-----BEGIN PGP SIGNATURE-----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=zo4w
-----END PGP SIGNATURE-----
Merge tag 'for-5.11/block-2020-12-14' of git://git.kernel.dk/linux-block
Pull block updates from Jens Axboe:
"Another series of killing more code than what is being added, again
thanks to Christoph's relentless cleanups and tech debt tackling.
This contains:
- blk-iocost improvements (Baolin Wang)
- part0 iostat fix (Jeffle Xu)
- Disable iopoll for split bios (Jeffle Xu)
- block tracepoint cleanups (Christoph Hellwig)
- Merging of struct block_device and hd_struct (Christoph Hellwig)
- Rework/cleanup of how block device sizes are updated (Christoph
Hellwig)
- Simplification of gendisk lookup and removal of block device
aliasing (Christoph Hellwig)
- Block device ioctl cleanups (Christoph Hellwig)
- Removal of bdget()/blkdev_get() as exported API (Christoph Hellwig)
- Disk change rework, avoid ->revalidate_disk() (Christoph Hellwig)
- sbitmap improvements (Pavel Begunkov)
- Hybrid polling fix (Pavel Begunkov)
- bvec iteration improvements (Pavel Begunkov)
- Zone revalidation fixes (Damien Le Moal)
- blk-throttle limit fix (Yu Kuai)
- Various little fixes"
* tag 'for-5.11/block-2020-12-14' of git://git.kernel.dk/linux-block: (126 commits)
blk-mq: fix msec comment from micro to milli seconds
blk-mq: update arg in comment of blk_mq_map_queue
blk-mq: add helper allocating tagset->tags
Revert "block: Fix a lockdep complaint triggered by request queue flushing"
nvme-loop: use blk_mq_hctx_set_fq_lock_class to set loop's lock class
blk-mq: add new API of blk_mq_hctx_set_fq_lock_class
block: disable iopoll for split bio
block: Improve blk_revalidate_disk_zones() checks
sbitmap: simplify wrap check
sbitmap: replace CAS with atomic and
sbitmap: remove swap_lock
sbitmap: optimise sbitmap_deferred_clear()
blk-mq: skip hybrid polling if iopoll doesn't spin
blk-iocost: Factor out the base vrate change into a separate function
blk-iocost: Factor out the active iocgs' state check into a separate function
blk-iocost: Move the usage ratio calculation to the correct place
blk-iocost: Remove unnecessary advance declaration
blk-iocost: Fix some typos in comments
blktrace: fix up a kerneldoc comment
block: remove the request_queue to argument request based tracepoints
...
-----BEGIN PGP SIGNATURE-----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=5YAS
-----END PGP SIGNATURE-----
Merge tag 'for-5.11-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
Pull btrfs updates from David Sterba:
"We have a mix of all kinds of changes, feature updates, core stuff,
performance improvements and lots of cleanups and preparatory changes.
User visible:
- export filesystem generation in sysfs
- new features for mount option 'rescue':
- what's currently supported is exported in sysfs
- 'ignorebadroots'/'ibadroots' - continue even if some essential
tree roots are not usable (extent, uuid, data reloc, device,
csum, free space)
- 'ignoredatacsums'/'idatacsums' - skip checksum verification on
data
- 'all' - now enables 'ignorebadroots' + 'ignoredatacsums' +
'nologreplay'
- export read mirror policy settings to sysfs, new policies will be
added in the future
- remove inode number cache feature (mount -o inode_cache), obsoleted
in 5.9
User visible fixes:
- async discard scheduling fixes on high loads
- update inode byte counter atomically so stat() does not report
wrong value in some cases
- free space tree fixes:
- correctly report status of v2 after remount
- clear v1 cache inodes when v2 is newly enabled after remount
Core:
- switch own tree lock implementation to standard rw semaphore:
- one-level lock nesting is not required anymore, the last use of
this was in free space that's now loaded asynchronously
- own implementation of adaptive spinning before taking mutex has
been part of rwsem
- performance seems to be better in general, much better (+tens
of percents) for some workloads
- lockdep does not complain
- finish direct IO conversion to iomap infrastructure, remove
temporary workaround for DSYNC after iomap API updates
- preparatory work to support data and metadata blocks smaller than
page:
- generalize code that assumes sectorsize == PAGE_SIZE, lots of
refactoring
- planned namely for 64K pages (eg. arm64, ppc64)
- scrub read-only support
- preparatory work for zoned allocation mode (SMR/ZBC/ZNS friendly):
- disable incompatible features
- round-robin superblock write
- free space cache (v1) is loaded asynchronously, remove tree path
recursion
- slightly improved time tacking for transaction kthread wake ups
Performance improvements (note that the numbers depend on load type or
other features and weren't run on the same machine):
- skip unnecessary work:
- do not start readahead for csum tree when scrubbing non-data
block groups
- do not start and wait for delalloc on snapshot roots on
transaction commit
- fix race when defragmenting leads to unnecessary IO
- dbench speedups (+throughput%/-max latency%):
- skip unnecessary searches for xattrs when logging an inode
(+10.8/-8.2)
- stop incrementing log batch when joining log transaction (1-2)
- unlock path before checking if extent is shared during nocow
writeback (+5.0/-20.5), on fio load +9.7% throughput/-9.8%
runtime
- several tree log improvements, eg. removing unnecessary
operations, fixing races that lead to additional work
(+12.7/-8.2)
- tree-checker error branches annotated with unlikely() (+3%
throughput)
Other:
- cleanups
- lockdep fixes
- more btrfs_inode conversions
- error variable cleanups"
* tag 'for-5.11-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (198 commits)
btrfs: scrub: allow scrub to work with subpage sectorsize
btrfs: scrub: support subpage data scrub
btrfs: scrub: support subpage tree block scrub
btrfs: scrub: always allocate one full page for one sector for RAID56
btrfs: scrub: reduce width of extent_len/stripe_len from 64 to 32 bits
btrfs: refactor btrfs_lookup_bio_sums to handle out-of-order bvecs
btrfs: remove btrfs_find_ordered_sum call from btrfs_lookup_bio_sums
btrfs: handle sectorsize < PAGE_SIZE case for extent buffer accessors
btrfs: update num_extent_pages to support subpage sized extent buffer
btrfs: don't allow tree block to cross page boundary for subpage support
btrfs: calculate inline extent buffer page size based on page size
btrfs: factor out btree page submission code to a helper
btrfs: make btrfs_verify_data_csum follow sector size
btrfs: pass bio_offset to check_data_csum() directly
btrfs: rename bio_offset of extent_submit_bio_start_t to dio_file_offset
btrfs: fix lockdep warning when creating free space tree
btrfs: skip space_cache v1 setup when not using it
btrfs: remove free space items when disabling space cache v1
btrfs: warn when remount will not change the free space tree
btrfs: use superblock state to print space_cache mount option
...
- Consolidate all kmap_atomic() internals into a generic implementation
which builds the base for the kmap_local() API and make the
kmap_atomic() interface wrappers which handle the disabling/enabling of
preemption and pagefaults.
- Switch the storage from per-CPU to per task and provide scheduler
support for clearing mapping when scheduling out and restoring them
when scheduling back in.
- Merge the migrate_disable/enable() code, which is also part of the
scheduler pull request. This was required to make the kmap_local()
interface available which does not disable preemption when a mapping
is established. It has to disable migration instead to guarantee that
the virtual address of the mapped slot is the same accross preemption.
- Provide better debug facilities: guard pages and enforced utilization
of the mapping mechanics on 64bit systems when the architecture allows
it.
- Provide the new kmap_local() API which can now be used to cleanup the
kmap_atomic() usage sites all over the place. Most of the usage sites
do not require the implicit disabling of preemption and pagefaults so
the penalty on 64bit and 32bit non-highmem systems is removed and quite
some of the code can be simplified. A wholesale conversion is not
possible because some usage depends on the implicit side effects and
some need to be cleaned up because they work around these side effects.
The migrate disable side effect is only effective on highmem systems
and when enforced debugging is enabled. On 64bit and 32bit non-highmem
systems the overhead is completely avoided.
-----BEGIN PGP SIGNATURE-----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=n71I
-----END PGP SIGNATURE-----
Merge tag 'core-mm-2020-12-14' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull kmap updates from Thomas Gleixner:
"The new preemtible kmap_local() implementation:
- Consolidate all kmap_atomic() internals into a generic
implementation which builds the base for the kmap_local() API and
make the kmap_atomic() interface wrappers which handle the
disabling/enabling of preemption and pagefaults.
- Switch the storage from per-CPU to per task and provide scheduler
support for clearing mapping when scheduling out and restoring them
when scheduling back in.
- Merge the migrate_disable/enable() code, which is also part of the
scheduler pull request. This was required to make the kmap_local()
interface available which does not disable preemption when a
mapping is established. It has to disable migration instead to
guarantee that the virtual address of the mapped slot is the same
across preemption.
- Provide better debug facilities: guard pages and enforced
utilization of the mapping mechanics on 64bit systems when the
architecture allows it.
- Provide the new kmap_local() API which can now be used to cleanup
the kmap_atomic() usage sites all over the place. Most of the usage
sites do not require the implicit disabling of preemption and
pagefaults so the penalty on 64bit and 32bit non-highmem systems is
removed and quite some of the code can be simplified. A wholesale
conversion is not possible because some usage depends on the
implicit side effects and some need to be cleaned up because they
work around these side effects.
The migrate disable side effect is only effective on highmem
systems and when enforced debugging is enabled. On 64bit and 32bit
non-highmem systems the overhead is completely avoided"
* tag 'core-mm-2020-12-14' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (33 commits)
ARM: highmem: Fix cache_is_vivt() reference
x86/crashdump/32: Simplify copy_oldmem_page()
io-mapping: Provide iomap_local variant
mm/highmem: Provide kmap_local*
sched: highmem: Store local kmaps in task struct
x86: Support kmap_local() forced debugging
mm/highmem: Provide CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP
mm/highmem: Provide and use CONFIG_DEBUG_KMAP_LOCAL
microblaze/mm/highmem: Add dropped #ifdef back
xtensa/mm/highmem: Make generic kmap_atomic() work correctly
mm/highmem: Take kmap_high_get() properly into account
highmem: High implementation details and document API
Documentation/io-mapping: Remove outdated blurb
io-mapping: Cleanup atomic iomap
mm/highmem: Remove the old kmap_atomic cruft
highmem: Get rid of kmap_types.h
xtensa/mm/highmem: Switch to generic kmap atomic
sparc/mm/highmem: Switch to generic kmap atomic
powerpc/mm/highmem: Switch to generic kmap atomic
nds32/mm/highmem: Switch to generic kmap atomic
...
Since btrfs scrub is utilizing its own infrastructure to submit
read/write, scrub is independent from all other routines.
This brings one very neat feature, allow us to read 4K data into offset
0 of a 64K page. So is the writeback routine.
This makes scrub on subpage sector size much easier to implement, and
thanks to previous commits which just changed the implementation to
always do scrub based on sector size, now scrub can handle subpage
filesystem without any problem.
This patch will just remove the restriction on
(sectorsize != PAGE_SIZE), to make scrub finally work on subpage
filesystems.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Btrfs scrub is more flexible than buffered data write path, as we can
read an unaligned subpage data into page offset 0.
This ability makes subpage support much easier, we just need to check
each scrub_page::page_len and ensure we only calculate hash for [0,
page_len) of a page.
There is a small thing to notice: for subpage case, we still do sector
by sector scrub. This means we will submit a read bio for each sector
to scrub, resulting in the same amount of read bios, just like on the 4K
page systems.
This behavior can be considered as a good thing, if we want everything
to be the same as 4K page systems. But this also means, we're wasting
the possibility to submit larger bio using 64K page size. This is
another problem to consider in the future.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
To support subpage tree block scrub, scrub_checksum_tree_block() only
needs to learn 2 new tricks:
- Follow sector size
Now scrub_page only represents one sector, we need to follow it
properly.
- Run checksum on all sectors
Since scrub_page only represents one sector, we need to run checksum
on all sectors, not only (nodesize >> PAGE_SIZE).
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For scrub_pages() and scrub_pages_for_parity(), we currently allocate
one scrub_page structure for one page.
This is fine if we only read/write one sector one time. But for cases
like scrubbing RAID56, we need to read/write the full stripe, which is
in 64K size for now.
For subpage size, we will submit the read in just one page, which is
normally a good thing, but for RAID56 case, it only expects to see one
sector, not the full stripe in its endio function.
This could lead to wrong parity checksum for RAID56 on subpage.
To make the existing code work well for subpage case, here we take a
shortcut by always allocating a full page for one sector.
This should provide the base to make RAID56 work for subpage case.
The cost is pretty obvious now, for one RAID56 stripe now we always need
16 pages. For support subpage situation (64K page size, 4K sector size),
this means we need full one megabyte to scrub just one RAID56 stripe.
And for data scrub, each 4K sector will also need one 64K page.
This is mostly just a workaround, the proper fix for this is a much
larger project, using scrub_block to replace scrub_page, and allow
scrub_block to handle multi pages, csums, and csum_bitmap to avoid
allocating one page for each sector.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Btrfs on-disk format chose to use u64 for almost everything, but there
are a other restrictions that won't let us use more than u32 for things
like extent length (the maximum length is 128MiB for non-hole extents),
or stripe length (we have device number limit).
This means if we don't have extra handling to convert u64 to u32, we
will always have some questionable operations like
"u32 = u64 >> sectorsize_bits" in the code.
This patch will try to address the problem by reducing the width for the
following members/parameters:
- scrub_parity::stripe_len
- @len of scrub_pages()
- @extent_len of scrub_remap_extent()
- @len of scrub_parity_mark_sectors_error()
- @len of scrub_parity_mark_sectors_data()
- @len of scrub_extent()
- @len of scrub_pages_for_parity()
- @len of scrub_extent_for_parity()
For members extracted from on-disk structure, like map->stripe_len, they
will be kept as is. Since that modification would require on-disk format
change.
There will be cases like "u32 = u64 - u64" or "u32 = u64", for such call
sites, extra ASSERT() is added to be extra safe for debug builds.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Refactor btrfs_lookup_bio_sums() by:
- Remove the @file_offset parameter
There are two factors making the @file_offset parameter useless:
* For csum lookup in csum tree, file offset makes no sense
We only need disk_bytenr, which is unrelated to file_offset
* page_offset (file offset) of each bvec is not contiguous.
Pages can be added to the same bio as long as their on-disk bytenr
is contiguous, meaning we could have pages at different file offsets
in the same bio.
Thus passing file_offset makes no sense any more.
The only user of file_offset is for data reloc inode, we will use
a new function, search_file_offset_in_bio(), to handle it.
- Extract the csum tree lookup into search_csum_tree()
The new function will handle the csum search in csum tree.
The return value is the same as btrfs_find_ordered_sum(), returning
the number of found sectors which have checksum.
- Change how we do the main loop
The only needed info from bio is:
* the on-disk bytenr
* the length
After extracting the above info, we can do the search without bio
at all, which makes the main loop much simpler:
for (cur_disk_bytenr = orig_disk_bytenr;
cur_disk_bytenr < orig_disk_bytenr + orig_len;
cur_disk_bytenr += count * sectorsize) {
/* Lookup csum tree */
count = search_csum_tree(fs_info, path, cur_disk_bytenr,
search_len, csum_dst);
if (!count) {
/* Csum hole handling */
}
}
- Use single variable as the source to calculate all other offsets
Instead of all different type of variables, we use only one main
variable, cur_disk_bytenr, which represents the current disk bytenr.
All involved values can be calculated from that variable, and
all those variable will only be visible in the inner loop.
The above refactoring makes btrfs_lookup_bio_sums() way more robust than
it used to be, especially related to the file offset lookup. Now
file_offset lookup is only related to data reloc inode, otherwise we
don't need to bother file_offset at all.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The function btrfs_lookup_bio_sums() is only called for read bios.
While btrfs_find_ordered_sum() is to search ordered extent sums, which
is only for write path.
This means to read a page we either:
- Submit read bio if it's not uptodate
This means we only need to search csum tree for checksums.
- The page is already uptodate
It can be marked uptodate for previous read, or being marked dirty.
As we always mark page uptodate for dirty page.
In that case, we don't need to submit read bio at all, thus no need
to search any checksums.
Remove the btrfs_find_ordered_sum() call in btrfs_lookup_bio_sums().
And since btrfs_lookup_bio_sums() is the only caller for
btrfs_find_ordered_sum(), also remove the implementation.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
To support sectorsize < PAGE_SIZE case, we need to take extra care of
extent buffer accessors.
Since sectorsize is smaller than PAGE_SIZE, one page can contain
multiple tree blocks, we must use eb->start to determine the real offset
to read/write for extent buffer accessors.
This patch introduces two helpers to do this:
- get_eb_page_index()
This is to calculate the index to access extent_buffer::pages.
It's just a simple wrapper around "start >> PAGE_SHIFT".
For sectorsize == PAGE_SIZE case, nothing is changed.
For sectorsize < PAGE_SIZE case, we always get index as 0, and
the existing page shift also works.
- get_eb_offset_in_page()
This is to calculate the offset to access extent_buffer::pages.
This needs to take extent_buffer::start into consideration.
For sectorsize == PAGE_SIZE case, extent_buffer::start is always
aligned to PAGE_SIZE, thus adding extent_buffer::start to
offset_in_page() won't change the result.
For sectorsize < PAGE_SIZE case, adding extent_buffer::start gives
us the correct offset to access.
This patch will touch the following parts to cover all extent buffer
accessors:
- BTRFS_SETGET_HEADER_FUNCS()
- read_extent_buffer()
- read_extent_buffer_to_user()
- memcmp_extent_buffer()
- write_extent_buffer_chunk_tree_uuid()
- write_extent_buffer_fsid()
- write_extent_buffer()
- memzero_extent_buffer()
- copy_extent_buffer_full()
- copy_extent_buffer()
- memcpy_extent_buffer()
- memmove_extent_buffer()
- btrfs_get_token_##bits()
- btrfs_get_##bits()
- btrfs_set_token_##bits()
- btrfs_set_##bits()
- generic_bin_search()
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For subpage sized extent buffer, we have ensured no extent buffer will
cross page boundary, thus we would only need one page for any extent
buffer.
Update function num_extent_pages to handle such case. Now
num_extent_pages() returns 1 for subpage sized extent buffer.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
As a preparation for subpage sector size support (allowing filesystem
with sector size smaller than page size to be mounted) if the sector
size is smaller than page size, we don't allow tree block to be read if
it crosses 64K(*) boundary.
The 64K is selected because:
- we are only going to support 64K page size for subpage for now
- 64K is also the maximum supported node size
This ensures that tree blocks are always contained in one page for a
system with 64K page size, which can greatly simplify the handling.
Otherwise we would have to do complex multi-page handling of tree
blocks. Currently there is no way to create such tree blocks.
In kernel we have avoided such tree blocks allocation even on 4K page
size, as it can lead to RAID56 stripe scrubbing.
While btrfs-progs have fixed its chunk allocator since 2016 for convert,
and has extra checks to do the same behavior as the kernel.
Just add such graceful checks in case of an ancient filesystem.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Btrfs only support 64K as maximum node size, thus for 4K page system, we
would have at most 16 pages for one extent buffer.
For a system using 64K page size, we would really have just one page.
While we always use 16 pages for extent_buffer::pages, this means for
systems using 64K pages, we are wasting memory for 15 page pointers
which will never be used.
Calculate the array size based on page size and the node size maximum.
- for systems using 4K page size, it will stay 16 pages
- for systems using 64K page size, it will be 1 page
Move the definition of BTRFS_MAX_METADATA_BLOCKSIZE to btrfs_tree.h, to
avoid circular inclusion of ctree.h.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In btree_write_cache_pages() we have a btree page submission routine
buried deeply in a nested loop.
This patch will extract that part of code into a helper function,
submit_eb_page(), to do the same work.
Since submit_eb_page() now can return >0 for successful extent
buffer submission, remove the "ASSERT(ret <= 0);" line.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently btrfs_verify_data_csum() just passes the whole page to
check_data_csum(), which is fine since we only support sectorsize ==
PAGE_SIZE.
To support subpage, we need to properly honor per-sector
checksum verification, just like what we did in dio read path.
This patch will do the csum verification in a for loop, starts with
pg_off == start - page_offset(page), with sectorsize increase for
each loop.
For sectorsize == PAGE_SIZE case, the pg_off will always be 0, and we
will only loop once.
For subpage case, we do the iterate over each sector and if we found any
error, we return error.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Parameter icsum for check_data_csum() is a little hard to understand.
So is the phy_offset for btrfs_verify_data_csum().
Both parameters are calculated values for csum lookup.
Instead of some calculated value, just pass bio_offset and let the
final and only user, check_data_csum(), calculate whatever it needs.
Since we are here, also make the bio_offset parameter and some related
variables to be u32 (unsigned int).
As bio size is limited by its bi_size, which is unsigned int, and has
extra size limit check during various bio operations.
Thus we are ensured that bio_offset won't overflow u32.
Thus for all involved functions, not only rename the parameter from
@phy_offset to @bio_offset, but also reduce its width to u32, so we
won't have suspicious "u32 = u64 >> sector_bits;" lines anymore.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The parameter bio_offset of extent_submit_bio_start_t is very confusing.
If it's really bio_offset (offset to bio), then it should be u32. But
in fact, it's only utilized by dio read, and that member is used as file
offset, which must be u64.
Rename it to dio_file_offset since the only user uses it as file offset,
and add comment for who is using it.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
A lock dependency loop exists between the root tree lock, the extent tree
lock, and the free space tree lock.
The root tree lock depends on the free space tree lock because
btrfs_create_tree holds the new tree's lock while adding it to the root
tree.
The extent tree lock depends on the root tree lock because during
umount, we write out space cache v1, which writes inodes in the root
tree, which results in holding the root tree lock while doing a lookup
in the extent tree.
Finally, the free space tree depends on the extent tree because
populate_free_space_tree holds a locked path in the extent tree and then
does a lookup in the free space tree to add the new item.
The simplest of the three to break is the one during tree creation: we
unlock the leaf before inserting the tree node into the root tree, which
fixes the lockdep warning.
[30.480136] ======================================================
[30.480830] WARNING: possible circular locking dependency detected
[30.481457] 5.9.0-rc8+ #76 Not tainted
[30.481897] ------------------------------------------------------
[30.482500] mount/520 is trying to acquire lock:
[30.483064] ffff9babebe03908 (btrfs-free-space-00){++++}-{3:3}, at: __btrfs_tree_read_lock+0x39/0x180
[30.484054]
but task is already holding lock:
[30.484637] ffff9babebe24468 (btrfs-extent-01#2){++++}-{3:3}, at: __btrfs_tree_read_lock+0x39/0x180
[30.485581]
which lock already depends on the new lock.
[30.486397]
the existing dependency chain (in reverse order) is:
[30.487205]
-> #2 (btrfs-extent-01#2){++++}-{3:3}:
[30.487825] down_read_nested+0x43/0x150
[30.488306] __btrfs_tree_read_lock+0x39/0x180
[30.488868] __btrfs_read_lock_root_node+0x3a/0x50
[30.489477] btrfs_search_slot+0x464/0x9b0
[30.490009] check_committed_ref+0x59/0x1d0
[30.490603] btrfs_cross_ref_exist+0x65/0xb0
[30.491108] run_delalloc_nocow+0x405/0x930
[30.491651] btrfs_run_delalloc_range+0x60/0x6b0
[30.492203] writepage_delalloc+0xd4/0x150
[30.492688] __extent_writepage+0x18d/0x3a0
[30.493199] extent_write_cache_pages+0x2af/0x450
[30.493743] extent_writepages+0x34/0x70
[30.494231] do_writepages+0x31/0xd0
[30.494642] __filemap_fdatawrite_range+0xad/0xe0
[30.495194] btrfs_fdatawrite_range+0x1b/0x50
[30.495677] __btrfs_write_out_cache+0x40d/0x460
[30.496227] btrfs_write_out_cache+0x8b/0x110
[30.496716] btrfs_start_dirty_block_groups+0x211/0x4e0
[30.497317] btrfs_commit_transaction+0xc0/0xba0
[30.497861] sync_filesystem+0x71/0x90
[30.498303] btrfs_remount+0x81/0x433
[30.498767] reconfigure_super+0x9f/0x210
[30.499261] path_mount+0x9d1/0xa30
[30.499722] do_mount+0x55/0x70
[30.500158] __x64_sys_mount+0xc4/0xe0
[30.500616] do_syscall_64+0x33/0x40
[30.501091] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[30.501629]
-> #1 (btrfs-root-00){++++}-{3:3}:
[30.502241] down_read_nested+0x43/0x150
[30.502727] __btrfs_tree_read_lock+0x39/0x180
[30.503291] __btrfs_read_lock_root_node+0x3a/0x50
[30.503903] btrfs_search_slot+0x464/0x9b0
[30.504405] btrfs_insert_empty_items+0x60/0xa0
[30.504973] btrfs_insert_item+0x60/0xd0
[30.505412] btrfs_create_tree+0x1b6/0x210
[30.505913] btrfs_create_free_space_tree+0x54/0x110
[30.506460] btrfs_mount_rw+0x15d/0x20f
[30.506937] btrfs_remount+0x356/0x433
[30.507369] reconfigure_super+0x9f/0x210
[30.507868] path_mount+0x9d1/0xa30
[30.508264] do_mount+0x55/0x70
[30.508668] __x64_sys_mount+0xc4/0xe0
[30.509186] do_syscall_64+0x33/0x40
[30.509652] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[30.510271]
-> #0 (btrfs-free-space-00){++++}-{3:3}:
[30.510972] __lock_acquire+0x11ad/0x1b60
[30.511432] lock_acquire+0xa2/0x360
[30.511917] down_read_nested+0x43/0x150
[30.512383] __btrfs_tree_read_lock+0x39/0x180
[30.512947] __btrfs_read_lock_root_node+0x3a/0x50
[30.513455] btrfs_search_slot+0x464/0x9b0
[30.513947] search_free_space_info+0x45/0x90
[30.514465] __add_to_free_space_tree+0x92/0x39d
[30.515010] btrfs_create_free_space_tree.cold.22+0x1ee/0x45d
[30.515639] btrfs_mount_rw+0x15d/0x20f
[30.516142] btrfs_remount+0x356/0x433
[30.516538] reconfigure_super+0x9f/0x210
[30.517065] path_mount+0x9d1/0xa30
[30.517438] do_mount+0x55/0x70
[30.517824] __x64_sys_mount+0xc4/0xe0
[30.518293] do_syscall_64+0x33/0x40
[30.518776] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[30.519335]
other info that might help us debug this:
[30.520210] Chain exists of:
btrfs-free-space-00 --> btrfs-root-00 --> btrfs-extent-01#2
[30.521407] Possible unsafe locking scenario:
[30.522037] CPU0 CPU1
[30.522456] ---- ----
[30.522941] lock(btrfs-extent-01#2);
[30.523311] lock(btrfs-root-00);
[30.523952] lock(btrfs-extent-01#2);
[30.524620] lock(btrfs-free-space-00);
[30.525068]
*** DEADLOCK ***
[30.525669] 5 locks held by mount/520:
[30.526116] #0: ffff9babebc520e0 (&type->s_umount_key#37){+.+.}-{3:3}, at: path_mount+0x7ef/0xa30
[30.527056] #1: ffff9babebc52640 (sb_internal#2){.+.+}-{0:0}, at: start_transaction+0x3d5/0x5c0
[30.527960] #2: ffff9babeae8f2e8 (&cache->free_space_lock#2){+.+.}-{3:3}, at: btrfs_create_free_space_tree.cold.22+0x101/0x45d
[30.529118] #3: ffff9babebe24468 (btrfs-extent-01#2){++++}-{3:3}, at: __btrfs_tree_read_lock+0x39/0x180
[30.530113] #4: ffff9babebd52eb8 (btrfs-extent-00){++++}-{3:3}, at: btrfs_try_tree_read_lock+0x16/0x100
[30.531124]
stack backtrace:
[30.531528] CPU: 0 PID: 520 Comm: mount Not tainted 5.9.0-rc8+ #76
[30.532166] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.1-4.module_el8.1.0+248+298dec18 04/01/2014
[30.533215] Call Trace:
[30.533452] dump_stack+0x8d/0xc0
[30.533797] check_noncircular+0x13c/0x150
[30.534233] __lock_acquire+0x11ad/0x1b60
[30.534667] lock_acquire+0xa2/0x360
[30.535063] ? __btrfs_tree_read_lock+0x39/0x180
[30.535525] down_read_nested+0x43/0x150
[30.535939] ? __btrfs_tree_read_lock+0x39/0x180
[30.536400] __btrfs_tree_read_lock+0x39/0x180
[30.536862] __btrfs_read_lock_root_node+0x3a/0x50
[30.537304] btrfs_search_slot+0x464/0x9b0
[30.537713] ? trace_hardirqs_on+0x1c/0xf0
[30.538148] search_free_space_info+0x45/0x90
[30.538572] __add_to_free_space_tree+0x92/0x39d
[30.539071] ? printk+0x48/0x4a
[30.539367] btrfs_create_free_space_tree.cold.22+0x1ee/0x45d
[30.539972] btrfs_mount_rw+0x15d/0x20f
[30.540350] btrfs_remount+0x356/0x433
[30.540773] ? shrink_dcache_sb+0xd9/0x100
[30.541203] reconfigure_super+0x9f/0x210
[30.541642] path_mount+0x9d1/0xa30
[30.542040] do_mount+0x55/0x70
[30.542366] __x64_sys_mount+0xc4/0xe0
[30.542822] do_syscall_64+0x33/0x40
[30.543197] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[30.543691] RIP: 0033:0x7f109f7ab93a
[30.546042] RSP: 002b:00007ffc47c4f858 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5
[30.546770] RAX: ffffffffffffffda RBX: 00007f109f8cf264 RCX: 00007f109f7ab93a
[30.547485] RDX: 0000557e6fc10770 RSI: 0000557e6fc19cf0 RDI: 0000557e6fc19cd0
[30.548185] RBP: 0000557e6fc10520 R08: 0000557e6fc18e30 R09: 0000557e6fc18cb0
[30.548911] R10: 0000000000200020 R11: 0000000000000246 R12: 0000000000000000
[30.549606] R13: 0000557e6fc19cd0 R14: 0000557e6fc10770 R15: 0000557e6fc10520
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Signed-off-by: David Sterba <dsterba@suse.com>
If we are not using space cache v1, we should not create the free space
object or free space inodes. This comes up when we delete the existing
free space objects/inodes when migrating to v2, only to see them get
recreated for every dirtied block group.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When the filesystem transitions from space cache v1 to v2 or to
nospace_cache, it removes the old cached data, but does not remove
the FREE_SPACE items nor the free space inodes they point to. This
doesn't cause any issues besides being a bit inefficient, since these
items no longer do anything useful.
To fix it, when we are mounting, and plan to disable the space cache,
destroy each block group's free space item and free space inode.
The code to remove the items is lifted from the existing use case of
removing the block group, with a light adaptation to handle whether or
not we have already looked up the free space inode.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Signed-off-by: David Sterba <dsterba@suse.com>
If the remount is ro->ro, rw->ro, or rw->rw, we will not create or
clear the free space tree. This can be surprising, so print a warning
to dmesg to make the failure more visible. It is also important to
ensure that the space cache options (SPACE_CACHE, FREE_SPACE_TREE) are
consistent, so ensure those are set to properly match the current on
disk state (which won't be changing).
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
To make the contents of /proc/mounts better match the actual state of
the filesystem, base the display of the space cache mount options off
the contents of the super block rather than the last mount options
passed in. Since there are many scenarios where the mount will ignore a
space cache option, simply showing the passed in option is misleading.
For example, if we mount with -o remount,space_cache=v2 on a read-write
file system without an existing free space tree, we won't build a free
space tree, but /proc/mounts will read space_cache=v2 (until we mount
again and it goes away)
cache_generation is set iff space_cache=v1, FREE_SPACE_TREE is set iff
space_cache=v2, and if neither is the case, we print nospace_cache.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When mounting, btrfs uses the cache_generation in the super block to
determine if space cache v1 is in use. However, by mounting with
nospace_cache or space_cache=v2, it is possible to disable space cache
v1, which does not result in un-setting cache_generation back to 0.
In order to base some logic, like mount option printing in /proc/mounts,
on the current state of the space cache rather than just the values of
the mount option, keep the value of cache_generation consistent with the
status of space cache v1.
We ensure that cache_generation > 0 iff the file system is using
space_cache v1. This requires committing a transaction on any mount
which changes whether we are using v1. (v1->nospace_cache, v1->v2,
nospace_cache->v1, v2->v1).
Since the mechanism for writing out the cache generation is transaction
commit, but we want some finer grained control over when we un-set it,
we can't just rely on the SPACE_CACHE mount option, and introduce an
fs_info flag that mount can use when it wants to unset the generation.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Signed-off-by: David Sterba <dsterba@suse.com>
A user might want to revert to v1 or nospace_cache on a root filesystem,
and much like turning on the free space tree, that can only be done
remounting from ro->rw. Support clearing the free space tree on such
mounts by moving it into the shared remount logic.
Since the CLEAR_CACHE option sticks around across remounts, this change
would result in clearing the tree for ever on every remount, which is
not desirable. To fix that, add CLEAR_CACHE to the oneshot options we
clear at mount end, which has the other bonus of not cluttering the
/proc/mounts output with clear_cache.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Some options only apply during mount time and are cleared at the end
of mount. For now, the example is USEBACKUPROOT, but CLEAR_CACHE also
fits the bill, and this is a preparation patch for also clearing that
option.
One subtlety is that the current code only resets USEBACKUPROOT on rw
mounts, but the option is meaningfully "consumed" by a ro mount, so it
feels appropriate to clear in that case as well. A subsequent read-write
remount would not go through open_ctree, which is the only place that
checks the option, so the change should be benign.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When a user attempts to remount a btrfs filesystem with
'mount -o remount,space_cache=v2', that operation silently succeeds.
Unfortunately, this is misleading, because the remount does not create
the free space tree. /proc/mounts will incorrectly show space_cache=v2,
but on the next mount, the file system will revert to the old
space_cache.
For now, we handle only the easier case, where the existing mount is
read-only and the new mount is read-write. In that case, we can create
the free space tree without contending with the block groups changing
as we go.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If we attempt to create a free space tree while any block groups have
needs_free_space set, we will double add the new free space item
and hit EEXIST. Previously, we only created the free space tree on a new
mount, so we never hit the case, but if we try to create it on a
remount, such block groups could exist and trip us up.
We don't do anything with this field unless the free space tree is
enabled, so there is no harm in not setting it.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Signed-off-by: David Sterba <dsterba@suse.com>
When we mount a rw filesystem, we start the orphan cleanup process in
tree root and filesystem tree. However, when we remount a ro file system
rw, we only clean the former. Move the calls to btrfs_orphan_cleanup()
on tree_root and fs_root to the shared rw mount routine to effectively
add them on ro->rw remount.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Mounting rw and remounting from ro to rw naturally share invariants and
functionality which result in a correctly setup rw filesystem. Luckily,
there is even a strong unity in the code which implements them. In
mount's open_ctree, these operations mostly happen after an early return
for ro file systems, and in remount, they happen in a section devoted to
remounting ro->rw, after some remount specific validation passes.
However, there are unfortunately a few differences. There are small
deviations in the order of some of the operations, remount does not
start orphan cleanup in root_tree or fs_tree, remount does not create
the free space tree, and remount does not handle "one-shot" mount
options like clear_cache and uuid tree rescan.
Since we want to add building the free space tree to remount, and also
to start the same orphan cleanup process on a filesystem mounted as ro
then remounted rw, we would benefit from unifying the logic between the
two code paths.
This patch only lifts the existing common functionality, and leaves a
natural path for fixing the discrepancies.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Boris Burkov <boris@bur.io>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Early on during a transaction commit we acquire the tree_log_mutex and
hold it until after we write the super blocks. But before writing the
extent buffers dirtied by the transaction and the super blocks we unblock
the transaction by setting its state to TRANS_STATE_UNBLOCKED and setting
fs_info->running_transaction to NULL.
This means that after that and before writing the super blocks, new
transactions can start. However if any transaction wants to log an inode,
it will block waiting for the transaction commit to write its dirty
extent buffers and the super blocks because the tree_log_mutex is only
released after those operations are complete, and starting a new log
transaction blocks on that mutex (at start_log_trans()).
Writing the dirty extent buffers and the super blocks can take a very
significant amount of time to complete, but we could allow the tasks
wanting to log an inode to proceed with most of their steps:
1) create the log trees
2) log metadata in the trees
3) write their dirty extent buffers
They only need to wait for the previous transaction commit to complete
(write its super blocks) before they attempt to write their super blocks,
otherwise we could end up with a corrupt filesystem after a crash.
So change start_log_trans() to use the root tree's log_mutex to serialize
for the creation of the log root tree instead of using the tree_log_mutex,
and make btrfs_sync_log() acquire the tree_log_mutex before writing the
super blocks. This allows for inode logging to wait much less time when
there is a previous transaction that is still committing, often not having
to wait at all, as by the time when we try to sync the log the previous
transaction already wrote its super blocks.
This patch belongs to a patch set that is comprised of the following
patches:
btrfs: fix race causing unnecessary inode logging during link and rename
btrfs: fix race that results in logging old extents during a fast fsync
btrfs: fix race that causes unnecessary logging of ancestor inodes
btrfs: fix race that makes inode logging fallback to transaction commit
btrfs: fix race leading to unnecessary transaction commit when logging inode
btrfs: do not block inode logging for so long during transaction commit
The following script that uses dbench was used to measure the impact of
the whole patchset:
$ cat test-dbench.sh
#!/bin/bash
DEV=/dev/nvme0n1
MNT=/mnt/btrfs
MOUNT_OPTIONS="-o ssd"
echo "performance" | \
tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
mkfs.btrfs -f -m single -d single $DEV
mount $MOUNT_OPTIONS $DEV $MNT
dbench -D $MNT -t 300 64
umount $MNT
The test was run on a machine with 12 cores, 64G of ram, using a NVMe
device and a non-debug kernel configuration (Debian's default).
Before patch set:
Operation Count AvgLat MaxLat
----------------------------------------
NTCreateX 11277211 0.250 85.340
Close 8283172 0.002 6.479
Rename 477515 1.935 86.026
Unlink 2277936 0.770 87.071
Deltree 256 15.732 81.379
Mkdir 128 0.003 0.009
Qpathinfo 10221180 0.056 44.404
Qfileinfo 1789967 0.002 4.066
Qfsinfo 1874399 0.003 9.176
Sfileinfo 918589 0.061 10.247
Find 3951758 0.341 54.040
WriteX 5616547 0.047 85.079
ReadX 17676028 0.005 9.704
LockX 36704 0.003 1.800
UnlockX 36704 0.002 0.687
Flush 790541 14.115 676.236
Throughput 1179.19 MB/sec 64 clients 64 procs max_latency=676.240 ms
After patch set:
Operation Count AvgLat MaxLat
----------------------------------------
NTCreateX 12687926 0.171 86.526
Close 9320780 0.002 8.063
Rename 537253 1.444 78.576
Unlink 2561827 0.559 87.228
Deltree 374 11.499 73.549
Mkdir 187 0.003 0.005
Qpathinfo 11500300 0.061 36.801
Qfileinfo 2017118 0.002 7.189
Qfsinfo 2108641 0.003 4.825
Sfileinfo 1033574 0.008 8.065
Find 4446553 0.408 47.835
WriteX 6335667 0.045 84.388
ReadX 19887312 0.003 9.215
LockX 41312 0.003 1.394
UnlockX 41312 0.002 1.425
Flush 889233 13.014 623.259
Throughput 1339.32 MB/sec 64 clients 64 procs max_latency=623.265 ms
+12.7% throughput, -8.2% max latency
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When logging an inode we may often have to fallback to a full transaction
commit, either because a new block group was allocated, there is some case
we can not deal with without a transaction commit or some error like an
ENOMEM happened. However after we fallback to a transaction commit, we
have a time window where we can make the next attempt to log any inode
commit the next transaction unnecessarily, adding additional overhead and
increasing latency.
A sequence of steps that leads to this issue is the following:
1) The current open transaction has a generation of 1000;
2) A new block group is allocated, and as a consequence we must make sure
any attempts to commit a log fallback to a transaction commit, so
btrfs_set_log_full_commit() is called from btrfs_make_block_group().
This sets fs_info->last_trans_log_full_commit to 1000;
3) Task A is holding a handle on transaction 1000 and tries to log inode X.
Once it gets to start_log_trans(), it calls btrfs_need_log_full_commit()
which returns true, since fs_info->last_trans_log_full_commit has a
value of 1000. So we end up returning EAGAIN and propagating it up to
btrfs_sync_file(), where we commit transaction 1000;
4) The transaction commit task (task A) sets the transaction state to
unblocked (TRANS_STATE_UNBLOCKED);
5) Some other task, task B, starts a new transaction with a generation of
1001;
6) Some stuff is done with transaction 1001, some btree blocks COWed, etc;
7) Transaction 1000 has not fully committed yet, we are still writing all
the extent buffers it created;
8) Some new task, task C, starts an fsync of inode Y, gets a handle for
transaction 1001, and it gets to btrfs_log_inode_parent() which does
the following check:
if (fs_info->last_trans_log_full_commit > last_committed) {
ret = 1;
goto end_no_trans;
}
At that point last_trans_log_full_commit has a value of 1000 and
last_committed (value of fs_info->last_trans_committed) has a value of
999, since transaction 1000 has not yet committed - it is either still
writing out dirty extent buffers, its super blocks or unpinning
extents.
As a consequence we return 1, which gets propagated up to
btrfs_sync_file(), which will then call btrfs_commit_transaction()
for transaction 1001.
As a consequence we have an unnecessary second transaction commit, we
previously committed transaction 1000 and now commit transaction 1001
as well, resulting in more overhead and increased latency.
So fix this double transaction commit issue simply by removing that check,
because all we need to do is wait for the previous transaction to finish
its commit, which we already do later when starting the log transaction at
start_log_trans(), because there we acquire the tree_log_mutex lock, which
is held by a transaction commit and only released after the transaction
commits its super blocks.
Another issue that check has is that it reads last_trans_log_full_commit
without using READ_ONCE(), which is incorrect since that member of
struct btrfs_fs_info is always updated with WRITE_ONCE() through the
helper btrfs_set_log_full_commit().
This double transaction commit issue can actually be triggered quite often
in long runs of dbench, since besides the creation of new block groups
that force inode logging to fallback to a transaction commit, there are
cases where dbench asks to fsync a directory which had files in it that
were previously renamed or subdirectories that were removed, resulting in
the inode logging to fallback to a full transaction commit.
This patch belongs to a patch set that is comprised of the following
patches:
btrfs: fix race causing unnecessary inode logging during link and rename
btrfs: fix race that results in logging old extents during a fast fsync
btrfs: fix race that causes unnecessary logging of ancestor inodes
btrfs: fix race that makes inode logging fallback to transaction commit
btrfs: fix race leading to unnecessary transaction commit when logging inode
btrfs: do not block inode logging for so long during transaction commit
Performance results are mentioned in the change log of the last patch.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When logging an inode and the previous transaction is still committing, we
have a time window where we can end up incorrectly think an inode has its
last_unlink_trans field with a value greater than the last transaction
committed, which results in the logging to fallback to a full transaction
commit, which is usually much more expensive than doing a log commit.
The race is described by the following steps:
1) We are at transaction 1000;
2) We modify an inode X (a directory) using transaction 1000 and set its
last_unlink_trans field to 1000, because for example we removed one
of its subdirectories;
3) We create a new inode Y with a dentry in inode X using transaction 1000,
so its generation field is set to 1000;
4) The commit for transaction 1000 is started by task A;
5) The task committing transaction 1000 sets the transaction state to
unblocked, writes the dirty extent buffers and the super blocks, then
unlocks tree_log_mutex;
6) Some task starts a new transaction with a generation of 1001;
7) We do some modification to inode Y (using transaction 1001);
8) The transaction 1000 commit starts unpinning extents. At this point
fs_info->last_trans_committed still has a value of 999;
9) Task B starts an fsync on inode Y, and gets a handle for transaction
1001. When it gets to check_parent_dirs_for_sync() it does the checking
of the ancestor dentries because the following check does not evaluate
to true:
if (S_ISREG(inode->vfs_inode.i_mode) &&
inode->generation <= last_committed &&
inode->last_unlink_trans <= last_committed)
goto out;
The generation value for inode Y is 1000 and last_committed, which has
the value read from fs_info->last_trans_committed, has a value of 999,
so that check evaluates to false and we proceed to check the ancestor
inodes.
Once we get to the first ancestor, inode X, we call
btrfs_must_commit_transaction() on it, which evaluates to true:
static bool btrfs_must_commit_transaction(...)
{
struct btrfs_fs_info *fs_info = inode->root->fs_info;
bool ret = false;
mutex_lock(&inode->log_mutex);
if (inode->last_unlink_trans > fs_info->last_trans_committed) {
/*
* Make sure any commits to the log are forced to be full
* commits.
*/
btrfs_set_log_full_commit(trans);
ret = true;
}
(...)
because inode's X last_unlink_trans has a value of 1000 and
fs_info->last_trans_committed still has a value of 999, it returns
true to check_parent_dirs_for_sync(), making it return 1 which is
propagated up to btrfs_sync_file(), causing it to fallback to a full
transaction commit of transaction 1001.
We should have not fallen back to commit transaction 1001, since inode
X had last_unlink_trans set to 1000 and the super blocks for
transaction 1000 were already written. So while not resulting in a
functional problem, it leads to a lot more work and higher latencies
for a fsync since committing a transaction is usually more expensive
than committing a log (if other filesystem changes happened under that
transaction).
Similar problem happens when logging directories, for the same reason as
btrfs_must_commit_transaction() returns true on an inode with its
last_unlink_trans having the generation of the previous transaction and
that transaction is still committing, unpinning its freed extents.
So fix this by comparing last_unlink_trans with the id of the current
transaction instead of fs_info->last_trans_committed.
This case is often hit when running dbench for a long enough duration, as
it does lots of rename and rmdir operations (both update the field
last_unlink_trans of an inode) and fsyncs of files and directories.
This patch belongs to a patch set that is comprised of the following
patches:
btrfs: fix race causing unnecessary inode logging during link and rename
btrfs: fix race that results in logging old extents during a fast fsync
btrfs: fix race that causes unnecessary logging of ancestor inodes
btrfs: fix race that makes inode logging fallback to transaction commit
btrfs: fix race leading to unnecessary transaction commit when logging inode
btrfs: do not block inode logging for so long during transaction commit
Performance results are mentioned in the change log of the last patch.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When logging an inode and we are checking if we need to log ancestors that
are new, if the previous transaction is still committing we have a time
window where we can unnecessarily log ancestor inodes that were created in
the previous transaction.
The race is described by the following steps:
1) We are at transaction 1000;
2) Directory inode X is created, its generation is set to 1000;
3) The commit for transaction 1000 is started by task A;
4) The task committing transaction 1000 sets the transaction state to
unblocked, writes the dirty extent buffers and the super blocks, then
unlocks tree_log_mutex;
5) Inode Y, a regular file, is created under directory inode X, this
results in starting a new transaction with a generation of 1001;
6) The transaction 1000 commit is unpinning extents. At this point
fs_info->last_trans_committed still has a value of 999;
7) Task B calls fsync on inode Y and gets a handle for transaction 1001;
8) Task B ends up at log_all_new_ancestors() and then because inode Y has
only one hard link, ends up at log_new_ancestors_fast(). There it reads
a value of 999 from fs_info->last_trans_committed, and sees that the
parent inode X has a generation of 1000, so we end up logging inode X:
if (inode->generation > fs_info->last_trans_committed) {
ret = btrfs_log_inode(trans, root, inode,
LOG_INODE_EXISTS, ctx);
(...)
which is not necessary since it was created in the past transaction,
with a generation of 1000, and that transaction has already committed
its super blocks - it's still unpinning extents so it has not yet
updated fs_info->last_trans_committed from 999 to 1000.
So this just causes us to spend more time logging and allocating and
writing more tree blocks for the log tree.
So fix this by comparing an inode's generation with the generation of the
transaction our transaction handle refers to - if the inode's generation
matches the generation of the current transaction than we know it is a
new inode we need to log, otherwise don't log it.
This case is often hit when running dbench for a long enough duration.
This patch belongs to a patch set that is comprised of the following
patches:
btrfs: fix race causing unnecessary inode logging during link and rename
btrfs: fix race that results in logging old extents during a fast fsync
btrfs: fix race that causes unnecessary logging of ancestor inodes
btrfs: fix race that makes inode logging fallback to transaction commit
btrfs: fix race leading to unnecessary transaction commit when logging inode
btrfs: do not block inode logging for so long during transaction commit
Performance results are mentioned in the change log of the last patch.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When logging the extents of an inode during a fast fsync, we have a time
window where we can log extents that are from the previous transaction and
already persisted. This only makes us waste time unnecessarily.
The following sequence of steps shows how this can happen:
1) We are at transaction 1000;
2) An ordered extent E from inode I completes, that is it has gone through
btrfs_finish_ordered_io(), and it set the extent maps' generation to
1000 when we unpin the extent, which is the generation of the current
transaction;
3) The commit for transaction 1000 starts by task A;
4) The task committing transaction 1000 sets the transaction state to
unblocked, writes the dirty extent buffers and the super blocks, then
unlocks tree_log_mutex;
5) Some change is made to inode I, resulting in creation of a new
transaction with a generation of 1001;
6) The transaction 1000 commit starts unpinning extents. At this point
fs_info->last_trans_committed still has a value of 999;
7) Task B starts an fsync on inode I, and when it gets to
btrfs_log_changed_extents() sees the extent map for extent E in the
list of modified extents. It sees the extent map has a generation of
1000 and fs_info->last_trans_committed has a value of 999, so it
proceeds to logging the respective file extent item and all the
checksums covering its range.
So we end up wasting time since the extent was already persisted and
is reachable through the trees pointed to by the super block committed
by transaction 1000.
So just fix this by comparing the extent maps generation against the
generation of the transaction handle - if it is smaller then the id in the
handle, we know the extent was already persisted and we do not need to log
it.
This patch belongs to a patch set that is comprised of the following
patches:
btrfs: fix race causing unnecessary inode logging during link and rename
btrfs: fix race that results in logging old extents during a fast fsync
btrfs: fix race that causes unnecessary logging of ancestor inodes
btrfs: fix race that makes inode logging fallback to transaction commit
btrfs: fix race leading to unnecessary transaction commit when logging inode
btrfs: do not block inode logging for so long during transaction commit
Performance results are mentioned in the change log of the last patch.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When we are doing a rename or a link operation for an inode that was logged
in the previous transaction and that transaction is still committing, we
have a time window where we incorrectly consider that the inode was logged
previously in the current transaction and therefore decide to log it to
update it in the log. The following steps give an example on how this
happens during a link operation:
1) Inode X is logged in transaction 1000, so its logged_trans field is set
to 1000;
2) Task A starts to commit transaction 1000;
3) The state of transaction 1000 is changed to TRANS_STATE_UNBLOCKED;
4) Task B starts a link operation for inode X, and as a consequence it
starts transaction 1001;
5) Task A is still committing transaction 1000, therefore the value stored
at fs_info->last_trans_committed is still 999;
6) Task B calls btrfs_log_new_name(), it reads a value of 999 from
fs_info->last_trans_committed and because the logged_trans field of
inode X has a value of 1000, the function does not return immediately,
instead it proceeds to logging the inode, which should not happen
because the inode was logged in the previous transaction (1000) and
not in the current one (1001).
This is not a functional problem, just wasted time and space logging an
inode that does not need to be logged, contributing to higher latency
for link and rename operations.
So fix this by comparing the inodes' logged_trans field with the
generation of the current transaction instead of comparing with the value
stored in fs_info->last_trans_committed.
This case is often hit when running dbench for a long enough duration, as
it does lots of rename operations.
This patch belongs to a patch set that is comprised of the following
patches:
btrfs: fix race causing unnecessary inode logging during link and rename
btrfs: fix race that results in logging old extents during a fast fsync
btrfs: fix race that causes unnecessary logging of ancestor inodes
btrfs: fix race that makes inode logging fallback to transaction commit
btrfs: fix race leading to unnecessary transaction commit when logging inode
btrfs: do not block inode logging for so long during transaction commit
Performance results are mentioned in the change log of the last patch.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
After removing the inode number cache that was using the free space
cache code, we can remove at least the recalc_thresholds callback from
the ops. Both code and tests use the same callback function. It's moved
before its first use.
The use_bitmaps callback is still needed by tests to create some
extents/bitmap setup.
Signed-off-by: David Sterba <dsterba@suse.com>
Since it's being used solely for the freespace cache unconditionally
set the flags required for it.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Following removal of the ino cache io_ctl_init will be called only on
behalf of the freespace inode. In this case we always want to check
CRCs so conditional code that depended on io_ctl::check_crc can be
removed.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
It's been deprecated since commit b547a88ea5 ("btrfs: start
deprecation of mount option inode_cache") which enumerates the reasons.
A filesystem that uses the feature (mount -o inode_cache) tracks the
inode numbers in bitmaps, that data stay on the filesystem after this
patch. The size is roughly 5MiB for 1M inodes [1], which is considered
small enough to be left there. Removal of the change can be implemented
in btrfs-progs if needed.
[1] https://lore.kernel.org/linux-btrfs/20201127145836.GZ6430@twin.jikos.cz/
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
The former is going away as part of the inode map removal so switch
callers to btrfs_find_free_objectid. No functional changes since with
INODE_MAP disabled (default) find_free_objectid was called anyway.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Those functions are going to be used even after inode cache is removed
so moved them to a more appropriate place.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Since commit 72deb455b5 ("block: remove CONFIG_LBDAF") (5.2) the
sector_t type is u64 on all arches and configs so we don't need to
typecast it. It used to be unsigned long and the result of sector size
shifts were not guaranteed to fit in the type.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Superblock (and its copies) is the only data structure in btrfs which
has a fixed location on a device. Since we cannot overwrite in a
sequential write required zone, we cannot place superblock in the zone.
One easy solution is limiting superblock and copies to be placed only in
conventional zones. However, this method has two downsides: one is
reduced number of superblock copies. The location of the second copy of
superblock is 256GB, which is in a sequential write required zone on
typical devices in the market today. So, the number of superblock and
copies is limited to be two. Second downside is that we cannot support
devices which have no conventional zones at all.
To solve these two problems, we employ superblock log writing. It uses
two adjacent zones as a circular buffer to write updated superblocks.
Once the first zone is filled up, start writing into the second one.
Then, when both zones are filled up and before starting to write to the
first zone again, it reset the first zone.
We can determine the position of the latest superblock by reading write
pointer information from a device. One corner case is when both zones
are full. For this situation, we read out the last superblock of each
zone, and compare them to determine which zone is older.
The following zones are reserved as the circular buffer on ZONED btrfs.
- The primary superblock: zones 0 and 1
- The first copy: zones 16 and 17
- The second copy: zones 1024 or zone at 256GB which is minimum, and
next to it
If these reserved zones are conventional, superblock is written fixed at
the start of the zone without logging.
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Placing both data and metadata in a block group is impossible in ZONED
mode. For data, we can allocate a space for it and write it immediately
after the allocation. For metadata, however, we cannot do that, because
the logical addresses are recorded in other metadata buffers to build up
the trees. As a result, a data buffer can be placed after a metadata
buffer, which is not written yet. Writing out the data buffer will break
the sequential write rule.
Check and disallow MIXED_BG with ZONED mode.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fallocate() is implemented by reserving actual extent instead of
reservations. This can result in exposing the sequential write
constraint of host-managed zoned block devices to the application, which
would break the POSIX semantic for the fallocated file. To avoid this,
report fallocate() as not supported when in ZONED mode for now.
In the future, we may be able to implement "in-memory" fallocate() in
ZONED mode by utilizing space_info->bytes_may_use or similar, so this
returns EOPNOTSUPP.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
NODATACOW implies overwriting the file data on a device, which is
impossible in sequential required zones. Disable NODATACOW globally with
mount option and per-file NODATACOW attribute by masking FS_NOCOW_FL.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
As updates to the space cache v1 are in-place, the space cache cannot be
located over sequential zones and there is no guarantees that the device
will have enough conventional zones to store this cache. Resolve this
problem by disabling completely the space cache v1. This does not
introduce any problems with sequential block groups: all the free space
is located after the allocation pointer and no free space before the
pointer. There is no need to have such cache.
Note: we can technically use free-space-tree (space cache v2) on ZONED
mode. But, since ZONED mode now always allocates extents in a block
group sequentially regardless of underlying device zone type, it's no
use to enable and maintain the tree.
For the same reason, NODATACOW is also disabled.
In summary, ZONED will disable:
| Disabled features | Reason |
|-------------------+-----------------------------------------------------|
| RAID/DUP | Cannot handle two zone append writes to different |
| | zones |
|-------------------+-----------------------------------------------------|
| space_cache (v1) | In-place updating |
| NODATACOW | In-place updating |
|-------------------+-----------------------------------------------------|
| fallocate | Reserved extent will be a write hole |
|-------------------+-----------------------------------------------------|
| MIXED_BG | Allocated metadata region will be write holes for |
| | data writes |
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The zone append write command has a maximum IO size restriction it
accepts. This is because a zone append write command cannot be split, as
we ask the device to place the data into a specific target zone and the
device responds with the actual written location of the data.
Introduce max_zone_append_size to zone_info and fs_info to track the
value, so we can limit all I/O to a zoned block device that we want to
write using the zone append command to the device's limits.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Introduce function btrfs_check_zoned_mode() to check if ZONED flag is
enabled on the file system and if the file system consists of zoned
devices with equal zone size.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Linus Torvalds