bcachefs: Nocow support

This adds support for nocow mode, where we do writes in-place when
possible. Patch components:

 - New boolean filesystem and inode option, nocow: note that when nocow
   is enabled, data checksumming and compression are implicitly disabled

 - To prevent in-place writes from racing with data moves
   (data_update.c) or bucket reuse (i.e. a bucket being reused and
   re-allocated while a nocow write is in flight, we have a new locking
   mechanism.

   Buckets can be locked for either data update or data move, using a
   fixed size hash table of two_state_shared locks. We don't have any
   chaining, meaning updates and moves to different buckets that hash to
   the same lock will wait unnecessarily - we'll want to watch for this
   becoming an issue.

 - The allocator path also needs to check for in-place writes in flight
   to a given bucket before giving it out: thus we add another counter
   to bucket_alloc_state so we can track this.

 - Fsync now may need to issue cache flushes to block devices instead of
   flushing the journal. We add a device bitmask to bch_inode_info,
   ei_devs_need_flush, which tracks devices that need to have flushes
   issued - note that this will lead to unnecessary flushes when other
   codepaths have already issued flushes, we may want to replace this with
   a sequence number.

 - New nocow write path: look up extents, and if they're writable write
   to them - otherwise fall back to the normal COW write path.

XXX: switch to sequence numbers instead of bitmask for devs needing
journal flush

XXX: ei_quota_lock being a mutex means bch2_nocow_write_done() needs to
run in process context - see if we can improve this

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

1 files changed, 23 insertions, 16 deletions

diff --git a/fs/bcachefs/extents.c b/fs/bcachefs/extents.c
index 627edba24900..55a8879dc4fe 100644
--- a/fs/bcachefs/extents.c
+++ b/fs/bcachefs/extents.c
@@ -664,22 +664,21 @@ unsigned bch2_bkey_replicas(struct bch_fs *c, struct bkey_s_c k)
 	return replicas;
 }
 
-static unsigned bch2_extent_ptr_durability(struct bch_fs *c,
-					   struct extent_ptr_decoded p)
+unsigned bch2_extent_ptr_durability(struct bch_fs *c, struct extent_ptr_decoded *p)
 {
 	unsigned durability = 0;
 	struct bch_dev *ca;
 
-	if (p.ptr.cached)
+	if (p->ptr.cached)
 		return 0;
 
-	ca = bch_dev_bkey_exists(c, p.ptr.dev);
+	ca = bch_dev_bkey_exists(c, p->ptr.dev);
 
 	if (ca->mi.state != BCH_MEMBER_STATE_failed)
 		durability = max_t(unsigned, durability, ca->mi.durability);
 
-	if (p.has_ec)
-		durability += p.ec.redundancy;
+	if (p->has_ec)
+		durability += p->ec.redundancy;
 
 	return durability;
 }
@@ -692,7 +691,7 @@ unsigned bch2_bkey_durability(struct bch_fs *c, struct bkey_s_c k)
 	unsigned durability = 0;
 
 	bkey_for_each_ptr_decode(k.k, ptrs, p, entry)
-		durability += bch2_extent_ptr_durability(c, p);
+		durability += bch2_extent_ptr_durability(c,& p);
 
 	return durability;
 }
@@ -907,23 +906,31 @@ bool bch2_bkey_matches_ptr(struct bch_fs *c, struct bkey_s_c k,
  */
 bool bch2_extents_match(struct bkey_s_c k1, struct bkey_s_c k2)
 {
-	struct bkey_ptrs_c ptrs1 = bch2_bkey_ptrs_c(k1);
-	struct bkey_ptrs_c ptrs2 = bch2_bkey_ptrs_c(k2);
-	const union bch_extent_entry *entry1, *entry2;
-	struct extent_ptr_decoded p1, p2;
-
-	if (bkey_extent_is_unwritten(k1) != bkey_extent_is_unwritten(k2))
+	if (k1.k->type != k2.k->type)
 		return false;
 
-	bkey_for_each_ptr_decode(k1.k, ptrs1, p1, entry1)
-		bkey_for_each_ptr_decode(k2.k, ptrs2, p2, entry2)
+	if (bkey_extent_is_direct_data(k1.k)) {
+		struct bkey_ptrs_c ptrs1 = bch2_bkey_ptrs_c(k1);
+		struct bkey_ptrs_c ptrs2 = bch2_bkey_ptrs_c(k2);
+		const union bch_extent_entry *entry1, *entry2;
+		struct extent_ptr_decoded p1, p2;
+
+		if (bkey_extent_is_unwritten(k1) != bkey_extent_is_unwritten(k2))
+			return false;
+
+		bkey_for_each_ptr_decode(k1.k, ptrs1, p1, entry1)
+			bkey_for_each_ptr_decode(k2.k, ptrs2, p2, entry2)
 			if (p1.ptr.dev		== p2.ptr.dev &&
 			    p1.ptr.gen		== p2.ptr.gen &&
 			    (s64) p1.ptr.offset + p1.crc.offset - bkey_start_offset(k1.k) ==
 			    (s64) p2.ptr.offset + p2.crc.offset - bkey_start_offset(k2.k))
 				return true;
 
-	return false;
+		return false;
+	} else {
+		/* KEY_TYPE_deleted, etc. */
+		return true;
+	}
 }
 
 bool bch2_extent_has_ptr(struct bkey_s_c k1, struct extent_ptr_decoded p1,