linux.git
18 months agobcachefs: Scale down number of writepoints when low on space
Kent Overstreet [Mon, 5 Nov 2018 02:55:35 +0000 (21:55 -0500)]
bcachefs: Scale down number of writepoints when low on space

this means we don't have to reserve space for them when calculating
filesystem capacity

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix an assertion when rebuilding replicas
Kent Overstreet [Sun, 4 Nov 2018 02:00:50 +0000 (22:00 -0400)]
bcachefs: Fix an assertion when rebuilding replicas

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Rename nofsck opt to fsck
Kent Overstreet [Sun, 4 Nov 2018 01:52:52 +0000 (21:52 -0400)]
bcachefs: Rename nofsck opt to fsck

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix journal replay when replicas sb section missing
Kent Overstreet [Sun, 4 Nov 2018 01:51:31 +0000 (21:51 -0400)]
bcachefs: Fix journal replay when replicas sb section missing

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix bounds checks in bch2_bio_map()
Kent Overstreet [Sun, 4 Nov 2018 00:19:04 +0000 (20:19 -0400)]
bcachefs: fix bounds checks in bch2_bio_map()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Some fixes for building in userspace
Kent Overstreet [Sun, 4 Nov 2018 00:04:54 +0000 (20:04 -0400)]
bcachefs: Some fixes for building in userspace

userspace allocators don't align allocations as nicely as kernel
allocators, which meant that in some cases we weren't allocating big
enough bvec arrays - just make the calculations more rigorous and
explicit to fix it.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix bch2_bkey_print_bfloat
Kent Overstreet [Thu, 1 Nov 2018 20:02:02 +0000 (16:02 -0400)]
bcachefs: fix bch2_bkey_print_bfloat

was popping an assertion in the eytzinger code

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: new avoid mechanism for io retries
Kent Overstreet [Thu, 1 Nov 2018 19:28:45 +0000 (15:28 -0400)]
bcachefs: new avoid mechanism for io retries

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: more key marking refactoring
Kent Overstreet [Thu, 1 Nov 2018 19:21:48 +0000 (15:21 -0400)]
bcachefs: more key marking refactoring

prep work for erasure coding

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: replicas: prep work for stripes
Kent Overstreet [Tue, 30 Oct 2018 18:32:21 +0000 (14:32 -0400)]
bcachefs: replicas: prep work for stripes

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: kill struct bch_replicas_cpu_entry
Kent Overstreet [Tue, 30 Oct 2018 18:14:19 +0000 (14:14 -0400)]
bcachefs: kill struct bch_replicas_cpu_entry

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: add functionality for heaps to update backpointers
Kent Overstreet [Sun, 21 Oct 2018 20:32:51 +0000 (16:32 -0400)]
bcachefs: add functionality for heaps to update backpointers

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: btree gc refactoring
Kent Overstreet [Sun, 21 Oct 2018 14:56:11 +0000 (10:56 -0400)]
bcachefs: btree gc refactoring

prep work for erasure coding

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: BCH_EXTENT_ENTRY_TYPES()
Kent Overstreet [Sun, 30 Sep 2018 22:39:20 +0000 (18:39 -0400)]
bcachefs: BCH_EXTENT_ENTRY_TYPES()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: bch2_extent_ptr_decoded_append()
Kent Overstreet [Fri, 28 Sep 2018 01:08:39 +0000 (21:08 -0400)]
bcachefs: bch2_extent_ptr_decoded_append()

This new helper for the move path avoids creating a new CRC entry when
we already have one that matches the pointer being added.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: bch2_extent_drop_ptrs()
Kent Overstreet [Sun, 30 Sep 2018 22:28:23 +0000 (18:28 -0400)]
bcachefs: bch2_extent_drop_ptrs()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: extent_for_each_ptr_decode()
Kent Overstreet [Fri, 28 Sep 2018 01:08:39 +0000 (21:08 -0400)]
bcachefs: extent_for_each_ptr_decode()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: kill bch_extent_crc_type
Kent Overstreet [Tue, 2 Oct 2018 20:40:12 +0000 (16:40 -0400)]
bcachefs: kill bch_extent_crc_type

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: extent_ptr_decoded
Kent Overstreet [Tue, 2 Oct 2018 15:03:39 +0000 (11:03 -0400)]
bcachefs: extent_ptr_decoded

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix missing include
Kent Overstreet [Fri, 12 Oct 2018 18:57:57 +0000 (14:57 -0400)]
bcachefs: fix missing include

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix a spurious gcc warning
Kent Overstreet [Fri, 12 Oct 2018 18:53:25 +0000 (14:53 -0400)]
bcachefs: fix a spurious gcc warning

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Allocation code refactoring
Kent Overstreet [Sat, 6 Oct 2018 08:12:42 +0000 (04:12 -0400)]
bcachefs: Allocation code refactoring

bch2_alloc_sectors_start() was a nightmare to work with - it's got some
tricky stuff to do, since it wants to use the buckets the writepoint
already has, unless they're not in the target it wants to write to,
unless it can't allocate from any other devices in which case it will
use those buckets if it has to - et cetera.

This restructures the code to start with a new empty list of open
buckets we're going to use for the new allocation, pulling buckets from
the write point's list as we decide that we really are going to use
them - making the code somewhat more functional and drastically easier
to understand.

Also fixes a bug where we could end up waiting on c->freelist_wait
(because allocating from one device failed) but return success from
bch2_bucket_alloc(), because allocating from a different device
succeeded.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Split out alloc_background.c
Kent Overstreet [Sat, 6 Oct 2018 04:46:55 +0000 (00:46 -0400)]
bcachefs: Split out alloc_background.c

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix failure to suspend
Kent Overstreet [Mon, 1 Oct 2018 04:33:42 +0000 (00:33 -0400)]
bcachefs: Fix failure to suspend

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix suspend when moving data faster than ratelimit
Kent Overstreet [Wed, 26 Sep 2018 03:27:57 +0000 (23:27 -0400)]
bcachefs: Fix suspend when moving data faster than ratelimit

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix bch2_acl_chmod()
Kent Overstreet [Fri, 21 Sep 2018 21:37:13 +0000 (17:37 -0400)]
bcachefs: fix bch2_acl_chmod()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix a deadlock
Kent Overstreet [Sat, 15 Sep 2018 21:57:22 +0000 (17:57 -0400)]
bcachefs: Fix a deadlock

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix a divide
Kent Overstreet [Thu, 6 Sep 2018 21:09:07 +0000 (17:09 -0400)]
bcachefs: fix a divide

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: make fsck spew less
Kent Overstreet [Tue, 28 Aug 2018 22:54:42 +0000 (18:54 -0400)]
bcachefs: make fsck spew less

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Dirent repair code
Kent Overstreet [Tue, 21 Aug 2018 23:42:00 +0000 (19:42 -0400)]
bcachefs: Dirent repair code

There was a bug for awhile in previous kernels where we weren't
computing dirent name lengths correctly and we weren't zeroing out
padding at the end of dirents (due to struct bch_dirent changing size by
adding __attribute__((aligned)), and not updating other code to use
offsetof).

This patch fixes dirents with junk at the end, by going off of the
dirent's hash.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix a btree iter bug when iter pos == POS_MAX
Kent Overstreet [Tue, 21 Aug 2018 21:38:41 +0000 (17:38 -0400)]
bcachefs: Fix a btree iter bug when iter pos == POS_MAX

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Comparison function cleanups
Kent Overstreet [Tue, 21 Aug 2018 20:30:14 +0000 (16:30 -0400)]
bcachefs: Comparison function cleanups

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Prioritize fragmentation in bucket allocator
Kent Overstreet [Tue, 21 Aug 2018 19:19:33 +0000 (15:19 -0400)]
bcachefs: Prioritize fragmentation in bucket allocator

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Pass around bset_tree less
Kent Overstreet [Sat, 11 Aug 2018 23:12:05 +0000 (19:12 -0400)]
bcachefs: Pass around bset_tree less

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: kill extent_insert_hook
Kent Overstreet [Wed, 8 Aug 2018 23:53:30 +0000 (19:53 -0400)]
bcachefs: kill extent_insert_hook

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: kill i_sectors_hook
Kent Overstreet [Sun, 5 Aug 2018 21:48:00 +0000 (17:48 -0400)]
bcachefs: kill i_sectors_hook

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: convert fcollapse to bch2_extent_update()
Kent Overstreet [Sat, 11 Aug 2018 21:26:11 +0000 (17:26 -0400)]
bcachefs: convert fcollapse to bch2_extent_update()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: convert fpunch to bch2_extent_update()
Kent Overstreet [Thu, 9 Aug 2018 01:11:43 +0000 (21:11 -0400)]
bcachefs: convert fpunch to bch2_extent_update()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: convert truncate to bch2_extent_update()
Kent Overstreet [Thu, 9 Aug 2018 01:09:31 +0000 (21:09 -0400)]
bcachefs: convert truncate to bch2_extent_update()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: convert bchfs_write_index_update() to bch2_extent_update()
Kent Overstreet [Wed, 8 Aug 2018 22:42:04 +0000 (18:42 -0400)]
bcachefs: convert bchfs_write_index_update() to bch2_extent_update()

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: bch2_extent_trim_atomic()
Kent Overstreet [Sun, 5 Aug 2018 21:46:41 +0000 (17:46 -0400)]
bcachefs: bch2_extent_trim_atomic()

Prep work for extents insert hook removal

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: mempoolify btree_trans
Kent Overstreet [Thu, 9 Aug 2018 01:22:46 +0000 (21:22 -0400)]
bcachefs: mempoolify btree_trans

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: BTREE_INSERT_JOURNAL_RES_FULL is no longer possible
Kent Overstreet [Sun, 5 Aug 2018 19:21:52 +0000 (15:21 -0400)]
bcachefs: BTREE_INSERT_JOURNAL_RES_FULL is no longer possible

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: extent_squash() can no longer fail
Kent Overstreet [Sun, 5 Aug 2018 19:28:29 +0000 (15:28 -0400)]
bcachefs: extent_squash() can no longer fail

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: make struct btree_iter a bit smaller
Kent Overstreet [Sun, 5 Aug 2018 18:41:29 +0000 (14:41 -0400)]
bcachefs: make struct btree_iter a bit smaller

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: lift ordering restriction on 0 size extents
Kent Overstreet [Fri, 22 Jul 2016 03:05:06 +0000 (19:05 -0800)]
bcachefs: lift ordering restriction on 0 size extents

This lifts the restriction that 0 size extents must not overlap with
other extents, which means we can now sort extents and non extents the
same way, and will let us simplify a bunch of other stuff as well.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: extent unit tests
Kent Overstreet [Thu, 2 Aug 2018 03:03:41 +0000 (23:03 -0400)]
bcachefs: extent unit tests

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: bkey_written()
Kent Overstreet [Mon, 6 Aug 2018 02:23:44 +0000 (22:23 -0400)]
bcachefs: bkey_written()

also cleanups of btree node offsets

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: improved rw_aux_tree_bsearch()
Kent Overstreet [Mon, 6 Aug 2018 02:34:03 +0000 (22:34 -0400)]
bcachefs: improved rw_aux_tree_bsearch()

shouldn't be any reason for an actual binary search here

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Factor out btree_key_can_insert()
Kent Overstreet [Fri, 3 Aug 2018 23:41:44 +0000 (19:41 -0400)]
bcachefs: Factor out btree_key_can_insert()

working on getting rid of all the reasons bch2_insert_fixup_extent() can
fail/stop partway, which is needed for other refactorings.

One of the reasons we could have to bail out is if we're splitting a
compressed extent we might need to add to our disk reservation - but we
can check that before actually starting the insert.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: BCH_SB_RESERVE_BYTES
Kent Overstreet [Tue, 24 Jul 2018 18:55:05 +0000 (14:55 -0400)]
bcachefs: BCH_SB_RESERVE_BYTES

Add an option, gc_reserve_bytes, to set the copygc reserve as a size
instead of a percent

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Better calculation of copygc threshold
Kent Overstreet [Wed, 1 Aug 2018 18:26:55 +0000 (14:26 -0400)]
bcachefs: Better calculation of copygc threshold

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Change how replicated data is accounted
Kent Overstreet [Tue, 24 Jul 2018 20:42:49 +0000 (16:42 -0400)]
bcachefs: Change how replicated data is accounted

Due to compression, the different replicas of a replicated extent don't
necessarily have to take up the same amount of space - so replicated
data sector counts shouldn't be stored divided by the number of
replicas.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Account for internal fragmentation better
Kent Overstreet [Tue, 24 Jul 2018 18:54:39 +0000 (14:54 -0400)]
bcachefs: Account for internal fragmentation better

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: kill s_alloc, use bch_data_type
Kent Overstreet [Tue, 24 Jul 2018 17:33:07 +0000 (13:33 -0400)]
bcachefs: kill s_alloc, use bch_data_type

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: bch2_mark_key() now takes bch_data_type
Kent Overstreet [Tue, 24 Jul 2018 16:59:13 +0000 (12:59 -0400)]
bcachefs: bch2_mark_key() now takes bch_data_type

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix an assertion in the btree node merge path
Kent Overstreet [Tue, 24 Jul 2018 20:42:27 +0000 (16:42 -0400)]
bcachefs: Fix an assertion in the btree node merge path

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix locking in allocator thread
Kent Overstreet [Tue, 24 Jul 2018 23:45:22 +0000 (19:45 -0400)]
bcachefs: Fix locking in allocator thread

gc lock must be held while invalidating buckets - fixes
"1f7a95698e bcachefs: Invalidate buckets when writing to alloc btree"

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix bch2_val_to_text()
Kent Overstreet [Mon, 23 Jul 2018 13:13:07 +0000 (09:13 -0400)]
bcachefs: fix bch2_val_to_text()

was returning wrong value

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: minor fsync fix
Kent Overstreet [Mon, 23 Jul 2018 11:53:29 +0000 (07:53 -0400)]
bcachefs: minor fsync fix

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Assorted journal refactoring
Kent Overstreet [Mon, 23 Jul 2018 11:52:00 +0000 (07:52 -0400)]
bcachefs: Assorted journal refactoring

Also improve error reporting - only return an error from
bch2_journal_flush_seq() if we had an error writing that entry (i.e. not
if there was an error with a newer entry).

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix last_seq_ondisk
Kent Overstreet [Mon, 23 Jul 2018 11:38:06 +0000 (07:38 -0400)]
bcachefs: fix last_seq_ondisk

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix mtime/ctime update on truncate
Kent Overstreet [Mon, 23 Jul 2018 09:48:53 +0000 (05:48 -0400)]
bcachefs: fix mtime/ctime update on truncate

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix fsync after create
Kent Overstreet [Mon, 23 Jul 2018 09:48:35 +0000 (05:48 -0400)]
bcachefs: fix fsync after create

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix nbuckets usage on device resize
Kent Overstreet [Mon, 23 Jul 2018 09:28:40 +0000 (05:28 -0400)]
bcachefs: fix nbuckets usage on device resize

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Invalidate buckets when writing to alloc btree
Kent Overstreet [Sun, 22 Jul 2018 14:43:01 +0000 (10:43 -0400)]
bcachefs: Invalidate buckets when writing to alloc btree

Prep work for persistent alloc information. Refactoring also lets us
make free_inc much smaller, which means a lot fewer buckets stranded on
freelists.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: kill bucket mark sector count saturation
Kent Overstreet [Sun, 22 Jul 2018 10:10:52 +0000 (06:10 -0400)]
bcachefs: kill bucket mark sector count saturation

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: don't call bch2_bucket_seq_cleanup from journal_buf_switch
Kent Overstreet [Sun, 22 Jul 2018 02:57:20 +0000 (22:57 -0400)]
bcachefs: don't call bch2_bucket_seq_cleanup from journal_buf_switch

journal_buf_switch is called from the foreground when getting a journal
reservation and thus is somewhat latency sensitive;
bch2_bucket_seq_cleanup has to run infrequently but is a bit expensive
when it does run.

Call it from the journal write path instead, and punt the journal write
to worqueue context.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix an assertion
Kent Overstreet [Sun, 22 Jul 2018 17:04:00 +0000 (13:04 -0400)]
bcachefs: Fix an assertion

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: fix rename + fsync
Kent Overstreet [Sun, 22 Jul 2018 17:15:51 +0000 (13:15 -0400)]
bcachefs: fix rename + fsync

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Use ei_update_lock consistently
Kent Overstreet [Sat, 21 Jul 2018 02:23:42 +0000 (22:23 -0400)]
bcachefs: Use ei_update_lock consistently

This is prep work for using deferred btree updates for inode updates -
the way inodes are done now we're relying on btree locking for ei_inode
and ei_update_lock could probably be removed, but it'll actually be
needed when we switch to deferred updates.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: bch2_trans_update() now takes struct btree_insert_entry
Kent Overstreet [Tue, 17 Jul 2018 19:28:11 +0000 (15:28 -0400)]
bcachefs: bch2_trans_update() now takes struct btree_insert_entry

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix mtime/ctime updates
Kent Overstreet [Tue, 17 Jul 2018 18:12:42 +0000 (14:12 -0400)]
bcachefs: Fix mtime/ctime updates

Also make inode flags consistent with how the rest of the inode is
updated

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Simplify bch2_write_inode_trans, fix lockdep splat
Kent Overstreet [Tue, 17 Jul 2018 18:03:47 +0000 (14:03 -0400)]
bcachefs: Simplify bch2_write_inode_trans, fix lockdep splat

ei_update_lock isn't currently needed for write inode (but it will be
needed again when deferred btree updates are used for inode updates)

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: add bch_verbose() statements for shutdown
Kent Overstreet [Sat, 21 Jul 2018 07:56:57 +0000 (03:56 -0400)]
bcachefs: add bch_verbose() statements for shutdown

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix a use after free in the journal code
Kent Overstreet [Tue, 17 Jul 2018 16:19:14 +0000 (12:19 -0400)]
bcachefs: Fix a use after free in the journal code

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Fix device add
Kent Overstreet [Sat, 21 Jul 2018 02:08:17 +0000 (22:08 -0400)]
bcachefs: Fix device add

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: trace transaction restarts
Kent Overstreet [Fri, 13 Jul 2018 03:30:45 +0000 (23:30 -0400)]
bcachefs: trace transaction restarts

exceptionally crappy "tracing", but it's a start at documenting the
places restarts can be triggered

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Convert raw uses of bch2_btree_iter_link() to new transactions
Kent Overstreet [Thu, 12 Jul 2018 23:19:41 +0000 (19:19 -0400)]
bcachefs: Convert raw uses of bch2_btree_iter_link() to new transactions

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Only check inode i_nlink during full fsck
Kent Overstreet [Sun, 15 Jul 2018 01:06:51 +0000 (21:06 -0400)]
bcachefs: Only check inode i_nlink during full fsck

Now that all filesystem operatinos that manipulate the filesystem
heirachy and i_nlink are fully atomic, we can add a feature bit to
indicate i_nlink doesn't need to be checked.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcachefs: Initial commit
Kent Overstreet [Fri, 17 Mar 2017 06:18:50 +0000 (22:18 -0800)]
bcachefs: Initial commit

Initially forked from drivers/md/bcache, bcachefs is a new copy-on-write
filesystem with every feature you could possibly want.

Website: https://bcachefs.org

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agoMAINTAINERS: Add entry for bcachefs
Kent Overstreet [Wed, 26 Apr 2023 16:27:51 +0000 (12:27 -0400)]
MAINTAINERS: Add entry for bcachefs

bcachefs is a new copy-on-write filesystem; add a MAINTAINERS entry for
it.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agoobjtool: Add bcachefs noreturns
Kent Overstreet [Sun, 10 Sep 2023 00:56:00 +0000 (20:56 -0400)]
objtool: Add bcachefs noreturns

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agolib/generic-radix-tree.c: Add peek_prev()
Kent Overstreet [Tue, 12 Sep 2023 05:17:22 +0000 (01:17 -0400)]
lib/generic-radix-tree.c: Add peek_prev()

This patch adds genradix_peek_prev(), genradix_iter_rewind(), and
genradix_for_each_reverse(), for iterating backwards over a generic
radix tree.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agolib/generic-radix-tree.c: Don't overflow in peek()
Kent Overstreet [Sat, 13 Feb 2021 01:11:25 +0000 (20:11 -0500)]
lib/generic-radix-tree.c: Don't overflow in peek()

When we started spreading new inode numbers throughout most of the 64
bit inode space, that triggered some corner case bugs, in particular
some integer overflows related to the radix tree code. Oops.

Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
18 months agoMAINTAINERS: Add entry for generic-radix-tree
Kent Overstreet [Wed, 26 Apr 2023 16:27:51 +0000 (12:27 -0400)]
MAINTAINERS: Add entry for generic-radix-tree

lib/generic-radix-tree.c is a simple radix tree that supports storing
arbitrary types. Add a maintainers entry for it.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agoclosures: Add a missing include
Kent Overstreet [Sun, 5 Mar 2023 03:45:27 +0000 (22:45 -0500)]
closures: Add a missing include

Fixes building in userspace.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agoclosures: closure_nr_remaining()
Kent Overstreet [Sat, 4 Mar 2023 07:39:39 +0000 (02:39 -0500)]
closures: closure_nr_remaining()

Factor out a new helper, which returns the number of events outstanding.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agoclosures: closure_wait_event()
Kent Overstreet [Sat, 9 Dec 2017 17:42:44 +0000 (12:42 -0500)]
closures: closure_wait_event()

Like wait_event() - except, because it uses closures and closure
waitlists it doesn't have the restriction on modifying task state inside
the condition check, like wait_event() does.

Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Acked-by: Coly Li <colyli@suse.de>
18 months agoMAINTAINERS: Add entry for closures
Kent Overstreet [Wed, 26 Apr 2023 16:27:51 +0000 (12:27 -0400)]
MAINTAINERS: Add entry for closures

closures, from bcache, are async widgets with a variety of uses.
bcachefs also uses them, so they're being moved to lib/; mark them as
maintained.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Acked-by: Coly Li <colyli@suse.de>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
18 months agobcache: move closures to lib/
Kent Overstreet [Sat, 18 Mar 2017 00:35:23 +0000 (16:35 -0800)]
bcache: move closures to lib/

Prep work for bcachefs - being a fork of bcache it also uses closures

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Acked-by: Coly Li <colyli@suse.de>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
18 months agolocking: export contention tracepoints for bcachefs six locks
Brian Foster [Mon, 14 Aug 2023 13:04:50 +0000 (09:04 -0400)]
locking: export contention tracepoints for bcachefs six locks

The bcachefs implementation of six locks is intended to land in
generic locking code in the long term, but has been pulled into the
bcachefs subsystem for internal use for the time being. This code
lift breaks the bcachefs module build as six locks depend a couple
of the generic locking tracepoints. Export these tracepoint symbols
for bcachefs.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
19 months agolib: Export errname
Kent Overstreet [Tue, 25 Apr 2023 18:45:28 +0000 (14:45 -0400)]
lib: Export errname

errname() returns the name of an errcode; this functionality is
otherwise only available for error pointers via %pE - bcachefs uses this
for better error messages.

Signed-off-by: Christopher James Halse Rogers <raof@ubuntu.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
19 months agolib/string_helpers: string_get_size() now returns characters wrote
Kent Overstreet [Mon, 25 Apr 2022 19:26:28 +0000 (15:26 -0400)]
lib/string_helpers: string_get_size() now returns characters wrote

printbuf now needs to know the number of characters that would have been
written if the buffer was too small, like snprintf(); this changes
string_get_size() to return the the return value of snprintf().

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
19 months agostacktrace: Export stack_trace_save_tsk
Christopher James Halse Rogers [Mon, 27 Jun 2022 00:45:12 +0000 (10:45 +1000)]
stacktrace: Export stack_trace_save_tsk

The bcachefs module wants it, and there doesn't seem to be any
reason it shouldn't be exported like the other functions.

Signed-off-by: Christopher James Halse Rogers <raof@ubuntu.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
19 months agofs: factor out d_mark_tmpfile()
Kent Overstreet [Tue, 10 Jul 2018 03:27:33 +0000 (23:27 -0400)]
fs: factor out d_mark_tmpfile()

New helper for bcachefs - bcachefs doesn't want the
inode_dec_link_count() call that d_tmpfile does, it handles i_nlink on
its own atomically with other btree updates

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <brauner@kernel.org>
Cc: linux-fsdevel@vger.kernel.org
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christian Brauner <brauner@kernel.org>
19 months agosched: Add task_struct->faults_disabled_mapping
Kent Overstreet [Wed, 16 Oct 2019 19:03:50 +0000 (15:03 -0400)]
sched: Add task_struct->faults_disabled_mapping

There has been a long standing page cache coherence bug with direct IO.
This provides part of a mechanism to fix it, currently just used by
bcachefs but potentially worth promoting to the VFS.

Direct IO evicts the range of the pagecache being read or written to.

For reads, we need dirty pages to be written to disk, so that the read
doesn't return stale data. For writes, we need to evict that range of
the pagecache so that it's not stale after the write completes.

However, without a locking mechanism to prevent those pages from being
re-added to the pagecache - by a buffered read or page fault - page
cache inconsistency is still possible.

This isn't necessarily just an issue for userspace when they're playing
games; filesystems may hang arbitrary state off the pagecache, and so
page cache inconsistency may cause real filesystem bugs, depending on
the filesystem. This is less of an issue for iomap based filesystems,
but e.g. buffer heads caches disk block mappings (!) and attaches them
to the pagecache, and bcachefs attaches disk reservations to pagecache
pages.

This issue has been hard to fix, because
 - we need to add a lock (henceforth called pagecache_add_lock), which
   would be held for the duration of the direct IO
 - page faults add pages to the page cache, thus need to take the same
   lock
 - dio -> gup -> page fault thus can deadlock

And we cannot enforce a lock ordering with this lock, since userspace
will be controlling the lock ordering (via the fd and buffer arguments
to direct IOs), so we need a different method of deadlock avoidance.

We need to tell the page fault handler that we're already holding a
pagecache_add_lock, and since plumbing it through the entire gup() path
would be highly impractical this adds a field to task_struct.

Then the full method is:
 - in the dio path, when we first take the pagecache_add_lock, note the
   mapping in the current task_struct
 - in the page fault handler, if faults_disabled_mapping is set, we
   check if it's the same mapping as the one we're taking a page fault
   for, and if so return an error.

   Then we check lock ordering: if there's a lock ordering violation and
   trylock fails, we'll have to cycle the locks and return an error that
   tells the DIO path to retry: faults_disabled_mapping is also used for
   signalling "locks were dropped, please retry".

Also relevant to this patch: mapping->invalidate_lock.
mapping->invalidate_lock provides most of the required semantics - it's
used by truncate/fallocate to block pages being added to the pagecache.
However, since it's a rwsem, direct IOs would need to take the write
side in order to block page cache adds, and would then be exclusive with
each other - we'll need a new type of lock to pair with this approach.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Jan Kara <jack@suse.cz>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: linux-fsdevel@vger.kernel.org
Cc: Andreas Grünbacher <andreas.gruenbacher@gmail.com>
19 months agoLinux 6.6-rc1
Linus Torvalds [Sun, 10 Sep 2023 23:28:41 +0000 (16:28 -0700)]
Linux 6.6-rc1

19 months agoMerge tag 'topic/drm-ci-2023-08-31-1' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Sun, 10 Sep 2023 18:55:26 +0000 (11:55 -0700)]
Merge tag 'topic/drm-ci-2023-08-31-1' of git://anongit.freedesktop.org/drm/drm

Pull drm ci scripts from Dave Airlie:
 "This is a bunch of ci integration for the freedesktop gitlab instance
  where we currently do upstream userspace testing on diverse sets of
  GPU hardware. From my perspective I think it's an experiment worth
  going with and seeing how the benefits/noise playout keeping these
  files useful.

  Ideally I'd like to get this so we can do pre-merge testing on PRs
  eventually.

  Below is some info from danvet on why we've ended up making the
  decision and how we can roll it back if we decide it was a bad plan.

  Why in upstream?

   - like documentation, testcases, tools CI integration is one of these
     things where you can waste endless amounts of time if you
     accidentally have a version that doesn't match your source code

   - but also like the above, there's a balance, this is the initial cut
     of what we think makes sense to keep in sync vs out-of-tree,
     probably needs adjustment

   - gitlab supports out-of-repo gitlab integration and that's what's
     been used for the kernel in drm, but it results in per-driver
     fragmentation and lots of duplicated effort. the simple act of
     smashing an arbitrary winner into a topic branch already started
     surfacing patches on dri-devel and sparking good cross driver team
     discussions

  Why gitlab?

   - it's not any more shit than any of the other CI

   - drm userspace uses it extensively for everything in userspace, we
     have a lot of people and experience with this, including
     integration of hw testing labs

   - media userspace like gstreamer is also on gitlab.fd.o, and there's
     discussion to extend this to the media subsystem in some fashion

  Can this be shared?

   - there's definitely a pile of code that could move to scripts/ if
     other subsystem adopt ci integration in upstream kernel git. other
     bits are more drm/gpu specific like the igt-gpu-tests/tools
     integration

   - docker images can be run locally or in other CI runners

  Will we regret this?

   - it's all in one directory, intentionally, for easy deletion

   - probably 1-2 years in upstream to see whether this is worth it or a
     Big Mistake. that's roughly what it took to _really_ roll out solid
     CI in the bigger userspace projects we have on gitlab.fd.o like
     mesa3d"

* tag 'topic/drm-ci-2023-08-31-1' of git://anongit.freedesktop.org/drm/drm:
  drm: ci: docs: fix build warning - add missing escape
  drm: Add initial ci/ subdirectory

19 months agoMerge tag 'x86-urgent-2023-09-10' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 10 Sep 2023 17:39:31 +0000 (10:39 -0700)]
Merge tag 'x86-urgent-2023-09-10' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:
 "Fix preemption delays in the SGX code, remove unnecessarily
  UAPI-exported code, fix a ld.lld linker (in)compatibility quirk and
  make the x86 SMP init code a bit more conservative to fix kexec()
  lockups"

* tag 'x86-urgent-2023-09-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/sgx: Break up long non-preemptible delays in sgx_vepc_release()
  x86: Remove the arch_calc_vm_prot_bits() macro from the UAPI
  x86/build: Fix linker fill bytes quirk/incompatibility for ld.lld
  x86/smp: Don't send INIT to non-present and non-booted CPUs