summaryrefslogtreecommitdiff
path: root/fs/btrfs/disk-io.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/btrfs/disk-io.c')
-rw-r--r--fs/btrfs/disk-io.c213
1 files changed, 160 insertions, 53 deletions
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index eff0dd1ae62f..321feb99c179 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -182,13 +182,12 @@ static int btrfs_repair_eb_io_failure(const struct extent_buffer *eb,
int mirror_num)
{
struct btrfs_fs_info *fs_info = eb->fs_info;
- int num_folios = num_extent_folios(eb);
int ret = 0;
if (sb_rdonly(fs_info->sb))
return -EROFS;
- for (int i = 0; i < num_folios; i++) {
+ for (int i = 0; i < num_extent_folios(eb); i++) {
struct folio *folio = eb->folios[i];
u64 start = max_t(u64, eb->start, folio_pos(folio));
u64 end = min_t(u64, eb->start + eb->len,
@@ -226,7 +225,7 @@ int btrfs_read_extent_buffer(struct extent_buffer *eb,
while (1) {
clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
- ret = read_extent_buffer_pages(eb, WAIT_COMPLETE, mirror_num, check);
+ ret = read_extent_buffer_pages(eb, mirror_num, check);
if (!ret)
break;
@@ -284,8 +283,7 @@ blk_status_t btree_csum_one_bio(struct btrfs_bio *bbio)
if (WARN_ON_ONCE(found_start != eb->start))
return BLK_STS_IOERR;
- if (WARN_ON(!btrfs_folio_test_uptodate(fs_info, eb->folios[0],
- eb->start, eb->len)))
+ if (WARN_ON(!btrfs_meta_folio_test_uptodate(eb->folios[0], eb)))
return BLK_STS_IOERR;
ASSERT(memcmp_extent_buffer(eb, fs_info->fs_devices->metadata_uuid,
@@ -1089,21 +1087,22 @@ struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
const struct btrfs_key *key)
{
struct btrfs_root *root;
- struct btrfs_path *path;
+ BTRFS_PATH_AUTO_FREE(path);
path = btrfs_alloc_path();
if (!path)
return ERR_PTR(-ENOMEM);
root = read_tree_root_path(tree_root, path, key);
- btrfs_free_path(path);
return root;
}
/*
- * Initialize subvolume root in-memory structure
+ * Initialize subvolume root in-memory structure.
*
* @anon_dev: anonymous device to attach to the root, if zero, allocate new
+ *
+ * In case of failure the caller is responsible to call btrfs_free_fs_root()
*/
static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
{
@@ -1127,7 +1126,7 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
if (!anon_dev) {
ret = get_anon_bdev(&root->anon_dev);
if (ret)
- goto fail;
+ return ret;
} else {
root->anon_dev = anon_dev;
}
@@ -1137,7 +1136,7 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
ret = btrfs_init_root_free_objectid(root);
if (ret) {
mutex_unlock(&root->objectid_mutex);
- goto fail;
+ return ret;
}
ASSERT(root->free_objectid <= BTRFS_LAST_FREE_OBJECTID);
@@ -1145,9 +1144,6 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
mutex_unlock(&root->objectid_mutex);
return 0;
-fail:
- /* The caller is responsible to call btrfs_free_fs_root */
- return ret;
}
static struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
@@ -1258,6 +1254,7 @@ void btrfs_free_fs_info(struct btrfs_fs_info *fs_info)
{
struct percpu_counter *em_counter = &fs_info->evictable_extent_maps;
+ percpu_counter_destroy(&fs_info->stats_read_blocks);
percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
percpu_counter_destroy(&fs_info->delalloc_bytes);
percpu_counter_destroy(&fs_info->ordered_bytes);
@@ -1567,7 +1564,7 @@ static int transaction_kthread(void *arg)
do {
cannot_commit = false;
- delay = msecs_to_jiffies(fs_info->commit_interval * 1000);
+ delay = secs_to_jiffies(fs_info->commit_interval);
mutex_lock(&fs_info->transaction_kthread_mutex);
spin_lock(&fs_info->trans_lock);
@@ -1582,9 +1579,9 @@ static int transaction_kthread(void *arg)
cur->state < TRANS_STATE_COMMIT_PREP &&
delta < fs_info->commit_interval) {
spin_unlock(&fs_info->trans_lock);
- delay -= msecs_to_jiffies((delta - 1) * 1000);
+ delay -= secs_to_jiffies(delta - 1);
delay = min(delay,
- msecs_to_jiffies(fs_info->commit_interval * 1000));
+ secs_to_jiffies(fs_info->commit_interval));
goto sleep;
}
transid = cur->transid;
@@ -2167,8 +2164,7 @@ static int load_global_roots_objectid(struct btrfs_root *tree_root,
found = true;
root = read_tree_root_path(tree_root, path, &key);
if (IS_ERR(root)) {
- if (!btrfs_test_opt(fs_info, IGNOREBADROOTS))
- ret = PTR_ERR(root);
+ ret = PTR_ERR(root);
break;
}
set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
@@ -2199,8 +2195,8 @@ static int load_global_roots_objectid(struct btrfs_root *tree_root,
static int load_global_roots(struct btrfs_root *tree_root)
{
- struct btrfs_path *path;
- int ret = 0;
+ BTRFS_PATH_AUTO_FREE(path);
+ int ret;
path = btrfs_alloc_path();
if (!path)
@@ -2209,18 +2205,17 @@ static int load_global_roots(struct btrfs_root *tree_root)
ret = load_global_roots_objectid(tree_root, path,
BTRFS_EXTENT_TREE_OBJECTID, "extent");
if (ret)
- goto out;
+ return ret;
ret = load_global_roots_objectid(tree_root, path,
BTRFS_CSUM_TREE_OBJECTID, "csum");
if (ret)
- goto out;
+ return ret;
if (!btrfs_fs_compat_ro(tree_root->fs_info, FREE_SPACE_TREE))
- goto out;
+ return ret;
ret = load_global_roots_objectid(tree_root, path,
BTRFS_FREE_SPACE_TREE_OBJECTID,
"free space");
-out:
- btrfs_free_path(path);
+
return ret;
}
@@ -2327,6 +2322,71 @@ out:
return ret;
}
+static int validate_sys_chunk_array(const struct btrfs_fs_info *fs_info,
+ const struct btrfs_super_block *sb)
+{
+ unsigned int cur = 0; /* Offset inside the sys chunk array */
+ /*
+ * At sb read time, fs_info is not fully initialized. Thus we have
+ * to use super block sectorsize, which should have been validated.
+ */
+ const u32 sectorsize = btrfs_super_sectorsize(sb);
+ u32 sys_array_size = btrfs_super_sys_array_size(sb);
+
+ if (sys_array_size > BTRFS_SYSTEM_CHUNK_ARRAY_SIZE) {
+ btrfs_err(fs_info, "system chunk array too big %u > %u",
+ sys_array_size, BTRFS_SYSTEM_CHUNK_ARRAY_SIZE);
+ return -EUCLEAN;
+ }
+
+ while (cur < sys_array_size) {
+ struct btrfs_disk_key *disk_key;
+ struct btrfs_chunk *chunk;
+ struct btrfs_key key;
+ u64 type;
+ u16 num_stripes;
+ u32 len;
+ int ret;
+
+ disk_key = (struct btrfs_disk_key *)(sb->sys_chunk_array + cur);
+ len = sizeof(*disk_key);
+
+ if (cur + len > sys_array_size)
+ goto short_read;
+ cur += len;
+
+ btrfs_disk_key_to_cpu(&key, disk_key);
+ if (key.type != BTRFS_CHUNK_ITEM_KEY) {
+ btrfs_err(fs_info,
+ "unexpected item type %u in sys_array at offset %u",
+ key.type, cur);
+ return -EUCLEAN;
+ }
+ chunk = (struct btrfs_chunk *)(sb->sys_chunk_array + cur);
+ num_stripes = btrfs_stack_chunk_num_stripes(chunk);
+ if (cur + btrfs_chunk_item_size(num_stripes) > sys_array_size)
+ goto short_read;
+ type = btrfs_stack_chunk_type(chunk);
+ if (!(type & BTRFS_BLOCK_GROUP_SYSTEM)) {
+ btrfs_err(fs_info,
+ "invalid chunk type %llu in sys_array at offset %u",
+ type, cur);
+ return -EUCLEAN;
+ }
+ ret = btrfs_check_chunk_valid(fs_info, NULL, chunk, key.offset,
+ sectorsize);
+ if (ret < 0)
+ return ret;
+ cur += btrfs_chunk_item_size(num_stripes);
+ }
+ return 0;
+short_read:
+ btrfs_err(fs_info,
+ "super block sys chunk array short read, cur=%u sys_array_size=%u",
+ cur, sys_array_size);
+ return -EUCLEAN;
+}
+
/*
* Real super block validation
* NOTE: super csum type and incompat features will not be checked here.
@@ -2381,21 +2441,27 @@ int btrfs_validate_super(const struct btrfs_fs_info *fs_info,
* Check sectorsize and nodesize first, other check will need it.
* Check all possible sectorsize(4K, 8K, 16K, 32K, 64K) here.
*/
- if (!is_power_of_2(sectorsize) || sectorsize < 4096 ||
+ if (!is_power_of_2(sectorsize) || sectorsize < BTRFS_MIN_BLOCKSIZE ||
sectorsize > BTRFS_MAX_METADATA_BLOCKSIZE) {
btrfs_err(fs_info, "invalid sectorsize %llu", sectorsize);
ret = -EINVAL;
}
/*
- * We only support at most two sectorsizes: 4K and PAGE_SIZE.
+ * We only support at most 3 sectorsizes: 4K, PAGE_SIZE, MIN_BLOCKSIZE.
+ *
+ * For 4K page sized systems with non-debug builds, all 3 matches (4K).
+ * For 4K page sized systems with debug builds, there are two block sizes
+ * supported. (4K and 2K)
*
* We can support 16K sectorsize with 64K page size without problem,
* but such sectorsize/pagesize combination doesn't make much sense.
* 4K will be our future standard, PAGE_SIZE is supported from the very
* beginning.
*/
- if (sectorsize > PAGE_SIZE || (sectorsize != SZ_4K && sectorsize != PAGE_SIZE)) {
+ if (sectorsize > PAGE_SIZE || (sectorsize != SZ_4K &&
+ sectorsize != PAGE_SIZE &&
+ sectorsize != BTRFS_MIN_BLOCKSIZE)) {
btrfs_err(fs_info,
"sectorsize %llu not yet supported for page size %lu",
sectorsize, PAGE_SIZE);
@@ -2495,6 +2561,11 @@ int btrfs_validate_super(const struct btrfs_fs_info *fs_info,
ret = -EINVAL;
}
+ if (ret)
+ return ret;
+
+ ret = validate_sys_chunk_array(fs_info, sb);
+
/*
* Obvious sys_chunk_array corruptions, it must hold at least one key
* and one chunk
@@ -2856,6 +2927,10 @@ static int init_mount_fs_info(struct btrfs_fs_info *fs_info, struct super_block
if (ret)
return ret;
+ ret = percpu_counter_init(&fs_info->stats_read_blocks, 0, GFP_KERNEL);
+ if (ret)
+ return ret;
+
fs_info->dirty_metadata_batch = PAGE_SIZE *
(1 + ilog2(nr_cpu_ids));
@@ -3318,9 +3393,9 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
fs_info->nodesize = nodesize;
fs_info->sectorsize = sectorsize;
fs_info->sectorsize_bits = ilog2(sectorsize);
- fs_info->sectors_per_page = (PAGE_SIZE >> fs_info->sectorsize_bits);
fs_info->csums_per_leaf = BTRFS_MAX_ITEM_SIZE(fs_info) / fs_info->csum_size;
fs_info->stripesize = stripesize;
+ fs_info->fs_devices->fs_info = fs_info;
/*
* Handle the space caching options appropriately now that we have the
@@ -3343,11 +3418,6 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
*/
fs_info->max_inline = min_t(u64, fs_info->max_inline, fs_info->sectorsize);
- if (sectorsize < PAGE_SIZE)
- btrfs_warn(fs_info,
- "read-write for sector size %u with page size %lu is experimental",
- sectorsize, PAGE_SIZE);
-
ret = btrfs_init_workqueues(fs_info);
if (ret)
goto fail_sb_buffer;
@@ -3782,7 +3852,6 @@ static int write_dev_supers(struct btrfs_device *device,
atomic_inc(&device->sb_write_errors);
continue;
}
- ASSERT(folio_order(folio) == 0);
offset = offset_in_folio(folio, bytenr);
disk_super = folio_address(folio) + offset;
@@ -3855,7 +3924,6 @@ static int wait_dev_supers(struct btrfs_device *device, int max_mirrors)
/* If the folio has been removed, then we know it completed. */
if (IS_ERR(folio))
continue;
- ASSERT(folio_order(folio) == 0);
/* Folio will be unlocked once the write completes. */
folio_wait_locked(folio);
@@ -4253,6 +4321,14 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
btrfs_cleanup_defrag_inodes(fs_info);
/*
+ * Handle the error fs first, as it will flush and wait for all ordered
+ * extents. This will generate delayed iputs, thus we want to handle
+ * it first.
+ */
+ if (unlikely(BTRFS_FS_ERROR(fs_info)))
+ btrfs_error_commit_super(fs_info);
+
+ /*
* Wait for any fixup workers to complete.
* If we don't wait for them here and they are still running by the time
* we call kthread_stop() against the cleaner kthread further below, we
@@ -4273,6 +4349,31 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
btrfs_flush_workqueue(fs_info->delalloc_workers);
/*
+ * We can have ordered extents getting their last reference dropped from
+ * the fs_info->workers queue because for async writes for data bios we
+ * queue a work for that queue, at btrfs_wq_submit_bio(), that runs
+ * run_one_async_done() which calls btrfs_bio_end_io() in case the bio
+ * has an error, and that later function can do the final
+ * btrfs_put_ordered_extent() on the ordered extent attached to the bio,
+ * which adds a delayed iput for the inode. So we must flush the queue
+ * so that we don't have delayed iputs after committing the current
+ * transaction below and stopping the cleaner and transaction kthreads.
+ */
+ btrfs_flush_workqueue(fs_info->workers);
+
+ /*
+ * When finishing a compressed write bio we schedule a work queue item
+ * to finish an ordered extent - btrfs_finish_compressed_write_work()
+ * calls btrfs_finish_ordered_extent() which in turns does a call to
+ * btrfs_queue_ordered_fn(), and that queues the ordered extent
+ * completion either in the endio_write_workers work queue or in the
+ * fs_info->endio_freespace_worker work queue. We flush those queues
+ * below, so before we flush them we must flush this queue for the
+ * workers of compressed writes.
+ */
+ flush_workqueue(fs_info->compressed_write_workers);
+
+ /*
* After we parked the cleaner kthread, ordered extents may have
* completed and created new delayed iputs. If one of the async reclaim
* tasks is running and in the RUN_DELAYED_IPUTS flush state, then we
@@ -4283,8 +4384,8 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
*
* So wait for all ongoing ordered extents to complete and then run
* delayed iputs. This works because once we reach this point no one
- * can either create new ordered extents nor create delayed iputs
- * through some other means.
+ * can create new ordered extents, but delayed iputs can still be added
+ * by a reclaim worker (see comments further below).
*
* Also note that btrfs_wait_ordered_roots() is not safe here, because
* it waits for BTRFS_ORDERED_COMPLETE to be set on an ordered extent,
@@ -4295,6 +4396,10 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
btrfs_flush_workqueue(fs_info->endio_write_workers);
/* Ordered extents for free space inodes. */
btrfs_flush_workqueue(fs_info->endio_freespace_worker);
+ /*
+ * Run delayed iputs in case an async reclaim worker is waiting for them
+ * to be run as mentioned above.
+ */
btrfs_run_delayed_iputs(fs_info);
cancel_work_sync(&fs_info->async_reclaim_work);
@@ -4302,6 +4407,18 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
cancel_work_sync(&fs_info->preempt_reclaim_work);
cancel_work_sync(&fs_info->em_shrinker_work);
+ /*
+ * Run delayed iputs again because an async reclaim worker may have
+ * added new ones if it was flushing delalloc:
+ *
+ * shrink_delalloc() -> btrfs_start_delalloc_roots() ->
+ * start_delalloc_inodes() -> btrfs_add_delayed_iput()
+ */
+ btrfs_run_delayed_iputs(fs_info);
+
+ /* There should be no more workload to generate new delayed iputs. */
+ set_bit(BTRFS_FS_STATE_NO_DELAYED_IPUT, &fs_info->fs_state);
+
/* Cancel or finish ongoing discard work */
btrfs_discard_cleanup(fs_info);
@@ -4330,9 +4447,6 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
btrfs_err(fs_info, "commit super ret %d", ret);
}
- if (BTRFS_FS_ERROR(fs_info))
- btrfs_error_commit_super(fs_info);
-
kthread_stop(fs_info->transaction_kthread);
kthread_stop(fs_info->cleaner_kthread);
@@ -4455,10 +4569,6 @@ static void btrfs_error_commit_super(struct btrfs_fs_info *fs_info)
/* cleanup FS via transaction */
btrfs_cleanup_transaction(fs_info);
- mutex_lock(&fs_info->cleaner_mutex);
- btrfs_run_delayed_iputs(fs_info);
- mutex_unlock(&fs_info->cleaner_mutex);
-
down_write(&fs_info->cleanup_work_sem);
up_write(&fs_info->cleanup_work_sem);
}
@@ -4829,7 +4939,7 @@ static int btrfs_cleanup_transaction(struct btrfs_fs_info *fs_info)
int btrfs_init_root_free_objectid(struct btrfs_root *root)
{
- struct btrfs_path *path;
+ BTRFS_PATH_AUTO_FREE(path);
int ret;
struct extent_buffer *l;
struct btrfs_key search_key;
@@ -4845,14 +4955,13 @@ int btrfs_init_root_free_objectid(struct btrfs_root *root)
search_key.offset = (u64)-1;
ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
if (ret < 0)
- goto error;
+ return ret;
if (ret == 0) {
/*
* Key with offset -1 found, there would have to exist a root
* with such id, but this is out of valid range.
*/
- ret = -EUCLEAN;
- goto error;
+ return -EUCLEAN;
}
if (path->slots[0] > 0) {
slot = path->slots[0] - 1;
@@ -4863,10 +4972,8 @@ int btrfs_init_root_free_objectid(struct btrfs_root *root)
} else {
root->free_objectid = BTRFS_FIRST_FREE_OBJECTID;
}
- ret = 0;
-error:
- btrfs_free_path(path);
- return ret;
+
+ return 0;
}
int btrfs_get_free_objectid(struct btrfs_root *root, u64 *objectid)