]> git.sesse.net Git - bcachefs-tools-debian/blobdiff - libbcachefs/super.c
Update bcachefs sources to 171da96d76 bcachefs: Drop some anonymous structs, unions
[bcachefs-tools-debian] / libbcachefs / super.c
index 95c16f70512f1c06bb849ce65b3dac7144eb5000..359ca164e0889764d03cd930449debc89327b59a 100644 (file)
@@ -16,6 +16,7 @@
 #include "btree_key_cache.h"
 #include "btree_update_interior.h"
 #include "btree_io.h"
+#include "btree_write_buffer.h"
 #include "buckets_waiting_for_journal.h"
 #include "chardev.h"
 #include "checksum.h"
@@ -55,7 +56,6 @@
 #include <linux/idr.h>
 #include <linux/module.h>
 #include <linux/percpu.h>
-#include <linux/pretty-printers.h>
 #include <linux/random.h>
 #include <linux/sysfs.h>
 #include <crypto/hash.h>
@@ -110,7 +110,7 @@ static struct kset *bcachefs_kset;
 static LIST_HEAD(bch_fs_list);
 static DEFINE_MUTEX(bch_fs_list_lock);
 
-static DECLARE_WAIT_QUEUE_HEAD(bch_read_only_wait);
+DECLARE_WAIT_QUEUE_HEAD(bch2_read_only_wait);
 
 static void bch2_dev_free(struct bch_dev *);
 static int bch2_dev_alloc(struct bch_fs *, unsigned);
@@ -238,13 +238,15 @@ static void __bch2_fs_read_only(struct bch_fs *c)
                bch2_dev_allocator_remove(c, ca);
 }
 
+#ifndef BCH_WRITE_REF_DEBUG
 static void bch2_writes_disabled(struct percpu_ref *writes)
 {
        struct bch_fs *c = container_of(writes, struct bch_fs, writes);
 
        set_bit(BCH_FS_WRITE_DISABLE_COMPLETE, &c->flags);
-       wake_up(&bch_read_only_wait);
+       wake_up(&bch2_read_only_wait);
 }
+#endif
 
 void bch2_fs_read_only(struct bch_fs *c)
 {
@@ -259,9 +261,13 @@ void bch2_fs_read_only(struct bch_fs *c)
         * Block new foreground-end write operations from starting - any new
         * writes will return -EROFS:
         */
+       set_bit(BCH_FS_GOING_RO, &c->flags);
+#ifndef BCH_WRITE_REF_DEBUG
        percpu_ref_kill(&c->writes);
-
-       cancel_work_sync(&c->ec_stripe_delete_work);
+#else
+       for (unsigned i = 0; i < BCH_WRITE_REF_NR; i++)
+               bch2_write_ref_put(c, i);
+#endif
 
        /*
         * If we're not doing an emergency shutdown, we want to wait on
@@ -274,16 +280,17 @@ void bch2_fs_read_only(struct bch_fs *c)
         * we do need to wait on them before returning and signalling
         * that going RO is complete:
         */
-       wait_event(bch_read_only_wait,
+       wait_event(bch2_read_only_wait,
                   test_bit(BCH_FS_WRITE_DISABLE_COMPLETE, &c->flags) ||
                   test_bit(BCH_FS_EMERGENCY_RO, &c->flags));
 
        __bch2_fs_read_only(c);
 
-       wait_event(bch_read_only_wait,
+       wait_event(bch2_read_only_wait,
                   test_bit(BCH_FS_WRITE_DISABLE_COMPLETE, &c->flags));
 
        clear_bit(BCH_FS_WRITE_DISABLE_COMPLETE, &c->flags);
+       clear_bit(BCH_FS_GOING_RO, &c->flags);
 
        if (!bch2_journal_error(&c->journal) &&
            !test_bit(BCH_FS_ERROR, &c->flags) &&
@@ -320,7 +327,7 @@ bool bch2_fs_emergency_read_only(struct bch_fs *c)
        bch2_journal_halt(&c->journal);
        bch2_fs_read_only_async(c);
 
-       wake_up(&bch_read_only_wait);
+       wake_up(&bch2_read_only_wait);
        return ret;
 }
 
@@ -392,20 +399,27 @@ static int __bch2_fs_read_write(struct bch_fs *c, bool early)
                return ret;
        }
 
-       schedule_work(&c->ec_stripe_delete_work);
-
-       bch2_do_discards(c);
-       bch2_do_invalidates(c);
-
        if (!early) {
                ret = bch2_fs_read_write_late(c);
                if (ret)
                        goto err;
        }
 
+#ifndef BCH_WRITE_REF_DEBUG
        percpu_ref_reinit(&c->writes);
+#else
+       for (unsigned i = 0; i < BCH_WRITE_REF_NR; i++) {
+               BUG_ON(atomic_long_read(&c->writes[i]));
+               atomic_long_inc(&c->writes[i]);
+       }
+#endif
        set_bit(BCH_FS_RW, &c->flags);
        set_bit(BCH_FS_WAS_RW, &c->flags);
+
+       bch2_do_discards(c);
+       bch2_do_invalidates(c);
+       bch2_do_stripe_deletes(c);
+       bch2_do_pending_node_rewrites(c);
        return 0;
 err:
        __bch2_fs_read_only(c);
@@ -434,6 +448,7 @@ static void __bch2_fs_free(struct bch_fs *c)
        for (i = 0; i < BCH_TIME_STAT_NR; i++)
                bch2_time_stats_exit(&c->times[i]);
 
+       bch2_free_pending_node_rewrites(c);
        bch2_fs_counters_exit(c);
        bch2_fs_snapshots_exit(c);
        bch2_fs_quota_exit(c);
@@ -453,24 +468,26 @@ static void __bch2_fs_free(struct bch_fs *c)
        bch2_fs_compress_exit(c);
        bch2_journal_keys_free(&c->journal_keys);
        bch2_journal_entries_free(c);
+       bch2_fs_btree_write_buffer_exit(c);
        percpu_free_rwsem(&c->mark_lock);
+       free_percpu(c->online_reserved);
 
        if (c->btree_paths_bufs)
                for_each_possible_cpu(cpu)
                        kfree(per_cpu_ptr(c->btree_paths_bufs, cpu)->path);
 
-       free_percpu(c->online_reserved);
        free_percpu(c->btree_paths_bufs);
        free_percpu(c->pcpu);
        mempool_exit(&c->large_bkey_pool);
        mempool_exit(&c->btree_bounce_pool);
        bioset_exit(&c->btree_bio);
        mempool_exit(&c->fill_iter);
+#ifndef BCH_WRITE_REF_DEBUG
        percpu_ref_exit(&c->writes);
+#endif
        kfree(rcu_dereference_protected(c->disk_groups, 1));
        kfree(c->journal_seq_blacklist_table);
        kfree(c->unused_inode_hints);
-       free_heap(&c->copygc_heap);
 
        if (c->io_complete_wq)
                destroy_workqueue(c->io_complete_wq);
@@ -689,12 +706,14 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
        INIT_LIST_HEAD(&c->data_progress_list);
        mutex_init(&c->data_progress_lock);
 
-       spin_lock_init(&c->ec_stripes_heap_lock);
+       mutex_init(&c->ec_stripes_heap_lock);
 
        seqcount_init(&c->gc_pos_lock);
 
        seqcount_init(&c->usage_lock);
 
+       sema_init(&c->io_in_flight, 128);
+
        c->copy_gc_enabled              = 1;
        c->rebalance.enabled            = 1;
        c->promote_whole_extents        = true;
@@ -743,9 +762,6 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 
        bch2_opts_apply(&c->opts, opts);
 
-       /* key cache currently disabled for inodes, because of snapshots: */
-       c->opts.inodes_use_key_cache = 0;
-
        c->btree_key_cache_btrees |= 1U << BTREE_ID_alloc;
        if (c->opts.inodes_use_key_cache)
                c->btree_key_cache_btrees |= 1U << BTREE_ID_inodes;
@@ -766,23 +782,25 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
        c->inode_shard_bits = ilog2(roundup_pow_of_two(num_possible_cpus()));
 
        if (!(c->btree_update_wq = alloc_workqueue("bcachefs",
-                               WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) ||
+                               WQ_FREEZABLE|WQ_UNBOUND|WQ_MEM_RECLAIM, 512)) ||
            !(c->btree_io_complete_wq = alloc_workqueue("bcachefs_btree_io",
-                               WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) ||
+                               WQ_FREEZABLE|WQ_MEM_RECLAIM, 1)) ||
            !(c->copygc_wq = alloc_workqueue("bcachefs_copygc",
                                WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) ||
            !(c->io_complete_wq = alloc_workqueue("bcachefs_io",
                                WQ_FREEZABLE|WQ_HIGHPRI|WQ_MEM_RECLAIM, 1)) ||
+#ifndef BCH_WRITE_REF_DEBUG
            percpu_ref_init(&c->writes, bch2_writes_disabled,
                            PERCPU_REF_INIT_DEAD, GFP_KERNEL) ||
+#endif
            mempool_init_kmalloc_pool(&c->fill_iter, 1, iter_size) ||
            bioset_init(&c->btree_bio, 1,
                        max(offsetof(struct btree_read_bio, bio),
                            offsetof(struct btree_write_bio, wbio.bio)),
                        BIOSET_NEED_BVECS) ||
            !(c->pcpu = alloc_percpu(struct bch_fs_pcpu)) ||
-           !(c->btree_paths_bufs = alloc_percpu(struct btree_path_buf)) ||
            !(c->online_reserved = alloc_percpu(u64)) ||
+           !(c->btree_paths_bufs = alloc_percpu(struct btree_path_buf)) ||
            mempool_init_kvpmalloc_pool(&c->btree_bounce_pool, 1,
                                        btree_bytes(c)) ||
            mempool_init_kmalloc_pool(&c->large_bkey_pool, 1, 2048) ||
@@ -802,6 +820,7 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
            bch2_fs_btree_iter_init(c) ?:
            bch2_fs_btree_interior_update_init(c) ?:
            bch2_fs_buckets_waiting_for_journal_init(c) ?:
+           bch2_fs_btree_write_buffer_init(c) ?:
            bch2_fs_subvolumes_init(c) ?:
            bch2_fs_io_init(c) ?:
            bch2_fs_nocow_locking_init(c) ?:
@@ -850,9 +869,12 @@ static void print_mount_opts(struct bch_fs *c)
        struct printbuf p = PRINTBUF;
        bool first = true;
 
+       prt_printf(&p, "mounted version=%s", bch2_metadata_versions[c->sb.version]);
+
        if (c->opts.read_only) {
-               prt_printf(&p, "ro");
+               prt_str(&p, " opts=");
                first = false;
+               prt_printf(&p, "ro");
        }
 
        for (i = 0; i < bch2_opts_nr; i++) {
@@ -865,16 +887,12 @@ static void print_mount_opts(struct bch_fs *c)
                if (v == bch2_opt_get_by_id(&bch2_opts_default, i))
                        continue;
 
-               if (!first)
-                       prt_printf(&p, ",");
+               prt_str(&p, first ? " opts=" : ",");
                first = false;
                bch2_opt_to_text(&p, c, c->disk_sb.sb, opt, v, OPT_SHOW_MOUNT_STYLE);
        }
 
-       if (!p.pos)
-               prt_printf(&p, "(null)");
-
-       bch_info(c, "mounted version=%s opts=%s", bch2_metadata_versions[c->sb.version], p.buf);
+       bch_info(c, "%s", p.buf);
        printbuf_exit(&p);
 }
 
@@ -1904,7 +1922,8 @@ out:
        kfree(sb);
        printbuf_exit(&errbuf);
        module_put(THIS_MODULE);
-       pr_verbose_init(opts, "ret %i", PTR_ERR_OR_ZERO(c));
+       pr_verbose_init(opts, "ret %s (%i)", bch2_err_str(PTR_ERR_OR_ZERO(c)),
+                       PTR_ERR_OR_ZERO(c));
        return c;
 err_print:
        pr_err("bch_fs_open err opening %s: %s",
@@ -1955,5 +1974,8 @@ err:
 BCH_DEBUG_PARAMS()
 #undef BCH_DEBUG_PARAM
 
+unsigned bch2_metadata_version = bcachefs_metadata_version_current;
+module_param_named(version, bch2_metadata_version, uint, 0400);
+
 module_exit(bcachefs_exit);
 module_init(bcachefs_init);