Commit 27c15ed2 authored by Kent Overstreet's avatar Kent Overstreet
Browse files

bcachefs: bch_member.btree_allocated_bitmap 



This adds a small (64 bit) per-device bitmap that tracks ranges that
have btree nodes, for accelerating btree node scan if it is ever needed.

- New helpers, bch2_dev_btree_bitmap_marked() and
  bch2_dev_bitmap_mark(), for checking and updating the bitmap

- Interior btree update path updates the bitmaps when required

- The check_allocations pass has a new fsck_err check,
  btree_bitmap_not_marked

- New on disk format version, mi_btree_mitmap, which indicates the new
  bitmap is present

- Upgrade table lists the required recovery pass and expected fsck error

- Btree node scan uses the bitmap to skip ranges if we're on the new
  version

Signed-off-by: default avatarKent Overstreet <kent.overstreet@linux.dev>
parent bdae2a7e

fs/bcachefs/bcachefs_format.h

+5 −2
Original line number Diff line number Diff line
@@ -578,7 +578,8 @@ struct bch_member { 
	__le64			nbuckets;	/* device size */

	__le16			first_bucket;   /* index of first bucket used */

	__le16			bucket_size;	/* sectors */

	__le32			pad;

	__u8			btree_bitmap_shift;

	__u8			pad[3];

	__le64			last_mount;	/* time_t */



	__le64			flags;
@@ -587,6 +588,7 @@ struct bch_member { 
	__le64			errors_at_reset[BCH_MEMBER_ERROR_NR];

	__le64			errors_reset_time;

	__le64			seq;

	__le64			btree_allocated_bitmap;

};



#define BCH_MEMBER_V1_BYTES	56
@@ -876,7 +878,8 @@ struct bch_sb_field_downgrade { 
	x(rebalance_work,		BCH_VERSION(1,  3))		\

	x(member_seq,			BCH_VERSION(1,  4))		\

	x(subvolume_fs_parent,		BCH_VERSION(1,  5))		\

	x(btree_subvolume_children,	BCH_VERSION(1,  6))

	x(btree_subvolume_children,	BCH_VERSION(1,  6))		\

	x(mi_btree_bitmap,		BCH_VERSION(1,  7))



enum bcachefs_metadata_version {

	bcachefs_metadata_version_min = 9,

fs/bcachefs/btree_gc.c

+13 −0
Original line number Diff line number Diff line
@@ -828,6 +828,7 @@ static int bch2_gc_mark_key(struct btree_trans *trans, enum btree_id btree_id, 
	struct bch_fs *c = trans->c;

	struct bkey deleted = KEY(0, 0, 0);

	struct bkey_s_c old = (struct bkey_s_c) { &deleted, NULL };

	struct printbuf buf = PRINTBUF;

	int ret = 0;



	deleted.p = k->k->p;
@@ -848,11 +849,23 @@ static int bch2_gc_mark_key(struct btree_trans *trans, enum btree_id btree_id, 
	if (ret)

		goto err;



	if (mustfix_fsck_err_on(level && !bch2_dev_btree_bitmap_marked(c, *k),

				c, btree_bitmap_not_marked,

				"btree ptr not marked in member info btree allocated bitmap\n  %s",

				(bch2_bkey_val_to_text(&buf, c, *k),

				 buf.buf))) {

		mutex_lock(&c->sb_lock);

		bch2_dev_btree_bitmap_mark(c, *k);

		bch2_write_super(c);

		mutex_unlock(&c->sb_lock);

	}



	ret = commit_do(trans, NULL, NULL, 0,

			bch2_key_trigger(trans, btree_id, level, old,

					 unsafe_bkey_s_c_to_s(*k), BTREE_TRIGGER_GC));

fsck_err:

err:

	printbuf_exit(&buf);

	bch_err_fn(c, ret);

	return ret;

}

fs/bcachefs/btree_node_scan.c

+7 −2
Original line number Diff line number Diff line
@@ -205,8 +205,13 @@ static int read_btree_nodes_worker(void *p) 
				last_print = jiffies;

			}



			try_read_btree_node(w->f, ca, bio, buf,

					    bucket * ca->mi.bucket_size + bucket_offset);

			u64 sector = bucket * ca->mi.bucket_size + bucket_offset;



			if (c->sb.version_upgrade_complete >= bcachefs_metadata_version_mi_btree_bitmap &&

			    !bch2_dev_btree_bitmap_marked_sectors(ca, sector, btree_sectors(c)))

				continue;



			try_read_btree_node(w->f, ca, bio, buf, sector);

		}

err:

	bio_put(bio);

fs/bcachefs/btree_update_interior.c

+24 −0
Original line number Diff line number Diff line
@@ -21,6 +21,7 @@ 
#include "keylist.h"

#include "recovery_passes.h"

#include "replicas.h"

#include "sb-members.h"

#include "super-io.h"

#include "trace.h"
@@ -605,6 +606,26 @@ static void btree_update_add_key(struct btree_update *as, 
	bch2_keylist_push(keys);

}



static bool btree_update_new_nodes_marked_sb(struct btree_update *as)

{

	for_each_keylist_key(&as->new_keys, k)

		if (!bch2_dev_btree_bitmap_marked(as->c, bkey_i_to_s_c(k)))

			return false;

	return true;

}



static void btree_update_new_nodes_mark_sb(struct btree_update *as)

{

	struct bch_fs *c = as->c;



	mutex_lock(&c->sb_lock);

	for_each_keylist_key(&as->new_keys, k)

		bch2_dev_btree_bitmap_mark(c, bkey_i_to_s_c(k));



	bch2_write_super(c);

	mutex_unlock(&c->sb_lock);

}



/*

 * The transactional part of an interior btree node update, where we journal the

 * update we did to the interior node and update alloc info:
@@ -662,6 +683,9 @@ static void btree_update_nodes_written(struct btree_update *as) 
	if (ret)

		goto err;



	if (!btree_update_new_nodes_marked_sb(as))

		btree_update_new_nodes_mark_sb(as);



	/*

	 * Wait for any in flight writes to finish before we free the old nodes

	 * on disk:

fs/bcachefs/sb-downgrade.c

+4 −1
Original line number Diff line number Diff line
@@ -51,7 +51,10 @@ 
	  BCH_FSCK_ERR_subvol_fs_path_parent_wrong)		\

	x(btree_subvolume_children,				\

	  BIT_ULL(BCH_RECOVERY_PASS_check_subvols),		\

	  BCH_FSCK_ERR_subvol_children_not_set)

	  BCH_FSCK_ERR_subvol_children_not_set)			\

	x(mi_btree_bitmap,					\

	  BIT_ULL(BCH_RECOVERY_PASS_check_allocations),		\

	  BCH_FSCK_ERR_btree_bitmap_not_marked)



#define DOWNGRADE_TABLE()