| Age | Commit message (Collapse) | Author |
|---|
|
git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
Pull btrfs updates from David Sterba:
"Apart from numerous cleanups, there are some performance improvements
and one minor mount option update. There's one more radix-tree
conversion (one remaining), and continued work towards enabling large
folios (almost finished).
Performance:
- extent buffer conversion to xarray gains throughput and runtime
improvements on metadata heavy operations doing writeback (sample
test shows +50% throughput, -33% runtime)
- extent io tree cleanups lead to performance improvements by
avoiding unnecessary searches or repeated searches
- more efficient extent unpinning when committing transaction
(estimated run time improvement 3-5%)
User visible changes:
- remove standalone mount option 'nologreplay', deprecated in 5.9,
replacement is 'rescue=nologreplay'
- in scrub, update reporting, add back device stats message after
detected errors (accidentally removed during recent refactoring)
Core:
- convert extent buffer radix tree to xarray
- in subpage mode, move block perfect compression out of experimental
build
- in zoned mode, introduce sub block groups to allow managing special
block groups, like the one for relocation or tree-log, to handle
some corner cases of ENOSPC
- in scrub, simplify bitmaps for block tracking status
- continued preparations for large folios:
- remove assertions for folio order 0
- add support where missing: compression, buffered write, defrag,
hole punching, subpage, send
- fix fsync of files with no hard links not persisting deletion
- reject tree blocks which are not nodesize aligned, a precaution
from 4.9 times
- move transaction abort calls closer to the error sites
- remove usage of some struct bio_vec internals
- simplifications in extent map
- extent IO cleanups and optimizations
- error handling improvements
- enhanced ASSERT() macro with optional format strings
- cleanups:
- remove unused code
- naming unifications, dropped __, added prefix
- merge similar functions
- use common helpers for various data structures"
* tag 'for-6.16-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (198 commits)
btrfs: move misplaced comment of btrfs_path::keep_locks
btrfs: remove standalone "nologreplay" mount option
btrfs: use a single variable to track return value at btrfs_page_mkwrite()
btrfs: don't return VM_FAULT_SIGBUS on failure to set delalloc for mmap write
btrfs: simplify early error checking in btrfs_page_mkwrite()
btrfs: pass true to btrfs_delalloc_release_space() at btrfs_page_mkwrite()
btrfs: fix wrong start offset for delalloc space release during mmap write
btrfs: fix harmless race getting delayed ref head count when running delayed refs
btrfs: log error codes during failures when writing super blocks
btrfs: simplify error return logic when getting folio at prepare_one_folio()
btrfs: return real error from __filemap_get_folio() calls
btrfs: remove superfluous return value check at btrfs_dio_iomap_begin()
btrfs: fix invalid data space release when truncating block in NOCOW mode
btrfs: update Kconfig option descriptions
btrfs: update list of features built under experimental config
btrfs: send: remove btrfs_debug() calls
btrfs: use boolean for delalloc argument to btrfs_free_reserved_extent()
btrfs: use boolean for delalloc argument to btrfs_free_reserved_bytes()
btrfs: fold error checks when allocating ordered extent and update comments
btrfs: check we grabbed inode reference when allocating an ordered extent
...
|
|
Pull block updates from Jens Axboe:
- ublk updates:
- Add support for updating the size of a ublk instance
- Zero-copy improvements
- Auto-registering of buffers for zero-copy
- Series simplifying and improving GET_DATA and request lookup
- Series adding quiesce support
- Lots of selftests additions
- Various cleanups
- NVMe updates via Christoph:
- add per-node DMA pools and use them for PRP/SGL allocations
(Caleb Sander Mateos, Keith Busch)
- nvme-fcloop refcounting fixes (Daniel Wagner)
- support delayed removal of the multipath node and optionally
support the multipath node for private namespaces (Nilay Shroff)
- support shared CQs in the PCI endpoint target code (Wilfred
Mallawa)
- support admin-queue only authentication (Hannes Reinecke)
- use the crc32c library instead of the crypto API (Eric Biggers)
- misc cleanups (Christoph Hellwig, Marcelo Moreira, Hannes
Reinecke, Leon Romanovsky, Gustavo A. R. Silva)
- MD updates via Yu:
- Fix that normal IO can be starved by sync IO, found by mkfs on
newly created large raid5, with some clean up patches for bdev
inflight counters
- Clean up brd, getting rid of atomic kmaps and bvec poking
- Add loop driver specifically for zoned IO testing
- Eliminate blk-rq-qos calls with a static key, if not enabled
- Improve hctx locking for when a plug has IO for multiple queues
pending
- Remove block layer bouncing support, which in turn means we can
remove the per-node bounce stat as well
- Improve blk-throttle support
- Improve delay support for blk-throttle
- Improve brd discard support
- Unify IO scheduler switching. This should also fix a bunch of lockdep
warnings we've been seeing, after enabling lockdep support for queue
freezing/unfreezeing
- Add support for block write streams via FDP (flexible data placement)
on NVMe
- Add a bunch of block helpers, facilitating the removal of a bunch of
duplicated boilerplate code
- Remove obsolete BLK_MQ pci and virtio Kconfig options
- Add atomic/untorn write support to blktrace
- Various little cleanups and fixes
* tag 'for-6.16/block-20250523' of git://git.kernel.dk/linux: (186 commits)
selftests: ublk: add test for UBLK_F_QUIESCE
ublk: add feature UBLK_F_QUIESCE
selftests: ublk: add test case for UBLK_U_CMD_UPDATE_SIZE
traceevent/block: Add REQ_ATOMIC flag to block trace events
ublk: run auto buf unregisgering in same io_ring_ctx with registering
io_uring: add helper io_uring_cmd_ctx_handle()
ublk: remove io argument from ublk_auto_buf_reg_fallback()
ublk: handle ublk_set_auto_buf_reg() failure correctly in ublk_fetch()
selftests: ublk: add test for covering UBLK_AUTO_BUF_REG_FALLBACK
selftests: ublk: support UBLK_F_AUTO_BUF_REG
ublk: support UBLK_AUTO_BUF_REG_FALLBACK
ublk: register buffer to local io_uring with provided buf index via UBLK_F_AUTO_BUF_REG
ublk: prepare for supporting to register request buffer automatically
ublk: convert to refcount_t
selftests: ublk: make IO & device removal test more stressful
nvme: rename nvme_mpath_shutdown_disk to nvme_mpath_remove_disk
nvme: introduce multipath_always_on module param
nvme-multipath: introduce delayed removal of the multipath head node
nvme-pci: derive and better document max segments limits
nvme-pci: use struct_size for allocation struct nvme_dev
...
|
|
That structure records needed info for block verification (either data
checksum pointer, or expected tree block generation).
But there is also a boolean to tell if this block belongs to a metadata
or not, as the data checksum pointer and expected tree block generation
is already a union, we need a dedicated bit to tell if this block is a
metadata or not.
However such layout means we're wasting 63 bits for x86_64, which is a
huge memory waste.
Thanks to the recent bitmap aggregation, we can easily move this
single-bit-per-block member to a new sub-bitmap.
And since we already have six 16 bits long bitmaps, adding another
bitmap won't even increase any memory usage for x86_64, as we need two
64 bits long anyway.
This will reduce the following memory usages:
- sizeof(struct scrub_sector_verification)
From 16 bytes to 8 bytes on x86_64.
- scrub_stripe::sectors
From 16 * 16 to 16 * 8 bytes.
- Per-device scrub_ctx memory usage
From 128 * (16 * 16) to 128 * (16 * 8), which saves 16KiB memory.
Reviewed-by: Filipe Manana <fdmanana@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>
|
|
Currently we have several small bitmaps inside scrub_stripe:
- extent_sector_bitmap
- error_bitmap
- io_error_bitmap
- csum_error_bitmap
- meta_error_bitmap
- meta_gen_error_bitmap
All those bitmaps are at most 16 bits long, but unsigned long is
either 32 or 64 (more common) bits.
This means we're wasting 1/2 or 3/4 space for each bitmap.
And we can have 128 scrub_stripe for each device, such wasted space adds up
quickly.
Instead of using a single unsigned long for each bitmap, aggregate them
into a larger bitmap, just like what we're doing for subpage support.
This reduces 24 bytes from each scrub_stripe structure on x86_64
systems.
This will need a lot of macros converting direct bitmap/bit operations into
our scrub_stripe specific helpers, but all those helpers are very small
and can be inlined.
So overall the overhead shouldn't be that huge, and we save quite some
memory space.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
When the bytenr doesn't match for a metadata tree block, we will report
it as an csum error, which is incorrect and should be reported as a
metadata error instead.
Fixes: a3ddbaebc7c9 ("btrfs: scrub: introduce a helper to verify one metadata block")
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Currently the following members of scrub_stripe are only utilized for
error reporting:
- init_error_bitmap
- init_nr_io_errors
- init_nr_csum_errors
- init_nr_meta_errors
- init_nr_meta_gen_errors
There is no need to put all those members into scrub_stripe, which take
24 bytes for each stripe, and we have 128 stripes for each device.
Instead introduce a structure, scrub_error_records, and move all above
members into that structure.
And allocate such structure from stack inside
scrub_stripe_read_repair_worker().
Since that function is called from a workqueue context, we have more
than enough stack space for just 24 bytes.
Reviewed-by: Filipe Manana <fdmanana@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>
|
|
[BUG]
Since the migration to the new scrub_stripe interface, scrub no longer
updates the device stats when hitting an error, no matter if it's a read
or checksum mismatch error. E.g:
BTRFS info (device dm-2): scrub: started on devid 1
BTRFS error (device dm-2): unable to fixup (regular) error at logical 13631488 on dev /dev/mapper/test-scratch1 physical 13631488
BTRFS warning (device dm-2): checksum error at logical 13631488 on dev /dev/mapper/test-scratch1, physical 13631488, root 5, inode 257, offset 0, length 4096, links 1 (path: file)
BTRFS error (device dm-2): unable to fixup (regular) error at logical 13631488 on dev /dev/mapper/test-scratch1 physical 13631488
BTRFS warning (device dm-2): checksum error at logical 13631488 on dev /dev/mapper/test-scratch1, physical 13631488, root 5, inode 257, offset 0, length 4096, links 1 (path: file)
BTRFS info (device dm-2): scrub: finished on devid 1 with status: 0
Note there is no line showing the device stats error update.
[CAUSE]
In the migration to the new scrub_stripe interface, we no longer call
btrfs_dev_stat_inc_and_print().
[FIX]
- Introduce a new bitmap for metadata generation errors
* A new bitmap
@meta_gen_error_bitmap is introduced to record which blocks have
metadata generation mismatch errors.
* A new counter for that bitmap
@init_nr_meta_gen_errors, is also introduced to store the number of
generation mismatch errors that are found during the initial read.
This is for the error reporting at scrub_stripe_report_errors().
* New dedicated error message for unrepaired generation mismatches
* Update @meta_gen_error_bitmap if a transid mismatch is hit
- Add btrfs_dev_stat_inc_and_print() calls to the following call sites
* scrub_stripe_report_errors()
* scrub_write_endio()
This is only for the write errors.
This means there is a minor behavior change:
- The timing of device stats error message
Since we concentrate the error messages at
scrub_stripe_report_errors(), the device stats error messages will all
show up in one go, after the detailed scrub error messages:
BTRFS error (device dm-2): unable to fixup (regular) error at logical 13631488 on dev /dev/mapper/test-scratch1 physical 13631488
BTRFS warning (device dm-2): checksum error at logical 13631488 on dev /dev/mapper/test-scratch1, physical 13631488, root 5, inode 257, offset 0, length 4096, links 1 (path: file)
BTRFS error (device dm-2): unable to fixup (regular) error at logical 13631488 on dev /dev/mapper/test-scratch1 physical 13631488
BTRFS warning (device dm-2): checksum error at logical 13631488 on dev /dev/mapper/test-scratch1, physical 13631488, root 5, inode 257, offset 0, length 4096, links 1 (path: file)
BTRFS error (device dm-2): bdev /dev/mapper/test-scratch1 errs: wr 0, rd 0, flush 0, corrupt 1, gen 0
BTRFS error (device dm-2): bdev /dev/mapper/test-scratch1 errs: wr 0, rd 0, flush 0, corrupt 2, gen 0
Fixes: e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror() to scrub_stripe infrastructure")
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Instead of the old @page and @page_offset pair inside scrub, here we can
directly use the virtual address for a sector.
This has the following benefit:
- Simplified parameters
A single @kaddr will repair @page and @page_offset.
- No more unnecessary kmap/kunmap calls
Since all pages utilized by scrub is allocated by scrub, and no
highmem is allowed, we do not need to do any kmap/kunmap.
And add an ASSERT() inside the new scrub_stripe_get_kaddr() to
catch any unexpected highmem page.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Qu Wenruo <wqu@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>
|
|
Move kmapping the page out of btrfs_check_sector_csum().
This allows using bvec_kmap_local() where suitable and reduces the number
of kmap*() calls in the raid56 code.
This also means btrfs_check_sector_csum() will only accept a properly
kmapped address.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Replace the code building a bio from a kernel direct map address and
submitting it synchronously with the bdev_rw_virt helper.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: David Sterba <dsterba@suse.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Link: https://lore.kernel.org/r/20250507120451.4000627-19-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
[BUG]
When trying read-only scrub on a btrfs with rescue=idatacsums mount
option, it will crash with the following call trace:
BUG: kernel NULL pointer dereference, address: 0000000000000208
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
CPU: 1 UID: 0 PID: 835 Comm: btrfs Tainted: G O 6.15.0-rc3-custom+ #236 PREEMPT(full)
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS unknown 02/02/2022
RIP: 0010:btrfs_lookup_csums_bitmap+0x49/0x480 [btrfs]
Call Trace:
<TASK>
scrub_find_fill_first_stripe+0x35b/0x3d0 [btrfs]
scrub_simple_mirror+0x175/0x290 [btrfs]
scrub_stripe+0x5f7/0x6f0 [btrfs]
scrub_chunk+0x9a/0x150 [btrfs]
scrub_enumerate_chunks+0x333/0x660 [btrfs]
btrfs_scrub_dev+0x23e/0x600 [btrfs]
btrfs_ioctl+0x1dcf/0x2f80 [btrfs]
__x64_sys_ioctl+0x97/0xc0
do_syscall_64+0x4f/0x120
entry_SYSCALL_64_after_hwframe+0x76/0x7e
[CAUSE]
Mount option "rescue=idatacsums" will completely skip loading the csum
tree, so that any data read will not find any data csum thus we will
ignore data checksum verification.
Normally call sites utilizing csum tree will check the fs state flag
NO_DATA_CSUMS bit, but unfortunately scrub does not check that bit at all.
This results in scrub to call btrfs_search_slot() on a NULL pointer
and triggered above crash.
[FIX]
Check both extent and csum tree root before doing any tree search.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.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 btrfs_key is defined as objectid/type/offset and the keys are also
printed like that. For better readability, update all key
initializations to match this order.
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
[BUG]
Syzbot reported a crash with the following call trace:
BTRFS info (device loop0): scrub: started on devid 1
BUG: kernel NULL pointer dereference, address: 0000000000000208
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 106e70067 P4D 106e70067 PUD 107143067 PMD 0
Oops: Oops: 0000 [#1] PREEMPT SMP NOPTI
CPU: 1 UID: 0 PID: 689 Comm: repro Kdump: loaded Tainted: G O 6.13.0-rc4-custom+ #206
Tainted: [O]=OOT_MODULE
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS unknown 02/02/2022
RIP: 0010:find_first_extent_item+0x26/0x1f0 [btrfs]
Call Trace:
<TASK>
scrub_find_fill_first_stripe+0x13d/0x3b0 [btrfs]
scrub_simple_mirror+0x175/0x260 [btrfs]
scrub_stripe+0x5d4/0x6c0 [btrfs]
scrub_chunk+0xbb/0x170 [btrfs]
scrub_enumerate_chunks+0x2f4/0x5f0 [btrfs]
btrfs_scrub_dev+0x240/0x600 [btrfs]
btrfs_ioctl+0x1dc8/0x2fa0 [btrfs]
? do_sys_openat2+0xa5/0xf0
__x64_sys_ioctl+0x97/0xc0
do_syscall_64+0x4f/0x120
entry_SYSCALL_64_after_hwframe+0x76/0x7e
</TASK>
[CAUSE]
The reproducer is using a corrupted image where extent tree root is
corrupted, thus forcing to use "rescue=all,ro" mount option to mount the
image.
Then it triggered a scrub, but since scrub relies on extent tree to find
where the data/metadata extents are, scrub_find_fill_first_stripe()
relies on an non-empty extent root.
But unfortunately scrub_find_fill_first_stripe() doesn't really expect
an NULL pointer for extent root, it use extent_root to grab fs_info and
triggered a NULL pointer dereference.
[FIX]
Add an extra check for a valid extent root at the beginning of
scrub_find_fill_first_stripe().
The new error path is introduced by 42437a6386ff ("btrfs: introduce
mount option rescue=ignorebadroots"), but that's pretty old, and later
commit b979547513ff ("btrfs: scrub: introduce helper to find and fill
sector info for a scrub_stripe") changed how we do scrub.
So for kernels older than 6.6, the fix will need manual backport.
Reported-by: syzbot+339e9dbe3a2ca419b85d@syzkaller.appspotmail.com
Link: https://lore.kernel.org/linux-btrfs/67756935.050a0220.25abdd.0a12.GAE@google.com/
Fixes: 42437a6386ff ("btrfs: introduce mount option rescue=ignorebadroots")
Reviewed-by: Anand Jain <anand.jain@oracle.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 map used to be passed to scrub_extent() until
e02ee89baa66c4 ("btrfs: scrub: switch scrub_simple_mirror() to
scrub_stripe infrastructure"), where the scrub implementation was
completely reworked.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The parameter is unused and we can reach sctx from scrub stripe if
needed.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Performing the initial extent sector read on a RAID stripe-tree backed
filesystem with pre-allocated extents will cause the RAID stripe-tree
lookup code to return ENODATA, as pre-allocated extents do not have any
on-disk bytes and thus no RAID stripe-tree entries.
But the current scrub read code marks these extents as errors, because
the lookup fails.
If btrfs_map_block() returns -ENODATA, it means that the call to
btrfs_get_raid_extent_offset() returned -ENODATA, because there is no
entry for the corresponding range in the RAID stripe-tree. But as this
range is in the extent tree it means we've hit a pre-allocated extent. In
this case, don't mark the sector in the stripe's error bitmaps as faulty
and carry on to the next.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Fix some confusing spelling errors that were currently identified,
the details are as follows:
block-group.c: 2800: uncompressible ==> incompressible
extent-tree.c: 3131: EXTEMT ==> EXTENT
extent_io.c: 3124: utlizing ==> utilizing
extent_map.c: 1323: ealier ==> earlier
extent_map.c: 1325: possiblity ==> possibility
fiemap.c: 189: emmitted ==> emitted
fiemap.c: 197: emmitted ==> emitted
fiemap.c: 203: emmitted ==> emitted
transaction.h: 36: trasaction ==> transaction
volumes.c: 5312: filesysmte ==> filesystem
zoned.c: 1977: trasnsaction ==> transaction
Signed-off-by: Shen Lichuan <shenlichuan@vivo.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The variable stop_loop was originally introduced in commit 625f1c8dc66d7
("Btrfs: improve the loop of scrub_stripe"). It was initialized to 0 in
commit 3b080b2564287 ("Btrfs: scrub raid56 stripes in the right way").
However, in a later commit 18d30ab961497 ("btrfs: scrub: use
scrub_simple_mirror() to handle RAID56 data stripe scrub"), the code
that modified stop_loop was removed, making the variable redundant.
Currently, stop_loop is only initialized with 0 and is never used or
modified within the scrub_stripe() function. As a result, this patch
removes the stop_loop variable to clean up the code and eliminate
unnecessary redundancy.
This change has no impact on functionality, as stop_loop was never
utilized in any meaningful way in the final version of the code.
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Riyan Dhiman <riyandhiman14@gmail.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The function name is a bit misleading as it submits the btrfs_bio
(bbio), rename it so we can use btrfs_submit_bio() when an actual bio is
submitted.
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Rename 'btrfs_io_stripe::is_scrub' to 'rst_search_commit_root'. While
'is_scrub' describes the state of the io_stripe (it is a stripe submitted
by scrub) it does not describe the purpose, namely looking at the commit
root when searching RAID stripe-tree entries.
Renaming the stripe to rst_search_commit_root describes this purpose.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Currently sctx->stat.last_physical only got updated in the following
cases:
- When the last stripe of a non-RAID56 chunk is scrubbed
This implies a pitfall, if the last stripe is at the chunk boundary,
and we finished the scrub of the whole chunk, we won't update
last_physical at all until the next chunk.
- When a P/Q stripe of a RAID56 chunk is scrubbed
This leads the following two problems:
- sctx->stat.last_physical is not updated for a almost full chunk
This is especially bad, affecting scrub resume, as the resume would
start from last_physical, causing unnecessary re-scrub.
- "btrfs scrub status" will not report any progress for a long time
Fix the problem by properly updating @last_physical after each stripe is
scrubbed.
And since we're here, for the sake of consistency, use spin lock to
protect the update of @last_physical, just like all the remaining
call sites touching sctx->stat.
Reported-by: Michel Palleau <michel.palleau@gmail.com>
Link: https://lore.kernel.org/linux-btrfs/CAMFk-+igFTv2E8svg=cQ6o3e6CrR5QwgQ3Ok9EyRaEvvthpqCQ@mail.gmail.com/
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.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 there are two locations which need to calculate the real
length of a stripe (which can be at the end of a chunk, and the chunk
size may not always be 64K aligned).
Factor them into a helper as we're going to have a third user soon.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
There is only one caller utilizing the @extra_gfp parameter,
alloc_eb_folio_array(). And in that case the extra_gfp is only assigned
to __GFP_NOFAIL.
Rename the @extra_gfp parameter to @nofail to indicate that.
Reviewed-by: Filipe Manana <fdmanana@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>
|
|
We have several places that attach to the current transaction with
btrfs_attach_transaction_barrier() and then commit the transaction if
there is one. Add a helper and use it to deduplicate this pattern.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
finish_extent_writes_for_zoned()
At finish_extent_writes_for_zoned() we use btrfs_join_transaction() to
catch any running transaction and then commit it. This will however create
a new and empty transaction in case there's no running transaction anymore
(got committed by the transaction kthread or other task for example) or
there's a running transaction finishing its commit and with a state >=
TRANS_STATE_UNBLOCKED. In the former case we don't need to do anything
while in the second case we just need to wait for the transaction to
complete its commit.
So improve this by using btrfs_attach_transaction_barrier() instead, which
does not create a new transaction if there's none running, and if there's
a current transaction that is committing, it will wait for it to fully
commit and not create a new transaction. This helps avoiding creating and
committing empty transactions, saving IO, time and unnecessary rotation of
the backup roots in the super block.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The range is specified only in two ways, we can simplify the case for
the whole filesystem range as a NULL block group parameter.
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
[BUG]
When running btrfs/060 with forced RST feature, it would crash the
following ASSERT() inside scrub_read_endio():
ASSERT(sector_nr < stripe->nr_sectors);
Before that, we would have tree dump from
btrfs_get_raid_extent_offset(), as we failed to find the RST entry for
the range.
[CAUSE]
Inside scrub_submit_extent_sector_read() every time we allocated a new
bbio we immediately called btrfs_map_block() to make sure there was some
RST range covering the scrub target.
But if btrfs_map_block() fails, we immediately call endio for the bbio,
while the bbio is newly allocated, it's completely empty.
Then inside scrub_read_endio(), we go through the bvecs to find
the sector number (as bi_sector is no longer reliable if the bio is
submitted to lower layers).
And since the bio is empty, such bvecs iteration would not find any
sector matching the sector, and return sector_nr == stripe->nr_sectors,
triggering the ASSERT().
[FIX]
Instead of calling btrfs_map_block() after allocating a new bbio, call
btrfs_map_block() first.
Since our only objective of calling btrfs_map_block() is only to update
stripe_len, there is really no need to do that after btrfs_alloc_bio().
This new timing would avoid the problem of handling empty bbio
completely, and in fact fixes a possible race window for the old code,
where if the submission thread is the only owner of the pending_io, the
scrub would never finish (since we didn't decrease the pending_io
counter).
Although the root cause of RST lookup failure still needs to be
addressed.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The following error message is displayed:
../fs/btrfs/scrub.c:2152:9: error: ‘ret’ may be used uninitialized
in this function [-Werror=maybe-uninitialized]"
Compiler version: gcc version: (Debian 10.2.1-6) 10.2.1 20210110
Reviewed-by: Boris Burkov <boris@bur.io>
Signed-off-by: Lu Yao <yaolu@kylinos.cn>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
When btrfs scrub finds an error, it reads mirrors to find correct data. If
all the errors are fixed, sctx->error_bitmap is cleared for the stripe
range. However, in the zoned mode, it runs relocation to repair scrub
errors when the bitmap is *not* empty, which is a flipped condition.
Also, it runs the relocation even if the scrub is read-only. This was
missed by a fix in commit 1f2030ff6e49 ("btrfs: scrub: respect the
read-only flag during repair").
The repair is only necessary when there is a repaired sector and should be
done on read-write scrub. So, tweak the condition for both regular and
zoned case.
Fixes: 54765392a1b9 ("btrfs: scrub: introduce helper to queue a stripe for scrub")
Fixes: 1f2030ff6e49 ("btrfs: scrub: respect the read-only flag during repair")
CC: stable@vger.kernel.org # 6.6+
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
At the moment scrub_supers() doesn't grab the super block's location via
the zoned device aware btrfs_sb_log_location() but via btrfs_sb_offset().
This leads to checksum errors on 'scrub' as we're not accessing the
correct location of the super block.
So use btrfs_sb_log_location() for getting the super blocks location on
scrub.
Reported-by: WA AM <waautomata@gmail.com>
Link: http://lore.kernel.org/linux-btrfs/CANU2Z0EvUzfYxczLgGUiREoMndE9WdQnbaawV5Fv5gNXptPUKw@mail.gmail.com
CC: stable@vger.kernel.org # 5.15+
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The find_first_extent_item() helper looks up an extent item by a key,
allowing to do an inexact search when key->offset is -1. It's never
expected to find such item, as it would break the allowed range of a
extent item offset.
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
[BUG]
If there is an extent beyond chunk boundary, currently RST scrub would
error out.
[CAUSE]
In scrub_submit_extent_sector_read(), we completely rely on
extent_sector_bitmap, which is populated using extent tree.
The extent tree can be corrupted that there is an extent item beyond a
chunk.
In that case, RST scrub would fail and error out.
[FIX]
Despite the extent_sector_bitmap usage, also limit the read to chunk
boundary.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
[BUG]
There is a bug report that, on a ext4-converted btrfs, scrub leads to
various problems, including:
- "unable to find chunk map" errors
BTRFS info (device vdb): scrub: started on devid 1
BTRFS critical (device vdb): unable to find chunk map for logical 2214744064 length 4096
BTRFS critical (device vdb): unable to find chunk map for logical 2214744064 length 45056
This would lead to unrepariable errors.
- Use-after-free KASAN reports:
==================================================================
BUG: KASAN: slab-use-after-free in __blk_rq_map_sg+0x18f/0x7c0
Read of size 8 at addr ffff8881013c9040 by task btrfs/909
CPU: 0 PID: 909 Comm: btrfs Not tainted 6.7.0-x64v3-dbg #11 c50636e9419a8354555555245df535e380563b2b
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 2023.11-2 12/24/2023
Call Trace:
<TASK>
dump_stack_lvl+0x43/0x60
print_report+0xcf/0x640
kasan_report+0xa6/0xd0
__blk_rq_map_sg+0x18f/0x7c0
virtblk_prep_rq.isra.0+0x215/0x6a0 [virtio_blk 19a65eeee9ae6fcf02edfad39bb9ddee07dcdaff]
virtio_queue_rqs+0xc4/0x310 [virtio_blk 19a65eeee9ae6fcf02edfad39bb9ddee07dcdaff]
blk_mq_flush_plug_list.part.0+0x780/0x860
__blk_flush_plug+0x1ba/0x220
blk_finish_plug+0x3b/0x60
submit_initial_group_read+0x10a/0x290 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
flush_scrub_stripes+0x38e/0x430 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
scrub_stripe+0x82a/0xae0 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
scrub_chunk+0x178/0x200 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
scrub_enumerate_chunks+0x4bc/0xa30 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
btrfs_scrub_dev+0x398/0x810 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
btrfs_ioctl+0x4b9/0x3020 [btrfs e57987a360bed82fe8756dcd3e0de5406ccfe965]
__x64_sys_ioctl+0xbd/0x100
do_syscall_64+0x5d/0xe0
entry_SYSCALL_64_after_hwframe+0x63/0x6b
RIP: 0033:0x7f47e5e0952b
- Crash, mostly due to above use-after-free
[CAUSE]
The converted fs has the following data chunk layout:
item 2 key (FIRST_CHUNK_TREE CHUNK_ITEM 2214658048) itemoff 16025 itemsize 80
length 86016 owner 2 stripe_len 65536 type DATA|single
For above logical bytenr 2214744064, it's at the chunk end
(2214658048 + 86016 = 2214744064).
This means btrfs_submit_bio() would split the bio, and trigger endio
function for both of the two halves.
However scrub_submit_initial_read() would only expect the endio function
to be called once, not any more.
This means the first endio function would already free the bbio::bio,
leaving the bvec freed, thus the 2nd endio call would lead to
use-after-free.
[FIX]
- Make sure scrub_read_endio() only updates bits in its range
Since we may read less than 64K at the end of the chunk, we should not
touch the bits beyond chunk boundary.
- Make sure scrub_submit_initial_read() only to read the chunk range
This is done by calculating the real number of sectors we need to
read, and add sector-by-sector to the bio.
Thankfully the scrub read repair path won't need extra fixes:
- scrub_stripe_submit_repair_read()
With above fixes, we won't update error bit for range beyond chunk,
thus scrub_stripe_submit_repair_read() should never submit any read
beyond the chunk.
Reported-by: Rongrong <i@rong.moe>
Fixes: e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror() to scrub_stripe infrastructure")
Tested-by: Rongrong <i@rong.moe>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Currently alloc_extent_buffer() utilizes find_or_create_page() to
allocate one page a time for an extent buffer.
This method has the following disadvantages:
- find_or_create_page() is the legacy way of allocating new pages
With the new folio infrastructure, find_or_create_page() is just
redirected to filemap_get_folio().
- Lacks the way to support higher order (order >= 1) folios
As we can not yet let filemap give us a higher order folio.
This patch would change the workflow by the following way:
Old | new
-----------------------------------+-------------------------------------
| ret = btrfs_alloc_page_array();
for (i = 0; i < num_pages; i++) { | for (i = 0; i < num_pages; i++) {
p = find_or_create_page(); | ret = filemap_add_folio();
/* Attach page private */ | /* Reuse page cache if needed */
/* Reused eb if needed */ |
| /* Attach page private and
| reuse eb if needed */
| }
By this we split the page allocation and private attaching into two
parts, allowing future updates to each part more easily, and migrate to
folio interfaces (especially for possible higher order folios).
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Currently we abuse the extent_map structure for two purposes:
1) To actually represent extents for inodes;
2) To represent chunk mappings.
This is odd and has several disadvantages:
1) To create a chunk map, we need to do two memory allocations: one for
an extent_map structure and another one for a map_lookup structure, so
more potential for an allocation failure and more complicated code to
manage and link two structures;
2) For a chunk map we actually only use 3 fields (24 bytes) of the
respective extent map structure: the 'start' field to have the logical
start address of the chunk, the 'len' field to have the chunk's size,
and the 'orig_block_len' field to contain the chunk's stripe size.
Besides wasting a memory, it's also odd and not intuitive at all to
have the stripe size in a field named 'orig_block_len'.
We are also using 'block_len' of the extent_map structure to contain
the chunk size, so we have 2 fields for the same value, 'len' and
'block_len', which is pointless;
3) When an extent map is associated to a chunk mapping, we set the bit
EXTENT_FLAG_FS_MAPPING on its flags and then make its member named
'map_lookup' point to the associated map_lookup structure. This means
that for an extent map associated to an inode extent, we are not using
this 'map_lookup' pointer, so wasting 8 bytes (on a 64 bits platform);
4) Extent maps associated to a chunk mapping are never merged or split so
it's pointless to use the existing extent map infrastructure.
So add a dedicated data structure named 'btrfs_chunk_map' to represent
chunk mappings, this is basically the existing map_lookup structure with
some extra fields:
1) 'start' to contain the chunk logical address;
2) 'chunk_len' to contain the chunk's length;
3) 'stripe_size' for the stripe size;
4) 'rb_node' for insertion into a rb tree;
5) 'refs' for reference counting.
This way we do a single memory allocation for chunk mappings and we don't
waste memory for them with unused/unnecessary fields from an extent_map.
We also save 8 bytes from the extent_map structure by removing the
'map_lookup' pointer, so the size of struct extent_map is reduced from
144 bytes down to 136 bytes, and we can now have 30 extents map per 4K
page instead of 28.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
There's no reason to open code what btrfs_next_item() does when searching
for extent items at scrub.c:scrub.c:find_first_extent_item(), so remove
the logic to find the next item and use btrfs_next_item() instead, making
the code shorter and less nested code blocks. While at it also fix the
comment to the plural "items" instead of "item" and end it with proper
punctuation.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The recent scrub rewrite forgot to remove the sectors_per_bio in
6.3 in 13a62fd997f0 ("btrfs: scrub: remove scrub_bio structure").
This was found by tool https://github.com/jirislaby/clang-struct .
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
queue_scrub_stripe()
[BUG]
There is a compilation warning reported on commit ae76d8e3e135 ("btrfs:
scrub: fix grouping of read IO"), where gcc (14.0.0 20231022 experimental)
is reporting the following uninitialized variable:
fs/btrfs/scrub.c: In function ‘scrub_simple_mirror.isra’:
fs/btrfs/scrub.c:2075:29: error: ‘found_logical’ may be used uninitialized [-Werror=maybe-uninitialized[https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html#index-Wmaybe-uninitialized]]
2075 | cur_logical = found_logical + BTRFS_STRIPE_LEN;
fs/btrfs/scrub.c:2040:21: note: ‘found_logical’ was declared here
2040 | u64 found_logical;
| ^~~~~~~~~~~~~
[CAUSE]
This is a false alert, as @found_logical is passed as parameter
@found_logical_ret of function queue_scrub_stripe().
As long as queue_scrub_stripe() returned 0, we would update
@found_logical_ret. And if queue_scrub_stripe() returned >0 or <0, the
caller would not utilized @found_logical, thus there should be nothing
wrong.
Although the triggering gcc is still experimental, it looks like the
extra check on "if (found_logical_ret)" can sometimes confuse the
compiler.
Meanwhile the only caller of queue_scrub_stripe() is always passing a
valid pointer, there is no need for such check at all.
[FIX]
Although the report itself is a false alert, we can still make it more
explicit by:
- Replace the check for @found_logical_ret with ASSERT()
- Initialize @found_logical to U64_MAX
- Add one extra ASSERT() to make sure @found_logical got updated
Link: https://lore.kernel.org/linux-btrfs/87fs1x1p93.fsf@gentoo.org/
Fixes: ae76d8e3e135 ("btrfs: scrub: fix grouping of read IO")
Reviewed-by: Anand Jain <anand.jain@oracle.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 the last_trans_committed field of struct btrfs_fs_info is
modified and read without any locking or other protection. For example
early in the fsync path, skip_inode_logging() is called which reads
fs_info->last_trans_committed, but at the same time we can have a
transaction commit completing and updating that field.
In the case of an fsync this is harmless and any data race should be
rare and at most cause an unnecessary logging of an inode.
To avoid data race warnings from tools like KCSAN and other issues such
as load and store tearing (amongst others, see [1]), create helpers to
access the last_trans_committed field of struct btrfs_fs_info using
READ_ONCE() and WRITE_ONCE(), and use these helpers everywhere.
[1] https://lwn.net/Articles/793253/
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
A filesystem that uses the raid stripe tree for logical to physical
address translation can't use the regular scrub path, that reads all
stripes and then checks if a sector is unused afterwards.
When using the raid stripe tree, this will result in lookup errors, as
the stripe tree doesn't know the requested logical addresses.
In case we're scrubbing a filesystem which uses the RAID stripe tree for
multi-device logical to physical address translation, perform an extra
block mapping step to get the real on-disk stripe length from the stripe
tree when scrubbing the sectors.
This prevents a double completion of the btrfs_bio caused by splitting the
underlying bio and ultimately a use-after-free.
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The parameter @need_raid_map is mostly a legacy from the old days where
we don't yet have a solid definition on the @mirror_num, and only
check-integrity was using that parameter, while all other call sites
just pass 1 for that parameter.
Now since we have removed check-integrity functionality, we can also
remove the @need_raid_map parameter.
This change will also remove the ability to read P/Q stripe directly
when passing 0 as @need_raid_map.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The function btrfsic_mount() is part of the deprecated check-integrity
functionality.
Now let's remove the main entry point of check-integrity, and thankfully
most of the check-integrity code is self-contained inside
check-integrity.c, we can safely remove the function without huge
changes to btrfs code base.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
scrub_stripe_read_repair_worker()
Currently the scrub_stripe_read_repair_worker() only does reads to
rebuild the corrupted sectors, it doesn't do any writeback.
The design is mostly to put writeback into a more ordered manner, to
co-operate with dev-replace with zoned mode, which requires every write
to be submitted in their bytenr order.
However the writeback for repaired sectors into the original mirror
doesn't need such strong sync requirement, as it can only happen for
non-zoned devices.
This patch would move the writeback for repaired sectors into
scrub_stripe_read_repair_worker(), which removes two calls sites for
repaired sectors writeback. (one from flush_scrub_stripes(), one from
scrub_raid56_parity_stripe())
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The workqueue fs_info->scrub_worker would go ordered workqueue if it's a
device replace operation.
However the scrub is relying on multiple workers to do data csum
verification, and we always submit several read requests in a row.
Thus there is no need to use ordered workqueue just for dev-replace.
We have extra synchronization (the main thread will always
submit-and-wait for dev-replace writes) to handle it for zoned devices.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
[REGRESSION]
There are several regression reports about the scrub performance with
v6.4 kernel.
On a PCIe 3.0 device, the old v6.3 kernel can go 3GB/s scrub speed, but
v6.4 can only go 1GB/s, an obvious 66% performance drop.
[CAUSE]
Iostat shows a very different behavior between v6.3 and v6.4 kernel:
Device r/s rkB/s rrqm/s %rrqm r_await rareq-sz aqu-sz %util
nvme0n1p3 9731.00 3425544.00 17237.00 63.92 2.18 352.02 21.18 100.00
nvme0n1p3 15578.00 993616.00 5.00 0.03 0.09 63.78 1.32 100.00
The upper one is v6.3 while the lower one is v6.4.
There are several obvious differences:
- Very few read merges
This turns out to be a behavior change that we no longer do bio
plug/unplug.
- Very low aqu-sz
This is due to the submit-and-wait behavior of flush_scrub_stripes(),
and extra extent/csum tree search.
Both behaviors are not that obvious on SATA SSDs, as SATA SSDs have NCQ
to merge the reads, while SATA SSDs can not handle high queue depth well
either.
[FIX]
For now this patch focuses on the read speed fix. Dev-replace replace
speed needs more work.
For the read part, we go two directions to fix the problems:
- Re-introduce blk plug/unplug to merge read requests
This is pretty simple, and the behavior is pretty easy to observe.
This would enlarge the average read request size to 512K.
- Introduce multi-group reads and no longer wait for each group
Instead of the old behavior, which submits 8 stripes and waits for
them, here we would enlarge the total number of stripes to 16 * 8.
Which is 8M per device, the same limit as the old scrub in-flight
bios size limit.
Now every time we fill a group (8 stripes), we submit them and
continue to next stripes.
Only when the full 16 * 8 stripes are all filled, we submit the
remaining ones (the last group), and wait for all groups to finish.
Then submit the repair writes and dev-replace writes.
This should enlarge the queue depth.
This would greatly improve the merge rate (thus read block size) and
queue depth:
Before (with regression, and cached extent/csum path):
Device r/s rkB/s rrqm/s %rrqm r_await rareq-sz aqu-sz %util
nvme0n1p3 20666.00 1318240.00 10.00 0.05 0.08 63.79 1.63 100.00
After (with all patches applied):
nvme0n1p3 5165.00 2278304.00 30557.00 85.54 0.55 441.10 2.81 100.00
i.e. 1287 to 2224 MB/s.
CC: stable@vger.kernel.org # 6.4+
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
One of the bottleneck of the new scrub code is the extra csum tree
search.
The old code would only do the csum tree search for each scrub bio,
which can be as large as 512KiB, thus they can afford to allocate a new
path each time.
But the new scrub code is doing csum tree search for each stripe, which
is only 64KiB, this means we'd better re-use the same csum path during
each search.
This patch would introduce a per-sctx path for csum tree search, as we
don't need to re-allocate the path every time we need to do a csum tree
search.
With this change we can further improve the queue depth and improve the
scrub read performance:
Before (with regression and cached extent tree path):
Device r/s rkB/s rrqm/s %rrqm r_await rareq-sz aqu-sz %util
nvme0n1p3 15875.00 1013328.00 12.00 0.08 0.08 63.83 1.35 100.00
After (with both cached extent/csum tree path):
nvme0n1p3 17759.00 1133280.00 10.00 0.06 0.08 63.81 1.50 100.00
Fixes: e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror() to scrub_stripe infrastructure")
CC: stable@vger.kernel.org # 6.4+
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Since commit e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror()
to scrub_stripe infrastructure"), scrub no longer re-use the same path
for extent tree search.
This can lead to unnecessary extent tree search, especially for the new
stripe based scrub, as we have way more stripes to prepare.
This patch would re-introduce a shared path for extent tree search, and
properly release it when the block group is scrubbed.
This change alone can improve scrub performance slightly by reducing the
time spend preparing the stripe thus improving the queue depth.
Before (with regression):
Device r/s rkB/s rrqm/s %rrqm r_await rareq-sz aqu-sz %util
nvme0n1p3 15578.00 993616.00 5.00 0.03 0.09 63.78 1.32 100.00
After (with this patch):
nvme0n1p3 15875.00 1013328.00 12.00 0.08 0.08 63.83 1.35 100.00
Fixes: e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror() to scrub_stripe infrastructure")
CC: stable@vger.kernel.org # 6.4+
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
The @path in scrub_simple_mirror() is no longer utilized after commit
e02ee89baa66 ("btrfs: scrub: switch scrub_simple_mirror() to scrub_stripe
infrastructure").
Before that commit, we call find_first_extent_item() directly, which
needs a path and that path can be reused. But after that switch commit,
the extent search is done inside queue_scrub_stripe(), which will no
longer accept a path from outside.
So the @path variable can be safely removed.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ remove the stale comment ]
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Variable stripe_nr is being divided by map->num_stripes however the
result is never read. The division and assignment are redundant and
can be removed. Cleans up clang scan build warning:
fs/btrfs/scrub.c:1264:3: warning: Value stored to 'stripe_nr' is
never read [deadcode.DeadStores]
The code is a leftover from 6ded22c1bfe6 ("btrfs: reduce div64 calls by
limiting the number of stripes of a chunk to u32") that converted div64
to normal division, it's the same but previous version did not trigger a
warning.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
