes and should
not change other behaviour.
Signed-off-by: David Sterba <dsterba@suse.com>
The pinned chunks might be left over so we clean them but at this point
of close_ctree, there's noone to race with, the locking can be removed.
Signed-off-by: David Sterba <dsterba@suse.com>
The return value of flush_space was used to have significance in the
early days when the code was first introduced and before the ticketed
enospc rework. Since the latter got introduced the return value lost any
significance whatsoever to its callers. So let's remove it. While at it
also remove the unused ticket variable in
btrfs_async_reclaim_metadata_space. It was used in the initial version
of the ticketed ENOSPC work, however Wang Xiaoguang detected a problem
with this and fixed it in ce129655c9 ("btrfs: introduce tickets_id to
determine whether asynchronous metadata reclaim work makes progress").
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ add comment ]
Signed-off-by: David Sterba <dsterba@suse.com>
Userspace transactions were introduced in commit 6bf13c0cc8 ("Btrfs:
transaction ioctls") to provide semantics that Ceph's object store
required. However, things have changed significantly since then, to the
point where btrfs is no longer suitable as a backend for ceph and in
fact it's actively advised against such usages. Considering this, there
doesn't seem to be a widespread, legit use case of userspace
transaction. They also clutter the file->private pointer.
So to end the agony let's nuke the userspace transaction ioctls. As a
first step let's give time for people to voice their objection by just
WARN()ining when the userspace transaction is used.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ move the warning past perm checks, keep the has-been-printed state;
we're ok with just one warning over all filesystems ]
Signed-off-by: David Sterba <dsterba@suse.com>
Superblock is read and written using buffer heads, we need to set the
bdev blocksize. The magic constant has been hardcoded in several places,
so replace it with a named constant.
Signed-off-by: David Sterba <dsterba@suse.com>
There are two independent parts, one that writes the superblocks and
another that waits for completion. No functional changes, but cleanups,
reformatting and comment updates.
Signed-off-by: David Sterba <dsterba@suse.com>
Polish the helper:
* drop underscores, no special meaning here
* pass fs_devices, as this is what the API implements
* drop noinline, no apparent reason for such simple helper
* constify uuid
* add comment
Signed-off-by: David Sterba <dsterba@suse.com>
There are two helpers called in chain from one location, we can merge the
functionaliy.
Originally, alloc_fs_devices could fill the device uuid randomly if we
we didn't give the uuid buffer. This happens for seed devices but the
fsid is generated in btrfs_prepare_sprout, so we can remove it.
Signed-off-by: David Sterba <dsterba@suse.com>
The function submit_extent_page has 15(!) parameters right now, op and
op_flags are effectively one value stored to bio::bi_opf, no need to
pass them separately. So it's 14 parameters now.
Signed-off-by: David Sterba <dsterba@suse.com>
This function prints an informative message and then continues
dev-replace. The message contains a progress percentage which is read
from the status. The status is allocated dynamically, about 2600 bytes,
just to read the single value. That's an overkill. We'll use the new
helper and drop the allocation.
Signed-off-by: David Sterba <dsterba@suse.com>
We'll want to read the percentage value from dev_replace elsewhere, move
the logic to a separate helper.
Signed-off-by: David Sterba <dsterba@suse.com>
All sorts of readahead errors are not considered fatal. We can continue
defragmentation without it, with some potential slow down, which will
last only for the current inode.
Signed-off-by: David Sterba <dsterba@suse.com>
We can safely use GFP_KERNEL, the function is called from two contexts:
- ioctl handler, called directly, no locks taken
- cleaner thread, running all queued defrag work, outside of any locks
Signed-off-by: David Sterba <dsterba@suse.com>
We don't need to restrict the allocation flags in btrfs_mount or
_remount. No big filesystem locks are held (possibly s_umount but that
does no count here).
Signed-off-by: David Sterba <dsterba@suse.com>
One of the error handling paths in __add_reloc_root contains btrfs_panic()
followed by some other code. As the name implies what it does is print
some error message and call BUG, naturally what follow afterwards is not
invoked. So remove this extra code.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This also adjusts the respective callers in other files. Those were
found with -Wunused-parameter.
btrfs_full_stripe_len's mapping_tree - introduced by 53b381b3ab
("Btrfs: RAID5 and RAID6") but it was never really used even in that
commit
btrfs_is_parity_mirror's mirror_num - same as above
chunk_drange_filter's chunk_offset - introduced by 94e60d5a5c ("Btrfs:
devid subset filter") and never used.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
clear_super - usage was removed in commit cea67ab92d ("btrfs: clean
the old superblocks before freeing the device") but that change forgot
to remove the actual variable.
max_key - commit 6174d3cb43 ("Btrfs: remove unused max_key arg from
btrfs_search_forward") removed the max_key parameter but it forgot to
remove references from callers.
stripe_len - this one was added by e06cd3dd7c ("Btrfs: add validadtion
checks for chunk loading") but even then it wasn't used.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
find_raid56_stripe_len statically returns SZ_64K which equals BTRFS_STRIPE_LEN.
It's sole caller is __btrfs_alloc_chunk and it assigns the return value to ai
variable which is already set to BTRFS_STRIPE_LEN. So remove the function
invocation altogether and remove the function itself. Also remove the variable
since it's only aliasing BTRFS_STRIPE_LEN and use the define directly. Use
the occassion to simplify the rounding down of stripe_size now that the value
we want it to align is a power of 2.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
No functional changes, just make the code more self-explanatory.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_new_inode() is the only consumer move it to inode.c,
from ioctl.c.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
find_workspace() allocates up to num_online_cpus() + 1 workspaces.
free_workspace() will only keep num_online_cpus() workspaces. When
(de)compressing we will allocate num_online_cpus() + 1 workspaces, then
free one, and repeat. Instead, we can just keep num_online_cpus() + 1
workspaces around, and never have to allocate/free another workspace in the
common case.
I tested on a Ubuntu 14.04 VM with 2 cores and 4 GiB of RAM. I mounted a
BtrFS partition with -o compress-force={lzo,zlib,zstd} and logged whenever
a workspace was allocated of freed. Then I copied vmlinux (527 MB) to the
partition. Before the patch, during the copy it would allocate and free 5-6
workspaces. After, it only allocated the initial 3. This held true for lzo,
zlib, and zstd. The time it took to execute cp vmlinux /mnt/btrfs && sync
dropped from 1.70s to 1.44s with lzo compression, and from 2.04s to 1.80s
for zstd compression.
Signed-off-by: Nick Terrell <terrelln@fb.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The helpers append "\n" so we can keep the actual strings shorter. The
extra newline will print an empty line. Some messages have been
slightly modified to be more consistent with the rest (lowercase first
letter).
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The current code was erroneously checking for
root_level > BTRFS_MAX_LEVEL. If we had a root_level of 8 then the check
won't trigger and we could potentially hit a buffer overflow. The
correct check should be root_level >= BTRFS_MAX_LEVEL .
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For a missing device, btrfs will just refuse to mount with almost
meaningless kernel message like:
BTRFS info (device vdb6): disk space caching is enabled
BTRFS info (device vdb6): has skinny extents
BTRFS error (device vdb6): failed to read the system array: -5
BTRFS error (device vdb6): open_ctree failed
This patch will print a new message about the missing device:
BTRFS info (device vdb6): disk space caching is enabled
BTRFS info (device vdb6): has skinny extents
BTRFS warning (device vdb6): devid 2 uuid 80470722-cad2-4b90-b7c3-fee294552f1b is missing
BTRFS error (device vdb6): failed to read the system array: -5
BTRFS error (device vdb6): open_ctree failed
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
As we use per-chunk degradable check, the global
num_tolerated_disk_barrier_failures is of no use.
We can now remove it.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The last user of num_tolerated_disk_barrier_failures is
barrier_all_devices().
But it can be easily changed to the new per-chunk degradable check
framework.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Just the same for mount time check, use btrfs_check_rw_degradable() to
check if we are OK to be remounted rw.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Now use the btrfs_check_rw_degradable() to check if we can mount in the
degraded mode.
With this patch, we can mount in the following case:
# mkfs.btrfs -f -m raid1 -d single /dev/sdb /dev/sdc
# wipefs -a /dev/sdc
# mount /dev/sdb /mnt/btrfs -o degraded
As the single data chunk is only on sdb, so it's OK to mount as
degraded, as missing one device is OK for RAID1.
But still fail in the following case as expected:
# mkfs.btrfs -f -m raid1 -d single /dev/sdb /dev/sdc
# wipefs -a /dev/sdb
# mount /dev/sdc /mnt/btrfs -o degraded
As the data chunk is only in sdb, so it's not OK to mount it as
degraded.
Reported-by: Zhao Lei <zhaolei@cn.fujitsu.com>
Reported-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Introduce a new function, btrfs_check_rw_degradable(), to check if all
chunks in btrfs is OK for degraded rw mount.
It provides the new basis for accurate btrfs mount/remount and even
runtime degraded mount check other than old one-size-fit-all method.
Btrfs currently uses num_tolerated_disk_barrier_failures to do global
check for tolerated missing device.
Although the one-size-fit-all solution is quite safe, it's too strict
if data and metadata has different duplication level.
For example, if one use Single data and RAID1 metadata for 2 disks, it
means any missing device will make the fs unable to be degraded
mounted.
But in fact, some times all single chunks may be in the existing
device and in that case, we should allow it to be rw degraded mounted.
Such case can be easily reproduced using the following script:
# mkfs.btrfs -f -m raid1 -d sing /dev/sdb /dev/sdc
# wipefs -f /dev/sdc
# mount /dev/sdb -o degraded,rw
If using btrfs-debug-tree to check /dev/sdb, one should find that the
data chunk is only in sdb, so in fact it should allow degraded mount.
This patchset will introduce a new per-chunk degradable check for
btrfs, allow above case to succeed, and it's quite small anyway.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ copied text from cover letter with more details about the problem being
solved ]
Signed-off-by: David Sterba <dsterba@suse.com>
When btrfs fails the checksum check, it'll fill the whole page with
"1".
However, if %csum_expected is 0 (which means there is no checksum), then
for some unknown reason, we just pretend that the read is correct, so
userspace would be confused about the dilemma that read is successful but
getting a page with all content being "1".
This can happen due to a bug in btrfs-convert.
This fixes it by always returning errors if checksum doesn't match.
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>
In btrfs_full_stripe_len/btrfs_is_parity_mirror we have similar code which
gets the chunk map for a particular range via get_chunk_map. However,
get_chunk_map can return an ERR_PTR value and while the 2 callers do catch
this with a WARN_ON they then proceed to indiscriminately dereference the
extent map. This of course leads to a crash. Fix the offenders by making the
dereference conditional on IS_ERR.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Many commits ago the data space_info in alloc_data_chunk_ondemand used to be
acquired from the inode. At that point commit
33b4d47f5e ("Btrfs: deal with NULL space info") got introduced to deal with
spurios cases where the space info could be null, following a rebalance.
Nowadays, however, the space info is referenced directly from the btrfs_fs_info
struct which is initialised at filesystem mount time. This makes the null
checks redundant, so remove them.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
All callers of flush_space pass the same number for orig/num_bytes
arguments. Let's remove one of the numbers and also modify the trace
point to show only a single number - bytes requested.
Seems that last point where the two parameters were treated differently
is before the ticketed enospc rework.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Several distributions mount the "proper root" as ro during initrd and
then remount it as rw before pivot_root(2). Thus, if a rescan had been
aborted by a previous shutdown, the rescan would never be resumed.
This issue would manifest itself as several btrfs ioctl(2)s causing the
entire machine to hang when btrfs_qgroup_wait_for_completion was hit
(due to the fs_info->qgroup_rescan_running flag being set but the rescan
itself not being resumed). Notably, Docker's btrfs storage driver makes
regular use of BTRFS_QUOTA_CTL_DISABLE and BTRFS_IOC_QUOTA_RESCAN_WAIT
(causing this problem to be manifested on boot for some machines).
Cc: <stable@vger.kernel.org> # v3.11+
Cc: Jeff Mahoney <jeffm@suse.com>
Fixes: b382a324b6 ("Btrfs: fix qgroup rescan resume on mount")
Signed-off-by: Aleksa Sarai <asarai@suse.de>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Tested-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Repeating the same computation in multiple places is not
necessary.
Signed-off-by: Edmund Nadolski <enadolski@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When called with a struct share_check, find_parent_nodes()
will detect a shared extent and immediately return with
BACKREF_SHARED_FOUND.
Signed-off-by: Edmund Nadolski <enadolski@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Since backref resolution is CPU-intensive, the cond_resched calls
should help alleviate soft lockup occurences.
Signed-off-by: Edmund Nadolski <enadolski@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This patch adds a tracepoint event for prelim_ref insertion and
merging. For each, the ref being inserted or merged and the count
of tree nodes is issued.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This patch adds counters to each of the rbtrees so that we can tell
how large they are growing for a given workload. These counters
will be exported by tracepoints in the next patch.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
It's been known for a while that the use of multiple lists
that are periodically merged was an algorithmic problem within
btrfs. There are several workloads that don't complete in any
reasonable amount of time (e.g. btrfs/130) and others that cause
soft lockups.
The solution is to use a set of rbtrees that do insertion merging
for both indirect and direct refs, with the former converting
refs into the latter. The result is a btrfs/130 workload that
used to take several hours now takes about half of that. This
runtime still isn't acceptable and a future patch will address that
by moving the rbtrees higher in the stack so the lookups can be
shared across multiple calls to find_parent_nodes.
Signed-off-by: Edmund Nadolski <enadolski@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit afce772e87 ("btrfs: fix check_shared for fiemap ioctl") added
the ref_tree code in backref.c to reduce backref searching for
shared extents under the FIEMAP ioctl. This code will not be
compatible with the upcoming rbtree changes for improved backref
searching, so this patch removes the ref_tree code. The rbtree
changes will provide the equivalent functionality for FIEMAP.
The above commit also introduced transaction semantics around calls to
btrfs_check_shared() in order to accurately account for delayed refs.
This functionality needs to be retained, so a complete revert of the
above commit is not desirable. This patch therefore removes the
ref_tree portion of the commit as above, however it does not remove
the transaction portion.
Signed-off-by: Edmund Nadolski <enadolski@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit afce772e87 ("btrfs: fix check_shared for fiemap ioctl") added
transaction semantics around calls to btrfs_check_shared() in order to
provide accurate accounting of delayed refs. The transaction management
should be done inside btrfs_check_shared(), so that callers do not need
to manage transactions individually.
Signed-off-by: Edmund Nadolski <enadolski@suse.com>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We typically use __ to indicate a helper routine that shouldn't be
called directly without understanding the proper context required
to do so. We use static functions to indicate that a function is
private to a particular C file. The backref code uses static
function and __ prefixes on nearly everything, which makes the code
difficult to read and establishes a pattern for future code that
shouldn't be followed. This patch drops all the unnecessary prefixes.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Replacing the double cast and ternary conditional with a helper makes
the code easier on the eyes.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Tracepoint arguments are all read-only. If we mark the arguments
as const, we're able to keep or convert those arguments to const
where appropriate.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have reader helpers for most of the on-disk structures that use
an extent_buffer and pointer as offset into the buffer that are
read-only. We should mark them as const and, in turn, allow consumers
of these interfaces to mark the buffers const as well.
No impact on code, but serves as documentation that a buffer is intended
not to be modified.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The sectorsize member of btrfs_block_group_cache is unused. So remove it, this
reduces the number of holes in the struct.
With patch:
/* size: 856, cachelines: 14, members: 40 */
/* sum members: 837, holes: 4, sum holes: 19 */
/* bit holes: 1, sum bit holes: 29 bits */
/* last cacheline: 24 bytes */
Without patch:
/* size: 864, cachelines: 14, members: 41 */
/* sum members: 841, holes: 5, sum holes: 23 */
/* bit holes: 1, sum bit holes: 29 bits */
/* last cacheline: 32 bytes */
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
__btrfs_alloc_chunk contains code which boils down to:
ndevs = min(ndevs, devs_max)
It's conditional upon devs_max not being 0. However, it cannot really be 0
since it's always set to either BTRFS_MAX_DEVS_SYS_CHUNK or
BTRFS_MAX_DEVS(fs_info->chunk_root). So eliminate the condition check and use
min explicitly. This has no functional changes.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
No functional changes.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
No functional changes, just make the loop a bit more readable
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add zstd compression and decompression support to BtrFS. zstd at its
fastest level compresses almost as well as zlib, while offering much
faster compression and decompression, approaching lzo speeds.
I benchmarked btrfs with zstd compression against no compression, lzo
compression, and zlib compression. I benchmarked two scenarios. Copying
a set of files to btrfs, and then reading the files. Copying a tarball
to btrfs, extracting it to btrfs, and then reading the extracted files.
After every operation, I call `sync` and include the sync time.
Between every pair of operations I unmount and remount the filesystem
to avoid caching. The benchmark files can be found in the upstream
zstd source repository under
`contrib/linux-kernel/{btrfs-benchmark.sh,btrfs-extract-benchmark.sh}`
[1] [2].
I ran the benchmarks on a Ubuntu 14.04 VM with 2 cores and 4 GiB of RAM.
The VM is running on a MacBook Pro with a 3.1 GHz Intel Core i7 processor,
16 GB of RAM, and a SSD.
The first compression benchmark is copying 10 copies of the unzipped
Silesia corpus [3] into a BtrFS filesystem mounted with
`-o compress-force=Method`. The decompression benchmark times how long
it takes to `tar` all 10 copies into `/dev/null`. The compression ratio is
measured by comparing the output of `df` and `du`. See the benchmark file
[1] for details. I benchmarked multiple zstd compression levels, although
the patch uses zstd level 1.
| Method | Ratio | Compression MB/s | Decompression speed |
|---------|-------|------------------|---------------------|
| None | 0.99 | 504 | 686 |
| lzo | 1.66 | 398 | 442 |
| zlib | 2.58 | 65 | 241 |
| zstd 1 | 2.57 | 260 | 383 |
| zstd 3 | 2.71 | 174 | 408 |
| zstd 6 | 2.87 | 70 | 398 |
| zstd 9 | 2.92 | 43 | 406 |
| zstd 12 | 2.93 | 21 | 408 |
| zstd 15 | 3.01 | 11 | 354 |
The next benchmark first copies `linux-4.11.6.tar` [4] to btrfs. Then it
measures the compression ratio, extracts the tar, and deletes the tar.
Then it measures the compression ratio again, and `tar`s the extracted
files into `/dev/null`. See the benchmark file [2] for details.
| Method | Tar Ratio | Extract Ratio | Copy (s) | Extract (s)| Read (s) |
|--------|-----------|---------------|----------|------------|----------|
| None | 0.97 | 0.78 | 0.981 | 5.501 | 8.807 |
| lzo | 2.06 | 1.38 | 1.631 | 8.458 | 8.585 |
| zlib | 3.40 | 1.86 | 7.750 | 21.544 | 11.744 |
| zstd 1 | 3.57 | 1.85 | 2.579 | 11.479 | 9.389 |
[1] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/btrfs-benchmark.sh
[2] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/btrfs-extract-benchmark.sh
[3] http://sun.aei.polsl.pl/~sdeor/index.php?page=silesia
[4] https://cdn.kernel.org/pub/linux/kernel/v4.x/linux-4.11.6.tar.xz
zstd source repository: https://github.com/facebook/zstd
Signed-off-by: Nick Terrell <terrelln@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
Pull btrfs fixes from David Sterba:
"Fixes addressing problems reported by users, and there's one more
regression fix"
* 'for-4.13-part3' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: round down size diff when shrinking/growing device
Btrfs: fix early ENOSPC due to delalloc
btrfs: fix lockup in find_free_extent with read-only block groups
Btrfs: fix dir item validation when replaying xattr deletes
Further testing showed that the fix introduced in 7dfb8be11b ("btrfs:
Round down values which are written for total_bytes_size") was
insufficient and it could still lead to discrepancies between the
total_bytes in the super block and the device total bytes. So this patch
also ensures that the difference between old/new sizes when
shrinking/growing is also rounded down. This ensure that we won't be
subtracting/adding a non-sectorsize multiples to the superblock/device
total sizees.
Fixes: 7dfb8be11b ("btrfs: Round down values which are written for total_bytes_size")
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If a lot of metadata is reserved for outstanding delayed allocations, we
rely on shrink_delalloc() to reclaim metadata space in order to fulfill
reservation tickets. However, shrink_delalloc() has a shortcut where if
it determines that space can be overcommitted, it will stop early. This
made sense before the ticketed enospc system, but now it means that
shrink_delalloc() will often not reclaim enough space to fulfill any
tickets, leading to an early ENOSPC. (Reservation tickets don't care
about being able to overcommit, they need every byte accounted for.)
Fix it by getting rid of the shortcut so that shrink_delalloc() reclaims
all of the metadata it is supposed to. This fixes early ENOSPCs we were
seeing when doing a btrfs receive to populate a new filesystem, as well
as early ENOSPCs Christoph saw when doing a big cp -r onto Btrfs.
Fixes: 957780eb27 ("Btrfs: introduce ticketed enospc infrastructure")
Tested-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Cc: stable@vger.kernel.org
Reviewed-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If we have a block group that is all of the following:
1) uncached in memory
2) is read-only
3) has a disk cache state that indicates we need to recreate the cache
AND the file system has enough free space fragmentation such that the
request for an extent of a given size can't be honored;
AND have a single CPU core;
AND it's the block group with the highest starting offset such that
there are no opportunities (like reading from disk) for the loop to
yield the CPU;
We can end up with a lockup.
The root cause is simple. Once we're in the position that we've read in
all of the other block groups directly and none of those block groups
can honor the request, there are no more opportunities to sleep. We end
up trying to start a caching thread which never gets run if we only have
one core. This *should* present as a hung task waiting on the caching
thread to make some progress, but it doesn't. Instead, it degrades into
a busy loop because of the placement of the read-only check.
During the first pass through the loop, block_group->cached will be set
to BTRFS_CACHE_STARTED and have_caching_bg will be set. Then we hit the
read-only check and short circuit the loop. We're not yet in
LOOP_CACHING_WAIT, so we skip that loop back before going through the
loop again for other raid groups.
Then we move to LOOP_CACHING_WAIT state.
During the this pass through the loop, ->cached will still be
BTRFS_CACHE_STARTED, which means it's not cached, so we'll enter
cache_block_group, do a lot of nothing, and return, and also set
have_caching_bg again. Then we hit the read-only check and short circuit
the loop. The same thing happens as before except now we DO trigger
the LOOP_CACHING_WAIT && have_caching_bg check and loop back up to the
top. We do this forever.
There are two fixes in this patch since they address the same underlying
bug.
The first is to add a cond_resched to the end of the loop to ensure
that the caching thread always has an opportunity to run. This will
fix the soft lockup issue, but find_free_extent will still loop doing
nothing until the thread has completed.
The second is to move the read-only check to the top of the loop. We're
never going to return an allocation within a read-only block group so
we may as well skip it early. The check for ->cached == BTRFS_CACHE_ERROR
would cause the same problem except that BTRFS_CACHE_ERROR is considered
a "done" state and we won't re-set have_caching_bg again.
Many thanks to Stephan Kulow <coolo@suse.de> for his excellent help in
the testing process.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We were passing an incorrect slot number to the function that validates
directory items when we are replaying xattr deletes from a log tree. The
correct slot is stored at variable 'i' and not at 'path->slots[0]', so
the call to the validation function was only correct for the first
iteration of the loop, when 'i == path->slots[0]'.
After this fix, the fstest generic/066 passes again.
Fixes: 8ee8c2d62d ("btrfs: Verify dir_item in replay_xattr_deletes")
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Firstly by applying the following with coccinelle's spatch:
@@ expression SB; @@
-SB->s_flags & MS_RDONLY
+sb_rdonly(SB)
to effect the conversion to sb_rdonly(sb), then by applying:
@@ expression A, SB; @@
(
-(!sb_rdonly(SB)) && A
+!sb_rdonly(SB) && A
|
-A != (sb_rdonly(SB))
+A != sb_rdonly(SB)
|
-A == (sb_rdonly(SB))
+A == sb_rdonly(SB)
|
-!(sb_rdonly(SB))
+!sb_rdonly(SB)
|
-A && (sb_rdonly(SB))
+A && sb_rdonly(SB)
|
-A || (sb_rdonly(SB))
+A || sb_rdonly(SB)
|
-(sb_rdonly(SB)) != A
+sb_rdonly(SB) != A
|
-(sb_rdonly(SB)) == A
+sb_rdonly(SB) == A
|
-(sb_rdonly(SB)) && A
+sb_rdonly(SB) && A
|
-(sb_rdonly(SB)) || A
+sb_rdonly(SB) || A
)
@@ expression A, B, SB; @@
(
-(sb_rdonly(SB)) ? 1 : 0
+sb_rdonly(SB)
|
-(sb_rdonly(SB)) ? A : B
+sb_rdonly(SB) ? A : B
)
to remove left over excess bracketage and finally by applying:
@@ expression A, SB; @@
(
-(A & MS_RDONLY) != sb_rdonly(SB)
+(bool)(A & MS_RDONLY) != sb_rdonly(SB)
|
-(A & MS_RDONLY) == sb_rdonly(SB)
+(bool)(A & MS_RDONLY) == sb_rdonly(SB)
)
to make comparisons against the result of sb_rdonly() (which is a bool)
work correctly.
Signed-off-by: David Howells <dhowells@redhat.com>
Pull ->s_options removal from Al Viro:
"Preparations for fsmount/fsopen stuff (coming next cycle). Everything
gets moved to explicit ->show_options(), killing ->s_options off +
some cosmetic bits around fs/namespace.c and friends. Basically, the
stuff needed to work with fsmount series with minimum of conflicts
with other work.
It's not strictly required for this merge window, but it would reduce
the PITA during the coming cycle, so it would be nice to have those
bits and pieces out of the way"
* 'work.mount' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
isofs: Fix isofs_show_options()
VFS: Kill off s_options and helpers
orangefs: Implement show_options
9p: Implement show_options
isofs: Implement show_options
afs: Implement show_options
affs: Implement show_options
befs: Implement show_options
spufs: Implement show_options
bpf: Implement show_options
ramfs: Implement show_options
pstore: Implement show_options
omfs: Implement show_options
hugetlbfs: Implement show_options
VFS: Don't use save/replace_mount_options if not using generic_show_options
VFS: Provide empty name qstr
VFS: Make get_filesystem() return the affected filesystem
VFS: Clean up whitespace in fs/namespace.c and fs/super.c
Provide a function to create a NUL-terminated string from unterminated data
Pull btrfs fixes from David Sterba:
"We've identified and fixed a silent corruption (introduced by code in
the first pull), a fixup after the blk_status_t merge and two fixes to
incremental send that Filipe has been hunting for some time"
* 'for-4.13-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
Btrfs: fix unexpected return value of bio_readpage_error
btrfs: btrfs_create_repair_bio never fails, skip error handling
btrfs: cloned bios must not be iterated by bio_for_each_segment_all
Btrfs: fix write corruption due to bio cloning on raid5/6
Btrfs: incremental send, fix invalid memory access
Btrfs: incremental send, fix invalid path for link commands
With blk_status_t conversion (that are now present in master),
bio_readpage_error() may return 1 as now ->submit_bio_hook() may not set
%ret if it runs without problems.
This fixes that unexpected return value by changing
btrfs_check_repairable() to return a bool instead of updating %ret, and
patch is applicable to both codebases with and without blk_status_t.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
As the function uses the non-failing bio allocation, we can remove error
handling from the callers as well.
Signed-off-by: David Sterba <dsterba@suse.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We've started using cloned bios more in 4.13, there are some specifics
regarding the iteration. Filipe found [1] that the raid56 iterated a
cloned bio using bio_for_each_segment_all, which is incorrect. The
cloned bios have wrong bi_vcnt and this could lead to silent
corruptions. This patch adds assertions to all remaining
bio_for_each_segment_all cases.
[1] https://patchwork.kernel.org/patch/9838535/
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Pull btrfs fix from David Sterba:
"This fixes a user-visible bug introduced by the nowait-aio patches
merged in this cycle"
* 'nowait-aio-btrfs-fixup' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: nowait aio: Correct assignment of pos
Assigning pos for usage early messes up in append mode, where the pos is
re-assigned in generic_write_checks(). Assign pos later to get the
correct position to write from iocb->ki_pos.
Since check_can_nocow also uses the value of pos, we shift
generic_write_checks() before check_can_nocow(). Checks with IOCB_DIRECT
are present in generic_write_checks(), so checking for IOCB_NOWAIT is
enough.
Also, put locking sequence in the fast path.
This fixes a user visible bug, as reported:
"apparently breaks several shell related features on my system.
In zsh history stopped working, because no new entries are added
anymore.
I fist noticed the issue when I tried to build mplayer. It uses a shell
script to generate a help_mp.h file:
[...]
Here is a simple testcase:
% echo "foo" >> test
% echo "foo" >> test
% cat test
foo
%
"
Fixes: edf064e7c6 ("btrfs: nowait aio support")
CC: Jens Axboe <axboe@kernel.dk>
Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Link: https://lkml.kernel.org/r/20170704042306.GA274@x4
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
-----BEGIN PGP SIGNATURE-----
iQIcBAABAgAGBQJZXhmCAAoJEAAOaEEZVoIVpRkP/1qlYn3pq6d5Kuz84pejOmlL
5jbkS/cOmeTxeUU4+B1xG8Lx7bAk8PfSXQOADbSJGiZd0ug95tJxplFYIGJzR/tG
aNMHeu/BVKKhUKORGuKR9rJKtwC839L/qao+yPBo5U3mU4L73rFWX8fxFuhSJ8HR
hvkgBu3Hx6GY59CzxJ8iJzj+B+uPSFrNweAk0+0UeWkBgTzEdiGqaXBX4cHIkq/5
hMoCG+xnmwHKbCBsQ5js+YJT+HedZ4lvfjOqGxgElUyjJ7Bkt/IFYOp8TUiu193T
tA4UinDjN8A7FImmIBIftrECmrAC9HIGhGZroYkMKbb8ReDR2ikE5FhKEpuAGU3a
BXBgX2mPQuArvZWM7qeJCkxV9QJ0u/8Ykbyzo30iPrICyrzbEvIubeB/mDA034+Z
Z0/z8C3v7826F3zP/NyaQEojUgRq30McMOIS8GMnx15HJwRsRKlzjfy9Wm4tWhl0
t3nH1jMqAZ7068s6rfh/oCwdgGOwr5o4hW/bnlITzxbjWQUOnZIe7KBxIezZJ2rv
OcIwd5qE8PNtpagGj5oUbnjGOTkERAgsMfvPk5tjUNt28/qUlVs2V0aeo47dlcsh
oYr8WMOIzw98Rl7Bo70mplLrqLD6nGl0LfXOyUlT4STgLWW4ksmLVuJjWIUxcO/0
yKWjj9wfYRQ0vSUqhsI5
=3Z93
-----END PGP SIGNATURE-----
Merge tag 'for-linus-v4.13-2' of git://git.kernel.org/pub/scm/linux/kernel/git/jlayton/linux
Pull Writeback error handling updates from Jeff Layton:
"This pile represents the bulk of the writeback error handling fixes
that I have for this cycle. Some of the earlier patches in this pile
may look trivial but they are prerequisites for later patches in the
series.
The aim of this set is to improve how we track and report writeback
errors to userland. Most applications that care about data integrity
will periodically call fsync/fdatasync/msync to ensure that their
writes have made it to the backing store.
For a very long time, we have tracked writeback errors using two flags
in the address_space: AS_EIO and AS_ENOSPC. Those flags are set when a
writeback error occurs (via mapping_set_error) and are cleared as a
side-effect of filemap_check_errors (as you noted yesterday). This
model really sucks for userland.
Only the first task to call fsync (or msync or fdatasync) will see the
error. Any subsequent task calling fsync on a file will get back 0
(unless another writeback error occurs in the interim). If I have
several tasks writing to a file and calling fsync to ensure that their
writes got stored, then I need to have them coordinate with one
another. That's difficult enough, but in a world of containerized
setups that coordination may even not be possible.
But wait...it gets worse!
The calls to filemap_check_errors can be buried pretty far down in the
call stack, and there are internal callers of filemap_write_and_wait
and the like that also end up clearing those errors. Many of those
callers ignore the error return from that function or return it to
userland at nonsensical times (e.g. truncate() or stat()). If I get
back -EIO on a truncate, there is no reason to think that it was
because some previous writeback failed, and a subsequent fsync() will
(incorrectly) return 0.
This pile aims to do three things:
1) ensure that when a writeback error occurs that that error will be
reported to userland on a subsequent fsync/fdatasync/msync call,
regardless of what internal callers are doing
2) report writeback errors on all file descriptions that were open at
the time that the error occurred. This is a user-visible change,
but I think most applications are written to assume this behavior
anyway. Those that aren't are unlikely to be hurt by it.
3) document what filesystems should do when there is a writeback
error. Today, there is very little consistency between them, and a
lot of cargo-cult copying. We need to make it very clear what
filesystems should do in this situation.
To achieve this, the set adds a new data type (errseq_t) and then
builds new writeback error tracking infrastructure around that. Once
all of that is in place, we change the filesystems to use the new
infrastructure for reporting wb errors to userland.
Note that this is just the initial foray into cleaning up this mess.
There is a lot of work remaining here:
1) convert the rest of the filesystems in a similar fashion. Once the
initial set is in, then I think most other fs' will be fairly
simple to convert. Hopefully most of those can in via individual
filesystem trees.
2) convert internal waiters on writeback to use errseq_t for
detecting errors instead of relying on the AS_* flags. I have some
draft patches for this for ext4, but they are not quite ready for
prime time yet.
This was a discussion topic this year at LSF/MM too. If you're
interested in the gory details, LWN has some good articles about this:
https://lwn.net/Articles/718734/https://lwn.net/Articles/724307/"
* tag 'for-linus-v4.13-2' of git://git.kernel.org/pub/scm/linux/kernel/git/jlayton/linux:
btrfs: minimal conversion to errseq_t writeback error reporting on fsync
xfs: minimal conversion to errseq_t writeback error reporting
ext4: use errseq_t based error handling for reporting data writeback errors
fs: convert __generic_file_fsync to use errseq_t based reporting
block: convert to errseq_t based writeback error tracking
dax: set errors in mapping when writeback fails
Documentation: flesh out the section in vfs.txt on storing and reporting writeback errors
mm: set both AS_EIO/AS_ENOSPC and errseq_t in mapping_set_error
fs: new infrastructure for writeback error handling and reporting
lib: add errseq_t type and infrastructure for handling it
mm: don't TestClearPageError in __filemap_fdatawait_range
mm: clear AS_EIO/AS_ENOSPC when writeback initiation fails
jbd2: don't clear and reset errors after waiting on writeback
buffer: set errors in mapping at the time that the error occurs
fs: check for writeback errors after syncing out buffers in generic_file_fsync
buffer: use mapping_set_error instead of setting the flag
mm: fix mapping_set_error call in me_pagecache_dirty
When doing an incremental send, while processing an extent that changed
between the parent and send snapshots and that extent was an inline extent
in the parent snapshot, it's possible to access a memory region beyond
the end of leaf if the inline extent is very small and it is the first
item in a leaf.
An example scenario is described below.
The send snapshot has the following leaf:
leaf 33865728 items 33 free space 773 generation 46 owner 5
fs uuid ab7090d8-dafd-4fb9-9246-723b6d2e2fb7
chunk uuid 2d16478c-c704-4ab9-b574-68bff2281b1f
(...)
item 14 key (335 EXTENT_DATA 0) itemoff 3052 itemsize 53
generation 36 type 1 (regular)
extent data disk byte 12791808 nr 4096
extent data offset 0 nr 4096 ram 4096
extent compression 0 (none)
item 15 key (335 EXTENT_DATA 8192) itemoff 2999 itemsize 53
generation 36 type 1 (regular)
extent data disk byte 138170368 nr 225280
extent data offset 0 nr 225280 ram 225280
extent compression 0 (none)
(...)
And the parent snapshot has the following leaf:
leaf 31272960 items 17 free space 17 generation 31 owner 5
fs uuid ab7090d8-dafd-4fb9-9246-723b6d2e2fb7
chunk uuid 2d16478c-c704-4ab9-b574-68bff2281b1f
item 0 key (335 EXTENT_DATA 0) itemoff 3951 itemsize 44
generation 31 type 0 (inline)
inline extent data size 23 ram_bytes 613 compression 1 (zlib)
(...)
When computing the send stream, it is detected that the extent of inode
335, at file offset 0, and at fs/btrfs/send.c:is_extent_unchanged() we
grab the leaf from the parent snapshot and access the inline extent item.
However, before jumping to the 'out' label, we access the 'offset' and
'disk_bytenr' fields of the extent item, which should not be done for
inline extents since the inlined data starts at the offset of the
'disk_bytenr' field and can be very small. For example accessing the
'offset' field of the file extent item results in the following trace:
[ 599.705368] general protection fault: 0000 [#1] PREEMPT SMP
[ 599.706296] Modules linked in: btrfs psmouse i2c_piix4 ppdev acpi_cpufreq serio_raw parport_pc i2c_core evdev tpm_tis tpm_tis_core sg pcspkr parport tpm button su$
[ 599.709340] CPU: 7 PID: 5283 Comm: btrfs Not tainted 4.10.0-rc8-btrfs-next-46+ #1
[ 599.709340] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.9.1-0-gb3ef39f-prebuilt.qemu-project.org 04/01/2014
[ 599.709340] task: ffff88023eedd040 task.stack: ffffc90006658000
[ 599.709340] RIP: 0010:read_extent_buffer+0xdb/0xf4 [btrfs]
[ 599.709340] RSP: 0018:ffffc9000665ba00 EFLAGS: 00010286
[ 599.709340] RAX: db73880000000000 RBX: 0000000000000000 RCX: 0000000000000001
[ 599.709340] RDX: ffffc9000665ba60 RSI: db73880000000000 RDI: ffffc9000665ba5f
[ 599.709340] RBP: ffffc9000665ba30 R08: 0000000000000001 R09: ffff88020dc5e098
[ 599.709340] R10: 0000000000001000 R11: 0000160000000000 R12: 6db6db6db6db6db7
[ 599.709340] R13: ffff880000000000 R14: 0000000000000000 R15: ffff88020dc5e088
[ 599.709340] FS: 00007f519555a8c0(0000) GS:ffff88023f3c0000(0000) knlGS:0000000000000000
[ 599.709340] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 599.709340] CR2: 00007f1411afd000 CR3: 0000000235f8e000 CR4: 00000000000006e0
[ 599.709340] Call Trace:
[ 599.709340] btrfs_get_token_64+0x93/0xce [btrfs]
[ 599.709340] ? printk+0x48/0x50
[ 599.709340] btrfs_get_64+0xb/0xd [btrfs]
[ 599.709340] process_extent+0x3a1/0x1106 [btrfs]
[ 599.709340] ? btree_read_extent_buffer_pages+0x5/0xef [btrfs]
[ 599.709340] changed_cb+0xb03/0xb3d [btrfs]
[ 599.709340] ? btrfs_get_token_32+0x7a/0xcc [btrfs]
[ 599.709340] btrfs_compare_trees+0x432/0x53d [btrfs]
[ 599.709340] ? process_extent+0x1106/0x1106 [btrfs]
[ 599.709340] btrfs_ioctl_send+0x960/0xe26 [btrfs]
[ 599.709340] btrfs_ioctl+0x181b/0x1fed [btrfs]
[ 599.709340] ? trace_hardirqs_on_caller+0x150/0x1ac
[ 599.709340] vfs_ioctl+0x21/0x38
[ 599.709340] ? vfs_ioctl+0x21/0x38
[ 599.709340] do_vfs_ioctl+0x611/0x645
[ 599.709340] ? rcu_read_unlock+0x5b/0x5d
[ 599.709340] ? __fget+0x6d/0x79
[ 599.709340] SyS_ioctl+0x57/0x7b
[ 599.709340] entry_SYSCALL_64_fastpath+0x18/0xad
[ 599.709340] RIP: 0033:0x7f51945eec47
[ 599.709340] RSP: 002b:00007ffc21c13e98 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
[ 599.709340] RAX: ffffffffffffffda RBX: ffffffff81096459 RCX: 00007f51945eec47
[ 599.709340] RDX: 00007ffc21c13f20 RSI: 0000000040489426 RDI: 0000000000000004
[ 599.709340] RBP: ffffc9000665bf98 R08: 00007f519450d700 R09: 00007f519450d700
[ 599.709340] R10: 00007f519450d9d0 R11: 0000000000000202 R12: 0000000000000046
[ 599.709340] R13: ffffc9000665bf78 R14: 0000000000000000 R15: 00007f5195574040
[ 599.709340] ? trace_hardirqs_off_caller+0x43/0xb1
[ 599.709340] Code: 29 f0 49 39 d8 4c 0f 47 c3 49 03 81 58 01 00 00 44 89 c1 4c 01 c2 4c 29 c3 48 c1 f8 03 49 0f af c4 48 c1 e0 0c 4c 01 e8 48 01 c6 <f3> a4 31 f6 4$
[ 599.709340] RIP: read_extent_buffer+0xdb/0xf4 [btrfs] RSP: ffffc9000665ba00
[ 599.762057] ---[ end trace fe00d7af61b9f49e ]---
This is because the 'offset' field starts at an offset of 37 bytes
(offsetof(struct btrfs_file_extent_item, offset)), has a length of 8
bytes and therefore attemping to read it causes a 1 byte access beyond
the end of the leaf, as the first item's content in a leaf is located
at the tail of the leaf, the item size is 44 bytes and the offset of
that field plus its length (37 + 8 = 45) goes beyond the item's size
by 1 byte.
So fix this by accessing the 'offset' and 'disk_bytenr' fields after
jumping to the 'out' label if we are processing an inline extent. We
move the reading operation of the 'disk_bytenr' field too because we
have the same problem as for the 'offset' field explained above when
the inline data is less then 8 bytes. The access to the 'generation'
field is also moved but just for the sake of grouping access to all
the fields.
Fixes: e1cbfd7bf6 ("Btrfs: send, fix file hole not being preserved due to inline extent")
Cc: <stable@vger.kernel.org> # v4.12+
Signed-off-by: Filipe Manana <fdmanana@suse.com>
In some scenarios an incremental send stream can contain link commands
with an invalid target path. Such scenarios happen after moving some
directory inode A, renaming a regular file inode B into the old name of
inode A and finally creating a new hard link for inode B at directory
inode A.
Consider the following example scenario where this issue happens.
Parent snapshot:
. (ino 256)
|
|--- dir1/ (ino 257)
| |--- dir2/ (ino 258)
| |--- dir3/ (ino 259)
| |--- file1 (ino 261)
| |--- dir4/ (ino 262)
|
|--- dir5/ (ino 260)
Send snapshot:
. (ino 256)
|
|--- dir1/ (ino 257)
|--- dir2/ (ino 258)
| |--- dir3/ (ino 259)
| |--- dir4 (ino 261)
|
|--- dir6/ (ino 263)
|--- dir44/ (ino 262)
|--- file11 (ino 261)
|--- dir55/ (ino 260)
When attempting to apply the corresponding incremental send stream, a
link command contains an invalid target path which makes the receiver
fail. The following is the verbose output of the btrfs receive command:
receiving snapshot mysnap2 uuid=90076fe6-5ba6-e64a-9321-9279670ed16b (...)
utimes
utimes dir1
utimes dir1/dir2/dir3
utimes
rename dir1/dir2/dir3/dir4 -> o262-7-0
link dir1/dir2/dir3/dir4 -> dir1/dir2/dir3/file1
link dir1/dir2/dir3/dir4/file11 -> dir1/dir2/dir3/file1
ERROR: link dir1/dir2/dir3/dir4/file11 -> dir1/dir2/dir3/file1 failed: Not a directory
The following steps happen during the computation of the incremental send
stream the lead to this issue:
1) When processing inode 261, we orphanize inode 262 due to a name/location
collision with one of the new hard links for inode 261 (created in the
second step below).
2) We create one of the 2 new hard links for inode 261, the one whose
location is at "dir1/dir2/dir3/dir4".
3) We then attempt to create the other new hard link for inode 261, which
has inode 262 as its parent directory. Because the path for this new
hard link was computed before we started processing the new references
(hard links), it reflects the old name/location of inode 262, that is,
it does not account for the orphanization step that happened when
we started processing the new references for inode 261, whence it is
no longer valid, causing the receiver to fail.
So fix this issue by recomputing the full path of new references if we
ended up orphanizing other inodes which are directories.
A test case for fstests follows soon.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Hans van Kranenburg
Lu Fengqi
Filipe Manana
Liu Bo
Nikolay Borisov
Colin Ian King
Anand Jain
Josef Bacik
Jeff Mahoney
David Sterba
Timofey Titovets
Nick Terrell
Qu Wenruo
Aleksa Sarai
Edmund Nadolski
Linus Torvalds
Omar Sandoval
David Howells
Goldwyn Rodrigues