Merge tag 'for-6.20-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
"User visible changes, feature updates:

- when using block size > pag

Merge tag 'for-6.20-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
"User visible changes, feature updates:

- when using block size > page size, enable direct IO

- fallback to buffered IO if the data profile has duplication,
workaround to avoid checksum mismatches on block group profiles
with redundancy, real direct IO is possible on single or RAID0

- redo export of zoned statistics, moved from sysfs to
/proc/pid/mountstats due to size limitations of the former

Experimental features:

- remove offload checksum tunable, intended to find best way to do it
but since we've switched to offload to thread for everything we
don't need it anymore

- initial support for remap-tree feature, a translation layer of
logical block addresses that allow changes without moving/rewriting
blocks to do eg. relocation, or other changes that require COW

Notable fixes:

- automatic removal of accidentally leftover chunks when
free-space-tree is enabled since mkfs.btrfs v6.16.1

- zoned mode:
- do not try to append to conventional zones when RAID is mixing
zoned and conventional drives
- fixup write pointers when mixing zoned and conventional on
DUP/RAID* profiles

- when using squota, relax deletion rules for qgroups with 0 members
to allow easier recovery from accounting bugs, also add more checks
to detect bad accounting

- fix periodic reclaim scanning, properly check boundary conditions
not to trigger it unexpectedly or miss the time to run it

- trim:
- continue after first error
- change reporting to the first detected error
- add more cancellation points
- reduce contention of big device lock that can block other
operations when there's lots of trimmed space

- when chunk allocation is forced (needs experimental build) fix
transaction abort when unexpected space layout is detected

Core:

- switch to crypto library API for checksumming, removed module
dependencies, pointer indirections, etc.

- error handling improvements

- adjust how and where transaction commit or abort are done and are
maybe not necessary

- minor compression optimization to skip single block ranges

- improve how compression folios are handled

- new and updated selftests

- cleanups, refactoring:
- auto-freeing and other automatic variable cleanup conversion
- structure size optimizations
- condition annotations"

* tag 'for-6.20-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (137 commits)
btrfs: get rid of compressed_bio::compressed_folios[]
btrfs: get rid of compressed_folios[] usage for encoded writes
btrfs: get rid of compressed_folios[] usage for compressed read
btrfs: remove the old btrfs_compress_folios() infrastructure
btrfs: switch to btrfs_compress_bio() interface for compressed writes
btrfs: introduce btrfs_compress_bio() helper
btrfs: zlib: introduce zlib_compress_bio() helper
btrfs: zstd: introduce zstd_compress_bio() helper
btrfs: lzo: introduce lzo_compress_bio() helper
btrfs: zoned: factor out the zone loading part into a testable function
btrfs: add cleanup function for btrfs_free_chunk_map
btrfs: tests: add cleanup functions for test specific functions
btrfs: raid56: fix memory leak of btrfs_raid_bio::stripe_uptodate_bitmap
btrfs: tests: add unit tests for pending extent walking functions
btrfs: fix EEXIST abort due to non-consecutive gaps in chunk allocation
btrfs: fix transaction commit blocking during trim of unallocated space
btrfs: handle user interrupt properly in btrfs_trim_fs()
btrfs: preserve first error in btrfs_trim_fs()
btrfs: continue trimming remaining devices on failure
btrfs: do not BUG_ON() in btrfs_remove_block_group()
...

show more ...

btrfs: tests: add unit tests for pending extent walking functions

I ran into another sort of trivial bug in v1 of the patch and concluded
that these functions really ought to be unit tested.

These

btrfs: tests: add unit tests for pending extent walking functions

I ran into another sort of trivial bug in v1 of the patch and concluded
that these functions really ought to be unit tested.

These two functions form the core of searching the chunk allocation
pending extent bitmap and have relatively easily definable semantics, so
unit testing them can help ensure the correctness of chunk allocation.

I also made a minor unrelated fix in volumes.h to properly forward
declare btrfs_space_info. Because of the order of the includes in the
new test, this was actually hitting a latent build warning.

Note:
This is an early example for me of a commit authored in part by an AI
agent, so I wanted to more clear about what I did. I defined a
trivial test and explained the set of tests I wanted to the agent and it
produced the large set of test cases seen here. I then checked each test
case to make sure it matched the description and simplified the
constants and numbers until they looked reasonable to me. I then checked
the looping logic to make sure it made sense to the original spirit of
the trivial test. Finally, carefully combed over all the lines it wrote
to loop over the tests it generated to make sure they followed our code
style guide.

Assisted-by: Claude:claude-opus-4-5
Signed-off-by: Boris Burkov <boris@bur.io>
Signed-off-by: David Sterba <dsterba@suse.com>

show more ...