aboutsummaryrefslogtreecommitdiff
path: root/fs/btrfs/extent_io.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/btrfs/extent_io.c')
-rw-r--r--fs/btrfs/extent_io.c187
1 files changed, 99 insertions, 88 deletions
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 96577e8bf9fd..a4080c21ec55 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -2,7 +2,6 @@
#include <linux/slab.h>
#include <linux/bio.h>
#include <linux/mm.h>
-#include <linux/gfp.h>
#include <linux/pagemap.h>
#include <linux/page-flags.h>
#include <linux/module.h>
@@ -104,8 +103,8 @@ void extent_io_exit(void)
void extent_io_tree_init(struct extent_io_tree *tree,
struct address_space *mapping, gfp_t mask)
{
- tree->state.rb_node = NULL;
- tree->buffer.rb_node = NULL;
+ tree->state = RB_ROOT;
+ tree->buffer = RB_ROOT;
tree->ops = NULL;
tree->dirty_bytes = 0;
spin_lock_init(&tree->lock);
@@ -136,7 +135,7 @@ static struct extent_state *alloc_extent_state(gfp_t mask)
return state;
}
-static void free_extent_state(struct extent_state *state)
+void free_extent_state(struct extent_state *state)
{
if (!state)
return;
@@ -336,21 +335,18 @@ static int merge_state(struct extent_io_tree *tree,
}
static int set_state_cb(struct extent_io_tree *tree,
- struct extent_state *state,
- unsigned long bits)
+ struct extent_state *state, int *bits)
{
if (tree->ops && tree->ops->set_bit_hook) {
return tree->ops->set_bit_hook(tree->mapping->host,
- state->start, state->end,
- state->state, bits);
+ state, bits);
}
return 0;
}
static void clear_state_cb(struct extent_io_tree *tree,
- struct extent_state *state,
- unsigned long bits)
+ struct extent_state *state, int *bits)
{
if (tree->ops && tree->ops->clear_bit_hook)
tree->ops->clear_bit_hook(tree->mapping->host, state, bits);
@@ -368,9 +364,10 @@ static void clear_state_cb(struct extent_io_tree *tree,
*/
static int insert_state(struct extent_io_tree *tree,
struct extent_state *state, u64 start, u64 end,
- int bits)
+ int *bits)
{
struct rb_node *node;
+ int bits_to_set = *bits & ~EXTENT_CTLBITS;
int ret;
if (end < start) {
@@ -385,9 +382,9 @@ static int insert_state(struct extent_io_tree *tree,
if (ret)
return ret;
- if (bits & EXTENT_DIRTY)
+ if (bits_to_set & EXTENT_DIRTY)
tree->dirty_bytes += end - start + 1;
- state->state |= bits;
+ state->state |= bits_to_set;
node = tree_insert(&tree->state, end, &state->rb_node);
if (node) {
struct extent_state *found;
@@ -457,13 +454,13 @@ static int split_state(struct extent_io_tree *tree, struct extent_state *orig,
* struct is freed and removed from the tree
*/
static int clear_state_bit(struct extent_io_tree *tree,
- struct extent_state *state, int bits, int wake,
- int delete)
+ struct extent_state *state,
+ int *bits, int wake)
{
- int bits_to_clear = bits & ~EXTENT_DO_ACCOUNTING;
+ int bits_to_clear = *bits & ~EXTENT_CTLBITS;
int ret = state->state & bits_to_clear;
- if ((bits & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
+ if ((bits_to_clear & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
u64 range = state->end - state->start + 1;
WARN_ON(range > tree->dirty_bytes);
tree->dirty_bytes -= range;
@@ -472,9 +469,8 @@ static int clear_state_bit(struct extent_io_tree *tree,
state->state &= ~bits_to_clear;
if (wake)
wake_up(&state->wq);
- if (delete || state->state == 0) {
+ if (state->state == 0) {
if (state->tree) {
- clear_state_cb(tree, state, state->state);
rb_erase(&state->rb_node, &tree->state);
state->tree = NULL;
free_extent_state(state);
@@ -513,7 +509,14 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
u64 last_end;
int err;
int set = 0;
+ int clear = 0;
+
+ if (delete)
+ bits |= ~EXTENT_CTLBITS;
+ bits |= EXTENT_FIRST_DELALLOC;
+ if (bits & (EXTENT_IOBITS | EXTENT_BOUNDARY))
+ clear = 1;
again:
if (!prealloc && (mask & __GFP_WAIT)) {
prealloc = alloc_extent_state(mask);
@@ -524,14 +527,20 @@ again:
spin_lock(&tree->lock);
if (cached_state) {
cached = *cached_state;
- *cached_state = NULL;
- cached_state = NULL;
+
+ if (clear) {
+ *cached_state = NULL;
+ cached_state = NULL;
+ }
+
if (cached && cached->tree && cached->start == start) {
- atomic_dec(&cached->refs);
+ if (clear)
+ atomic_dec(&cached->refs);
state = cached;
goto hit_next;
}
- free_extent_state(cached);
+ if (clear)
+ free_extent_state(cached);
}
/*
* this search will find the extents that end after
@@ -572,8 +581,7 @@ hit_next:
if (err)
goto out;
if (state->end <= end) {
- set |= clear_state_bit(tree, state, bits, wake,
- delete);
+ set |= clear_state_bit(tree, state, &bits, wake);
if (last_end == (u64)-1)
goto out;
start = last_end + 1;
@@ -594,7 +602,7 @@ hit_next:
if (wake)
wake_up(&state->wq);
- set |= clear_state_bit(tree, prealloc, bits, wake, delete);
+ set |= clear_state_bit(tree, prealloc, &bits, wake);
prealloc = NULL;
goto out;
@@ -605,7 +613,7 @@ hit_next:
else
next_node = NULL;
- set |= clear_state_bit(tree, state, bits, wake, delete);
+ set |= clear_state_bit(tree, state, &bits, wake);
if (last_end == (u64)-1)
goto out;
start = last_end + 1;
@@ -698,19 +706,19 @@ out:
static int set_state_bits(struct extent_io_tree *tree,
struct extent_state *state,
- int bits)
+ int *bits)
{
int ret;
+ int bits_to_set = *bits & ~EXTENT_CTLBITS;
ret = set_state_cb(tree, state, bits);
if (ret)
return ret;
-
- if ((bits & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
+ if ((bits_to_set & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
u64 range = state->end - state->start + 1;
tree->dirty_bytes += range;
}
- state->state |= bits;
+ state->state |= bits_to_set;
return 0;
}
@@ -737,10 +745,9 @@ static void cache_state(struct extent_state *state,
* [start, end] is inclusive This takes the tree lock.
*/
-static int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
- int bits, int exclusive_bits, u64 *failed_start,
- struct extent_state **cached_state,
- gfp_t mask)
+int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+ int bits, int exclusive_bits, u64 *failed_start,
+ struct extent_state **cached_state, gfp_t mask)
{
struct extent_state *state;
struct extent_state *prealloc = NULL;
@@ -749,6 +756,7 @@ static int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
u64 last_start;
u64 last_end;
+ bits |= EXTENT_FIRST_DELALLOC;
again:
if (!prealloc && (mask & __GFP_WAIT)) {
prealloc = alloc_extent_state(mask);
@@ -770,7 +778,7 @@ again:
*/
node = tree_search(tree, start);
if (!node) {
- err = insert_state(tree, prealloc, start, end, bits);
+ err = insert_state(tree, prealloc, start, end, &bits);
prealloc = NULL;
BUG_ON(err == -EEXIST);
goto out;
@@ -794,7 +802,7 @@ hit_next:
goto out;
}
- err = set_state_bits(tree, state, bits);
+ err = set_state_bits(tree, state, &bits);
if (err)
goto out;
@@ -844,7 +852,7 @@ hit_next:
if (err)
goto out;
if (state->end <= end) {
- err = set_state_bits(tree, state, bits);
+ err = set_state_bits(tree, state, &bits);
if (err)
goto out;
cache_state(state, cached_state);
@@ -869,7 +877,7 @@ hit_next:
else
this_end = last_start - 1;
err = insert_state(tree, prealloc, start, this_end,
- bits);
+ &bits);
BUG_ON(err == -EEXIST);
if (err) {
prealloc = NULL;
@@ -895,7 +903,7 @@ hit_next:
err = split_state(tree, state, prealloc, end + 1);
BUG_ON(err == -EEXIST);
- err = set_state_bits(tree, prealloc, bits);
+ err = set_state_bits(tree, prealloc, &bits);
if (err) {
prealloc = NULL;
goto out;
@@ -946,11 +954,11 @@ int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
}
int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
+ struct extent_state **cached_state, gfp_t mask)
{
return set_extent_bit(tree, start, end,
EXTENT_DELALLOC | EXTENT_DIRTY | EXTENT_UPTODATE,
- 0, NULL, NULL, mask);
+ 0, NULL, cached_state, mask);
}
int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
@@ -958,8 +966,7 @@ int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
{
return clear_extent_bit(tree, start, end,
EXTENT_DIRTY | EXTENT_DELALLOC |
- EXTENT_DO_ACCOUNTING, 0, 0,
- NULL, mask);
+ EXTENT_DO_ACCOUNTING, 0, 0, NULL, mask);
}
int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
@@ -984,10 +991,11 @@ int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
}
static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
- u64 end, gfp_t mask)
+ u64 end, struct extent_state **cached_state,
+ gfp_t mask)
{
return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0,
- NULL, mask);
+ cached_state, mask);
}
int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end)
@@ -1171,7 +1179,8 @@ out:
* 1 is returned if we find something, 0 if nothing was in the tree
*/
static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
- u64 *start, u64 *end, u64 max_bytes)
+ u64 *start, u64 *end, u64 max_bytes,
+ struct extent_state **cached_state)
{
struct rb_node *node;
struct extent_state *state;
@@ -1203,8 +1212,11 @@ static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
*end = state->end;
goto out;
}
- if (!found)
+ if (!found) {
*start = state->start;
+ *cached_state = state;
+ atomic_inc(&state->refs);
+ }
found++;
*end = state->end;
cur_start = state->end + 1;
@@ -1336,10 +1348,11 @@ again:
delalloc_start = *start;
delalloc_end = 0;
found = find_delalloc_range(tree, &delalloc_start, &delalloc_end,
- max_bytes);
+ max_bytes, &cached_state);
if (!found || delalloc_end <= *start) {
*start = delalloc_start;
*end = delalloc_end;
+ free_extent_state(cached_state);
return found;
}
@@ -1421,9 +1434,6 @@ int extent_clear_unlock_delalloc(struct inode *inode,
if (op & EXTENT_CLEAR_DELALLOC)
clear_bits |= EXTENT_DELALLOC;
- if (op & EXTENT_CLEAR_ACCOUNTING)
- clear_bits |= EXTENT_DO_ACCOUNTING;
-
clear_extent_bit(tree, start, end, clear_bits, 1, 0, NULL, GFP_NOFS);
if (!(op & (EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK |
@@ -1722,7 +1732,7 @@ static void end_bio_extent_writepage(struct bio *bio, int err)
}
if (!uptodate) {
- clear_extent_uptodate(tree, start, end, GFP_NOFS);
+ clear_extent_uptodate(tree, start, end, NULL, GFP_NOFS);
ClearPageUptodate(page);
SetPageError(page);
}
@@ -1750,7 +1760,8 @@ static void end_bio_extent_writepage(struct bio *bio, int err)
static void end_bio_extent_readpage(struct bio *bio, int err)
{
int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
- struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+ struct bio_vec *bvec_end = bio->bi_io_vec + bio->bi_vcnt - 1;
+ struct bio_vec *bvec = bio->bi_io_vec;
struct extent_io_tree *tree;
u64 start;
u64 end;
@@ -1773,7 +1784,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
else
whole_page = 0;
- if (--bvec >= bio->bi_io_vec)
+ if (++bvec <= bvec_end)
prefetchw(&bvec->bv_page->flags);
if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
@@ -1818,7 +1829,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
}
check_page_locked(tree, page);
}
- } while (bvec >= bio->bi_io_vec);
+ } while (bvec <= bvec_end);
bio_put(bio);
}
@@ -1901,7 +1912,7 @@ static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
if (tree->ops && tree->ops->submit_bio_hook)
tree->ops->submit_bio_hook(page->mapping->host, rw, bio,
- mirror_num, bio_flags);
+ mirror_num, bio_flags, start);
else
submit_bio(rw, bio);
if (bio_flagged(bio, BIO_EOPNOTSUPP))
@@ -2005,6 +2016,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
sector_t sector;
struct extent_map *em;
struct block_device *bdev;
+ struct btrfs_ordered_extent *ordered;
int ret;
int nr = 0;
size_t page_offset = 0;
@@ -2016,7 +2028,15 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
set_page_extent_mapped(page);
end = page_end;
- lock_extent(tree, start, end, GFP_NOFS);
+ while (1) {
+ lock_extent(tree, start, end, GFP_NOFS);
+ ordered = btrfs_lookup_ordered_extent(inode, start);
+ if (!ordered)
+ break;
+ unlock_extent(tree, start, end, GFP_NOFS);
+ btrfs_start_ordered_extent(inode, ordered, 1);
+ btrfs_put_ordered_extent(ordered);
+ }
if (page->index == last_byte >> PAGE_CACHE_SHIFT) {
char *userpage;
@@ -2663,33 +2683,20 @@ int extent_readpages(struct extent_io_tree *tree,
{
struct bio *bio = NULL;
unsigned page_idx;
- struct pagevec pvec;
unsigned long bio_flags = 0;
- pagevec_init(&pvec, 0);
for (page_idx = 0; page_idx < nr_pages; page_idx++) {
struct page *page = list_entry(pages->prev, struct page, lru);
prefetchw(&page->flags);
list_del(&page->lru);
- /*
- * what we want to do here is call add_to_page_cache_lru,
- * but that isn't exported, so we reproduce it here
- */
- if (!add_to_page_cache(page, mapping,
+ if (!add_to_page_cache_lru(page, mapping,
page->index, GFP_KERNEL)) {
-
- /* open coding of lru_cache_add, also not exported */
- page_cache_get(page);
- if (!pagevec_add(&pvec, page))
- __pagevec_lru_add_file(&pvec);
__extent_read_full_page(tree, page, get_extent,
&bio, 0, &bio_flags);
}
page_cache_release(page);
}
- if (pagevec_count(&pvec))
- __pagevec_lru_add_file(&pvec);
BUG_ON(!list_empty(pages));
if (bio)
submit_one_bio(READ, bio, 0, bio_flags);
@@ -2704,6 +2711,7 @@ int extent_readpages(struct extent_io_tree *tree,
int extent_invalidatepage(struct extent_io_tree *tree,
struct page *page, unsigned long offset)
{
+ struct extent_state *cached_state = NULL;
u64 start = ((u64)page->index << PAGE_CACHE_SHIFT);
u64 end = start + PAGE_CACHE_SIZE - 1;
size_t blocksize = page->mapping->host->i_sb->s_blocksize;
@@ -2712,12 +2720,12 @@ int extent_invalidatepage(struct extent_io_tree *tree,
if (start > end)
return 0;
- lock_extent(tree, start, end, GFP_NOFS);
+ lock_extent_bits(tree, start, end, 0, &cached_state, GFP_NOFS);
wait_on_page_writeback(page);
clear_extent_bit(tree, start, end,
EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
EXTENT_DO_ACCOUNTING,
- 1, 1, NULL, GFP_NOFS);
+ 1, 1, &cached_state, GFP_NOFS);
return 0;
}
@@ -2920,16 +2928,17 @@ sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
get_extent_t *get_extent)
{
struct inode *inode = mapping->host;
+ struct extent_state *cached_state = NULL;
u64 start = iblock << inode->i_blkbits;
sector_t sector = 0;
size_t blksize = (1 << inode->i_blkbits);
struct extent_map *em;
- lock_extent(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
- GFP_NOFS);
+ lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
+ 0, &cached_state, GFP_NOFS);
em = get_extent(inode, NULL, 0, start, blksize, 0);
- unlock_extent(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
- GFP_NOFS);
+ unlock_extent_cached(&BTRFS_I(inode)->io_tree, start,
+ start + blksize - 1, &cached_state, GFP_NOFS);
if (!em || IS_ERR(em))
return 0;
@@ -2951,6 +2960,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
u32 flags = 0;
u64 disko = 0;
struct extent_map *em = NULL;
+ struct extent_state *cached_state = NULL;
int end = 0;
u64 em_start = 0, em_len = 0;
unsigned long emflags;
@@ -2959,8 +2969,8 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
if (len == 0)
return -EINVAL;
- lock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
- GFP_NOFS);
+ lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0,
+ &cached_state, GFP_NOFS);
em = get_extent(inode, NULL, 0, off, max - off, 0);
if (!em)
goto out;
@@ -3023,8 +3033,8 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
out_free:
free_extent_map(em);
out:
- unlock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
- GFP_NOFS);
+ unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, start + len,
+ &cached_state, GFP_NOFS);
return ret;
}
@@ -3165,10 +3175,9 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
spin_unlock(&tree->buffer_lock);
goto free_eb;
}
- spin_unlock(&tree->buffer_lock);
-
/* add one reference for the tree */
atomic_inc(&eb->refs);
+ spin_unlock(&tree->buffer_lock);
return eb;
free_eb:
@@ -3265,7 +3274,8 @@ int set_extent_buffer_dirty(struct extent_io_tree *tree,
}
int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
- struct extent_buffer *eb)
+ struct extent_buffer *eb,
+ struct extent_state **cached_state)
{
unsigned long i;
struct page *page;
@@ -3275,7 +3285,7 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
- GFP_NOFS);
+ cached_state, GFP_NOFS);
for (i = 0; i < num_pages; i++) {
page = extent_buffer_page(eb, i);
if (page)
@@ -3335,7 +3345,8 @@ int extent_range_uptodate(struct extent_io_tree *tree,
}
int extent_buffer_uptodate(struct extent_io_tree *tree,
- struct extent_buffer *eb)
+ struct extent_buffer *eb,
+ struct extent_state *cached_state)
{
int ret = 0;
unsigned long num_pages;
@@ -3347,7 +3358,7 @@ int extent_buffer_uptodate(struct extent_io_tree *tree,
return 1;
ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
- EXTENT_UPTODATE, 1, NULL);
+ EXTENT_UPTODATE, 1, cached_state);
if (ret)
return ret;