#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
-#include <linux/smp_lock.h>
#include <linux/backing-dev.h>
#include <linux/mpage.h>
#include <linux/swap.h>
#include <linux/statfs.h>
#include <linux/compat.h>
#include <linux/bit_spinlock.h>
-#include <linux/version.h>
#include <linux/xattr.h>
#include <linux/posix_acl.h>
+#include <linux/falloc.h>
+#include "compat.h"
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "volumes.h"
#include "ordered-data.h"
#include "xattr.h"
-#include "compat.h"
#include "tree-log.h"
-#include "ref-cache.h"
#include "compression.h"
+#include "locking.h"
struct btrfs_iget_args {
u64 ino;
static struct kmem_cache *btrfs_inode_cachep;
struct kmem_cache *btrfs_trans_handle_cachep;
struct kmem_cache *btrfs_transaction_cachep;
-struct kmem_cache *btrfs_bit_radix_cachep;
struct kmem_cache *btrfs_path_cachep;
#define S_SHIFT 12
static void btrfs_truncate(struct inode *inode);
static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end);
+static noinline int cow_file_range(struct inode *inode,
+ struct page *locked_page,
+ u64 start, u64 end, int *page_started,
+ unsigned long *nr_written, int unlock);
-/*
- * a very lame attempt at stopping writes when the FS is 85% full. There
- * are countless ways this is incorrect, but it is better than nothing.
- */
-int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
- int for_del)
+static int btrfs_init_inode_security(struct inode *inode, struct inode *dir)
{
- u64 total;
- u64 used;
- u64 thresh;
- unsigned long flags;
- int ret = 0;
-
- spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
- total = btrfs_super_total_bytes(&root->fs_info->super_copy);
- used = btrfs_super_bytes_used(&root->fs_info->super_copy);
- if (for_del)
- thresh = total * 90;
- else
- thresh = total * 85;
-
- do_div(thresh, 100);
+ int err;
- if (used + root->fs_info->delalloc_bytes + num_required > thresh)
- ret = -ENOSPC;
- spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
- return ret;
+ err = btrfs_init_acl(inode, dir);
+ if (!err)
+ err = btrfs_xattr_security_init(inode, dir);
+ return err;
}
/*
* the btree. The caller should have done a btrfs_drop_extents so that
* no overlapping inline items exist in the btree
*/
-static int noinline insert_inline_extent(struct btrfs_trans_handle *trans,
+static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct inode *inode,
u64 start, size_t size, size_t compressed_size,
struct page **compressed_pages)
cur_size = compressed_size;
}
- path = btrfs_alloc_path(); if (!path)
+ path = btrfs_alloc_path();
+ if (!path)
return -ENOMEM;
+ path->leave_spinning = 1;
btrfs_set_trans_block_group(trans, inode);
key.objectid = inode->i_ino;
key.offset = start;
btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
- inode_add_bytes(inode, size);
datasize = btrfs_file_extent_calc_inline_size(cur_size);
inode_add_bytes(inode, size);
BUG_ON(ret);
if (ret) {
err = ret;
- printk("got bad ret %d\n", ret);
goto fail;
}
leaf = path->nodes[0];
if (use_compress) {
struct page *cpage;
int i = 0;
- while(compressed_size > 0) {
+ while (compressed_size > 0) {
cpage = compressed_pages[i];
- cur_size = min(compressed_size,
+ cur_size = min_t(unsigned long, compressed_size,
PAGE_CACHE_SIZE);
- kaddr = kmap(cpage);
+ kaddr = kmap_atomic(cpage, KM_USER0);
write_extent_buffer(leaf, kaddr, ptr, cur_size);
- kunmap(cpage);
+ kunmap_atomic(kaddr, KM_USER0);
i++;
ptr += cur_size;
* does the checks required to make sure the data is small enough
* to fit as an inline extent.
*/
-static int cow_file_range_inline(struct btrfs_trans_handle *trans,
+static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct inode *inode, u64 start, u64 end,
size_t compressed_size,
data_len = compressed_size;
if (start > 0 ||
+ actual_end >= PAGE_CACHE_SIZE ||
data_len >= BTRFS_MAX_INLINE_DATA_SIZE(root) ||
(!compressed_size &&
(actual_end & (root->sectorsize - 1)) == 0) ||
}
ret = btrfs_drop_extents(trans, root, inode, start,
- aligned_end, aligned_end, &hint_byte);
+ aligned_end, aligned_end, start,
+ &hint_byte, 1);
BUG_ON(ret);
if (isize > actual_end)
inline_len, compressed_size,
compressed_pages);
BUG_ON(ret);
- btrfs_drop_extent_cache(inode, start, aligned_end, 0);
+ btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0);
+ return 0;
+}
+
+struct async_extent {
+ u64 start;
+ u64 ram_size;
+ u64 compressed_size;
+ struct page **pages;
+ unsigned long nr_pages;
+ struct list_head list;
+};
+
+struct async_cow {
+ struct inode *inode;
+ struct btrfs_root *root;
+ struct page *locked_page;
+ u64 start;
+ u64 end;
+ struct list_head extents;
+ struct btrfs_work work;
+};
+
+static noinline int add_async_extent(struct async_cow *cow,
+ u64 start, u64 ram_size,
+ u64 compressed_size,
+ struct page **pages,
+ unsigned long nr_pages)
+{
+ struct async_extent *async_extent;
+
+ async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS);
+ async_extent->start = start;
+ async_extent->ram_size = ram_size;
+ async_extent->compressed_size = compressed_size;
+ async_extent->pages = pages;
+ async_extent->nr_pages = nr_pages;
+ list_add_tail(&async_extent->list, &cow->extents);
return 0;
}
/*
- * when extent_io.c finds a delayed allocation range in the file,
- * the call backs end up in this code. The basic idea is to
- * allocate extents on disk for the range, and create ordered data structs
- * in ram to track those extents.
+ * we create compressed extents in two phases. The first
+ * phase compresses a range of pages that have already been
+ * locked (both pages and state bits are locked).
*
- * locked_page is the page that writepage had locked already. We use
- * it to make sure we don't do extra locks or unlocks.
+ * This is done inside an ordered work queue, and the compression
+ * is spread across many cpus. The actual IO submission is step
+ * two, and the ordered work queue takes care of making sure that
+ * happens in the same order things were put onto the queue by
+ * writepages and friends.
*
- * *page_started is set to one if we unlock locked_page and do everything
- * required to start IO on it. It may be clean and already done with
- * IO when we return.
+ * If this code finds it can't get good compression, it puts an
+ * entry onto the work queue to write the uncompressed bytes. This
+ * makes sure that both compressed inodes and uncompressed inodes
+ * are written in the same order that pdflush sent them down.
*/
-static int cow_file_range(struct inode *inode, struct page *locked_page,
- u64 start, u64 end, int *page_started)
+static noinline int compress_file_range(struct inode *inode,
+ struct page *locked_page,
+ u64 start, u64 end,
+ struct async_cow *async_cow,
+ int *num_added)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
- u64 alloc_hint = 0;
u64 num_bytes;
- unsigned long ram_size;
u64 orig_start;
u64 disk_num_bytes;
- u64 cur_alloc_size;
u64 blocksize = root->sectorsize;
u64 actual_end;
- struct btrfs_key ins;
- struct extent_map *em;
- struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ u64 isize = i_size_read(inode);
int ret = 0;
struct page **pages = NULL;
unsigned long nr_pages;
unsigned long total_compressed = 0;
unsigned long total_in = 0;
unsigned long max_compressed = 128 * 1024;
- unsigned long max_uncompressed = 256 * 1024;
+ unsigned long max_uncompressed = 128 * 1024;
int i;
- int ordered_type;
int will_compress;
- trans = btrfs_join_transaction(root, 1);
- BUG_ON(!trans);
- btrfs_set_trans_block_group(trans, inode);
orig_start = start;
- /*
- * compression made this loop a bit ugly, but the basic idea is to
- * compress some pages but keep the total size of the compressed
- * extent relatively small. If compression is off, this goto target
- * is never used.
- */
+ actual_end = min_t(u64, isize, end + 1);
again:
will_compress = 0;
nr_pages = (end >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT) + 1;
nr_pages = min(nr_pages, (128 * 1024UL) / PAGE_CACHE_SIZE);
- actual_end = min_t(u64, i_size_read(inode), end + 1);
+ /*
+ * we don't want to send crud past the end of i_size through
+ * compression, that's just a waste of CPU time. So, if the
+ * end of the file is before the start of our current
+ * requested range of bytes, we bail out to the uncompressed
+ * cleanup code that can deal with all of this.
+ *
+ * It isn't really the fastest way to fix things, but this is a
+ * very uncommon corner.
+ */
+ if (actual_end <= start)
+ goto cleanup_and_bail_uncompressed;
+
total_compressed = actual_end - start;
/* we want to make sure that amount of ram required to uncompress
* an extent is reasonable, so we limit the total size in ram
- * of a compressed extent to 256k
+ * of a compressed extent to 128k. This is a crucial number
+ * because it also controls how easily we can spread reads across
+ * cpus for decompression.
+ *
+ * We also want to make sure the amount of IO required to do
+ * a random read is reasonably small, so we limit the size of
+ * a compressed extent to 128k.
*/
total_compressed = min(total_compressed, max_uncompressed);
num_bytes = (end - start + blocksize) & ~(blocksize - 1);
total_in = 0;
ret = 0;
- /* we do compression for mount -o compress and when the
- * inode has not been flagged as nocompress
+ /*
+ * we do compression for mount -o compress and when the
+ * inode has not been flagged as nocompress. This flag can
+ * change at any time if we discover bad compression ratios.
*/
- if (!btrfs_test_flag(inode, NOCOMPRESS) &&
+ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) &&
btrfs_test_opt(root, COMPRESS)) {
WARN_ON(pages);
pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS);
- /* we want to make sure the amount of IO required to satisfy
- * a random read is reasonably small, so we limit the size
- * of a compressed extent to 128k
- */
ret = btrfs_zlib_compress_pages(inode->i_mapping, start,
total_compressed, pages,
nr_pages, &nr_pages_ret,
}
}
if (start == 0) {
+ trans = btrfs_join_transaction(root, 1);
+ BUG_ON(!trans);
+ btrfs_set_trans_block_group(trans, inode);
+
/* lets try to make an inline extent */
- if (ret || total_in < (end - start + 1)) {
+ if (ret || total_in < (actual_end - start)) {
/* we didn't compress the entire range, try
- * to make an uncompressed inline extent. This
- * is almost sure to fail, but maybe inline sizes
- * will get bigger later
+ * to make an uncompressed inline extent.
*/
ret = cow_file_range_inline(trans, root, inode,
start, end, 0, NULL);
} else {
+ /* try making a compressed inline extent */
ret = cow_file_range_inline(trans, root, inode,
start, end,
total_compressed, pages);
}
+ btrfs_end_transaction(trans, root);
if (ret == 0) {
+ /*
+ * inline extent creation worked, we don't need
+ * to create any more async work items. Unlock
+ * and free up our temp pages.
+ */
extent_clear_unlock_delalloc(inode,
- &BTRFS_I(inode)->io_tree,
- start, end, NULL,
- 1, 1, 1);
- *page_started = 1;
+ &BTRFS_I(inode)->io_tree,
+ start, end, NULL,
+ EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
+ EXTENT_CLEAR_DELALLOC |
+ EXTENT_CLEAR_ACCOUNTING |
+ EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK);
ret = 0;
goto free_pages_out;
}
* free any pages it allocated and our page pointer array
*/
for (i = 0; i < nr_pages_ret; i++) {
+ WARN_ON(pages[i]->mapping);
page_cache_release(pages[i]);
}
kfree(pages);
nr_pages_ret = 0;
/* flag the file so we don't compress in the future */
- btrfs_set_flag(inode, NOCOMPRESS);
+ BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
}
+ if (will_compress) {
+ *num_added += 1;
- BUG_ON(disk_num_bytes >
- btrfs_super_total_bytes(&root->fs_info->super_copy));
+ /* the async work queues will take care of doing actual
+ * allocation on disk for these compressed pages,
+ * and will submit them to the elevator.
+ */
+ add_async_extent(async_cow, start, num_bytes,
+ total_compressed, pages, nr_pages_ret);
- btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
+ if (start + num_bytes < end && start + num_bytes < actual_end) {
+ start += num_bytes;
+ pages = NULL;
+ cond_resched();
+ goto again;
+ }
+ } else {
+cleanup_and_bail_uncompressed:
+ /*
+ * No compression, but we still need to write the pages in
+ * the file we've been given so far. redirty the locked
+ * page if it corresponds to our extent and set things up
+ * for the async work queue to run cow_file_range to do
+ * the normal delalloc dance
+ */
+ if (page_offset(locked_page) >= start &&
+ page_offset(locked_page) <= end) {
+ __set_page_dirty_nobuffers(locked_page);
+ /* unlocked later on in the async handlers */
+ }
+ add_async_extent(async_cow, start, end - start + 1, 0, NULL, 0);
+ *num_added += 1;
+ }
- while(disk_num_bytes > 0) {
- unsigned long min_bytes;
+out:
+ return 0;
+free_pages_out:
+ for (i = 0; i < nr_pages_ret; i++) {
+ WARN_ON(pages[i]->mapping);
+ page_cache_release(pages[i]);
+ }
+ kfree(pages);
+
+ goto out;
+}
+
+/*
+ * phase two of compressed writeback. This is the ordered portion
+ * of the code, which only gets called in the order the work was
+ * queued. We walk all the async extents created by compress_file_range
+ * and send them down to the disk.
+ */
+static noinline int submit_compressed_extents(struct inode *inode,
+ struct async_cow *async_cow)
+{
+ struct async_extent *async_extent;
+ u64 alloc_hint = 0;
+ struct btrfs_trans_handle *trans;
+ struct btrfs_key ins;
+ struct extent_map *em;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ struct extent_io_tree *io_tree;
+ int ret;
+
+ if (list_empty(&async_cow->extents))
+ return 0;
+
+ trans = btrfs_join_transaction(root, 1);
+
+ while (!list_empty(&async_cow->extents)) {
+ async_extent = list_entry(async_cow->extents.next,
+ struct async_extent, list);
+ list_del(&async_extent->list);
+
+ io_tree = &BTRFS_I(inode)->io_tree;
+
+ /* did the compression code fall back to uncompressed IO? */
+ if (!async_extent->pages) {
+ int page_started = 0;
+ unsigned long nr_written = 0;
+
+ lock_extent(io_tree, async_extent->start,
+ async_extent->start +
+ async_extent->ram_size - 1, GFP_NOFS);
+
+ /* allocate blocks */
+ cow_file_range(inode, async_cow->locked_page,
+ async_extent->start,
+ async_extent->start +
+ async_extent->ram_size - 1,
+ &page_started, &nr_written, 0);
+
+ /*
+ * if page_started, cow_file_range inserted an
+ * inline extent and took care of all the unlocking
+ * and IO for us. Otherwise, we need to submit
+ * all those pages down to the drive.
+ */
+ if (!page_started)
+ extent_write_locked_range(io_tree,
+ inode, async_extent->start,
+ async_extent->start +
+ async_extent->ram_size - 1,
+ btrfs_get_extent,
+ WB_SYNC_ALL);
+ kfree(async_extent);
+ cond_resched();
+ continue;
+ }
+
+ lock_extent(io_tree, async_extent->start,
+ async_extent->start + async_extent->ram_size - 1,
+ GFP_NOFS);
/*
- * the max size of a compressed extent is pretty small,
- * make the code a little less complex by forcing
- * the allocator to find a whole compressed extent at once
+ * here we're doing allocation and writeback of the
+ * compressed pages
*/
- if (will_compress)
- min_bytes = disk_num_bytes;
- else
- min_bytes = root->sectorsize;
+ btrfs_drop_extent_cache(inode, async_extent->start,
+ async_extent->start +
+ async_extent->ram_size - 1, 0);
+
+ ret = btrfs_reserve_extent(trans, root,
+ async_extent->compressed_size,
+ async_extent->compressed_size,
+ 0, alloc_hint,
+ (u64)-1, &ins, 1);
+ BUG_ON(ret);
+ em = alloc_extent_map(GFP_NOFS);
+ em->start = async_extent->start;
+ em->len = async_extent->ram_size;
+ em->orig_start = em->start;
+
+ em->block_start = ins.objectid;
+ em->block_len = ins.offset;
+ em->bdev = root->fs_info->fs_devices->latest_bdev;
+ set_bit(EXTENT_FLAG_PINNED, &em->flags);
+ set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+
+ while (1) {
+ write_lock(&em_tree->lock);
+ ret = add_extent_mapping(em_tree, em);
+ write_unlock(&em_tree->lock);
+ if (ret != -EEXIST) {
+ free_extent_map(em);
+ break;
+ }
+ btrfs_drop_extent_cache(inode, async_extent->start,
+ async_extent->start +
+ async_extent->ram_size - 1, 0);
+ }
+
+ ret = btrfs_add_ordered_extent(inode, async_extent->start,
+ ins.objectid,
+ async_extent->ram_size,
+ ins.offset,
+ BTRFS_ORDERED_COMPRESSED);
+ BUG_ON(ret);
+
+ btrfs_end_transaction(trans, root);
+
+ /*
+ * clear dirty, set writeback and unlock the pages.
+ */
+ extent_clear_unlock_delalloc(inode,
+ &BTRFS_I(inode)->io_tree,
+ async_extent->start,
+ async_extent->start +
+ async_extent->ram_size - 1,
+ NULL, EXTENT_CLEAR_UNLOCK_PAGE |
+ EXTENT_CLEAR_UNLOCK |
+ EXTENT_CLEAR_DELALLOC |
+ EXTENT_CLEAR_DIRTY | EXTENT_SET_WRITEBACK);
+
+ ret = btrfs_submit_compressed_write(inode,
+ async_extent->start,
+ async_extent->ram_size,
+ ins.objectid,
+ ins.offset, async_extent->pages,
+ async_extent->nr_pages);
+
+ BUG_ON(ret);
+ trans = btrfs_join_transaction(root, 1);
+ alloc_hint = ins.objectid + ins.offset;
+ kfree(async_extent);
+ cond_resched();
+ }
+
+ btrfs_end_transaction(trans, root);
+ return 0;
+}
+
+/*
+ * when extent_io.c finds a delayed allocation range in the file,
+ * the call backs end up in this code. The basic idea is to
+ * allocate extents on disk for the range, and create ordered data structs
+ * in ram to track those extents.
+ *
+ * locked_page is the page that writepage had locked already. We use
+ * it to make sure we don't do extra locks or unlocks.
+ *
+ * *page_started is set to one if we unlock locked_page and do everything
+ * required to start IO on it. It may be clean and already done with
+ * IO when we return.
+ */
+static noinline int cow_file_range(struct inode *inode,
+ struct page *locked_page,
+ u64 start, u64 end, int *page_started,
+ unsigned long *nr_written,
+ int unlock)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_trans_handle *trans;
+ u64 alloc_hint = 0;
+ u64 num_bytes;
+ unsigned long ram_size;
+ u64 disk_num_bytes;
+ u64 cur_alloc_size;
+ u64 blocksize = root->sectorsize;
+ u64 actual_end;
+ u64 isize = i_size_read(inode);
+ struct btrfs_key ins;
+ struct extent_map *em;
+ struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ int ret = 0;
+
+ trans = btrfs_join_transaction(root, 1);
+ BUG_ON(!trans);
+ btrfs_set_trans_block_group(trans, inode);
+
+ actual_end = min_t(u64, isize, end + 1);
+
+ num_bytes = (end - start + blocksize) & ~(blocksize - 1);
+ num_bytes = max(blocksize, num_bytes);
+ disk_num_bytes = num_bytes;
+ ret = 0;
+
+ if (start == 0) {
+ /* lets try to make an inline extent */
+ ret = cow_file_range_inline(trans, root, inode,
+ start, end, 0, NULL);
+ if (ret == 0) {
+ extent_clear_unlock_delalloc(inode,
+ &BTRFS_I(inode)->io_tree,
+ start, end, NULL,
+ EXTENT_CLEAR_UNLOCK_PAGE |
+ EXTENT_CLEAR_UNLOCK |
+ EXTENT_CLEAR_DELALLOC |
+ EXTENT_CLEAR_ACCOUNTING |
+ EXTENT_CLEAR_DIRTY |
+ EXTENT_SET_WRITEBACK |
+ EXTENT_END_WRITEBACK);
+ *nr_written = *nr_written +
+ (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE;
+ *page_started = 1;
+ ret = 0;
+ goto out;
+ }
+ }
+
+ BUG_ON(disk_num_bytes >
+ btrfs_super_total_bytes(&root->fs_info->super_copy));
+
+
+ read_lock(&BTRFS_I(inode)->extent_tree.lock);
+ em = search_extent_mapping(&BTRFS_I(inode)->extent_tree,
+ start, num_bytes);
+ if (em) {
+ alloc_hint = em->block_start;
+ free_extent_map(em);
+ }
+ read_unlock(&BTRFS_I(inode)->extent_tree.lock);
+ btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
+
+ while (disk_num_bytes > 0) {
+ unsigned long op;
cur_alloc_size = min(disk_num_bytes, root->fs_info->max_extent);
ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
- min_bytes, 0, alloc_hint,
+ root->sectorsize, 0, alloc_hint,
(u64)-1, &ins, 1);
- if (ret) {
- WARN_ON(1);
- goto free_pages_out_fail;
- }
+ BUG_ON(ret);
+
em = alloc_extent_map(GFP_NOFS);
em->start = start;
-
- if (will_compress) {
- ram_size = num_bytes;
- em->len = num_bytes;
- } else {
- /* ramsize == disk size */
- ram_size = ins.offset;
- em->len = ins.offset;
- }
+ em->orig_start = em->start;
+ ram_size = ins.offset;
+ em->len = ins.offset;
em->block_start = ins.objectid;
em->block_len = ins.offset;
em->bdev = root->fs_info->fs_devices->latest_bdev;
set_bit(EXTENT_FLAG_PINNED, &em->flags);
- if (will_compress)
- set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
-
- while(1) {
- spin_lock(&em_tree->lock);
+ while (1) {
+ write_lock(&em_tree->lock);
ret = add_extent_mapping(em_tree, em);
- spin_unlock(&em_tree->lock);
+ write_unlock(&em_tree->lock);
if (ret != -EEXIST) {
free_extent_map(em);
break;
}
cur_alloc_size = ins.offset;
- ordered_type = will_compress ? BTRFS_ORDERED_COMPRESSED : 0;
ret = btrfs_add_ordered_extent(inode, start, ins.objectid,
- ram_size, cur_alloc_size,
- ordered_type);
+ ram_size, cur_alloc_size, 0);
BUG_ON(ret);
- if (disk_num_bytes < cur_alloc_size) {
- printk("num_bytes %Lu cur_alloc %Lu\n", disk_num_bytes,
- cur_alloc_size);
- break;
+ if (root->root_key.objectid ==
+ BTRFS_DATA_RELOC_TREE_OBJECTID) {
+ ret = btrfs_reloc_clone_csums(inode, start,
+ cur_alloc_size);
+ BUG_ON(ret);
}
- if (will_compress) {
- /*
- * we're doing compression, we and we need to
- * submit the compressed extents down to the device.
- *
- * We lock down all the file pages, clearing their
- * dirty bits and setting them writeback. Everyone
- * that wants to modify the page will wait on the
- * ordered extent above.
- *
- * The writeback bits on the file pages are
- * cleared when the compressed pages are on disk
- */
- btrfs_end_transaction(trans, root);
-
- if (start <= page_offset(locked_page) &&
- page_offset(locked_page) < start + ram_size) {
- *page_started = 1;
- }
-
- extent_clear_unlock_delalloc(inode,
- &BTRFS_I(inode)->io_tree,
- start,
- start + ram_size - 1,
- NULL, 1, 1, 0);
-
- ret = btrfs_submit_compressed_write(inode, start,
- ram_size, ins.objectid,
- cur_alloc_size, pages,
- nr_pages_ret);
+ if (disk_num_bytes < cur_alloc_size)
+ break;
- BUG_ON(ret);
- trans = btrfs_join_transaction(root, 1);
- if (start + ram_size < end) {
- start += ram_size;
- alloc_hint = ins.objectid + ins.offset;
- /* pages will be freed at end_bio time */
- pages = NULL;
- goto again;
- } else {
- /* we've written everything, time to go */
- break;
- }
- }
/* we're not doing compressed IO, don't unlock the first
* page (which the caller expects to stay locked), don't
* clear any dirty bits and don't set any writeback bits
+ *
+ * Do set the Private2 bit so we know this page was properly
+ * setup for writepage
*/
+ op = unlock ? EXTENT_CLEAR_UNLOCK_PAGE : 0;
+ op |= EXTENT_CLEAR_UNLOCK | EXTENT_CLEAR_DELALLOC |
+ EXTENT_SET_PRIVATE2;
+
extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
start, start + ram_size - 1,
- locked_page, 0, 0, 0);
+ locked_page, op);
disk_num_bytes -= cur_alloc_size;
num_bytes -= cur_alloc_size;
alloc_hint = ins.objectid + ins.offset;
start += cur_alloc_size;
}
-
- ret = 0;
out:
+ ret = 0;
btrfs_end_transaction(trans, root);
return ret;
+}
-free_pages_out_fail:
- extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
- start, end, locked_page, 0, 0, 0);
-free_pages_out:
- for (i = 0; i < nr_pages_ret; i++)
- page_cache_release(pages[i]);
- if (pages)
- kfree(pages);
+/*
+ * work queue call back to started compression on a file and pages
+ */
+static noinline void async_cow_start(struct btrfs_work *work)
+{
+ struct async_cow *async_cow;
+ int num_added = 0;
+ async_cow = container_of(work, struct async_cow, work);
+
+ compress_file_range(async_cow->inode, async_cow->locked_page,
+ async_cow->start, async_cow->end, async_cow,
+ &num_added);
+ if (num_added == 0)
+ async_cow->inode = NULL;
+}
- goto out;
+/*
+ * work queue call back to submit previously compressed pages
+ */
+static noinline void async_cow_submit(struct btrfs_work *work)
+{
+ struct async_cow *async_cow;
+ struct btrfs_root *root;
+ unsigned long nr_pages;
+
+ async_cow = container_of(work, struct async_cow, work);
+
+ root = async_cow->root;
+ nr_pages = (async_cow->end - async_cow->start + PAGE_CACHE_SIZE) >>
+ PAGE_CACHE_SHIFT;
+
+ atomic_sub(nr_pages, &root->fs_info->async_delalloc_pages);
+
+ if (atomic_read(&root->fs_info->async_delalloc_pages) <
+ 5 * 1042 * 1024 &&
+ waitqueue_active(&root->fs_info->async_submit_wait))
+ wake_up(&root->fs_info->async_submit_wait);
+
+ if (async_cow->inode)
+ submit_compressed_extents(async_cow->inode, async_cow);
+}
+
+static noinline void async_cow_free(struct btrfs_work *work)
+{
+ struct async_cow *async_cow;
+ async_cow = container_of(work, struct async_cow, work);
+ kfree(async_cow);
+}
+
+static int cow_file_range_async(struct inode *inode, struct page *locked_page,
+ u64 start, u64 end, int *page_started,
+ unsigned long *nr_written)
+{
+ struct async_cow *async_cow;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ unsigned long nr_pages;
+ u64 cur_end;
+ int limit = 10 * 1024 * 1042;
+
+ clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED,
+ 1, 0, NULL, GFP_NOFS);
+ while (start < end) {
+ async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS);
+ async_cow->inode = inode;
+ async_cow->root = root;
+ async_cow->locked_page = locked_page;
+ async_cow->start = start;
+
+ if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS)
+ cur_end = end;
+ else
+ cur_end = min(end, start + 512 * 1024 - 1);
+
+ async_cow->end = cur_end;
+ INIT_LIST_HEAD(&async_cow->extents);
+
+ async_cow->work.func = async_cow_start;
+ async_cow->work.ordered_func = async_cow_submit;
+ async_cow->work.ordered_free = async_cow_free;
+ async_cow->work.flags = 0;
+
+ nr_pages = (cur_end - start + PAGE_CACHE_SIZE) >>
+ PAGE_CACHE_SHIFT;
+ atomic_add(nr_pages, &root->fs_info->async_delalloc_pages);
+
+ btrfs_queue_worker(&root->fs_info->delalloc_workers,
+ &async_cow->work);
+
+ if (atomic_read(&root->fs_info->async_delalloc_pages) > limit) {
+ wait_event(root->fs_info->async_submit_wait,
+ (atomic_read(&root->fs_info->async_delalloc_pages) <
+ limit));
+ }
+
+ while (atomic_read(&root->fs_info->async_submit_draining) &&
+ atomic_read(&root->fs_info->async_delalloc_pages)) {
+ wait_event(root->fs_info->async_submit_wait,
+ (atomic_read(&root->fs_info->async_delalloc_pages) ==
+ 0));
+ }
+
+ *nr_written += nr_pages;
+ start = cur_end + 1;
+ }
+ *page_started = 1;
+ return 0;
+}
+
+static noinline int csum_exist_in_range(struct btrfs_root *root,
+ u64 bytenr, u64 num_bytes)
+{
+ int ret;
+ struct btrfs_ordered_sum *sums;
+ LIST_HEAD(list);
+
+ ret = btrfs_lookup_csums_range(root->fs_info->csum_root, bytenr,
+ bytenr + num_bytes - 1, &list);
+ if (ret == 0 && list_empty(&list))
+ return 0;
+
+ while (!list_empty(&list)) {
+ sums = list_entry(list.next, struct btrfs_ordered_sum, list);
+ list_del(&sums->list);
+ kfree(sums);
+ }
+ return 1;
}
/*
* If no cow copies or snapshots exist, we write directly to the existing
* blocks on disk
*/
-static int run_delalloc_nocow(struct inode *inode, struct page *locked_page,
- u64 start, u64 end, int *page_started)
+static noinline int run_delalloc_nocow(struct inode *inode,
+ struct page *locked_page,
+ u64 start, u64 end, int *page_started, int force,
+ unsigned long *nr_written)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
u64 cow_start;
u64 cur_offset;
u64 extent_end;
+ u64 extent_offset;
u64 disk_bytenr;
u64 num_bytes;
int extent_type;
int ret;
+ int type;
int nocow;
int check_prev = 1;
nocow = 0;
disk_bytenr = 0;
+ num_bytes = 0;
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
if (found_key.objectid > inode->i_ino ||
struct btrfs_file_extent_item);
extent_type = btrfs_file_extent_type(leaf, fi);
- if (extent_type == BTRFS_FILE_EXTENT_REG) {
- struct btrfs_block_group_cache *block_group;
+ if (extent_type == BTRFS_FILE_EXTENT_REG ||
+ extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+ extent_offset = btrfs_file_extent_offset(leaf, fi);
extent_end = found_key.offset +
btrfs_file_extent_num_bytes(leaf, fi);
if (extent_end <= start) {
path->slots[0]++;
goto next_slot;
}
+ if (disk_bytenr == 0)
+ goto out_check;
if (btrfs_file_extent_compression(leaf, fi) ||
btrfs_file_extent_encryption(leaf, fi) ||
btrfs_file_extent_other_encoding(leaf, fi))
goto out_check;
- if (disk_bytenr == 0)
+ if (extent_type == BTRFS_FILE_EXTENT_REG && !force)
goto out_check;
- if (btrfs_cross_ref_exist(trans, root, disk_bytenr))
+ if (btrfs_extent_readonly(root, disk_bytenr))
goto out_check;
- block_group = btrfs_lookup_block_group(root->fs_info,
- disk_bytenr);
- if (!block_group || block_group->ro)
+ if (btrfs_cross_ref_exist(trans, root, inode->i_ino,
+ found_key.offset -
+ extent_offset, disk_bytenr))
+ goto out_check;
+ disk_bytenr += extent_offset;
+ disk_bytenr += cur_offset - found_key.offset;
+ num_bytes = min(end + 1, extent_end) - cur_offset;
+ /*
+ * force cow if csum exists in the range.
+ * this ensure that csum for a given extent are
+ * either valid or do not exist.
+ */
+ if (csum_exist_in_range(root, disk_bytenr, num_bytes))
goto out_check;
- disk_bytenr += btrfs_file_extent_offset(leaf, fi);
nocow = 1;
} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
extent_end = found_key.offset +
goto next_slot;
}
- btrfs_release_path(root, path);
- if (cow_start != (u64)-1) {
- ret = cow_file_range(inode, locked_page, cow_start,
- found_key.offset - 1, page_started);
- BUG_ON(ret);
- cow_start = (u64)-1;
- }
+ btrfs_release_path(root, path);
+ if (cow_start != (u64)-1) {
+ ret = cow_file_range(inode, locked_page, cow_start,
+ found_key.offset - 1, page_started,
+ nr_written, 1);
+ BUG_ON(ret);
+ cow_start = (u64)-1;
+ }
+
+ if (extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
+ struct extent_map *em;
+ struct extent_map_tree *em_tree;
+ em_tree = &BTRFS_I(inode)->extent_tree;
+ em = alloc_extent_map(GFP_NOFS);
+ em->start = cur_offset;
+ em->orig_start = em->start;
+ em->len = num_bytes;
+ em->block_len = num_bytes;
+ em->block_start = disk_bytenr;
+ em->bdev = root->fs_info->fs_devices->latest_bdev;
+ set_bit(EXTENT_FLAG_PINNED, &em->flags);
+ while (1) {
+ write_lock(&em_tree->lock);
+ ret = add_extent_mapping(em_tree, em);
+ write_unlock(&em_tree->lock);
+ if (ret != -EEXIST) {
+ free_extent_map(em);
+ break;
+ }
+ btrfs_drop_extent_cache(inode, em->start,
+ em->start + em->len - 1, 0);
+ }
+ type = BTRFS_ORDERED_PREALLOC;
+ } else {
+ type = BTRFS_ORDERED_NOCOW;
+ }
+
+ ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr,
+ num_bytes, num_bytes, type);
+ BUG_ON(ret);
+
+ extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
+ cur_offset, cur_offset + num_bytes - 1,
+ locked_page, EXTENT_CLEAR_UNLOCK_PAGE |
+ EXTENT_CLEAR_UNLOCK | EXTENT_CLEAR_DELALLOC |
+ EXTENT_SET_PRIVATE2);
+ cur_offset = extent_end;
+ if (cur_offset > end)
+ break;
+ }
+ btrfs_release_path(root, path);
+
+ if (cur_offset <= end && cow_start == (u64)-1)
+ cow_start = cur_offset;
+ if (cow_start != (u64)-1) {
+ ret = cow_file_range(inode, locked_page, cow_start, end,
+ page_started, nr_written, 1);
+ BUG_ON(ret);
+ }
+
+ ret = btrfs_end_transaction(trans, root);
+ BUG_ON(ret);
+ btrfs_free_path(path);
+ return 0;
+}
+
+/*
+ * extent_io.c call back to do delayed allocation processing
+ */
+static int run_delalloc_range(struct inode *inode, struct page *locked_page,
+ u64 start, u64 end, int *page_started,
+ unsigned long *nr_written)
+{
+ int ret;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+
+ if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW)
+ ret = run_delalloc_nocow(inode, locked_page, start, end,
+ page_started, 1, nr_written);
+ else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC)
+ ret = run_delalloc_nocow(inode, locked_page, start, end,
+ page_started, 0, nr_written);
+ else if (!btrfs_test_opt(root, COMPRESS))
+ ret = cow_file_range(inode, locked_page, start, end,
+ page_started, nr_written, 1);
+ else
+ ret = cow_file_range_async(inode, locked_page, start, end,
+ page_started, nr_written);
+ return ret;
+}
+
+static int btrfs_split_extent_hook(struct inode *inode,
+ struct extent_state *orig, u64 split)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ u64 size;
+
+ if (!(orig->state & EXTENT_DELALLOC))
+ return 0;
+
+ size = orig->end - orig->start + 1;
+ if (size > root->fs_info->max_extent) {
+ u64 num_extents;
+ u64 new_size;
- disk_bytenr += cur_offset - found_key.offset;
- num_bytes = min(end + 1, extent_end) - cur_offset;
+ new_size = orig->end - split + 1;
+ num_extents = div64_u64(size + root->fs_info->max_extent - 1,
+ root->fs_info->max_extent);
- ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr,
- num_bytes, num_bytes,
- BTRFS_ORDERED_NOCOW);
- cur_offset = extent_end;
- if (cur_offset > end)
- break;
+ /*
+ * if we break a large extent up then leave oustanding_extents
+ * be, since we've already accounted for the large extent.
+ */
+ if (div64_u64(new_size + root->fs_info->max_extent - 1,
+ root->fs_info->max_extent) < num_extents)
+ return 0;
}
- btrfs_release_path(root, path);
- if (cur_offset <= end && cow_start == (u64)-1)
- cow_start = cur_offset;
- if (cow_start != (u64)-1) {
- ret = cow_file_range(inode, locked_page, cow_start, end,
- page_started);
- BUG_ON(ret);
- }
+ spin_lock(&BTRFS_I(inode)->accounting_lock);
+ BTRFS_I(inode)->outstanding_extents++;
+ spin_unlock(&BTRFS_I(inode)->accounting_lock);
- ret = btrfs_end_transaction(trans, root);
- BUG_ON(ret);
- btrfs_free_path(path);
return 0;
}
/*
- * extent_io.c call back to do delayed allocation processing
+ * extent_io.c merge_extent_hook, used to track merged delayed allocation
+ * extents so we can keep track of new extents that are just merged onto old
+ * extents, such as when we are doing sequential writes, so we can properly
+ * account for the metadata space we'll need.
*/
-static int run_delalloc_range(struct inode *inode, struct page *locked_page,
- u64 start, u64 end, int *page_started)
+static int btrfs_merge_extent_hook(struct inode *inode,
+ struct extent_state *new,
+ struct extent_state *other)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
- int ret;
+ u64 new_size, old_size;
+ u64 num_extents;
- if (btrfs_test_opt(root, NODATACOW) ||
- btrfs_test_flag(inode, NODATACOW))
- ret = run_delalloc_nocow(inode, locked_page, start, end,
- page_started);
+ /* not delalloc, ignore it */
+ if (!(other->state & EXTENT_DELALLOC))
+ return 0;
+
+ old_size = other->end - other->start + 1;
+ if (new->start < other->start)
+ new_size = other->end - new->start + 1;
else
- ret = cow_file_range(inode, locked_page, start, end,
- page_started);
+ new_size = new->end - other->start + 1;
- return ret;
+ /* we're not bigger than the max, unreserve the space and go */
+ if (new_size <= root->fs_info->max_extent) {
+ spin_lock(&BTRFS_I(inode)->accounting_lock);
+ BTRFS_I(inode)->outstanding_extents--;
+ spin_unlock(&BTRFS_I(inode)->accounting_lock);
+ return 0;
+ }
+
+ /*
+ * If we grew by another max_extent, just return, we want to keep that
+ * reserved amount.
+ */
+ num_extents = div64_u64(old_size + root->fs_info->max_extent - 1,
+ root->fs_info->max_extent);
+ if (div64_u64(new_size + root->fs_info->max_extent - 1,
+ root->fs_info->max_extent) > num_extents)
+ return 0;
+
+ spin_lock(&BTRFS_I(inode)->accounting_lock);
+ BTRFS_I(inode)->outstanding_extents--;
+ spin_unlock(&BTRFS_I(inode)->accounting_lock);
+
+ return 0;
}
/*
* bytes in this file, and to maintain the list of inodes that
* have pending delalloc work to be done.
*/
-int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
+static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
unsigned long old, unsigned long bits)
{
- unsigned long flags;
+
+ /*
+ * set_bit and clear bit hooks normally require _irqsave/restore
+ * but in this case, we are only testeing for the DELALLOC
+ * bit, which is only set or cleared with irqs on
+ */
if (!(old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
struct btrfs_root *root = BTRFS_I(inode)->root;
- spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
+
+ spin_lock(&BTRFS_I(inode)->accounting_lock);
+ BTRFS_I(inode)->outstanding_extents++;
+ spin_unlock(&BTRFS_I(inode)->accounting_lock);
+ btrfs_delalloc_reserve_space(root, inode, end - start + 1);
+ spin_lock(&root->fs_info->delalloc_lock);
BTRFS_I(inode)->delalloc_bytes += end - start + 1;
root->fs_info->delalloc_bytes += end - start + 1;
if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
&root->fs_info->delalloc_inodes);
}
- spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
+ spin_unlock(&root->fs_info->delalloc_lock);
}
return 0;
}
/*
* extent_io.c clear_bit_hook, see set_bit_hook for why
*/
-int btrfs_clear_bit_hook(struct inode *inode, u64 start, u64 end,
- unsigned long old, unsigned long bits)
+static int btrfs_clear_bit_hook(struct inode *inode,
+ struct extent_state *state, unsigned long bits)
{
- if ((old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+ /*
+ * set_bit and clear bit hooks normally require _irqsave/restore
+ * but in this case, we are only testeing for the DELALLOC
+ * bit, which is only set or cleared with irqs on
+ */
+ if ((state->state & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
struct btrfs_root *root = BTRFS_I(inode)->root;
- unsigned long flags;
- spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
- if (end - start + 1 > root->fs_info->delalloc_bytes) {
- printk("warning: delalloc account %Lu %Lu\n",
- end - start + 1, root->fs_info->delalloc_bytes);
+ if (bits & EXTENT_DO_ACCOUNTING) {
+ spin_lock(&BTRFS_I(inode)->accounting_lock);
+ BTRFS_I(inode)->outstanding_extents--;
+ spin_unlock(&BTRFS_I(inode)->accounting_lock);
+ btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
+ }
+
+ spin_lock(&root->fs_info->delalloc_lock);
+ if (state->end - state->start + 1 >
+ root->fs_info->delalloc_bytes) {
+ printk(KERN_INFO "btrfs warning: delalloc account "
+ "%llu %llu\n",
+ (unsigned long long)
+ state->end - state->start + 1,
+ (unsigned long long)
+ root->fs_info->delalloc_bytes);
+ btrfs_delalloc_free_space(root, inode, (u64)-1);
root->fs_info->delalloc_bytes = 0;
BTRFS_I(inode)->delalloc_bytes = 0;
} else {
- root->fs_info->delalloc_bytes -= end - start + 1;
- BTRFS_I(inode)->delalloc_bytes -= end - start + 1;
+ btrfs_delalloc_free_space(root, inode,
+ state->end -
+ state->start + 1);
+ root->fs_info->delalloc_bytes -= state->end -
+ state->start + 1;
+ BTRFS_I(inode)->delalloc_bytes -= state->end -
+ state->start + 1;
}
if (BTRFS_I(inode)->delalloc_bytes == 0 &&
!list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
list_del_init(&BTRFS_I(inode)->delalloc_inodes);
}
- spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
+ spin_unlock(&root->fs_info->delalloc_lock);
}
return 0;
}
u64 map_length;
int ret;
+ if (bio_flags & EXTENT_BIO_COMPRESSED)
+ return 0;
+
length = bio->bi_size;
map_tree = &root->fs_info->mapping_tree;
map_length = length;
ret = btrfs_map_block(map_tree, READ, logical,
&map_length, NULL, 0);
- if (map_length < length + size) {
+ if (map_length < length + size)
return 1;
- }
return 0;
}
* At IO completion time the cums attached on the ordered extent record
* are inserted into the btree
*/
-int __btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
- int mirror_num, unsigned long bio_flags)
+static int __btrfs_submit_bio_start(struct inode *inode, int rw,
+ struct bio *bio, int mirror_num,
+ unsigned long bio_flags)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
int ret = 0;
- ret = btrfs_csum_one_bio(root, inode, bio);
+ ret = btrfs_csum_one_bio(root, inode, bio, 0, 0);
BUG_ON(ret);
+ return 0;
+}
+/*
+ * in order to insert checksums into the metadata in large chunks,
+ * we wait until bio submission time. All the pages in the bio are
+ * checksummed and sums are attached onto the ordered extent record.
+ *
+ * At IO completion time the cums attached on the ordered extent record
+ * are inserted into the btree
+ */
+static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
+ int mirror_num, unsigned long bio_flags)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
return btrfs_map_bio(root, rw, bio, mirror_num, 1);
}
/*
- * extent_io.c submission hook. This does the right thing for csum calculation on write,
- * or reading the csums from the tree before a read
+ * extent_io.c submission hook. This does the right thing for csum calculation
+ * on write, or reading the csums from the tree before a read
*/
-int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
+static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
int mirror_num, unsigned long bio_flags)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
int ret = 0;
int skip_sum;
+ skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
+
ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
BUG_ON(ret);
- skip_sum = btrfs_test_opt(root, NODATASUM) ||
- btrfs_test_flag(inode, NODATASUM);
-
if (!(rw & (1 << BIO_RW))) {
- if (!skip_sum)
- btrfs_lookup_bio_sums(root, inode, bio);
-
- if (bio_flags & EXTENT_BIO_COMPRESSED)
+ if (bio_flags & EXTENT_BIO_COMPRESSED) {
return btrfs_submit_compressed_read(inode, bio,
mirror_num, bio_flags);
+ } else if (!skip_sum)
+ btrfs_lookup_bio_sums(root, inode, bio, NULL);
goto mapit;
} else if (!skip_sum) {
+ /* csum items have already been cloned */
+ if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID)
+ goto mapit;
/* we're doing a write, do the async checksumming */
return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
inode, rw, bio, mirror_num,
- bio_flags, __btrfs_submit_bio_hook);
+ bio_flags, __btrfs_submit_bio_start,
+ __btrfs_submit_bio_done);
}
mapit:
struct inode *inode, u64 file_offset,
struct list_head *list)
{
- struct list_head *cur;
struct btrfs_ordered_sum *sum;
btrfs_set_trans_block_group(trans, inode);
- list_for_each(cur, list) {
- sum = list_entry(cur, struct btrfs_ordered_sum, list);
- btrfs_csum_file_blocks(trans, BTRFS_I(inode)->root,
- inode, sum);
+
+ list_for_each_entry(sum, list, list) {
+ btrfs_csum_file_blocks(trans,
+ BTRFS_I(inode)->root->fs_info->csum_root, sum);
}
return 0;
}
int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end)
{
+ if ((end & (PAGE_CACHE_SIZE - 1)) == 0)
+ WARN_ON(1);
return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
GFP_NOFS);
}
struct btrfs_work work;
};
-void btrfs_writepage_fixup_worker(struct btrfs_work *work)
+static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
{
struct btrfs_writepage_fixup *fixup;
struct btrfs_ordered_extent *ordered;
lock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
/* already ordered? We're done */
- if (test_range_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
- EXTENT_ORDERED, 0)) {
+ if (PagePrivate2(page))
goto out;
- }
ordered = btrfs_lookup_ordered_extent(inode, page_start);
if (ordered) {
* to fix it up. The async helper will wait for ordered extents, set
* the delalloc bit and make it safe to write the page.
*/
-int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end)
+static int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end)
{
struct inode *inode = page->mapping->host;
struct btrfs_writepage_fixup *fixup;
struct btrfs_root *root = BTRFS_I(inode)->root;
- int ret;
- ret = test_range_bit(&BTRFS_I(inode)->io_tree, start, end,
- EXTENT_ORDERED, 0);
- if (ret)
+ /* this page is properly in the ordered list */
+ if (TestClearPagePrivate2(page))
return 0;
if (PageChecked(page))
return -EAGAIN;
}
+static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
+ struct inode *inode, u64 file_pos,
+ u64 disk_bytenr, u64 disk_num_bytes,
+ u64 num_bytes, u64 ram_bytes,
+ u64 locked_end,
+ u8 compression, u8 encryption,
+ u16 other_encoding, int extent_type)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_file_extent_item *fi;
+ struct btrfs_path *path;
+ struct extent_buffer *leaf;
+ struct btrfs_key ins;
+ u64 hint;
+ int ret;
+
+ path = btrfs_alloc_path();
+ BUG_ON(!path);
+
+ path->leave_spinning = 1;
+
+ /*
+ * we may be replacing one extent in the tree with another.
+ * The new extent is pinned in the extent map, and we don't want
+ * to drop it from the cache until it is completely in the btree.
+ *
+ * So, tell btrfs_drop_extents to leave this extent in the cache.
+ * the caller is expected to unpin it and allow it to be merged
+ * with the others.
+ */
+ ret = btrfs_drop_extents(trans, root, inode, file_pos,
+ file_pos + num_bytes, locked_end,
+ file_pos, &hint, 0);
+ BUG_ON(ret);
+
+ ins.objectid = inode->i_ino;
+ ins.offset = file_pos;
+ ins.type = BTRFS_EXTENT_DATA_KEY;
+ ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi));
+ BUG_ON(ret);
+ leaf = path->nodes[0];
+ fi = btrfs_item_ptr(leaf, path->slots[0],
+ struct btrfs_file_extent_item);
+ btrfs_set_file_extent_generation(leaf, fi, trans->transid);
+ btrfs_set_file_extent_type(leaf, fi, extent_type);
+ btrfs_set_file_extent_disk_bytenr(leaf, fi, disk_bytenr);
+ btrfs_set_file_extent_disk_num_bytes(leaf, fi, disk_num_bytes);
+ btrfs_set_file_extent_offset(leaf, fi, 0);
+ btrfs_set_file_extent_num_bytes(leaf, fi, num_bytes);
+ btrfs_set_file_extent_ram_bytes(leaf, fi, ram_bytes);
+ btrfs_set_file_extent_compression(leaf, fi, compression);
+ btrfs_set_file_extent_encryption(leaf, fi, encryption);
+ btrfs_set_file_extent_other_encoding(leaf, fi, other_encoding);
+
+ btrfs_unlock_up_safe(path, 1);
+ btrfs_set_lock_blocking(leaf);
+
+ btrfs_mark_buffer_dirty(leaf);
+
+ inode_add_bytes(inode, num_bytes);
+
+ ins.objectid = disk_bytenr;
+ ins.offset = disk_num_bytes;
+ ins.type = BTRFS_EXTENT_ITEM_KEY;
+ ret = btrfs_alloc_reserved_file_extent(trans, root,
+ root->root_key.objectid,
+ inode->i_ino, file_pos, &ins);
+ BUG_ON(ret);
+ btrfs_free_path(path);
+
+ return 0;
+}
+
+/*
+ * helper function for btrfs_finish_ordered_io, this
+ * just reads in some of the csum leaves to prime them into ram
+ * before we start the transaction. It limits the amount of btree
+ * reads required while inside the transaction.
+ */
+static noinline void reada_csum(struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct btrfs_ordered_extent *ordered_extent)
+{
+ struct btrfs_ordered_sum *sum;
+ u64 bytenr;
+
+ sum = list_entry(ordered_extent->list.next, struct btrfs_ordered_sum,
+ list);
+ bytenr = sum->sums[0].bytenr;
+
+ /*
+ * we don't care about the results, the point of this search is
+ * just to get the btree leaves into ram
+ */
+ btrfs_lookup_csum(NULL, root->fs_info->csum_root, path, bytenr, 0);
+}
+
/* as ordered data IO finishes, this gets called so we can finish
* an ordered extent if the range of bytes in the file it covers are
* fully written.
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
- struct btrfs_ordered_extent *ordered_extent;
+ struct btrfs_ordered_extent *ordered_extent = NULL;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
- struct btrfs_file_extent_item *extent_item;
- struct btrfs_path *path = NULL;
- struct extent_buffer *leaf;
- u64 alloc_hint = 0;
- struct list_head list;
- struct btrfs_key ins;
+ struct btrfs_path *path;
+ int compressed = 0;
int ret;
ret = btrfs_dec_test_ordered_pending(inode, start, end - start + 1);
if (!ret)
return 0;
+ /*
+ * before we join the transaction, try to do some of our IO.
+ * This will limit the amount of IO that we have to do with
+ * the transaction running. We're unlikely to need to do any
+ * IO if the file extents are new, the disk_i_size checks
+ * covers the most common case.
+ */
+ if (start < BTRFS_I(inode)->disk_i_size) {
+ path = btrfs_alloc_path();
+ if (path) {
+ ret = btrfs_lookup_file_extent(NULL, root, path,
+ inode->i_ino,
+ start, 0);
+ ordered_extent = btrfs_lookup_ordered_extent(inode,
+ start);
+ if (!list_empty(&ordered_extent->list)) {
+ btrfs_release_path(root, path);
+ reada_csum(root, path, ordered_extent);
+ }
+ btrfs_free_path(path);
+ }
+ }
+
trans = btrfs_join_transaction(root, 1);
- ordered_extent = btrfs_lookup_ordered_extent(inode, start);
+ if (!ordered_extent)
+ ordered_extent = btrfs_lookup_ordered_extent(inode, start);
BUG_ON(!ordered_extent);
if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags))
goto nocow;
- path = btrfs_alloc_path();
- BUG_ON(!path);
-
lock_extent(io_tree, ordered_extent->file_offset,
ordered_extent->file_offset + ordered_extent->len - 1,
GFP_NOFS);
- INIT_LIST_HEAD(&list);
-
- ret = btrfs_drop_extents(trans, root, inode,
- ordered_extent->file_offset,
- ordered_extent->file_offset +
- ordered_extent->len,
- ordered_extent->file_offset, &alloc_hint);
- BUG_ON(ret);
-
- ins.objectid = inode->i_ino;
- ins.offset = ordered_extent->file_offset;
- ins.type = BTRFS_EXTENT_DATA_KEY;
- ret = btrfs_insert_empty_item(trans, root, path, &ins,
- sizeof(*extent_item));
- BUG_ON(ret);
- leaf = path->nodes[0];
- extent_item = btrfs_item_ptr(leaf, path->slots[0],
- struct btrfs_file_extent_item);
- btrfs_set_file_extent_generation(leaf, extent_item, trans->transid);
- btrfs_set_file_extent_type(leaf, extent_item, BTRFS_FILE_EXTENT_REG);
- btrfs_set_file_extent_disk_bytenr(leaf, extent_item,
- ordered_extent->start);
- btrfs_set_file_extent_disk_num_bytes(leaf, extent_item,
- ordered_extent->disk_len);
- btrfs_set_file_extent_offset(leaf, extent_item, 0);
-
if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
- btrfs_set_file_extent_compression(leaf, extent_item, 1);
- else
- btrfs_set_file_extent_compression(leaf, extent_item, 0);
- btrfs_set_file_extent_encryption(leaf, extent_item, 0);
- btrfs_set_file_extent_other_encoding(leaf, extent_item, 0);
-
- /* ram bytes = extent_num_bytes for now */
- btrfs_set_file_extent_num_bytes(leaf, extent_item,
- ordered_extent->len);
- btrfs_set_file_extent_ram_bytes(leaf, extent_item,
- ordered_extent->len);
- btrfs_mark_buffer_dirty(leaf);
-
- btrfs_drop_extent_cache(inode, ordered_extent->file_offset,
- ordered_extent->file_offset +
- ordered_extent->len - 1, 0);
-
- ins.objectid = ordered_extent->start;
- ins.offset = ordered_extent->disk_len;
- ins.type = BTRFS_EXTENT_ITEM_KEY;
- ret = btrfs_alloc_reserved_extent(trans, root, leaf->start,
- root->root_key.objectid,
- trans->transid, inode->i_ino, &ins);
- BUG_ON(ret);
- btrfs_release_path(root, path);
-
- inode_add_bytes(inode, ordered_extent->len);
+ compressed = 1;
+ if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) {
+ BUG_ON(compressed);
+ ret = btrfs_mark_extent_written(trans, root, inode,
+ ordered_extent->file_offset,
+ ordered_extent->file_offset +
+ ordered_extent->len);
+ BUG_ON(ret);
+ } else {
+ ret = insert_reserved_file_extent(trans, inode,
+ ordered_extent->file_offset,
+ ordered_extent->start,
+ ordered_extent->disk_len,
+ ordered_extent->len,
+ ordered_extent->len,
+ ordered_extent->file_offset +
+ ordered_extent->len,
+ compressed, 0, 0,
+ BTRFS_FILE_EXTENT_REG);
+ unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
+ ordered_extent->file_offset,
+ ordered_extent->len);
+ BUG_ON(ret);
+ }
unlock_extent(io_tree, ordered_extent->file_offset,
ordered_extent->file_offset + ordered_extent->len - 1,
GFP_NOFS);
btrfs_put_ordered_extent(ordered_extent);
btrfs_end_transaction(trans, root);
- if (path)
- btrfs_free_path(path);
return 0;
}
-int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
+static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
struct extent_state *state, int uptodate)
{
+ ClearPagePrivate2(page);
return btrfs_finish_ordered_io(page->mapping->host, start, end);
}
u64 start;
u64 len;
u64 logical;
+ unsigned long bio_flags;
int last_mirror;
};
-int btrfs_io_failed_hook(struct bio *failed_bio,
+static int btrfs_io_failed_hook(struct bio *failed_bio,
struct page *page, u64 start, u64 end,
struct extent_state *state)
{
int ret;
int rw;
u64 logical;
- unsigned long bio_flags = 0;
ret = get_state_private(failure_tree, start, &private);
if (ret) {
failrec->start = start;
failrec->len = end - start + 1;
failrec->last_mirror = 0;
+ failrec->bio_flags = 0;
- spin_lock(&em_tree->lock);
+ read_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree, start, failrec->len);
if (em->start > start || em->start + em->len < start) {
free_extent_map(em);
em = NULL;
}
- spin_unlock(&em_tree->lock);
+ read_unlock(&em_tree->lock);
if (!em || IS_ERR(em)) {
kfree(failrec);
}
logical = start - em->start;
logical = em->block_start + logical;
- if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
- bio_flags = EXTENT_BIO_COMPRESSED;
+ if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
+ logical = em->block_start;
+ failrec->bio_flags = EXTENT_BIO_COMPRESSED;
+ }
failrec->logical = logical;
free_extent_map(em);
set_extent_bits(failure_tree, start, end, EXTENT_LOCKED |
failrec->logical, failrec->len);
failrec->last_mirror++;
if (!state) {
- spin_lock_irq(&BTRFS_I(inode)->io_tree.lock);
+ spin_lock(&BTRFS_I(inode)->io_tree.lock);
state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree,
failrec->start,
EXTENT_LOCKED);
if (state && state->start != failrec->start)
state = NULL;
- spin_unlock_irq(&BTRFS_I(inode)->io_tree.lock);
+ spin_unlock(&BTRFS_I(inode)->io_tree.lock);
}
if (!state || failrec->last_mirror > num_copies) {
set_state_private(failure_tree, failrec->start, 0);
bio->bi_sector = failrec->logical >> 9;
bio->bi_bdev = failed_bio->bi_bdev;
bio->bi_size = 0;
+
bio_add_page(bio, page, failrec->len, start - page_offset(page));
if (failed_bio->bi_rw & (1 << BIO_RW))
rw = WRITE;
BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio,
failrec->last_mirror,
- bio_flags);
+ failrec->bio_flags);
return 0;
}
* each time an IO finishes, we do a fast check in the IO failure tree
* to see if we need to process or clean up an io_failure_record
*/
-int btrfs_clean_io_failures(struct inode *inode, u64 start)
+static int btrfs_clean_io_failures(struct inode *inode, u64 start)
{
u64 private;
u64 private_failure;
* if there's a match, we allow the bio to finish. If not, we go through
* the io_failure_record routines to find good copies
*/
-int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
+static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
struct extent_state *state)
{
size_t offset = start - ((u64)page->index << PAGE_CACHE_SHIFT);
int ret;
struct btrfs_root *root = BTRFS_I(inode)->root;
u32 csum = ~(u32)0;
- unsigned long flags;
- if (btrfs_test_opt(root, NODATASUM) ||
- btrfs_test_flag(inode, NODATASUM))
+ if (PageChecked(page)) {
+ ClearPageChecked(page);
+ goto good;
+ }
+
+ if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)
+ return 0;
+
+ if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID &&
+ test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) {
+ clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM,
+ GFP_NOFS);
return 0;
+ }
+
if (state && state->start == start) {
private = state->private;
ret = 0;
} else {
ret = get_state_private(io_tree, start, &private);
}
- local_irq_save(flags);
- kaddr = kmap_atomic(page, KM_IRQ0);
- if (ret) {
+ kaddr = kmap_atomic(page, KM_USER0);
+ if (ret)
goto zeroit;
- }
+
csum = btrfs_csum_data(root, kaddr + offset, csum, end - start + 1);
btrfs_csum_final(csum, (char *)&csum);
- if (csum != private) {
+ if (csum != private)
goto zeroit;
- }
- kunmap_atomic(kaddr, KM_IRQ0);
- local_irq_restore(flags);
+ kunmap_atomic(kaddr, KM_USER0);
+good:
/* if the io failure tree for this inode is non-empty,
* check to see if we've recovered from a failed IO
*/
return 0;
zeroit:
- printk("btrfs csum failed ino %lu off %llu csum %u private %Lu\n",
- page->mapping->host->i_ino, (unsigned long long)start, csum,
- private);
+ if (printk_ratelimit()) {
+ printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
+ "private %llu\n", page->mapping->host->i_ino,
+ (unsigned long long)start, csum,
+ (unsigned long long)private);
+ }
memset(kaddr + offset, 1, end - start + 1);
flush_dcache_page(page);
- kunmap_atomic(kaddr, KM_IRQ0);
- local_irq_restore(flags);
+ kunmap_atomic(kaddr, KM_USER0);
if (private == 0)
return 0;
return -EIO;
struct inode *inode;
int ret = 0, nr_unlink = 0, nr_truncate = 0;
- /* don't do orphan cleanup if the fs is readonly. */
- if (root->fs_info->sb->s_flags & MS_RDONLY)
- return;
-
path = btrfs_alloc_path();
if (!path)
return;
* crossing root thing. we store the inode number in the
* offset of the orphan item.
*/
- inode = btrfs_iget_locked(root->fs_info->sb,
- found_key.offset, root);
- if (!inode)
+ found_key.objectid = found_key.offset;
+ found_key.type = BTRFS_INODE_ITEM_KEY;
+ found_key.offset = 0;
+ inode = btrfs_iget(root->fs_info->sb, &found_key, root);
+ if (IS_ERR(inode))
break;
- if (inode->i_state & I_NEW) {
- BTRFS_I(inode)->root = root;
-
- /* have to set the location manually */
- BTRFS_I(inode)->location.objectid = inode->i_ino;
- BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
- BTRFS_I(inode)->location.offset = 0;
-
- btrfs_read_locked_inode(inode);
- unlock_new_inode(inode);
- }
-
/*
* add this inode to the orphan list so btrfs_orphan_del does
* the proper thing when we hit it
}
/*
+ * very simple check to peek ahead in the leaf looking for xattrs. If we
+ * don't find any xattrs, we know there can't be any acls.
+ *
+ * slot is the slot the inode is in, objectid is the objectid of the inode
+ */
+static noinline int acls_after_inode_item(struct extent_buffer *leaf,
+ int slot, u64 objectid)
+{
+ u32 nritems = btrfs_header_nritems(leaf);
+ struct btrfs_key found_key;
+ int scanned = 0;
+
+ slot++;
+ while (slot < nritems) {
+ btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+ /* we found a different objectid, there must not be acls */
+ if (found_key.objectid != objectid)
+ return 0;
+
+ /* we found an xattr, assume we've got an acl */
+ if (found_key.type == BTRFS_XATTR_ITEM_KEY)
+ return 1;
+
+ /*
+ * we found a key greater than an xattr key, there can't
+ * be any acls later on
+ */
+ if (found_key.type > BTRFS_XATTR_ITEM_KEY)
+ return 0;
+
+ slot++;
+ scanned++;
+
+ /*
+ * it goes inode, inode backrefs, xattrs, extents,
+ * so if there are a ton of hard links to an inode there can
+ * be a lot of backrefs. Don't waste time searching too hard,
+ * this is just an optimization
+ */
+ if (scanned >= 8)
+ break;
+ }
+ /* we hit the end of the leaf before we found an xattr or
+ * something larger than an xattr. We have to assume the inode
+ * has acls
+ */
+ return 1;
+}
+
+/*
* read an inode from the btree into the in-memory inode
*/
-void btrfs_read_locked_inode(struct inode *inode)
+static void btrfs_read_locked_inode(struct inode *inode)
{
struct btrfs_path *path;
struct extent_buffer *leaf;
struct btrfs_timespec *tspec;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_key location;
+ int maybe_acls;
u64 alloc_group_block;
u32 rdev;
int ret;
inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item));
BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
+ BTRFS_I(inode)->sequence = btrfs_inode_sequence(leaf, inode_item);
inode->i_generation = BTRFS_I(inode)->generation;
inode->i_rdev = 0;
rdev = btrfs_inode_rdev(leaf, inode_item);
BTRFS_I(inode)->index_cnt = (u64)-1;
+ BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item);
alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
- BTRFS_I(inode)->block_group = btrfs_lookup_block_group(root->fs_info,
- alloc_group_block);
- BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item);
- if (!BTRFS_I(inode)->block_group) {
- BTRFS_I(inode)->block_group = btrfs_find_block_group(root,
- NULL, 0,
- BTRFS_BLOCK_GROUP_METADATA, 0);
- }
+
+ /*
+ * try to precache a NULL acl entry for files that don't have
+ * any xattrs or acls
+ */
+ maybe_acls = acls_after_inode_item(leaf, path->slots[0], inode->i_ino);
+ if (!maybe_acls)
+ cache_no_acl(inode);
+
+ BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0,
+ alloc_group_block, 0);
btrfs_free_path(path);
inode_item = NULL;
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
break;
default:
+ inode->i_op = &btrfs_special_inode_operations;
init_special_inode(inode, inode->i_mode, rdev);
break;
}
+
+ btrfs_update_iflags(inode);
return;
make_bad:
btrfs_set_inode_nbytes(leaf, item, inode_get_bytes(inode));
btrfs_set_inode_generation(leaf, item, BTRFS_I(inode)->generation);
+ btrfs_set_inode_sequence(leaf, item, BTRFS_I(inode)->sequence);
btrfs_set_inode_transid(leaf, item, trans->transid);
btrfs_set_inode_rdev(leaf, item, inode->i_rdev);
btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags);
- btrfs_set_inode_block_group(leaf, item,
- BTRFS_I(inode)->block_group->key.objectid);
+ btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group);
}
/*
* copy everything in the in-memory inode into the btree.
*/
-int noinline btrfs_update_inode(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct inode *inode)
+noinline int btrfs_update_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode)
{
struct btrfs_inode_item *inode_item;
struct btrfs_path *path;
path = btrfs_alloc_path();
BUG_ON(!path);
+ path->leave_spinning = 1;
ret = btrfs_lookup_inode(trans, root, path,
&BTRFS_I(inode)->location, 1);
if (ret) {
goto failed;
}
+ btrfs_unlock_up_safe(path, 1);
leaf = path->nodes[0];
inode_item = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_inode_item);
goto err;
}
+ path->leave_spinning = 1;
di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
name, name_len, -1);
if (IS_ERR(di)) {
inode->i_ino,
dir->i_ino, &index);
if (ret) {
- printk("failed to delete reference to %.*s, "
+ printk(KERN_INFO "btrfs failed to delete reference to %.*s, "
"inode %lu parent %lu\n", name_len, name,
inode->i_ino, dir->i_ino);
goto err;
ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len,
inode, dir->i_ino);
BUG_ON(ret != 0 && ret != -ENOENT);
- if (ret != -ENOENT)
- BTRFS_I(dir)->log_dirty_trans = trans->transid;
ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len,
dir, index);
btrfs_update_inode(trans, root, dir);
btrfs_drop_nlink(inode);
ret = btrfs_update_inode(trans, root, inode);
- dir->i_sb->s_dirt = 1;
out:
return ret;
}
root = BTRFS_I(dir)->root;
- ret = btrfs_check_free_space(root, 1, 1);
- if (ret)
- goto fail;
-
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, dir);
+
+ btrfs_record_unlink_dir(trans, dir, dentry->d_inode, 0);
+
ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
dentry->d_name.name, dentry->d_name.len);
- if (inode->i_nlink == 0)
- ret = btrfs_orphan_add(trans, inode);
+ if (inode->i_nlink == 0)
+ ret = btrfs_orphan_add(trans, inode);
+
+ nr = trans->blocks_used;
+
+ btrfs_end_transaction_throttle(trans, root);
+ btrfs_btree_balance_dirty(root, nr);
+ return ret;
+}
+
+int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *dir, u64 objectid,
+ const char *name, int name_len)
+{
+ struct btrfs_path *path;
+ struct extent_buffer *leaf;
+ struct btrfs_dir_item *di;
+ struct btrfs_key key;
+ u64 index;
+ int ret;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+ name, name_len, -1);
+ BUG_ON(!di || IS_ERR(di));
+
+ leaf = path->nodes[0];
+ btrfs_dir_item_key_to_cpu(leaf, di, &key);
+ WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid);
+ ret = btrfs_delete_one_dir_name(trans, root, path, di);
+ BUG_ON(ret);
+ btrfs_release_path(root, path);
+
+ ret = btrfs_del_root_ref(trans, root->fs_info->tree_root,
+ objectid, root->root_key.objectid,
+ dir->i_ino, &index, name, name_len);
+ if (ret < 0) {
+ BUG_ON(ret != -ENOENT);
+ di = btrfs_search_dir_index_item(root, path, dir->i_ino,
+ name, name_len);
+ BUG_ON(!di || IS_ERR(di));
+
+ leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+ btrfs_release_path(root, path);
+ index = key.offset;
+ }
+
+ di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+ index, name, name_len, -1);
+ BUG_ON(!di || IS_ERR(di));
+
+ leaf = path->nodes[0];
+ btrfs_dir_item_key_to_cpu(leaf, di, &key);
+ WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid);
+ ret = btrfs_delete_one_dir_name(trans, root, path, di);
+ BUG_ON(ret);
+ btrfs_release_path(root, path);
- nr = trans->blocks_used;
+ btrfs_i_size_write(dir, dir->i_size - name_len * 2);
+ dir->i_mtime = dir->i_ctime = CURRENT_TIME;
+ ret = btrfs_update_inode(trans, root, dir);
+ BUG_ON(ret);
+ dir->i_sb->s_dirt = 1;
- btrfs_end_transaction_throttle(trans, root);
-fail:
- btrfs_btree_balance_dirty(root, nr);
- return ret;
+ btrfs_free_path(path);
+ return 0;
}
static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
struct btrfs_trans_handle *trans;
unsigned long nr = 0;
- if (inode->i_size > BTRFS_EMPTY_DIR_SIZE) {
+ if (inode->i_size > BTRFS_EMPTY_DIR_SIZE ||
+ inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
return -ENOTEMPTY;
- }
-
- ret = btrfs_check_free_space(root, 1, 1);
- if (ret)
- goto fail;
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, dir);
+ if (unlikely(inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
+ err = btrfs_unlink_subvol(trans, root, dir,
+ BTRFS_I(inode)->location.objectid,
+ dentry->d_name.name,
+ dentry->d_name.len);
+ goto out;
+ }
+
err = btrfs_orphan_add(trans, inode);
if (err)
- goto fail_trans;
+ goto out;
/* now the directory is empty */
err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
dentry->d_name.name, dentry->d_name.len);
- if (!err) {
+ if (!err)
btrfs_i_size_write(inode, 0);
- }
-
-fail_trans:
+out:
nr = trans->blocks_used;
ret = btrfs_end_transaction_throttle(trans, root);
-fail:
btrfs_btree_balance_dirty(root, nr);
if (ret && !err)
return err;
}
+#if 0
/*
* when truncating bytes in a file, it is possible to avoid reading
* the leaves that contain only checksum items. This can be the
ref->generation = leaf_gen;
ref->nritems = 0;
+ btrfs_sort_leaf_ref(ref);
+
ret = btrfs_add_leaf_ref(root, ref, 0);
WARN_ON(ret);
btrfs_free_leaf_ref(root, ref);
return ret;
}
+#endif
+
/*
* this can truncate away extent items, csum items and directory items.
* It starts at a high offset and removes keys until it can't find
struct btrfs_path *path;
struct btrfs_key key;
struct btrfs_key found_key;
- u32 found_type;
+ u32 found_type = (u8)-1;
struct extent_buffer *leaf;
struct btrfs_file_extent_item *fi;
u64 extent_start = 0;
u64 extent_num_bytes = 0;
+ u64 extent_offset = 0;
u64 item_end = 0;
- u64 root_gen = 0;
- u64 root_owner = 0;
int found_extent;
int del_item;
int pending_del_nr = 0;
int pending_del_slot = 0;
int extent_type = -1;
+ int encoding;
u64 mask = root->sectorsize - 1;
if (root->ref_cows)
btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0);
path = btrfs_alloc_path();
- path->reada = -1;
BUG_ON(!path);
+ path->reada = -1;
/* FIXME, add redo link to tree so we don't leak on crash */
key.objectid = inode->i_ino;
key.offset = (u64)-1;
key.type = (u8)-1;
- btrfs_init_path(path);
-
- ret = drop_csum_leaves(trans, root, path, inode, new_size);
- BUG_ON(ret);
-
search_again:
+ path->leave_spinning = 1;
ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
- if (ret < 0) {
+ if (ret < 0)
goto error;
- }
+
if (ret > 0) {
/* there are no items in the tree for us to truncate, we're
* done
path->slots[0]--;
}
- while(1) {
+ while (1) {
fi = NULL;
leaf = path->nodes[0];
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
found_type = btrfs_key_type(&found_key);
+ encoding = 0;
if (found_key.objectid != inode->i_ino)
break;
fi = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_file_extent_item);
extent_type = btrfs_file_extent_type(leaf, fi);
+ encoding = btrfs_file_extent_compression(leaf, fi);
+ encoding |= btrfs_file_extent_encryption(leaf, fi);
+ encoding |= btrfs_file_extent_other_encoding(leaf, fi);
+
if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
item_end +=
btrfs_file_extent_num_bytes(leaf, fi);
}
item_end--;
}
- if (found_type == BTRFS_CSUM_ITEM_KEY) {
- ret = btrfs_csum_truncate(trans, root, path,
- new_size);
- BUG_ON(ret);
- }
if (item_end < new_size) {
- if (found_type == BTRFS_DIR_ITEM_KEY) {
+ if (found_type == BTRFS_DIR_ITEM_KEY)
found_type = BTRFS_INODE_ITEM_KEY;
- } else if (found_type == BTRFS_EXTENT_ITEM_KEY) {
- found_type = BTRFS_CSUM_ITEM_KEY;
- } else if (found_type == BTRFS_EXTENT_DATA_KEY) {
+ else if (found_type == BTRFS_EXTENT_ITEM_KEY)
+ found_type = BTRFS_EXTENT_DATA_KEY;
+ else if (found_type == BTRFS_EXTENT_DATA_KEY)
found_type = BTRFS_XATTR_ITEM_KEY;
- } else if (found_type == BTRFS_XATTR_ITEM_KEY) {
+ else if (found_type == BTRFS_XATTR_ITEM_KEY)
found_type = BTRFS_INODE_REF_KEY;
- } else if (found_type) {
+ else if (found_type)
found_type--;
- } else {
+ else
break;
- }
btrfs_set_key_type(&key, found_type);
goto next;
}
if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
u64 num_dec;
extent_start = btrfs_file_extent_disk_bytenr(leaf, fi);
- if (!del_item) {
+ if (!del_item && !encoding) {
u64 orig_num_bytes =
btrfs_file_extent_num_bytes(leaf, fi);
extent_num_bytes = new_size -
extent_num_bytes =
btrfs_file_extent_disk_num_bytes(leaf,
fi);
+ extent_offset = found_key.offset -
+ btrfs_file_extent_offset(leaf, fi);
+
/* FIXME blocksize != 4096 */
num_dec = btrfs_file_extent_num_bytes(leaf, fi);
if (extent_start != 0) {
if (root->ref_cows)
inode_sub_bytes(inode, num_dec);
}
- root_gen = btrfs_header_generation(leaf);
- root_owner = btrfs_header_owner(leaf);
}
} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
/*
pending_del_nr++;
pending_del_slot = path->slots[0];
} else {
- printk("bad pending slot %d pending_del_nr %d pending_del_slot %d\n", path->slots[0], pending_del_nr, pending_del_slot);
+ BUG();
}
} else {
break;
}
- if (found_extent) {
+ if (found_extent && root->ref_cows) {
+ btrfs_set_path_blocking(path);
ret = btrfs_free_extent(trans, root, extent_start,
- extent_num_bytes,
- leaf->start, root_owner,
- root_gen, inode->i_ino, 0);
+ extent_num_bytes, 0,
+ btrfs_header_owner(leaf),
+ inode->i_ino, extent_offset);
BUG_ON(ret);
}
next:
if (pending_del_nr)
goto del_pending;
btrfs_release_path(root, path);
+ if (found_type == BTRFS_INODE_ITEM_KEY)
+ break;
goto search_again;
}
BUG_ON(ret);
pending_del_nr = 0;
btrfs_release_path(root, path);
+ if (found_type == BTRFS_INODE_ITEM_KEY)
+ break;
goto search_again;
}
}
pending_del_nr);
}
btrfs_free_path(path);
- inode->i_sb->s_dirt = 1;
return ret;
}
goto again;
}
- btrfs_set_extent_delalloc(inode, page_start, page_end);
+ ret = btrfs_set_extent_delalloc(inode, page_start, page_end);
+ if (ret) {
+ unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ goto out_unlock;
+ }
+
ret = 0;
if (offset != PAGE_CACHE_SIZE) {
kaddr = kmap(page);
u64 last_byte;
u64 cur_offset;
u64 hole_size;
- int err;
+ int err = 0;
if (size <= hole_start)
return 0;
- err = btrfs_check_free_space(root, 1, 0);
- if (err)
- return err;
-
btrfs_truncate_page(inode->i_mapping, inode->i_size);
while (1) {
last_byte = min(extent_map_end(em), block_end);
last_byte = (last_byte + mask) & ~mask;
if (test_bit(EXTENT_FLAG_VACANCY, &em->flags)) {
+ u64 hint_byte = 0;
hole_size = last_byte - cur_offset;
+ err = btrfs_drop_extents(trans, root, inode,
+ cur_offset,
+ cur_offset + hole_size,
+ block_end,
+ cur_offset, &hint_byte, 1);
+ if (err)
+ break;
+
+ err = btrfs_reserve_metadata_space(root, 1);
+ if (err)
+ break;
+
err = btrfs_insert_file_extent(trans, root,
inode->i_ino, cur_offset, 0,
0, hole_size, 0, hole_size,
0, 0, 0);
btrfs_drop_extent_cache(inode, hole_start,
last_byte - 1, 0);
+ btrfs_unreserve_metadata_space(root, 1);
}
free_extent_map(em);
cur_offset = last_byte;
if (err)
return err;
- if (S_ISREG(inode->i_mode) &&
- attr->ia_valid & ATTR_SIZE && attr->ia_size > inode->i_size) {
- err = btrfs_cont_expand(inode, attr->ia_size);
- if (err)
- return err;
+ if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
+ if (attr->ia_size > inode->i_size) {
+ err = btrfs_cont_expand(inode, attr->ia_size);
+ if (err)
+ return err;
+ } else if (inode->i_size > 0 &&
+ attr->ia_size == 0) {
+
+ /* we're truncating a file that used to have good
+ * data down to zero. Make sure it gets into
+ * the ordered flush list so that any new writes
+ * get down to disk quickly.
+ */
+ BTRFS_I(inode)->ordered_data_close = 1;
+ }
}
err = inode_setattr(inode, attr);
}
btrfs_wait_ordered_range(inode, 0, (u64)-1);
+ if (inode->i_nlink > 0) {
+ BUG_ON(btrfs_root_refs(&root->root_item) != 0);
+ goto no_delete;
+ }
+
btrfs_i_size_write(inode, 0);
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_join_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, 0);
namelen, 0);
if (IS_ERR(di))
ret = PTR_ERR(di);
- if (!di || IS_ERR(di)) {
+
+ if (!di || IS_ERR(di))
goto out_err;
- }
+
btrfs_dir_item_key_to_cpu(path->nodes[0], di, location);
out:
btrfs_free_path(path);
* is kind of like crossing a mount point.
*/
static int fixup_tree_root_location(struct btrfs_root *root,
- struct btrfs_key *location,
- struct btrfs_root **sub_root,
- struct dentry *dentry)
+ struct inode *dir,
+ struct dentry *dentry,
+ struct btrfs_key *location,
+ struct btrfs_root **sub_root)
{
- struct btrfs_root_item *ri;
+ struct btrfs_path *path;
+ struct btrfs_root *new_root;
+ struct btrfs_root_ref *ref;
+ struct extent_buffer *leaf;
+ int ret;
+ int err = 0;
- if (btrfs_key_type(location) != BTRFS_ROOT_ITEM_KEY)
- return 0;
- if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
- return 0;
+ path = btrfs_alloc_path();
+ if (!path) {
+ err = -ENOMEM;
+ goto out;
+ }
- *sub_root = btrfs_read_fs_root(root->fs_info, location,
- dentry->d_name.name,
- dentry->d_name.len);
- if (IS_ERR(*sub_root))
- return PTR_ERR(*sub_root);
+ err = -ENOENT;
+ ret = btrfs_find_root_ref(root->fs_info->tree_root, path,
+ BTRFS_I(dir)->root->root_key.objectid,
+ location->objectid);
+ if (ret) {
+ if (ret < 0)
+ err = ret;
+ goto out;
+ }
- ri = &(*sub_root)->root_item;
- location->objectid = btrfs_root_dirid(ri);
- btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
+ leaf = path->nodes[0];
+ ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
+ if (btrfs_root_ref_dirid(leaf, ref) != dir->i_ino ||
+ btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len)
+ goto out;
+
+ ret = memcmp_extent_buffer(leaf, dentry->d_name.name,
+ (unsigned long)(ref + 1),
+ dentry->d_name.len);
+ if (ret)
+ goto out;
+
+ btrfs_release_path(root->fs_info->tree_root, path);
+
+ new_root = btrfs_read_fs_root_no_name(root->fs_info, location);
+ if (IS_ERR(new_root)) {
+ err = PTR_ERR(new_root);
+ goto out;
+ }
+
+ if (btrfs_root_refs(&new_root->root_item) == 0) {
+ err = -ENOENT;
+ goto out;
+ }
+
+ *sub_root = new_root;
+ location->objectid = btrfs_root_dirid(&new_root->root_item);
+ location->type = BTRFS_INODE_ITEM_KEY;
location->offset = 0;
+ err = 0;
+out:
+ btrfs_free_path(path);
+ return err;
+}
+
+static void inode_tree_add(struct inode *inode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_inode *entry;
+ struct rb_node **p;
+ struct rb_node *parent;
+again:
+ p = &root->inode_tree.rb_node;
+ parent = NULL;
+
+ if (hlist_unhashed(&inode->i_hash))
+ return;
+
+ spin_lock(&root->inode_lock);
+ while (*p) {
+ parent = *p;
+ entry = rb_entry(parent, struct btrfs_inode, rb_node);
+
+ if (inode->i_ino < entry->vfs_inode.i_ino)
+ p = &parent->rb_left;
+ else if (inode->i_ino > entry->vfs_inode.i_ino)
+ p = &parent->rb_right;
+ else {
+ WARN_ON(!(entry->vfs_inode.i_state &
+ (I_WILL_FREE | I_FREEING | I_CLEAR)));
+ rb_erase(parent, &root->inode_tree);
+ RB_CLEAR_NODE(parent);
+ spin_unlock(&root->inode_lock);
+ goto again;
+ }
+ }
+ rb_link_node(&BTRFS_I(inode)->rb_node, parent, p);
+ rb_insert_color(&BTRFS_I(inode)->rb_node, &root->inode_tree);
+ spin_unlock(&root->inode_lock);
+}
+
+static void inode_tree_del(struct inode *inode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ int empty = 0;
+
+ spin_lock(&root->inode_lock);
+ if (!RB_EMPTY_NODE(&BTRFS_I(inode)->rb_node)) {
+ rb_erase(&BTRFS_I(inode)->rb_node, &root->inode_tree);
+ RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
+ empty = RB_EMPTY_ROOT(&root->inode_tree);
+ }
+ spin_unlock(&root->inode_lock);
+
+ if (empty && btrfs_root_refs(&root->root_item) == 0) {
+ synchronize_srcu(&root->fs_info->subvol_srcu);
+ spin_lock(&root->inode_lock);
+ empty = RB_EMPTY_ROOT(&root->inode_tree);
+ spin_unlock(&root->inode_lock);
+ if (empty)
+ btrfs_add_dead_root(root);
+ }
+}
+
+int btrfs_invalidate_inodes(struct btrfs_root *root)
+{
+ struct rb_node *node;
+ struct rb_node *prev;
+ struct btrfs_inode *entry;
+ struct inode *inode;
+ u64 objectid = 0;
+
+ WARN_ON(btrfs_root_refs(&root->root_item) != 0);
+
+ spin_lock(&root->inode_lock);
+again:
+ node = root->inode_tree.rb_node;
+ prev = NULL;
+ while (node) {
+ prev = node;
+ entry = rb_entry(node, struct btrfs_inode, rb_node);
+
+ if (objectid < entry->vfs_inode.i_ino)
+ node = node->rb_left;
+ else if (objectid > entry->vfs_inode.i_ino)
+ node = node->rb_right;
+ else
+ break;
+ }
+ if (!node) {
+ while (prev) {
+ entry = rb_entry(prev, struct btrfs_inode, rb_node);
+ if (objectid <= entry->vfs_inode.i_ino) {
+ node = prev;
+ break;
+ }
+ prev = rb_next(prev);
+ }
+ }
+ while (node) {
+ entry = rb_entry(node, struct btrfs_inode, rb_node);
+ objectid = entry->vfs_inode.i_ino + 1;
+ inode = igrab(&entry->vfs_inode);
+ if (inode) {
+ spin_unlock(&root->inode_lock);
+ if (atomic_read(&inode->i_count) > 1)
+ d_prune_aliases(inode);
+ /*
+ * btrfs_drop_inode will remove it from
+ * the inode cache when its usage count
+ * hits zero.
+ */
+ iput(inode);
+ cond_resched();
+ spin_lock(&root->inode_lock);
+ goto again;
+ }
+
+ if (cond_resched_lock(&root->inode_lock))
+ goto again;
+ node = rb_next(node);
+ }
+ spin_unlock(&root->inode_lock);
return 0;
}
{
struct btrfs_inode *bi = BTRFS_I(inode);
- bi->i_acl = NULL;
- bi->i_default_acl = NULL;
-
bi->generation = 0;
+ bi->sequence = 0;
bi->last_trans = 0;
bi->logged_trans = 0;
bi->delalloc_bytes = 0;
+ bi->reserved_bytes = 0;
bi->disk_i_size = 0;
bi->flags = 0;
bi->index_cnt = (u64)-1;
- bi->log_dirty_trans = 0;
+ bi->last_unlink_trans = 0;
+ bi->ordered_data_close = 0;
extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
extent_io_tree_init(&BTRFS_I(inode)->io_tree,
inode->i_mapping, GFP_NOFS);
extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
inode->i_mapping, GFP_NOFS);
INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
+ INIT_LIST_HEAD(&BTRFS_I(inode)->ordered_operations);
+ RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
mutex_init(&BTRFS_I(inode)->extent_mutex);
mutex_init(&BTRFS_I(inode)->log_mutex);
}
inode->i_ino = args->ino;
init_btrfs_i(inode);
BTRFS_I(inode)->root = args->root;
+ btrfs_set_inode_space_info(args->root, inode);
return 0;
}
static int btrfs_find_actor(struct inode *inode, void *opaque)
{
struct btrfs_iget_args *args = opaque;
- return (args->ino == inode->i_ino &&
- args->root == BTRFS_I(inode)->root);
-}
-
-struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
- struct btrfs_root *root, int wait)
-{
- struct inode *inode;
- struct btrfs_iget_args args;
- args.ino = objectid;
- args.root = root;
-
- if (wait) {
- inode = ilookup5(s, objectid, btrfs_find_actor,
- (void *)&args);
- } else {
- inode = ilookup5_nowait(s, objectid, btrfs_find_actor,
- (void *)&args);
- }
- return inode;
+ return args->ino == inode->i_ino &&
+ args->root == BTRFS_I(inode)->root;
}
-struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
- struct btrfs_root *root)
+static struct inode *btrfs_iget_locked(struct super_block *s,
+ u64 objectid,
+ struct btrfs_root *root)
{
struct inode *inode;
struct btrfs_iget_args args;
* Returns in *is_new if the inode was read from disk
*/
struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
- struct btrfs_root *root, int *is_new)
+ struct btrfs_root *root)
{
struct inode *inode;
inode = btrfs_iget_locked(s, location->objectid, root);
if (!inode)
- return ERR_PTR(-EACCES);
+ return ERR_PTR(-ENOMEM);
if (inode->i_state & I_NEW) {
BTRFS_I(inode)->root = root;
memcpy(&BTRFS_I(inode)->location, location, sizeof(*location));
btrfs_read_locked_inode(inode);
+
+ inode_tree_add(inode);
unlock_new_inode(inode);
- if (is_new)
- *is_new = 1;
- } else {
- if (is_new)
- *is_new = 0;
}
return inode;
}
-static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
- struct nameidata *nd)
+static struct inode *new_simple_dir(struct super_block *s,
+ struct btrfs_key *key,
+ struct btrfs_root *root)
+{
+ struct inode *inode = new_inode(s);
+
+ if (!inode)
+ return ERR_PTR(-ENOMEM);
+
+ init_btrfs_i(inode);
+
+ BTRFS_I(inode)->root = root;
+ memcpy(&BTRFS_I(inode)->location, key, sizeof(*key));
+ BTRFS_I(inode)->dummy_inode = 1;
+
+ inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID;
+ inode->i_op = &simple_dir_inode_operations;
+ inode->i_fop = &simple_dir_operations;
+ inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO;
+ inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
+
+ return inode;
+}
+
+struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
{
- struct inode * inode;
- struct btrfs_inode *bi = BTRFS_I(dir);
- struct btrfs_root *root = bi->root;
+ struct inode *inode;
+ struct btrfs_root *root = BTRFS_I(dir)->root;
struct btrfs_root *sub_root = root;
struct btrfs_key location;
- int ret, new, do_orphan = 0;
+ int index;
+ int ret;
+
+ dentry->d_op = &btrfs_dentry_operations;
if (dentry->d_name.len > BTRFS_NAME_LEN)
return ERR_PTR(-ENAMETOOLONG);
if (ret < 0)
return ERR_PTR(ret);
- inode = NULL;
- if (location.objectid) {
- ret = fixup_tree_root_location(root, &location, &sub_root,
- dentry);
- if (ret < 0)
- return ERR_PTR(ret);
- if (ret > 0)
- return ERR_PTR(-ENOENT);
- inode = btrfs_iget(dir->i_sb, &location, sub_root, &new);
- if (IS_ERR(inode))
- return ERR_CAST(inode);
+ if (location.objectid == 0)
+ return NULL;
- /* the inode and parent dir are two different roots */
- if (new && root != sub_root) {
- igrab(inode);
- sub_root->inode = inode;
- do_orphan = 1;
- }
+ if (location.type == BTRFS_INODE_ITEM_KEY) {
+ inode = btrfs_iget(dir->i_sb, &location, root);
+ return inode;
+ }
+
+ BUG_ON(location.type != BTRFS_ROOT_ITEM_KEY);
+
+ index = srcu_read_lock(&root->fs_info->subvol_srcu);
+ ret = fixup_tree_root_location(root, dir, dentry,
+ &location, &sub_root);
+ if (ret < 0) {
+ if (ret != -ENOENT)
+ inode = ERR_PTR(ret);
+ else
+ inode = new_simple_dir(dir->i_sb, &location, sub_root);
+ } else {
+ inode = btrfs_iget(dir->i_sb, &location, sub_root);
+ }
+ srcu_read_unlock(&root->fs_info->subvol_srcu, index);
+
+ return inode;
+}
+
+static int btrfs_dentry_delete(struct dentry *dentry)
+{
+ struct btrfs_root *root;
+
+ if (!dentry->d_inode && !IS_ROOT(dentry))
+ dentry = dentry->d_parent;
+
+ if (dentry->d_inode) {
+ root = BTRFS_I(dentry->d_inode)->root;
+ if (btrfs_root_refs(&root->root_item) == 0)
+ return 1;
}
+ return 0;
+}
+
+static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
+ struct nameidata *nd)
+{
+ struct inode *inode;
- if (unlikely(do_orphan))
- btrfs_orphan_cleanup(sub_root);
+ inode = btrfs_lookup_dentry(dir, dentry);
+ if (IS_ERR(inode))
+ return ERR_CAST(inode);
return d_splice_alias(inode, dentry);
}
return 0;
filp->f_pos = 2;
}
-
path = btrfs_alloc_path();
path->reada = 2;
path->slots[0]++;
}
}
+
advance = 1;
item = btrfs_item_nr(leaf, slot);
btrfs_item_key_to_cpu(leaf, &found_key, slot);
d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)];
btrfs_dir_item_key_to_cpu(leaf, di, &location);
+
+ /* is this a reference to our own snapshot? If so
+ * skip it
+ */
+ if (location.type == BTRFS_ROOT_ITEM_KEY &&
+ location.objectid == root->root_key.objectid) {
+ over = 0;
+ goto skip;
+ }
over = filldir(dirent, name_ptr, name_len,
found_key.offset, location.objectid,
d_type);
+skip:
if (name_ptr != tmp_name)
kfree(name_ptr);
if (over)
goto nopos;
-
di_len = btrfs_dir_name_len(leaf, di) +
btrfs_dir_data_len(leaf, di) + sizeof(*di);
di_cur += di_len;
/* Reached end of directory/root. Bump pos past the last item. */
if (key_type == BTRFS_DIR_INDEX_KEY)
- filp->f_pos = INT_LIMIT(typeof(filp->f_pos));
+ filp->f_pos = INT_LIMIT(off_t);
else
filp->f_pos++;
nopos:
struct btrfs_trans_handle *trans;
int ret = 0;
- if (root->fs_info->closing > 1)
+ if (root->fs_info->btree_inode == inode)
return 0;
if (wait) {
* helper to find a free sequence number in a given directory. This current
* code is very simple, later versions will do smarter things in the btree
*/
-static int btrfs_set_inode_index(struct inode *dir, struct inode *inode,
- u64 *index)
+int btrfs_set_inode_index(struct inode *dir, u64 *index)
{
int ret = 0;
if (BTRFS_I(dir)->index_cnt == (u64)-1) {
ret = btrfs_set_inode_index_count(dir);
- if (ret) {
+ if (ret)
return ret;
- }
}
*index = BTRFS_I(dir)->index_cnt;
struct btrfs_root *root,
struct inode *dir,
const char *name, int name_len,
- u64 ref_objectid,
- u64 objectid,
- struct btrfs_block_group_cache *group,
- int mode, u64 *index)
+ u64 ref_objectid, u64 objectid,
+ u64 alloc_hint, int mode, u64 *index)
{
struct inode *inode;
struct btrfs_inode_item *inode_item;
- struct btrfs_block_group_cache *new_inode_group;
struct btrfs_key *location;
struct btrfs_path *path;
struct btrfs_inode_ref *ref;
return ERR_PTR(-ENOMEM);
if (dir) {
- ret = btrfs_set_inode_index(dir, inode, index);
- if (ret)
+ ret = btrfs_set_inode_index(dir, index);
+ if (ret) {
+ iput(inode);
return ERR_PTR(ret);
+ }
}
/*
* index_cnt is ignored for everything but a dir,
BTRFS_I(inode)->index_cnt = 2;
BTRFS_I(inode)->root = root;
BTRFS_I(inode)->generation = trans->transid;
+ btrfs_set_inode_space_info(root, inode);
if (mode & S_IFDIR)
owner = 0;
else
owner = 1;
- new_inode_group = btrfs_find_block_group(root, group, 0,
- BTRFS_BLOCK_GROUP_METADATA, owner);
- if (!new_inode_group) {
- printk("find_block group failed\n");
- new_inode_group = group;
- }
- BTRFS_I(inode)->block_group = new_inode_group;
+ BTRFS_I(inode)->block_group =
+ btrfs_find_block_group(root, 0, alloc_hint, owner);
key[0].objectid = objectid;
btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY);
sizes[0] = sizeof(struct btrfs_inode_item);
sizes[1] = name_len + sizeof(*ref);
+ path->leave_spinning = 1;
ret = btrfs_insert_empty_items(trans, root, path, key, sizes, 2);
if (ret != 0)
goto fail;
- if (objectid > root->highest_inode)
- root->highest_inode = objectid;
+ inode->i_uid = current_fsuid();
+
+ if (dir && (dir->i_mode & S_ISGID)) {
+ inode->i_gid = dir->i_gid;
+ if (S_ISDIR(mode))
+ mode |= S_ISGID;
+ } else
+ inode->i_gid = current_fsgid();
- inode->i_uid = current->fsuid;
- inode->i_gid = current->fsgid;
inode->i_mode = mode;
inode->i_ino = objectid;
inode_set_bytes(inode, 0);
location->offset = 0;
btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
+ btrfs_inherit_iflags(inode, dir);
+
+ if ((mode & S_IFREG)) {
+ if (btrfs_test_opt(root, NODATASUM))
+ BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM;
+ if (btrfs_test_opt(root, NODATACOW))
+ BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW;
+ }
+
insert_inode_hash(inode);
+ inode_tree_add(inode);
return inode;
fail:
if (dir)
BTRFS_I(dir)->index_cnt--;
btrfs_free_path(path);
+ iput(inode);
return ERR_PTR(ret);
}
struct inode *parent_inode, struct inode *inode,
const char *name, int name_len, int add_backref, u64 index)
{
- int ret;
+ int ret = 0;
struct btrfs_key key;
struct btrfs_root *root = BTRFS_I(parent_inode)->root;
- key.objectid = inode->i_ino;
- btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
- key.offset = 0;
+ if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+ memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key));
+ } else {
+ key.objectid = inode->i_ino;
+ btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+ key.offset = 0;
+ }
+
+ if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+ ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
+ key.objectid, root->root_key.objectid,
+ parent_inode->i_ino,
+ index, name, name_len);
+ } else if (add_backref) {
+ ret = btrfs_insert_inode_ref(trans, root,
+ name, name_len, inode->i_ino,
+ parent_inode->i_ino, index);
+ }
- ret = btrfs_insert_dir_item(trans, root, name, name_len,
- parent_inode->i_ino,
- &key, btrfs_inode_type(inode),
- index);
if (ret == 0) {
- if (add_backref) {
- ret = btrfs_insert_inode_ref(trans, root,
- name, name_len,
- inode->i_ino,
- parent_inode->i_ino,
- index);
- }
+ ret = btrfs_insert_dir_item(trans, root, name, name_len,
+ parent_inode->i_ino, &key,
+ btrfs_inode_type(inode), index);
+ BUG_ON(ret);
+
btrfs_i_size_write(parent_inode, parent_inode->i_size +
name_len * 2);
parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
if (!new_valid_dev(rdev))
return -EINVAL;
- err = btrfs_check_free_space(root, 1, 0);
+ /*
+ * 2 for inode item and ref
+ * 2 for dir items
+ * 1 for xattr if selinux is on
+ */
+ err = btrfs_reserve_metadata_space(root, 5);
if (err)
- goto fail;
+ return err;
trans = btrfs_start_transaction(root, 1);
+ if (!trans)
+ goto fail;
btrfs_set_trans_block_group(trans, dir);
err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
if (IS_ERR(inode))
goto out_unlock;
- err = btrfs_init_acl(inode, dir);
+ err = btrfs_init_inode_security(inode, dir);
if (err) {
drop_inode = 1;
goto out_unlock;
init_special_inode(inode, inode->i_mode, rdev);
btrfs_update_inode(trans, root, inode);
}
- dir->i_sb->s_dirt = 1;
btrfs_update_inode_block_group(trans, inode);
btrfs_update_inode_block_group(trans, dir);
out_unlock:
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
fail:
+ btrfs_unreserve_metadata_space(root, 5);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
u64 objectid;
u64 index = 0;
- err = btrfs_check_free_space(root, 1, 0);
+ /*
+ * 2 for inode item and ref
+ * 2 for dir items
+ * 1 for xattr if selinux is on
+ */
+ err = btrfs_reserve_metadata_space(root, 5);
if (err)
- goto fail;
+ return err;
+
trans = btrfs_start_transaction(root, 1);
+ if (!trans)
+ goto fail;
btrfs_set_trans_block_group(trans, dir);
err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
if (IS_ERR(inode))
goto out_unlock;
- err = btrfs_init_acl(inode, dir);
+ err = btrfs_init_inode_security(inode, dir);
if (err) {
drop_inode = 1;
goto out_unlock;
inode->i_op = &btrfs_file_inode_operations;
BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
}
- dir->i_sb->s_dirt = 1;
btrfs_update_inode_block_group(trans, inode);
btrfs_update_inode_block_group(trans, dir);
out_unlock:
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
fail:
+ btrfs_unreserve_metadata_space(root, 5);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
if (inode->i_nlink == 0)
return -ENOENT;
- btrfs_inc_nlink(inode);
- err = btrfs_check_free_space(root, 1, 0);
+ /*
+ * 1 item for inode ref
+ * 2 items for dir items
+ */
+ err = btrfs_reserve_metadata_space(root, 3);
if (err)
- goto fail;
- err = btrfs_set_inode_index(dir, inode, &index);
+ return err;
+
+ btrfs_inc_nlink(inode);
+
+ err = btrfs_set_inode_index(dir, &index);
if (err)
goto fail;
err = btrfs_add_nondir(trans, dentry, inode, 1, index);
- if (err)
- drop_inode = 1;
-
- dir->i_sb->s_dirt = 1;
- btrfs_update_inode_block_group(trans, dir);
- err = btrfs_update_inode(trans, root, inode);
-
- if (err)
+ if (err) {
drop_inode = 1;
+ } else {
+ btrfs_update_inode_block_group(trans, dir);
+ err = btrfs_update_inode(trans, root, inode);
+ BUG_ON(err);
+ btrfs_log_new_name(trans, inode, NULL, dentry->d_parent);
+ }
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
fail:
+ btrfs_unreserve_metadata_space(root, 3);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
u64 index = 0;
unsigned long nr = 1;
- err = btrfs_check_free_space(root, 1, 0);
+ /*
+ * 2 items for inode and ref
+ * 2 items for dir items
+ * 1 for xattr if selinux is on
+ */
+ err = btrfs_reserve_metadata_space(root, 5);
if (err)
- goto out_unlock;
+ return err;
trans = btrfs_start_transaction(root, 1);
- btrfs_set_trans_block_group(trans, dir);
-
- if (IS_ERR(trans)) {
- err = PTR_ERR(trans);
+ if (!trans) {
+ err = -ENOMEM;
goto out_unlock;
}
+ btrfs_set_trans_block_group(trans, dir);
err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
if (err) {
drop_on_err = 1;
- err = btrfs_init_acl(inode, dir);
+ err = btrfs_init_inode_security(inode, dir);
if (err)
goto out_fail;
d_instantiate(dentry, inode);
drop_on_err = 0;
- dir->i_sb->s_dirt = 1;
btrfs_update_inode_block_group(trans, inode);
btrfs_update_inode_block_group(trans, dir);
btrfs_end_transaction_throttle(trans, root);
out_unlock:
+ btrfs_unreserve_metadata_space(root, 5);
if (drop_on_err)
iput(inode);
btrfs_btree_balance_dirty(root, nr);
read_extent_buffer(leaf, tmp, ptr, inline_size);
- max_size = min(PAGE_CACHE_SIZE, max_size);
+ max_size = min_t(unsigned long, PAGE_CACHE_SIZE, max_size);
ret = btrfs_zlib_decompress(tmp, page, extent_offset,
inline_size, max_size);
if (ret) {
/*
* a bit scary, this does extent mapping from logical file offset to the disk.
- * the ugly parts come from merging extents from the disk with the
- * in-ram representation. This gets more complex because of the data=ordered code,
+ * the ugly parts come from merging extents from the disk with the in-ram
+ * representation. This gets more complex because of the data=ordered code,
* where the in-ram extents might be locked pending data=ordered completion.
*
* This also copies inline extents directly into the page.
*/
+
struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
size_t pg_offset, u64 start, u64 len,
int create)
int compressed;
again:
- spin_lock(&em_tree->lock);
+ read_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree, start, len);
if (em)
em->bdev = root->fs_info->fs_devices->latest_bdev;
- spin_unlock(&em_tree->lock);
+ read_unlock(&em_tree->lock);
if (em) {
if (em->start > start || em->start + em->len <= start)
}
em->bdev = root->fs_info->fs_devices->latest_bdev;
em->start = EXTENT_MAP_HOLE;
+ em->orig_start = EXTENT_MAP_HOLE;
em->len = (u64)-1;
em->block_len = (u64)-1;
found_type = btrfs_file_extent_type(leaf, item);
extent_start = found_key.offset;
compressed = btrfs_file_extent_compression(leaf, item);
- if (found_type == BTRFS_FILE_EXTENT_REG) {
+ if (found_type == BTRFS_FILE_EXTENT_REG ||
+ found_type == BTRFS_FILE_EXTENT_PREALLOC) {
extent_end = extent_start +
btrfs_file_extent_num_bytes(leaf, item);
} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
goto not_found_em;
}
- if (found_type == BTRFS_FILE_EXTENT_REG) {
+ if (found_type == BTRFS_FILE_EXTENT_REG ||
+ found_type == BTRFS_FILE_EXTENT_PREALLOC) {
em->start = extent_start;
em->len = extent_end - extent_start;
+ em->orig_start = extent_start -
+ btrfs_file_extent_offset(leaf, item);
bytenr = btrfs_file_extent_disk_bytenr(leaf, item);
if (bytenr == 0) {
em->block_start = EXTENT_MAP_HOLE;
bytenr += btrfs_file_extent_offset(leaf, item);
em->block_start = bytenr;
em->block_len = em->len;
+ if (found_type == BTRFS_FILE_EXTENT_PREALLOC)
+ set_bit(EXTENT_FLAG_PREALLOC, &em->flags);
}
goto insert;
} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
em->start = extent_start + extent_offset;
em->len = (copy_size + root->sectorsize - 1) &
~((u64)root->sectorsize - 1);
+ em->orig_start = EXTENT_MAP_INLINE;
if (compressed)
set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
ptr = btrfs_file_extent_inline_start(item) + extent_offset;
map = kmap(page);
read_extent_buffer(leaf, map + pg_offset, ptr,
copy_size);
+ if (pg_offset + copy_size < PAGE_CACHE_SIZE) {
+ memset(map + pg_offset + copy_size, 0,
+ PAGE_CACHE_SIZE - pg_offset -
+ copy_size);
+ }
kunmap(page);
}
flush_dcache_page(page);
extent_map_end(em) - 1, GFP_NOFS);
goto insert;
} else {
- printk("unkknown found_type %d\n", found_type);
+ printk(KERN_ERR "btrfs unknown found_type %d\n", found_type);
WARN_ON(1);
}
not_found:
insert:
btrfs_release_path(root, path);
if (em->start > start || extent_map_end(em) <= start) {
- printk("bad extent! em: [%Lu %Lu] passed [%Lu %Lu]\n", em->start, em->len, start, len);
+ printk(KERN_ERR "Btrfs: bad extent! em: [%llu %llu] passed "
+ "[%llu %llu]\n", (unsigned long long)em->start,
+ (unsigned long long)em->len,
+ (unsigned long long)start,
+ (unsigned long long)len);
err = -EIO;
goto out;
}
err = 0;
- spin_lock(&em_tree->lock);
+ write_lock(&em_tree->lock);
ret = add_extent_mapping(em_tree, em);
/* it is possible that someone inserted the extent into the tree
* while we had the lock dropped. It is also possible that
}
} else {
err = -EIO;
- printk("failing to insert %Lu %Lu\n",
- start, len);
free_extent_map(em);
em = NULL;
}
err = 0;
}
}
- spin_unlock(&em_tree->lock);
+ write_unlock(&em_tree->lock);
out:
if (path)
btrfs_free_path(path);
if (trans) {
ret = btrfs_end_transaction(trans, root);
- if (!err) {
+ if (!err)
err = ret;
- }
}
if (err) {
free_extent_map(em);
- WARN_ON(1);
return ERR_PTR(err);
}
return em;
return -EINVAL;
}
-static sector_t btrfs_bmap(struct address_space *mapping, sector_t iblock)
+static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+ __u64 start, __u64 len)
{
- return extent_bmap(mapping, iblock, btrfs_get_extent);
+ return extent_fiemap(inode, fieinfo, start, len, btrfs_get_extent);
}
int btrfs_readpage(struct file *file, struct page *page)
struct writeback_control *wbc)
{
struct extent_io_tree *tree;
+
tree = &BTRFS_I(mapping->host)->io_tree;
return extent_writepages(tree, mapping, btrfs_get_extent, wbc);
}
{
if (PageWriteback(page) || PageDirty(page))
return 0;
- return __btrfs_releasepage(page, gfp_flags);
+ return __btrfs_releasepage(page, gfp_flags & GFP_NOFS);
}
static void btrfs_invalidatepage(struct page *page, unsigned long offset)
u64 page_start = page_offset(page);
u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
+
+ /*
+ * we have the page locked, so new writeback can't start,
+ * and the dirty bit won't be cleared while we are here.
+ *
+ * Wait for IO on this page so that we can safely clear
+ * the PagePrivate2 bit and do ordered accounting
+ */
wait_on_page_writeback(page);
+
tree = &BTRFS_I(page->mapping->host)->io_tree;
if (offset) {
btrfs_releasepage(page, GFP_NOFS);
return;
}
-
lock_extent(tree, page_start, page_end, GFP_NOFS);
ordered = btrfs_lookup_ordered_extent(page->mapping->host,
page_offset(page));
*/
clear_extent_bit(tree, page_start, page_end,
EXTENT_DIRTY | EXTENT_DELALLOC |
- EXTENT_LOCKED, 1, 0, GFP_NOFS);
- btrfs_finish_ordered_io(page->mapping->host,
- page_start, page_end);
+ EXTENT_LOCKED | EXTENT_DO_ACCOUNTING, 1, 0,
+ NULL, GFP_NOFS);
+ /*
+ * whoever cleared the private bit is responsible
+ * for the finish_ordered_io
+ */
+ if (TestClearPagePrivate2(page)) {
+ btrfs_finish_ordered_io(page->mapping->host,
+ page_start, page_end);
+ }
btrfs_put_ordered_extent(ordered);
lock_extent(tree, page_start, page_end, GFP_NOFS);
}
clear_extent_bit(tree, page_start, page_end,
EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
- EXTENT_ORDERED,
- 1, 1, GFP_NOFS);
+ EXTENT_DO_ACCOUNTING, 1, 1, NULL, GFP_NOFS);
__btrfs_releasepage(page, GFP_NOFS);
ClearPageChecked(page);
* beyond EOF, then the page is guaranteed safe against truncation until we
* unlock the page.
*/
-int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page)
+int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
+ struct page *page = vmf->page;
struct inode *inode = fdentry(vma->vm_file)->d_inode;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
u64 page_start;
u64 page_end;
- ret = btrfs_check_free_space(root, PAGE_CACHE_SIZE, 0);
- if (ret)
+ ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
+ if (ret) {
+ if (ret == -ENOMEM)
+ ret = VM_FAULT_OOM;
+ else /* -ENOSPC, -EIO, etc */
+ ret = VM_FAULT_SIGBUS;
+ goto out;
+ }
+
+ ret = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
+ if (ret) {
+ btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
+ ret = VM_FAULT_SIGBUS;
goto out;
+ }
- ret = -EINVAL;
+ ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */
again:
lock_page(page);
size = i_size_read(inode);
if ((page->mapping != inode->i_mapping) ||
(page_start >= size)) {
+ btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
/* page got truncated out from underneath us */
goto out_unlock;
}
goto again;
}
- btrfs_set_extent_delalloc(inode, page_start, page_end);
+ /*
+ * XXX - page_mkwrite gets called every time the page is dirtied, even
+ * if it was already dirty, so for space accounting reasons we need to
+ * clear any delalloc bits for the range we are fixing to save. There
+ * is probably a better way to do this, but for now keep consistent with
+ * prepare_pages in the normal write path.
+ */
+ clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
+ EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING,
+ GFP_NOFS);
+
+ ret = btrfs_set_extent_delalloc(inode, page_start, page_end);
+ if (ret) {
+ unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ ret = VM_FAULT_SIGBUS;
+ btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
+ goto out_unlock;
+ }
ret = 0;
/* page is wholly or partially inside EOF */
}
ClearPageChecked(page);
set_page_dirty(page);
+ SetPageUptodate(page);
+
+ BTRFS_I(inode)->last_trans = root->fs_info->generation + 1;
unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
out_unlock:
+ btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
+ if (!ret)
+ return VM_FAULT_LOCKED;
unlock_page(page);
out:
return ret;
btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
trans = btrfs_start_transaction(root, 1);
+
+ /*
+ * setattr is responsible for setting the ordered_data_close flag,
+ * but that is only tested during the last file release. That
+ * could happen well after the next commit, leaving a great big
+ * window where new writes may get lost if someone chooses to write
+ * to this file after truncating to zero
+ *
+ * The inode doesn't have any dirty data here, and so if we commit
+ * this is a noop. If someone immediately starts writing to the inode
+ * it is very likely we'll catch some of their writes in this
+ * transaction, and the commit will find this file on the ordered
+ * data list with good things to send down.
+ *
+ * This is a best effort solution, there is still a window where
+ * using truncate to replace the contents of the file will
+ * end up with a zero length file after a crash.
+ */
+ if (inode->i_size == 0 && BTRFS_I(inode)->ordered_data_close)
+ btrfs_add_ordered_operation(trans, root, inode);
+
btrfs_set_trans_block_group(trans, inode);
btrfs_i_size_write(inode, inode->i_size);
}
/*
- * Invalidate a single dcache entry at the root of the filesystem.
- * Needed after creation of snapshot or subvolume.
- */
-void btrfs_invalidate_dcache_root(struct btrfs_root *root, char *name,
- int namelen)
-{
- struct dentry *alias, *entry;
- struct qstr qstr;
-
- alias = d_find_alias(root->fs_info->sb->s_root->d_inode);
- if (alias) {
- qstr.name = name;
- qstr.len = namelen;
- /* change me if btrfs ever gets a d_hash operation */
- qstr.hash = full_name_hash(qstr.name, qstr.len);
- entry = d_lookup(alias, &qstr);
- dput(alias);
- if (entry) {
- d_invalidate(entry);
- dput(entry);
- }
- }
-}
-
-/*
* create a new subvolume directory/inode (helper for the ioctl).
*/
-int btrfs_create_subvol_root(struct btrfs_root *new_root, struct dentry *dentry,
- struct btrfs_trans_handle *trans, u64 new_dirid,
- struct btrfs_block_group_cache *block_group)
+int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
+ struct btrfs_root *new_root,
+ u64 new_dirid, u64 alloc_hint)
{
struct inode *inode;
- int error;
+ int err;
u64 index = 0;
inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, new_dirid,
- new_dirid, block_group, S_IFDIR | 0700, &index);
+ new_dirid, alloc_hint, S_IFDIR | 0700, &index);
if (IS_ERR(inode))
return PTR_ERR(inode);
inode->i_op = &btrfs_dir_inode_operations;
inode->i_fop = &btrfs_dir_file_operations;
- new_root->inode = inode;
inode->i_nlink = 1;
btrfs_i_size_write(inode, 0);
- error = btrfs_update_inode(trans, new_root, inode);
- if (error)
- return error;
+ err = btrfs_update_inode(trans, new_root, inode);
+ BUG_ON(err);
- d_instantiate(dentry, inode);
+ iput(inode);
return 0;
}
return NULL;
ei->last_trans = 0;
ei->logged_trans = 0;
+ ei->outstanding_extents = 0;
+ ei->reserved_extents = 0;
+ spin_lock_init(&ei->accounting_lock);
btrfs_ordered_inode_tree_init(&ei->ordered_tree);
- ei->i_acl = BTRFS_ACL_NOT_CACHED;
- ei->i_default_acl = BTRFS_ACL_NOT_CACHED;
INIT_LIST_HEAD(&ei->i_orphan);
+ INIT_LIST_HEAD(&ei->ordered_operations);
return &ei->vfs_inode;
}
void btrfs_destroy_inode(struct inode *inode)
{
struct btrfs_ordered_extent *ordered;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+
WARN_ON(!list_empty(&inode->i_dentry));
WARN_ON(inode->i_data.nrpages);
- if (BTRFS_I(inode)->i_acl &&
- BTRFS_I(inode)->i_acl != BTRFS_ACL_NOT_CACHED)
- posix_acl_release(BTRFS_I(inode)->i_acl);
- if (BTRFS_I(inode)->i_default_acl &&
- BTRFS_I(inode)->i_default_acl != BTRFS_ACL_NOT_CACHED)
- posix_acl_release(BTRFS_I(inode)->i_default_acl);
+ /*
+ * Make sure we're properly removed from the ordered operation
+ * lists.
+ */
+ smp_mb();
+ if (!list_empty(&BTRFS_I(inode)->ordered_operations)) {
+ spin_lock(&root->fs_info->ordered_extent_lock);
+ list_del_init(&BTRFS_I(inode)->ordered_operations);
+ spin_unlock(&root->fs_info->ordered_extent_lock);
+ }
- spin_lock(&BTRFS_I(inode)->root->list_lock);
+ spin_lock(&root->list_lock);
if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
printk(KERN_ERR "BTRFS: inode %lu: inode still on the orphan"
" list\n", inode->i_ino);
dump_stack();
}
- spin_unlock(&BTRFS_I(inode)->root->list_lock);
+ spin_unlock(&root->list_lock);
- while(1) {
+ while (1) {
ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1);
if (!ordered)
break;
else {
- printk("found ordered extent %Lu %Lu\n",
- ordered->file_offset, ordered->len);
+ printk(KERN_ERR "btrfs found ordered "
+ "extent %llu %llu on inode cleanup\n",
+ (unsigned long long)ordered->file_offset,
+ (unsigned long long)ordered->len);
btrfs_remove_ordered_extent(inode, ordered);
btrfs_put_ordered_extent(ordered);
btrfs_put_ordered_extent(ordered);
}
}
+ inode_tree_del(inode);
btrfs_drop_extent_cache(inode, 0, (u64)-1, 0);
kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
}
+void btrfs_drop_inode(struct inode *inode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+
+ if (inode->i_nlink > 0 && btrfs_root_refs(&root->root_item) == 0)
+ generic_delete_inode(inode);
+ else
+ generic_drop_inode(inode);
+}
+
static void init_once(void *foo)
{
struct btrfs_inode *ei = (struct btrfs_inode *) foo;
kmem_cache_destroy(btrfs_trans_handle_cachep);
if (btrfs_transaction_cachep)
kmem_cache_destroy(btrfs_transaction_cachep);
- if (btrfs_bit_radix_cachep)
- kmem_cache_destroy(btrfs_bit_radix_cachep);
if (btrfs_path_cachep)
kmem_cache_destroy(btrfs_path_cachep);
}
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
- unsigned long extra_flags,
- void (*ctor)(void *))
-{
- return kmem_cache_create(name, size, 0, (SLAB_RECLAIM_ACCOUNT |
- SLAB_MEM_SPREAD | extra_flags), ctor);
-}
-
int btrfs_init_cachep(void)
{
- btrfs_inode_cachep = btrfs_cache_create("btrfs_inode_cache",
- sizeof(struct btrfs_inode),
- 0, init_once);
+ btrfs_inode_cachep = kmem_cache_create("btrfs_inode_cache",
+ sizeof(struct btrfs_inode), 0,
+ SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, init_once);
if (!btrfs_inode_cachep)
goto fail;
- btrfs_trans_handle_cachep =
- btrfs_cache_create("btrfs_trans_handle_cache",
- sizeof(struct btrfs_trans_handle),
- 0, NULL);
+
+ btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle_cache",
+ sizeof(struct btrfs_trans_handle), 0,
+ SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
if (!btrfs_trans_handle_cachep)
goto fail;
- btrfs_transaction_cachep = btrfs_cache_create("btrfs_transaction_cache",
- sizeof(struct btrfs_transaction),
- 0, NULL);
+
+ btrfs_transaction_cachep = kmem_cache_create("btrfs_transaction_cache",
+ sizeof(struct btrfs_transaction), 0,
+ SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
if (!btrfs_transaction_cachep)
goto fail;
- btrfs_path_cachep = btrfs_cache_create("btrfs_path_cache",
- sizeof(struct btrfs_path),
- 0, NULL);
+
+ btrfs_path_cachep = kmem_cache_create("btrfs_path_cache",
+ sizeof(struct btrfs_path), 0,
+ SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
if (!btrfs_path_cachep)
goto fail;
- btrfs_bit_radix_cachep = btrfs_cache_create("btrfs_radix", 256,
- SLAB_DESTROY_BY_RCU, NULL);
- if (!btrfs_bit_radix_cachep)
- goto fail;
+
return 0;
fail:
btrfs_destroy_cachep();
{
struct inode *inode = dentry->d_inode;
generic_fillattr(inode, stat);
+ stat->dev = BTRFS_I(inode)->root->anon_super.s_dev;
stat->blksize = PAGE_CACHE_SIZE;
stat->blocks = (inode_get_bytes(inode) +
BTRFS_I(inode)->delalloc_bytes) >> 9;
return 0;
}
-static int btrfs_rename(struct inode * old_dir, struct dentry *old_dentry,
- struct inode * new_dir,struct dentry *new_dentry)
+static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
+ struct inode *new_dir, struct dentry *new_dentry)
{
struct btrfs_trans_handle *trans;
struct btrfs_root *root = BTRFS_I(old_dir)->root;
+ struct btrfs_root *dest = BTRFS_I(new_dir)->root;
struct inode *new_inode = new_dentry->d_inode;
struct inode *old_inode = old_dentry->d_inode;
struct timespec ctime = CURRENT_TIME;
u64 index = 0;
+ u64 root_objectid;
int ret;
+ if (new_dir->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
+ return -EPERM;
+
+ /* we only allow rename subvolume link between subvolumes */
+ if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest)
+ return -EXDEV;
+
+ if (old_inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID ||
+ (new_inode && new_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID))
+ return -ENOTEMPTY;
+
if (S_ISDIR(old_inode->i_mode) && new_inode &&
- new_inode->i_size > BTRFS_EMPTY_DIR_SIZE) {
+ new_inode->i_size > BTRFS_EMPTY_DIR_SIZE)
return -ENOTEMPTY;
- }
- ret = btrfs_check_free_space(root, 1, 0);
+ /*
+ * 2 items for dir items
+ * 1 item for orphan entry
+ * 1 item for ref
+ */
+ ret = btrfs_reserve_metadata_space(root, 4);
if (ret)
- goto out_unlock;
+ return ret;
- trans = btrfs_start_transaction(root, 1);
+ /*
+ * we're using rename to replace one file with another.
+ * and the replacement file is large. Start IO on it now so
+ * we don't add too much work to the end of the transaction
+ */
+ if (new_inode && S_ISREG(old_inode->i_mode) && new_inode->i_size &&
+ old_inode->i_size > BTRFS_ORDERED_OPERATIONS_FLUSH_LIMIT)
+ filemap_flush(old_inode->i_mapping);
+
+ /* close the racy window with snapshot create/destroy ioctl */
+ if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+ down_read(&root->fs_info->subvol_sem);
+ trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, new_dir);
- btrfs_inc_nlink(old_dentry->d_inode);
+ if (dest != root)
+ btrfs_record_root_in_trans(trans, dest);
+
+ ret = btrfs_set_inode_index(new_dir, &index);
+ if (ret)
+ goto out_fail;
+
+ if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+ /* force full log commit if subvolume involved. */
+ root->fs_info->last_trans_log_full_commit = trans->transid;
+ } else {
+ ret = btrfs_insert_inode_ref(trans, dest,
+ new_dentry->d_name.name,
+ new_dentry->d_name.len,
+ old_inode->i_ino,
+ new_dir->i_ino, index);
+ if (ret)
+ goto out_fail;
+ /*
+ * this is an ugly little race, but the rename is required
+ * to make sure that if we crash, the inode is either at the
+ * old name or the new one. pinning the log transaction lets
+ * us make sure we don't allow a log commit to come in after
+ * we unlink the name but before we add the new name back in.
+ */
+ btrfs_pin_log_trans(root);
+ }
+ /*
+ * make sure the inode gets flushed if it is replacing
+ * something.
+ */
+ if (new_inode && new_inode->i_size &&
+ old_inode && S_ISREG(old_inode->i_mode)) {
+ btrfs_add_ordered_operation(trans, root, old_inode);
+ }
+
old_dir->i_ctime = old_dir->i_mtime = ctime;
new_dir->i_ctime = new_dir->i_mtime = ctime;
old_inode->i_ctime = ctime;
- ret = btrfs_unlink_inode(trans, root, old_dir, old_dentry->d_inode,
- old_dentry->d_name.name,
- old_dentry->d_name.len);
- if (ret)
- goto out_fail;
+ if (old_dentry->d_parent != new_dentry->d_parent)
+ btrfs_record_unlink_dir(trans, old_dir, old_inode, 1);
+
+ if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+ root_objectid = BTRFS_I(old_inode)->root->root_key.objectid;
+ ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid,
+ old_dentry->d_name.name,
+ old_dentry->d_name.len);
+ } else {
+ btrfs_inc_nlink(old_dentry->d_inode);
+ ret = btrfs_unlink_inode(trans, root, old_dir,
+ old_dentry->d_inode,
+ old_dentry->d_name.name,
+ old_dentry->d_name.len);
+ }
+ BUG_ON(ret);
if (new_inode) {
new_inode->i_ctime = CURRENT_TIME;
- ret = btrfs_unlink_inode(trans, root, new_dir,
- new_dentry->d_inode,
- new_dentry->d_name.name,
- new_dentry->d_name.len);
- if (ret)
- goto out_fail;
+ if (unlikely(new_inode->i_ino ==
+ BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
+ root_objectid = BTRFS_I(new_inode)->location.objectid;
+ ret = btrfs_unlink_subvol(trans, dest, new_dir,
+ root_objectid,
+ new_dentry->d_name.name,
+ new_dentry->d_name.len);
+ BUG_ON(new_inode->i_nlink == 0);
+ } else {
+ ret = btrfs_unlink_inode(trans, dest, new_dir,
+ new_dentry->d_inode,
+ new_dentry->d_name.name,
+ new_dentry->d_name.len);
+ }
+ BUG_ON(ret);
if (new_inode->i_nlink == 0) {
ret = btrfs_orphan_add(trans, new_dentry->d_inode);
- if (ret)
- goto out_fail;
+ BUG_ON(ret);
}
-
}
- ret = btrfs_set_inode_index(new_dir, old_inode, &index);
- if (ret)
- goto out_fail;
- ret = btrfs_add_link(trans, new_dentry->d_parent->d_inode,
- old_inode, new_dentry->d_name.name,
- new_dentry->d_name.len, 1, index);
- if (ret)
- goto out_fail;
+ ret = btrfs_add_link(trans, new_dir, old_inode,
+ new_dentry->d_name.name,
+ new_dentry->d_name.len, 0, index);
+ BUG_ON(ret);
+ if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) {
+ btrfs_log_new_name(trans, old_inode, old_dir,
+ new_dentry->d_parent);
+ btrfs_end_log_trans(root);
+ }
out_fail:
btrfs_end_transaction_throttle(trans, root);
-out_unlock:
+
+ if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+ up_read(&root->fs_info->subvol_sem);
+
+ btrfs_unreserve_metadata_space(root, 4);
return ret;
}
struct list_head *head = &root->fs_info->delalloc_inodes;
struct btrfs_inode *binode;
struct inode *inode;
- unsigned long flags;
- spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
- while(!list_empty(head)) {
+ if (root->fs_info->sb->s_flags & MS_RDONLY)
+ return -EROFS;
+
+ spin_lock(&root->fs_info->delalloc_lock);
+ while (!list_empty(head)) {
binode = list_entry(head->next, struct btrfs_inode,
delalloc_inodes);
inode = igrab(&binode->vfs_inode);
if (!inode)
list_del_init(&binode->delalloc_inodes);
- spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
+ spin_unlock(&root->fs_info->delalloc_lock);
if (inode) {
filemap_flush(inode->i_mapping);
iput(inode);
}
cond_resched();
- spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
+ spin_lock(&root->fs_info->delalloc_lock);
}
- spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
+ spin_unlock(&root->fs_info->delalloc_lock);
/* the filemap_flush will queue IO into the worker threads, but
* we have to make sure the IO is actually started and that
* ordered extents get created before we return
*/
atomic_inc(&root->fs_info->async_submit_draining);
- while(atomic_read(&root->fs_info->nr_async_submits)) {
+ while (atomic_read(&root->fs_info->nr_async_submits) ||
+ atomic_read(&root->fs_info->async_delalloc_pages)) {
wait_event(root->fs_info->async_submit_wait,
- (atomic_read(&root->fs_info->nr_async_submits) == 0));
+ (atomic_read(&root->fs_info->nr_async_submits) == 0 &&
+ atomic_read(&root->fs_info->async_delalloc_pages) == 0));
}
atomic_dec(&root->fs_info->async_submit_draining);
return 0;
if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
return -ENAMETOOLONG;
- err = btrfs_check_free_space(root, 1, 0);
+ /*
+ * 2 items for inode item and ref
+ * 2 items for dir items
+ * 1 item for xattr if selinux is on
+ */
+ err = btrfs_reserve_metadata_space(root, 5);
if (err)
- goto out_fail;
+ return err;
trans = btrfs_start_transaction(root, 1);
+ if (!trans)
+ goto out_fail;
btrfs_set_trans_block_group(trans, dir);
err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
if (IS_ERR(inode))
goto out_unlock;
- err = btrfs_init_acl(inode, dir);
+ err = btrfs_init_inode_security(inode, dir);
if (err) {
drop_inode = 1;
goto out_unlock;
inode->i_op = &btrfs_file_inode_operations;
BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
}
- dir->i_sb->s_dirt = 1;
btrfs_update_inode_block_group(trans, inode);
btrfs_update_inode_block_group(trans, dir);
if (drop_inode)
inode->i_op = &btrfs_symlink_inode_operations;
inode->i_mapping->a_ops = &btrfs_symlink_aops;
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
+ inode_set_bytes(inode, name_len);
btrfs_i_size_write(inode, name_len - 1);
err = btrfs_update_inode(trans, root, inode);
if (err)
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
out_fail:
+ btrfs_unreserve_metadata_space(root, 5);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
return err;
}
+static int prealloc_file_range(struct btrfs_trans_handle *trans,
+ struct inode *inode, u64 start, u64 end,
+ u64 locked_end, u64 alloc_hint, int mode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_key ins;
+ u64 alloc_size;
+ u64 cur_offset = start;
+ u64 num_bytes = end - start;
+ int ret = 0;
+
+ while (num_bytes > 0) {
+ alloc_size = min(num_bytes, root->fs_info->max_extent);
+
+ ret = btrfs_reserve_metadata_space(root, 1);
+ if (ret)
+ goto out;
+
+ ret = btrfs_reserve_extent(trans, root, alloc_size,
+ root->sectorsize, 0, alloc_hint,
+ (u64)-1, &ins, 1);
+ if (ret) {
+ WARN_ON(1);
+ goto out;
+ }
+ ret = insert_reserved_file_extent(trans, inode,
+ cur_offset, ins.objectid,
+ ins.offset, ins.offset,
+ ins.offset, locked_end,
+ 0, 0, 0,
+ BTRFS_FILE_EXTENT_PREALLOC);
+ BUG_ON(ret);
+ btrfs_drop_extent_cache(inode, cur_offset,
+ cur_offset + ins.offset -1, 0);
+ num_bytes -= ins.offset;
+ cur_offset += ins.offset;
+ alloc_hint = ins.objectid + ins.offset;
+ btrfs_unreserve_metadata_space(root, 1);
+ }
+out:
+ if (cur_offset > start) {
+ inode->i_ctime = CURRENT_TIME;
+ BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC;
+ if (!(mode & FALLOC_FL_KEEP_SIZE) &&
+ cur_offset > i_size_read(inode))
+ btrfs_i_size_write(inode, cur_offset);
+ ret = btrfs_update_inode(trans, root, inode);
+ BUG_ON(ret);
+ }
+
+ return ret;
+}
+
+static long btrfs_fallocate(struct inode *inode, int mode,
+ loff_t offset, loff_t len)
+{
+ u64 cur_offset;
+ u64 last_byte;
+ u64 alloc_start;
+ u64 alloc_end;
+ u64 alloc_hint = 0;
+ u64 locked_end;
+ u64 mask = BTRFS_I(inode)->root->sectorsize - 1;
+ struct extent_map *em;
+ struct btrfs_trans_handle *trans;
+ struct btrfs_root *root;
+ int ret;
+
+ alloc_start = offset & ~mask;
+ alloc_end = (offset + len + mask) & ~mask;
+
+ /*
+ * wait for ordered IO before we have any locks. We'll loop again
+ * below with the locks held.
+ */
+ btrfs_wait_ordered_range(inode, alloc_start, alloc_end - alloc_start);
+
+ mutex_lock(&inode->i_mutex);
+ if (alloc_start > inode->i_size) {
+ ret = btrfs_cont_expand(inode, alloc_start);
+ if (ret)
+ goto out;
+ }
+
+ root = BTRFS_I(inode)->root;
+
+ ret = btrfs_check_data_free_space(root, inode,
+ alloc_end - alloc_start);
+ if (ret)
+ goto out;
+
+ locked_end = alloc_end - 1;
+ while (1) {
+ struct btrfs_ordered_extent *ordered;
+
+ trans = btrfs_start_transaction(BTRFS_I(inode)->root, 1);
+ if (!trans) {
+ ret = -EIO;
+ goto out_free;
+ }
+
+ /* the extent lock is ordered inside the running
+ * transaction
+ */
+ lock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
+ GFP_NOFS);
+ ordered = btrfs_lookup_first_ordered_extent(inode,
+ alloc_end - 1);
+ if (ordered &&
+ ordered->file_offset + ordered->len > alloc_start &&
+ ordered->file_offset < alloc_end) {
+ btrfs_put_ordered_extent(ordered);
+ unlock_extent(&BTRFS_I(inode)->io_tree,
+ alloc_start, locked_end, GFP_NOFS);
+ btrfs_end_transaction(trans, BTRFS_I(inode)->root);
+
+ /*
+ * we can't wait on the range with the transaction
+ * running or with the extent lock held
+ */
+ btrfs_wait_ordered_range(inode, alloc_start,
+ alloc_end - alloc_start);
+ } else {
+ if (ordered)
+ btrfs_put_ordered_extent(ordered);
+ break;
+ }
+ }
+
+ cur_offset = alloc_start;
+ while (1) {
+ em = btrfs_get_extent(inode, NULL, 0, cur_offset,
+ alloc_end - cur_offset, 0);
+ BUG_ON(IS_ERR(em) || !em);
+ last_byte = min(extent_map_end(em), alloc_end);
+ last_byte = (last_byte + mask) & ~mask;
+ if (em->block_start == EXTENT_MAP_HOLE) {
+ ret = prealloc_file_range(trans, inode, cur_offset,
+ last_byte, locked_end + 1,
+ alloc_hint, mode);
+ if (ret < 0) {
+ free_extent_map(em);
+ break;
+ }
+ }
+ if (em->block_start <= EXTENT_MAP_LAST_BYTE)
+ alloc_hint = em->block_start;
+ free_extent_map(em);
+
+ cur_offset = last_byte;
+ if (cur_offset >= alloc_end) {
+ ret = 0;
+ break;
+ }
+ }
+ unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
+ GFP_NOFS);
+
+ btrfs_end_transaction(trans, BTRFS_I(inode)->root);
+out_free:
+ btrfs_free_reserved_data_space(root, inode, alloc_end - alloc_start);
+out:
+ mutex_unlock(&inode->i_mutex);
+ return ret;
+}
+
static int btrfs_set_page_dirty(struct page *page)
{
return __set_page_dirty_nobuffers(page);
static int btrfs_permission(struct inode *inode, int mask)
{
- if (btrfs_test_flag(inode, READONLY) && (mask & MAY_WRITE))
+ if ((BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) && (mask & MAY_WRITE))
return -EACCES;
return generic_permission(inode, mask, btrfs_check_acl);
}
static struct inode_operations btrfs_dir_inode_operations = {
+ .getattr = btrfs_getattr,
.lookup = btrfs_lookup,
.create = btrfs_create,
.unlink = btrfs_unlink,
.lookup = btrfs_lookup,
.permission = btrfs_permission,
};
+
static struct file_operations btrfs_dir_file_operations = {
.llseek = generic_file_llseek,
.read = generic_read_dir,
.readpage_io_failed_hook = btrfs_io_failed_hook,
.set_bit_hook = btrfs_set_bit_hook,
.clear_bit_hook = btrfs_clear_bit_hook,
+ .merge_extent_hook = btrfs_merge_extent_hook,
+ .split_extent_hook = btrfs_split_extent_hook,
};
+/*
+ * btrfs doesn't support the bmap operation because swapfiles
+ * use bmap to make a mapping of extents in the file. They assume
+ * these extents won't change over the life of the file and they
+ * use the bmap result to do IO directly to the drive.
+ *
+ * the btrfs bmap call would return logical addresses that aren't
+ * suitable for IO and they also will change frequently as COW
+ * operations happen. So, swapfile + btrfs == corruption.
+ *
+ * For now we're avoiding this by dropping bmap.
+ */
static struct address_space_operations btrfs_aops = {
.readpage = btrfs_readpage,
.writepage = btrfs_writepage,
.writepages = btrfs_writepages,
.readpages = btrfs_readpages,
.sync_page = block_sync_page,
- .bmap = btrfs_bmap,
.direct_IO = btrfs_direct_IO,
.invalidatepage = btrfs_invalidatepage,
.releasepage = btrfs_releasepage,
.listxattr = btrfs_listxattr,
.removexattr = btrfs_removexattr,
.permission = btrfs_permission,
+ .fallocate = btrfs_fallocate,
+ .fiemap = btrfs_fiemap,
};
static struct inode_operations btrfs_special_inode_operations = {
.getattr = btrfs_getattr,
.follow_link = page_follow_link_light,
.put_link = page_put_link,
.permission = btrfs_permission,
+ .setxattr = btrfs_setxattr,
+ .getxattr = btrfs_getxattr,
+ .listxattr = btrfs_listxattr,
+ .removexattr = btrfs_removexattr,
+};
+
+const struct dentry_operations btrfs_dentry_operations = {
+ .d_delete = btrfs_dentry_delete,
};
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff