mirror of
https://github.com/torvalds/linux.git
synced 2024-12-12 14:12:51 +00:00
385a82f62a
bcachefs had a transient bug where freespace_initialized was not properly being set, which lead to unexpected use of the early bucket allocator at runtime. This issue has been fixed, but the existence of it uncovered a coherency issue in the early bucket allocation code that is somewhat related to how uncached iterators deal with the key cache. The problem itself manifests as occasional failure of generic/113 due to corruption, often seen as a duplicate backpointer or multiple data types per-bucket error. The immediate cause of the error is a racing bucket allocation along the lines of the following sequence: - Task 1 selects key A in bch2_bucket_alloc_early() and schedules. - Task 2 selects the same key A, but proceeds to complete the allocation and associated I/O, after which it releases the open_bucket. - Task 1 resumes with key A, but does not recognize the bucket is now allocated because the open_bucket has been removed from the hash when it was released in the previous step. This generally shouldn't happen because the allocating task updates the alloc btree key before releasing the bucket. This is not sufficient in this particular instance, however, because an uncached iterator for a cached btree doesn't actually lock the key cache slot when no key exists for a given slot in the cache. Thus the fact that the allocation side updates the cached key means that multiple uncached iters can stumble across the same alloc key and duplicate the bucket allocation as described above. This is something that probably needs a longer term fix in the iterator code. As a short term fix, close the race through explicit use of a cached iterator for likely allocation candidates. We don't want to scan the btree with a cached iterator because that would unnecessarily pollute the cache. This mitigates cache pollution by primarily scanning the tree with an uncached iterator, but closes the race by creating a key cache entry for any prospective slot prior to the bucket allocation attempt (also similar to how _alloc_freelist() works via try_alloc_bucket()). This survives many iterations of generic/113 on a kernel hacked to always use the early bucket allocator. Signed-off-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev> |
||
---|---|---|
.. | ||
acl.c | ||
acl.h | ||
alloc_background.c | ||
alloc_background.h | ||
alloc_foreground.c | ||
alloc_foreground.h | ||
alloc_types.h | ||
backpointers.c | ||
backpointers.h | ||
bbpos_types.h | ||
bbpos.h | ||
bcachefs_format.h | ||
bcachefs_ioctl.h | ||
bcachefs.h | ||
bkey_buf.h | ||
bkey_cmp.h | ||
bkey_methods.c | ||
bkey_methods.h | ||
bkey_sort.c | ||
bkey_sort.h | ||
bkey.c | ||
bkey.h | ||
bset.c | ||
bset.h | ||
btree_cache.c | ||
btree_cache.h | ||
btree_gc.c | ||
btree_gc.h | ||
btree_io.c | ||
btree_io.h | ||
btree_iter.c | ||
btree_iter.h | ||
btree_journal_iter.c | ||
btree_journal_iter.h | ||
btree_key_cache.c | ||
btree_key_cache.h | ||
btree_locking.c | ||
btree_locking.h | ||
btree_trans_commit.c | ||
btree_types.h | ||
btree_update_interior.c | ||
btree_update_interior.h | ||
btree_update.c | ||
btree_update.h | ||
btree_write_buffer_types.h | ||
btree_write_buffer.c | ||
btree_write_buffer.h | ||
buckets_types.h | ||
buckets_waiting_for_journal_types.h | ||
buckets_waiting_for_journal.c | ||
buckets_waiting_for_journal.h | ||
buckets.c | ||
buckets.h | ||
chardev.c | ||
chardev.h | ||
checksum.c | ||
checksum.h | ||
clock_types.h | ||
clock.c | ||
clock.h | ||
compress.c | ||
compress.h | ||
counters.c | ||
counters.h | ||
darray.h | ||
data_update.c | ||
data_update.h | ||
debug.c | ||
debug.h | ||
dirent.c | ||
dirent.h | ||
disk_groups_types.h | ||
disk_groups.c | ||
disk_groups.h | ||
ec_types.h | ||
ec.c | ||
ec.h | ||
errcode.c | ||
errcode.h | ||
error.c | ||
error.h | ||
extent_update.c | ||
extent_update.h | ||
extents_types.h | ||
extents.c | ||
extents.h | ||
eytzinger.h | ||
fifo.h | ||
fs-common.c | ||
fs-common.h | ||
fs-io-buffered.c | ||
fs-io-buffered.h | ||
fs-io-direct.c | ||
fs-io-direct.h | ||
fs-io-pagecache.c | ||
fs-io-pagecache.h | ||
fs-io.c | ||
fs-io.h | ||
fs-ioctl.c | ||
fs-ioctl.h | ||
fs.c | ||
fs.h | ||
fsck.c | ||
fsck.h | ||
inode.c | ||
inode.h | ||
io_misc.c | ||
io_misc.h | ||
io_read.c | ||
io_read.h | ||
io_write_types.h | ||
io_write.c | ||
io_write.h | ||
journal_io.c | ||
journal_io.h | ||
journal_reclaim.c | ||
journal_reclaim.h | ||
journal_sb.c | ||
journal_sb.h | ||
journal_seq_blacklist.c | ||
journal_seq_blacklist.h | ||
journal_types.h | ||
journal.c | ||
journal.h | ||
Kconfig | ||
keylist_types.h | ||
keylist.c | ||
keylist.h | ||
logged_ops.c | ||
logged_ops.h | ||
lru.c | ||
lru.h | ||
Makefile | ||
mean_and_variance_test.c | ||
mean_and_variance.c | ||
mean_and_variance.h | ||
migrate.c | ||
migrate.h | ||
move_types.h | ||
move.c | ||
move.h | ||
movinggc.c | ||
movinggc.h | ||
nocow_locking_types.h | ||
nocow_locking.c | ||
nocow_locking.h | ||
opts.c | ||
opts.h | ||
printbuf.c | ||
printbuf.h | ||
quota_types.h | ||
quota.c | ||
quota.h | ||
rebalance_types.h | ||
rebalance.c | ||
rebalance.h | ||
recovery_types.h | ||
recovery.c | ||
recovery.h | ||
reflink.c | ||
reflink.h | ||
replicas_types.h | ||
replicas.c | ||
replicas.h | ||
sb-clean.c | ||
sb-clean.h | ||
sb-errors_types.h | ||
sb-errors.c | ||
sb-errors.h | ||
sb-members.c | ||
sb-members.h | ||
seqmutex.h | ||
siphash.c | ||
siphash.h | ||
six.c | ||
six.h | ||
snapshot.c | ||
snapshot.h | ||
str_hash.h | ||
subvolume_types.h | ||
subvolume.c | ||
subvolume.h | ||
super_types.h | ||
super-io.c | ||
super-io.h | ||
super.c | ||
super.h | ||
sysfs.c | ||
sysfs.h | ||
tests.c | ||
tests.h | ||
trace.c | ||
trace.h | ||
two_state_shared_lock.c | ||
two_state_shared_lock.h | ||
util.c | ||
util.h | ||
varint.c | ||
varint.h | ||
vstructs.h | ||
xattr.c | ||
xattr.h |