bcachefs: journal seq blacklist gc no longer has to walk btree

Since btree_ptr_v2, we no longer require the journal seq blacklist table
for skipping blacklisted bsets (btree node entries); the pointer to a
given node indicates how much data is present.

Therefore there's no longer any need for journal seq blacklist gc to
walk the btree - we can prune entries older than journal last_seq.

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

diff --git a/fs/bcachefs/bcachefs.h b/fs/bcachefs/bcachefs.h
index 6e324c1fe924540c91eb369368947b1ed10b3ef9..d8eb5beb5977d312e7008ca783eab66e8879a702 100644 (file)
--- a/fs/bcachefs/bcachefs.h
+++ b/fs/bcachefs/bcachefs.h
@@ -930,7 +930,6 @@ struct bch_fs {
        /* JOURNAL SEQ BLACKLIST */
        struct journal_seq_blacklist_table *
                                journal_seq_blacklist_table;
-       struct work_struct      journal_seq_blacklist_gc_work;
 
        /* ALLOCATOR */
        spinlock_t              freelist_lock;

diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c
index 1d37eda8d7f006619a207c1a9c0eb246e5a378c9..98cf9a65216f6799dd262704d507fbeea8288b6e 100644 (file)
--- a/fs/bcachefs/journal_io.c
+++ b/fs/bcachefs/journal_io.c
@@ -122,6 +122,10 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca,
        struct printbuf buf = PRINTBUF;
        int ret = JOURNAL_ENTRY_ADD_OK;
 
+       if (!c->journal.oldest_seq_found_ondisk ||
+           le64_to_cpu(j->seq) < c->journal.oldest_seq_found_ondisk)
+               c->journal.oldest_seq_found_ondisk = le64_to_cpu(j->seq);
+
        /* Is this entry older than the range we need? */
        if (!c->opts.read_entire_journal &&
            le64_to_cpu(j->seq) < jlist->last_seq)

diff --git a/fs/bcachefs/journal_seq_blacklist.c b/fs/bcachefs/journal_seq_blacklist.c
index c87c68ae27f43fd1e8ffe055e6104f6b3008750a..d8d40d46a27a2aa28eb3b0eb17c69830408e3354 100644 (file)
--- a/fs/bcachefs/journal_seq_blacklist.c
+++ b/fs/bcachefs/journal_seq_blacklist.c
@@ -1,8 +1,8 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
-#include "btree_iter.h"
 #include "eytzinger.h"
+#include "journal.h"
 #include "journal_seq_blacklist.h"
 #include "super-io.h"
 
@@ -217,78 +217,40 @@ const struct bch_sb_field_ops bch_sb_field_ops_journal_seq_blacklist = {
        .to_text        = bch2_sb_journal_seq_blacklist_to_text
 };
 
-void bch2_blacklist_entries_gc(struct work_struct *work)
+bool bch2_blacklist_entries_gc(struct bch_fs *c)
 {
-       struct bch_fs *c = container_of(work, struct bch_fs,
-                                       journal_seq_blacklist_gc_work);
-       struct journal_seq_blacklist_table *t;
-       struct bch_sb_field_journal_seq_blacklist *bl;
        struct journal_seq_blacklist_entry *src, *dst;
-       struct btree_trans *trans = bch2_trans_get(c);
-       unsigned i, nr, new_nr;
-       int ret;
-
-       for (i = 0; i < BTREE_ID_NR; i++) {
-               struct btree_iter iter;
-               struct btree *b;
-
-               bch2_trans_node_iter_init(trans, &iter, i, POS_MIN,
-                                         0, 0, BTREE_ITER_prefetch);
-retry:
-               bch2_trans_begin(trans);
-
-               b = bch2_btree_iter_peek_node(&iter);
-
-               while (!(ret = PTR_ERR_OR_ZERO(b)) &&
-                      b &&
-                      !test_bit(BCH_FS_stopping, &c->flags))
-                       b = bch2_btree_iter_next_node(&iter);
-
-               if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
-                       goto retry;
 
-               bch2_trans_iter_exit(trans, &iter);
-       }
-
-       bch2_trans_put(trans);
-       if (ret)
-               return;
-
-       mutex_lock(&c->sb_lock);
-       bl = bch2_sb_field_get(c->disk_sb.sb, journal_seq_blacklist);
+       struct bch_sb_field_journal_seq_blacklist *bl =
+               bch2_sb_field_get(c->disk_sb.sb, journal_seq_blacklist);
        if (!bl)
-               goto out;
+               return false;
 
-       nr = blacklist_nr_entries(bl);
+       unsigned nr = blacklist_nr_entries(bl);
        dst = bl->start;
 
-       t = c->journal_seq_blacklist_table;
+       struct journal_seq_blacklist_table *t = c->journal_seq_blacklist_table;
        BUG_ON(nr != t->nr);
 
+       unsigned i;
        for (src = bl->start, i = eytzinger0_first(t->nr);
             src < bl->start + nr;
             src++, i = eytzinger0_next(i, nr)) {
                BUG_ON(t->entries[i].start      != le64_to_cpu(src->start));
                BUG_ON(t->entries[i].end        != le64_to_cpu(src->end));
 
-               if (t->entries[i].dirty)
+               if (t->entries[i].dirty || t->entries[i].end >= c->journal.oldest_seq_found_ondisk)
                        *dst++ = *src;
        }
 
-       new_nr = dst - bl->start;
-
-       bch_info(c, "nr blacklist entries was %u, now %u", nr, new_nr);
-
-       if (new_nr != nr) {
-               bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist,
-                               new_nr ? sb_blacklist_u64s(new_nr) : 0);
-               BUG_ON(new_nr && !bl);
+       unsigned new_nr = dst - bl->start;
+       if (new_nr == nr)
+               return false;
 
-               if (!new_nr)
-                       c->disk_sb.sb->features[0] &= cpu_to_le64(~(1ULL << BCH_FEATURE_journal_seq_blacklist_v3));
+       bch_verbose(c, "nr blacklist entries was %u, now %u", nr, new_nr);
 
-               bch2_write_super(c);
-       }
-out:
-       mutex_unlock(&c->sb_lock);
+       bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist,
+                                 new_nr ? sb_blacklist_u64s(new_nr) : 0);
+       BUG_ON(new_nr && !bl);
+       return true;
 }

diff --git a/fs/bcachefs/journal_seq_blacklist.h b/fs/bcachefs/journal_seq_blacklist.h
index afb886ec8e254b240f13c7a81313b26dd7da08a3..d47636f96fdc61f74b6aba5e7116e61c524c953f 100644 (file)
--- a/fs/bcachefs/journal_seq_blacklist.h
+++ b/fs/bcachefs/journal_seq_blacklist.h
@@ -17,6 +17,6 @@ int bch2_blacklist_table_initialize(struct bch_fs *);
 
 extern const struct bch_sb_field_ops bch_sb_field_ops_journal_seq_blacklist;
 
-void bch2_blacklist_entries_gc(struct work_struct *);
+bool bch2_blacklist_entries_gc(struct bch_fs *);
 
 #endif /* _BCACHEFS_JOURNAL_SEQ_BLACKLIST_H */

diff --git a/fs/bcachefs/journal_types.h b/fs/bcachefs/journal_types.h
index 08debe6bfeeff47a92f7a64d612c5bed91a1c404..3d2135e1d7a1cf020e95ae48cd8e6683f2e97245 100644 (file)
--- a/fs/bcachefs/journal_types.h
+++ b/fs/bcachefs/journal_types.h
@@ -229,6 +229,7 @@ struct journal {
        u64                     last_seq_ondisk;
        u64                     err_seq;
        u64                     last_empty_seq;
+       u64                     oldest_seq_found_ondisk;
 
        /*
         * FIFO of journal entries whose btree updates have not yet been

diff --git a/fs/bcachefs/recovery.c b/fs/bcachefs/recovery.c
index 27ee27b285bdeda91b5702d3b1616f6420a79a4f..c33e20aa56a2876724fb9e93e3ae7d618c937622 100644 (file)
--- a/fs/bcachefs/recovery.c
+++ b/fs/bcachefs/recovery.c
@@ -878,6 +878,9 @@ use_clean:
                write_sb = true;
        }
 
+       if (bch2_blacklist_entries_gc(c))
+               write_sb = true;
+
        if (write_sb)
                bch2_write_super(c);
        mutex_unlock(&c->sb_lock);
@@ -900,10 +903,6 @@ use_clean:
                bch_info(c, "scanning for old btree nodes done");
        }
 
-       if (c->journal_seq_blacklist_table &&
-           c->journal_seq_blacklist_table->nr > 128)
-               queue_work(system_long_wq, &c->journal_seq_blacklist_gc_work);
-
        ret = 0;
 out:
        bch2_flush_fsck_errs(c);

diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c
index 36833f86d50da532772d85dfbe4c0ea04fb05355..72dde1f1a3e67802b9c164138f94b2181922adf8 100644 (file)
--- a/fs/bcachefs/super.c
+++ b/fs/bcachefs/super.c
@@ -611,8 +611,6 @@ void __bch2_fs_stop(struct bch_fs *c)
 
        set_bit(BCH_FS_stopping, &c->flags);
 
-       cancel_work_sync(&c->journal_seq_blacklist_gc_work);
-
        down_write(&c->state_lock);
        bch2_fs_read_only(c);
        up_write(&c->state_lock);
@@ -796,9 +794,6 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 
        spin_lock_init(&c->btree_write_error_lock);
 
-       INIT_WORK(&c->journal_seq_blacklist_gc_work,
-                 bch2_blacklist_entries_gc);
-
        INIT_LIST_HEAD(&c->journal_iters);
 
        INIT_LIST_HEAD(&c->fsck_error_msgs);