summaryrefslogtreecommitdiff
path: root/drivers
AgeCommit message (Collapse)Author
2016-10-07bcache: Don't resort extent pointers when resorting btree nodesKent Overstreet
this is important if we ever decide to allow changing which tier a device belongs to
2016-10-07bcache: mempoolify compressionKent Overstreet
2016-10-07bcache: fix moving queue init for runtime device addKent Overstreet
2016-10-07fix bch_cache_set_add_cacheKent Overstreet
2016-10-07bcache: fix bch_bio_free_pages_pool()Kent Overstreet
2016-10-07bcachefs: fsck improvementsKent Overstreet
Check for various inconsistencies on every mount - it's fast, and we really want to know if anyone is seeing them. Doesn't fix anything yet, because that has its own risk and none of these inconsistencies have been observed in the wild yet.
2016-10-07bcachefs: don't mount incomplete cache setKent Overstreet
2016-10-07bcache: fix divide by zero in journal codeKent Overstreet
2016-10-07bcache: compression + tieringKent Overstreet
2016-10-07bcachefs: fix assertionKent Overstreet
2016-10-07bcache: less bouncing in lz4 pathKent Overstreet
2016-10-07bcache: fix promotionKent Overstreet
2016-10-07bcache: fix promotion locking bugKent Overstreet
promote path uses @k for replace key, but it pointed into unlocked btree node
2016-10-07bcache: fix orig_sectors in bch_read_extent_iter()Kent Overstreet
supposed to use the separate bvec_iter that was passed in
2016-10-07bcache: kill skip parameter to bch_read_extent()Kent Overstreet
2016-10-07bcache: have bch_write() verify bio->bi_sector/sizeKent Overstreet
2016-10-07bcache: fix disk reservations + gcKent Overstreet
2016-10-07bcache: fix a bucket marking thing with cached dataKent Overstreet
2016-10-07bcache: expose option for str hash typeKent Overstreet
2016-10-07bcachefs: avoid creating whiteout dirents unnecessarilyKent Overstreet
2016-10-07bcache: another closure microoptimizationKent Overstreet
2016-10-07bcache: Optimistic spinning for six locksKent Overstreet
Stolen from the mutex code
2016-10-07bcache: six locks: track ownerKent Overstreet
2016-10-07bcache: lockstat for six locksKent Overstreet
2016-10-07bcachefs: minor locking fixKent Overstreet
2016-10-07bcache: six lock improvementsKent Overstreet
Do barriers better, and do fewer atomic ops in lock path
2016-10-07bcache: don't use __six_unlock_type() outside six lock codeKent Overstreet
2016-10-07bcache: more efficient closure_sync()Kent Overstreet
2016-10-07bcache: get rid of a bio_get() in the bch_write() pathKent Overstreet
2016-10-07bcache: optimize bch_check_mark_super()Kent Overstreet
bch_extent_ptr_is_dirty() is expensive
2016-10-07bcache: optimize continue_at_nobarrier()Kent Overstreet
2016-10-07bcache: make sectors_written percpuKent Overstreet
2016-10-07bcachefs: dio write refactoringKent Overstreet
2016-10-07bcache: perf tweaksKent Overstreet
2016-10-07bcache: fix extent_cleanup_crcs()Kent Overstreet
2016-10-07bcachefs: page state improvementsKent Overstreet
2016-10-07bcachefs: transactional i_sectors/i_sizeKent Overstreet
2016-10-07bcache: make transactions work for extentsKent Overstreet
2016-10-07bcache: fix a bug with bch_insert_fixup_extent() racing with start of gcKent Overstreet
2016-10-07bcache: kill fail_if_staleKent Overstreet
2016-10-07bcache: fix pending btree node delete accountingKent Overstreet
2016-10-07bcache: fix a bug in bch_insert_fixup_extent()Kent Overstreet
When we break because we got to the end of the btree node, we'll end up calling extent_insert_advance_pos() when we should have called extent_insert_should_stop() first - we might not have room in the journal reservation anymore
2016-10-07bcache: fix a minor race with sysfs trigger_gcKent Overstreet
2016-10-07bcachefs: fix build error on CONFIG_BCACHE_DEBUG=nKent Overstreet
2016-10-07bcache: mark closure_sync() __schedKent Overstreet
2016-10-07bcache: fix setting bool/uint opts from sysfsKent Overstreet
ugh variable shadowing
2016-10-07bcachefs: fix a deadlockKent Overstreet
pagecache add lock/page lock locking inversion
2016-10-07bcache: better journal pipeliningKent Overstreet
previously, when a journal entry filled up or was otherwise ready to be written, we couldn't start on the next journal entry until all outstanding reservations had finished using the current journal entry. this fixes that - we'll start the next journal entry immediately, and the write will happen whenever the last res_put() happens on the old journal entry.
2016-10-07bcache: verify btree node lock orderingKent Overstreet
2016-10-07bcache: fix journal_reclaim_work() lockingKent Overstreet
it assumed it was only ever called by one thread, but that changed when freezig was fixed