spin_lock(&sinfo->lock);
spin_lock(&cache->lock);
+ if (cache->swap_extents) {
+ ret = -ETXTBSY;
+ goto out;
+ }
+
if (cache->ro) {
cache->ro++;
ret = 0;
}
ret = inc_block_group_ro(cache, 0);
- if (!do_chunk_alloc)
+ if (!do_chunk_alloc || ret == -ETXTBSY)
goto unlock_out;
if (!ret)
goto out;
if (ret < 0)
goto out;
ret = inc_block_group_ro(cache, 0);
+ if (ret == -ETXTBSY)
+ goto unlock_out;
out:
if (cache->flags & BTRFS_BLOCK_GROUP_SYSTEM) {
alloc_flags = btrfs_get_alloc_profile(fs_info, cache->flags);
ASSERT(list_empty(&block_group->io_list));
ASSERT(list_empty(&block_group->bg_list));
ASSERT(refcount_read(&block_group->refs) == 1);
+ ASSERT(block_group->swap_extents == 0);
btrfs_put_block_group(block_group);
spin_lock(&info->block_group_cache_lock);
__btrfs_remove_free_space_cache(block_group->free_space_ctl);
}
}
+
+bool btrfs_inc_block_group_swap_extents(struct btrfs_block_group *bg)
+{
+ bool ret = true;
+
+ spin_lock(&bg->lock);
+ if (bg->ro)
+ ret = false;
+ else
+ bg->swap_extents++;
+ spin_unlock(&bg->lock);
+
+ return ret;
+}
+
+void btrfs_dec_block_group_swap_extents(struct btrfs_block_group *bg, int amount)
+{
+ spin_lock(&bg->lock);
+ ASSERT(!bg->ro);
+ ASSERT(bg->swap_extents >= amount);
+ bg->swap_extents -= amount;
+ spin_unlock(&bg->lock);
+}
/* Flag indicating this block group is placed on a sequential zone */
bool seq_zone;
+ /*
+ * Number of extents in this block group used for swap files.
+ * All accesses protected by the spinlock 'lock'.
+ */
+ int swap_extents;
+
/* Record locked full stripes for RAID5/6 block group */
struct btrfs_full_stripe_locks_tree full_stripe_locks_root;
void btrfs_freeze_block_group(struct btrfs_block_group *cache);
void btrfs_unfreeze_block_group(struct btrfs_block_group *cache);
+bool btrfs_inc_block_group_swap_extents(struct btrfs_block_group *bg);
+void btrfs_dec_block_group_swap_extents(struct btrfs_block_group *bg, int amount);
+
#endif /* BTRFS_BLOCK_GROUP_H */
struct btrfs_fs_info *fs_info = inode->root->fs_info;
SHASH_DESC_ON_STACK(shash, fs_info->csum_shash);
const u32 csum_size = fs_info->csum_size;
+ const u32 sectorsize = fs_info->sectorsize;
struct page *page;
unsigned long i;
char *kaddr;
shash->tfm = fs_info->csum_shash;
for (i = 0; i < cb->nr_pages; i++) {
+ u32 pg_offset;
+ u32 bytes_left = PAGE_SIZE;
page = cb->compressed_pages[i];
- kaddr = kmap_atomic(page);
- crypto_shash_digest(shash, kaddr, PAGE_SIZE, csum);
- kunmap_atomic(kaddr);
-
- if (memcmp(&csum, cb_sum, csum_size)) {
- btrfs_print_data_csum_error(inode, disk_start,
- csum, cb_sum, cb->mirror_num);
- if (btrfs_io_bio(bio)->device)
- btrfs_dev_stat_inc_and_print(
- btrfs_io_bio(bio)->device,
- BTRFS_DEV_STAT_CORRUPTION_ERRS);
- return -EIO;
+ /* Determine the remaining bytes inside the page first */
+ if (i == cb->nr_pages - 1)
+ bytes_left = cb->compressed_len - i * PAGE_SIZE;
+
+ /* Hash through the page sector by sector */
+ for (pg_offset = 0; pg_offset < bytes_left;
+ pg_offset += sectorsize) {
+ kaddr = kmap_atomic(page);
+ crypto_shash_digest(shash, kaddr + pg_offset,
+ sectorsize, csum);
+ kunmap_atomic(kaddr);
+
+ if (memcmp(&csum, cb_sum, csum_size) != 0) {
+ btrfs_print_data_csum_error(inode, disk_start,
+ csum, cb_sum, cb->mirror_num);
+ if (btrfs_io_bio(bio)->device)
+ btrfs_dev_stat_inc_and_print(
+ btrfs_io_bio(bio)->device,
+ BTRFS_DEV_STAT_CORRUPTION_ERRS);
+ return -EIO;
+ }
+ cb_sum += csum_size;
+ disk_start += sectorsize;
}
- cb_sum += csum_size;
}
return 0;
}
read_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree,
page_offset(bio_first_page_all(bio)),
- PAGE_SIZE);
+ fs_info->sectorsize);
read_unlock(&em_tree->lock);
if (!em)
return BLK_STS_IOERR;
refcount_set(&cb->pending_bios, 1);
for (pg_index = 0; pg_index < nr_pages; pg_index++) {
+ u32 pg_len = PAGE_SIZE;
int submit = 0;
+ /*
+ * To handle subpage case, we need to make sure the bio only
+ * covers the range we need.
+ *
+ * If we're at the last page, truncate the length to only cover
+ * the remaining part.
+ */
+ if (pg_index == nr_pages - 1)
+ pg_len = min_t(u32, PAGE_SIZE,
+ compressed_len - pg_index * PAGE_SIZE);
+
page = cb->compressed_pages[pg_index];
page->mapping = inode->i_mapping;
page->index = em_start >> PAGE_SHIFT;
if (comp_bio->bi_iter.bi_size)
- submit = btrfs_bio_fits_in_stripe(page, PAGE_SIZE,
+ submit = btrfs_bio_fits_in_stripe(page, pg_len,
comp_bio, 0);
page->mapping = NULL;
- if (submit || bio_add_page(comp_bio, page, PAGE_SIZE, 0) <
- PAGE_SIZE) {
+ if (submit || bio_add_page(comp_bio, page, pg_len, 0) < pg_len) {
unsigned int nr_sectors;
ret = btrfs_bio_wq_end_io(fs_info, comp_bio,
comp_bio->bi_private = cb;
comp_bio->bi_end_io = end_compressed_bio_read;
- bio_add_page(comp_bio, page, PAGE_SIZE, 0);
+ bio_add_page(comp_bio, page, pg_len, 0);
}
- cur_disk_byte += PAGE_SIZE;
+ cur_disk_byte += pg_len;
}
ret = btrfs_bio_wq_end_io(fs_info, comp_bio, BTRFS_WQ_ENDIO_DATA);
* points to a struct btrfs_device.
*/
bool is_block_group;
+ /*
+ * Only used when 'is_block_group' is true and it is the number of
+ * extents used by a swapfile for this block group ('ptr' field).
+ */
+ int bg_extent_count;
};
bool btrfs_pinned_by_swapfile(struct btrfs_fs_info *fs_info, void *ptr);
struct rb_node *node;
spin_lock(&cluster->lock);
- if (cluster->block_group != block_group)
- goto out;
+ if (cluster->block_group != block_group) {
+ spin_unlock(&cluster->lock);
+ return;
+ }
cluster->block_group = NULL;
cluster->window_start = 0;
entry->offset, &entry->offset_index, bitmap);
}
cluster->root = RB_ROOT;
-
-out:
spin_unlock(&cluster->lock);
btrfs_put_block_group(block_group);
}
entry->bytes -= bytes;
}
- if (entry->bytes == 0)
- rb_erase(&entry->offset_index, &cluster->root);
break;
}
out:
ctl->free_space -= bytes;
if (!entry->bitmap && !btrfs_free_space_trimmed(entry))
ctl->discardable_bytes[BTRFS_STAT_CURR] -= bytes;
+
+ spin_lock(&cluster->lock);
if (entry->bytes == 0) {
+ rb_erase(&entry->offset_index, &cluster->root);
ctl->free_extents--;
if (entry->bitmap) {
kmem_cache_free(btrfs_free_space_bitmap_cachep,
kmem_cache_free(btrfs_free_space_cachep, entry);
}
+ spin_unlock(&cluster->lock);
spin_unlock(&ctl->tree_lock);
return ret;
*/
btrfs_release_path(path);
- /* If extent is RO, we must COW it */
- if (btrfs_extent_readonly(fs_info, disk_bytenr))
- goto out_check;
ret = btrfs_cross_ref_exist(root, ino,
found_key.offset -
extent_offset, disk_bytenr, false);
WARN_ON_ONCE(freespace_inode);
goto out_check;
}
+ /* If the extent's block group is RO, we must COW */
if (!btrfs_inc_nocow_writers(fs_info, disk_bytenr))
goto out_check;
nocow = true;
sp->ptr = ptr;
sp->inode = inode;
sp->is_block_group = is_block_group;
+ sp->bg_extent_count = 1;
spin_lock(&fs_info->swapfile_pins_lock);
p = &fs_info->swapfile_pins.rb_node;
(sp->ptr == entry->ptr && sp->inode > entry->inode)) {
p = &(*p)->rb_right;
} else {
+ if (is_block_group)
+ entry->bg_extent_count++;
spin_unlock(&fs_info->swapfile_pins_lock);
kfree(sp);
return 1;
sp = rb_entry(node, struct btrfs_swapfile_pin, node);
if (sp->inode == inode) {
rb_erase(&sp->node, &fs_info->swapfile_pins);
- if (sp->is_block_group)
+ if (sp->is_block_group) {
+ btrfs_dec_block_group_swap_extents(sp->ptr,
+ sp->bg_extent_count);
btrfs_put_block_group(sp->ptr);
+ }
kfree(sp);
}
node = next;
sector_t *span)
{
struct inode *inode = file_inode(file);
- struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_fs_info *fs_info = root->fs_info;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
struct extent_state *cached_state = NULL;
struct extent_map *em = NULL;
"cannot activate swapfile while exclusive operation is running");
return -EBUSY;
}
+
+ /*
+ * Prevent snapshot creation while we are activating the swap file.
+ * We do not want to race with snapshot creation. If snapshot creation
+ * already started before we bumped nr_swapfiles from 0 to 1 and
+ * completes before the first write into the swap file after it is
+ * activated, than that write would fallback to COW.
+ */
+ if (!btrfs_drew_try_write_lock(&root->snapshot_lock)) {
+ btrfs_exclop_finish(fs_info);
+ btrfs_warn(fs_info,
+ "cannot activate swapfile because snapshot creation is in progress");
+ return -EINVAL;
+ }
/*
* Snapshots can create extents which require COW even if NODATACOW is
* set. We use this counter to prevent snapshots. We must increment it
* before walking the extents because we don't want a concurrent
* snapshot to run after we've already checked the extents.
*/
- atomic_inc(&BTRFS_I(inode)->root->nr_swapfiles);
+ atomic_inc(&root->nr_swapfiles);
isize = ALIGN_DOWN(inode->i_size, fs_info->sectorsize);
goto out;
}
+ if (!btrfs_inc_block_group_swap_extents(bg)) {
+ btrfs_warn(fs_info,
+ "block group for swapfile at %llu is read-only%s",
+ bg->start,
+ atomic_read(&fs_info->scrubs_running) ?
+ " (scrub running)" : "");
+ btrfs_put_block_group(bg);
+ ret = -EINVAL;
+ goto out;
+ }
+
ret = btrfs_add_swapfile_pin(inode, bg, true);
if (ret) {
btrfs_put_block_group(bg);
if (ret)
btrfs_swap_deactivate(file);
+ btrfs_drew_write_unlock(&root->snapshot_lock);
+
btrfs_exclop_finish(fs_info);
if (ret)
SetPageUptodate(p_page);
if (has_qstripe) {
+ /* RAID6, allocate and map temp space for the Q stripe */
q_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
if (!q_page) {
__free_page(p_page);
goto cleanup;
}
SetPageUptodate(q_page);
+ pointers[rbio->real_stripes - 1] = kmap(q_page);
}
atomic_set(&rbio->error, 0);
+ /* Map the parity stripe just once */
+ pointers[nr_data] = kmap(p_page);
+
for_each_set_bit(pagenr, rbio->dbitmap, rbio->stripe_npages) {
struct page *p;
void *parity;
pointers[stripe] = kmap(p);
}
- /* then add the parity stripe */
- pointers[stripe++] = kmap(p_page);
-
if (has_qstripe) {
- /*
- * raid6, add the qstripe and call the
- * library function to fill in our p/q
- */
- pointers[stripe++] = kmap(q_page);
-
+ /* RAID6, call the library function to fill in our P/Q */
raid6_call.gen_syndrome(rbio->real_stripes, PAGE_SIZE,
pointers);
} else {
for (stripe = 0; stripe < nr_data; stripe++)
kunmap(page_in_rbio(rbio, stripe, pagenr, 0));
- kunmap(p_page);
}
+ kunmap(p_page);
__free_page(p_page);
- if (q_page)
+ if (q_page) {
+ kunmap(q_page);
__free_page(q_page);
+ }
writeback:
/*
*/
btrfs_release_path(path);
+ /*
+ * When using NO_HOLES and we are cloning a range that covers
+ * only a hole (no extents) into a range beyond the current
+ * i_size, punching a hole in the target range will not create
+ * an extent map defining a hole, because the range starts at or
+ * beyond current i_size. If the file previously had an i_size
+ * greater than the new i_size set by this clone operation, we
+ * need to make sure the next fsync is a full fsync, so that it
+ * detects and logs a hole covering a range from the current
+ * i_size to the new i_size. If the clone range covers extents,
+ * besides a hole, then we know the full sync flag was already
+ * set by previous calls to btrfs_replace_file_extents() that
+ * replaced file extent items.
+ */
+ if (last_dest_end >= i_size_read(inode))
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+
ret = btrfs_replace_file_extents(inode, path, last_dest_end,
destoff + len - 1, NULL, &trans);
if (ret)
* commit_transactions.
*/
ro_set = 0;
+ } else if (ret == -ETXTBSY) {
+ btrfs_warn(fs_info,
+ "skipping scrub of block group %llu due to active swapfile",
+ cache->start);
+ scrub_pause_off(fs_info);
+ ret = 0;
+ goto skip_unfreeze;
} else {
btrfs_warn(fs_info,
"failed setting block group ro: %d", ret);
} else {
spin_unlock(&cache->lock);
}
-
+skip_unfreeze:
btrfs_unfreeze_block_group(cache);
btrfs_put_block_group(cache);
if (ret)
return -EUCLEAN;
}
for (; ptr < end; ptr += sizeof(*dref)) {
- u64 root_objectid;
- u64 owner;
u64 offset;
- u64 hash;
+ /*
+ * We cannot check the extent_data_ref hash due to possible
+ * overflow from the leaf due to hash collisions.
+ */
dref = (struct btrfs_extent_data_ref *)ptr;
- root_objectid = btrfs_extent_data_ref_root(leaf, dref);
- owner = btrfs_extent_data_ref_objectid(leaf, dref);
offset = btrfs_extent_data_ref_offset(leaf, dref);
- hash = hash_extent_data_ref(root_objectid, owner, offset);
- if (unlikely(hash != key->offset)) {
- extent_err(leaf, slot,
- "invalid extent data ref hash, item has 0x%016llx key has 0x%016llx",
- hash, key->offset);
- return -EUCLEAN;
- }
if (unlikely(!IS_ALIGNED(offset, leaf->fs_info->sectorsize))) {
extent_err(leaf, slot,
"invalid extent data backref offset, have %llu expect aligned to %u",
root_log_ctx.log_transid = log_root_tree->log_transid;
if (btrfs_is_zoned(fs_info)) {
- mutex_lock(&fs_info->tree_root->log_mutex);
if (!log_root_tree->node) {
ret = btrfs_alloc_log_tree_node(trans, log_root_tree);
if (ret) {
- mutex_unlock(&fs_info->tree_log_mutex);
mutex_unlock(&log_root_tree->log_mutex);
goto out;
}
}
- mutex_unlock(&fs_info->tree_root->log_mutex);
}
/*
projects / linux-block.git / commitdiff