#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>
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);
return state;
}
-static void free_extent_state(struct extent_state *state)
+void free_extent_state(struct extent_state *state)
{
if (!state)
return;
return NULL;
}
+static void merge_cb(struct extent_io_tree *tree, struct extent_state *new,
+ struct extent_state *other)
+{
+ if (tree->ops && tree->ops->merge_extent_hook)
+ tree->ops->merge_extent_hook(tree->mapping->host, new,
+ other);
+}
+
/*
* utility function to look for merge candidates inside a given range.
* Any extents with matching state are merged together into a single
other = rb_entry(other_node, struct extent_state, rb_node);
if (other->end == state->start - 1 &&
other->state == state->state) {
+ merge_cb(tree, state, other);
state->start = other->start;
other->tree = NULL;
rb_erase(&other->rb_node, &tree->state);
other = rb_entry(other_node, struct extent_state, rb_node);
if (other->start == state->end + 1 &&
other->state == state->state) {
+ merge_cb(tree, state, other);
other->start = state->start;
state->tree = NULL;
rb_erase(&state->rb_node, &tree->state);
free_extent_state(state);
+ state = NULL;
}
}
+
return 0;
}
-static void set_state_cb(struct extent_io_tree *tree,
- struct extent_state *state,
- unsigned long bits)
+static int set_state_cb(struct extent_io_tree *tree,
+ struct extent_state *state, int *bits)
{
if (tree->ops && tree->ops->set_bit_hook) {
- tree->ops->set_bit_hook(tree->mapping->host, state->start,
- state->end, state->state, bits);
+ return tree->ops->set_bit_hook(tree->mapping->host,
+ 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->start,
- state->end, state->state, bits);
- }
+ if (tree->ops && tree->ops->clear_bit_hook)
+ tree->ops->clear_bit_hook(tree->mapping->host, state, bits);
}
/*
*/
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) {
printk(KERN_ERR "btrfs end < start %llu %llu\n",
(unsigned long long)start);
WARN_ON(1);
}
- if (bits & EXTENT_DIRTY)
- tree->dirty_bytes += end - start + 1;
- set_state_cb(tree, state, bits);
- state->state |= bits;
state->start = start;
state->end = end;
+ ret = set_state_cb(tree, state, bits);
+ if (ret)
+ return ret;
+
+ if (bits_to_set & EXTENT_DIRTY)
+ tree->dirty_bytes += end - start + 1;
+ state->state |= bits_to_set;
node = tree_insert(&tree->state, end, &state->rb_node);
if (node) {
struct extent_state *found;
return 0;
}
+static int split_cb(struct extent_io_tree *tree, struct extent_state *orig,
+ u64 split)
+{
+ if (tree->ops && tree->ops->split_extent_hook)
+ return tree->ops->split_extent_hook(tree->mapping->host,
+ orig, split);
+ return 0;
+}
+
/*
* split a given extent state struct in two, inserting the preallocated
* struct 'prealloc' as the newly created second half. 'split' indicates an
struct extent_state *prealloc, u64 split)
{
struct rb_node *node;
+
+ split_cb(tree, orig, split);
+
prealloc->start = orig->start;
prealloc->end = split - 1;
prealloc->state = orig->state;
* 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 ret = state->state & bits;
+ 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;
}
clear_state_cb(tree, state, bits);
- state->state &= ~bits;
+ 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);
* bits were already set, or zero if none of the bits were already set.
*/
int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
- int bits, int wake, int delete, gfp_t mask)
+ int bits, int wake, int delete,
+ struct extent_state **cached_state,
+ gfp_t mask)
{
struct extent_state *state;
+ struct extent_state *cached;
struct extent_state *prealloc = NULL;
+ struct rb_node *next_node;
struct rb_node *node;
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);
}
spin_lock(&tree->lock);
+ if (cached_state) {
+ cached = *cached_state;
+
+ if (clear) {
+ *cached_state = NULL;
+ cached_state = NULL;
+ }
+
+ if (cached && cached->tree && cached->start == start) {
+ if (clear)
+ atomic_dec(&cached->refs);
+ state = cached;
+ goto hit_next;
+ }
+ if (clear)
+ free_extent_state(cached);
+ }
/*
* this search will find the extents that end after
* our range starts
if (!node)
goto out;
state = rb_entry(node, struct extent_state, rb_node);
+hit_next:
if (state->start > end)
goto out;
WARN_ON(state->end < start);
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;
- } else {
- start = state->start;
}
goto search_again;
}
prealloc = alloc_extent_state(GFP_ATOMIC);
err = split_state(tree, state, prealloc, end + 1);
BUG_ON(err == -EEXIST);
-
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;
}
- set |= clear_state_bit(tree, state, bits, wake, delete);
+ if (state->end < end && prealloc && !need_resched())
+ next_node = rb_next(&state->rb_node);
+ else
+ next_node = NULL;
+
+ set |= clear_state_bit(tree, state, &bits, wake);
if (last_end == (u64)-1)
goto out;
start = last_end + 1;
+ if (start <= end && next_node) {
+ state = rb_entry(next_node, struct extent_state,
+ rb_node);
+ if (state->start == start)
+ goto hit_next;
+ }
goto search_again;
out:
return 0;
}
-static void set_state_bits(struct extent_io_tree *tree,
+static int set_state_bits(struct extent_io_tree *tree,
struct extent_state *state,
- int bits)
+ int *bits)
{
- if ((bits & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
+ int ret;
+ int bits_to_set = *bits & ~EXTENT_CTLBITS;
+
+ ret = set_state_cb(tree, state, bits);
+ if (ret)
+ return ret;
+ if ((bits_to_set & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
u64 range = state->end - state->start + 1;
tree->dirty_bytes += range;
}
- set_state_cb(tree, state, bits);
- state->state |= bits;
+ state->state |= bits_to_set;
+
+ return 0;
+}
+
+static void cache_state(struct extent_state *state,
+ struct extent_state **cached_ptr)
+{
+ if (cached_ptr && !(*cached_ptr)) {
+ if (state->state & (EXTENT_IOBITS | EXTENT_BOUNDARY)) {
+ *cached_ptr = state;
+ atomic_inc(&state->refs);
+ }
+ }
}
/*
- * set some bits on a range in the tree. This may require allocations
- * or sleeping, so the gfp mask is used to indicate what is allowed.
+ * set some bits on a range in the tree. This may require allocations or
+ * sleeping, so the gfp mask is used to indicate what is allowed.
*
- * If 'exclusive' == 1, this will fail with -EEXIST if some part of the
- * range already has the desired bits set. The start of the existing
- * range is returned in failed_start in this case.
+ * If any of the exclusive bits are set, this will fail with -EEXIST if some
+ * part of the range already has the desired bits set. The start of the
+ * existing range is returned in failed_start in this case.
*
- * [start, end] is inclusive
- * This takes the tree lock.
+ * [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, u64 *failed_start,
- 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;
struct rb_node *node;
int err = 0;
- int set;
u64 last_start;
u64 last_end;
+
+ bits |= EXTENT_FIRST_DELALLOC;
again:
if (!prealloc && (mask & __GFP_WAIT)) {
prealloc = alloc_extent_state(mask);
}
spin_lock(&tree->lock);
+ if (cached_state && *cached_state) {
+ state = *cached_state;
+ if (state->start == start && state->tree) {
+ node = &state->rb_node;
+ goto hit_next;
+ }
+ }
/*
* this search will find all the extents that end after
* our range starts.
*/
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;
*/
if (state->start == start && state->end <= end) {
struct rb_node *next_node;
- set = state->state & bits;
- if (set && exclusive) {
+ if (state->state & exclusive_bits) {
*failed_start = state->start;
err = -EEXIST;
goto out;
}
- set_state_bits(tree, state, bits);
+
+ err = set_state_bits(tree, state, &bits);
+ if (err)
+ goto out;
+
+ cache_state(state, cached_state);
merge_state(tree, state);
if (last_end == (u64)-1)
goto out;
* desired bit on it.
*/
if (state->start < start) {
- set = state->state & bits;
- if (exclusive && set) {
+ if (state->state & exclusive_bits) {
*failed_start = start;
err = -EEXIST;
goto out;
if (err)
goto out;
if (state->end <= end) {
- set_state_bits(tree, state, bits);
+ err = set_state_bits(tree, state, &bits);
+ if (err)
+ goto out;
+ cache_state(state, cached_state);
merge_state(tree, state);
if (last_end == (u64)-1)
goto out;
start = last_end + 1;
- } else {
- start = state->start;
}
goto search_again;
}
else
this_end = last_start - 1;
err = insert_state(tree, prealloc, start, this_end,
- bits);
- prealloc = NULL;
+ &bits);
BUG_ON(err == -EEXIST);
- if (err)
+ if (err) {
+ prealloc = NULL;
goto out;
+ }
+ cache_state(prealloc, cached_state);
+ prealloc = NULL;
start = this_end + 1;
goto search_again;
}
* on the first half
*/
if (state->start <= end && state->end > end) {
- set = state->state & bits;
- if (exclusive && set) {
+ if (state->state & exclusive_bits) {
*failed_start = start;
err = -EEXIST;
goto out;
err = split_state(tree, state, prealloc, end + 1);
BUG_ON(err == -EEXIST);
- set_state_bits(tree, prealloc, bits);
+ err = set_state_bits(tree, prealloc, &bits);
+ if (err) {
+ prealloc = NULL;
+ goto out;
+ }
+ cache_state(prealloc, cached_state);
merge_state(tree, prealloc);
prealloc = NULL;
goto out;
gfp_t mask)
{
return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
- mask);
-}
-
-int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
-{
- return set_extent_bit(tree, start, end, EXTENT_ORDERED, 0, NULL, mask);
+ NULL, mask);
}
int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
int bits, gfp_t mask)
{
return set_extent_bit(tree, start, end, bits, 0, NULL,
- mask);
+ NULL, mask);
}
int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
int bits, gfp_t mask)
{
- return clear_extent_bit(tree, start, end, bits, 0, 0, mask);
+ return clear_extent_bit(tree, start, end, bits, 0, 0, NULL, mask);
}
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, mask);
+ 0, NULL, cached_state, mask);
}
int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
return clear_extent_bit(tree, start, end,
- EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, mask);
-}
-
-int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
-{
- return clear_extent_bit(tree, start, end, EXTENT_ORDERED, 1, 0, mask);
+ EXTENT_DIRTY | EXTENT_DELALLOC |
+ EXTENT_DO_ACCOUNTING, 0, 0, NULL, mask);
}
int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL,
- mask);
+ NULL, mask);
}
static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
- return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, mask);
+ return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0,
+ NULL, mask);
}
int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
- mask);
+ NULL, mask);
}
static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
- u64 end, gfp_t mask)
-{
- return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0, mask);
-}
-
-static int set_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
+ u64 end, struct extent_state **cached_state,
+ gfp_t mask)
{
- return set_extent_bit(tree, start, end, EXTENT_WRITEBACK,
- 0, NULL, mask);
-}
-
-static int clear_extent_writeback(struct extent_io_tree *tree, u64 start,
- u64 end, gfp_t mask)
-{
- return clear_extent_bit(tree, start, end, EXTENT_WRITEBACK, 1, 0, mask);
+ return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0,
+ cached_state, mask);
}
int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end)
* either insert or lock state struct between start and end use mask to tell
* us if waiting is desired.
*/
-int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
+int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+ int bits, struct extent_state **cached_state, gfp_t mask)
{
int err;
u64 failed_start;
while (1) {
- err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
- &failed_start, mask);
+ err = set_extent_bit(tree, start, end, EXTENT_LOCKED | bits,
+ EXTENT_LOCKED, &failed_start,
+ cached_state, mask);
if (err == -EEXIST && (mask & __GFP_WAIT)) {
wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
start = failed_start;
return err;
}
+int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
+{
+ return lock_extent_bits(tree, start, end, 0, NULL, mask);
+}
+
int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
int err;
u64 failed_start;
- err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
- &failed_start, mask);
+ err = set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED,
+ &failed_start, NULL, mask);
if (err == -EEXIST) {
if (failed_start > start)
clear_extent_bit(tree, start, failed_start - 1,
- EXTENT_LOCKED, 1, 0, mask);
+ EXTENT_LOCKED, 1, 0, NULL, mask);
return 0;
}
return 1;
}
+int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
+ struct extent_state **cached, gfp_t mask)
+{
+ return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, cached,
+ mask);
+}
+
int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
- return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, mask);
+ return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL,
+ mask);
}
/*
page_cache_release(page);
index++;
}
- set_extent_dirty(tree, start, end, GFP_NOFS);
return 0;
}
page_cache_release(page);
index++;
}
- set_extent_writeback(tree, start, end, GFP_NOFS);
return 0;
}
* 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;
*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;
u64 delalloc_start;
u64 delalloc_end;
u64 found;
+ struct extent_state *cached_state = NULL;
int ret;
int loops = 0;
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;
}
/* some of the pages are gone, lets avoid looping by
* shortening the size of the delalloc range we're searching
*/
+ free_extent_state(cached_state);
if (!loops) {
unsigned long offset = (*start) & (PAGE_CACHE_SIZE - 1);
max_bytes = PAGE_CACHE_SIZE - offset;
BUG_ON(ret);
/* step three, lock the state bits for the whole range */
- lock_extent(tree, delalloc_start, delalloc_end, GFP_NOFS);
+ lock_extent_bits(tree, delalloc_start, delalloc_end,
+ 0, &cached_state, GFP_NOFS);
/* then test to make sure it is all still delalloc */
ret = test_range_bit(tree, delalloc_start, delalloc_end,
- EXTENT_DELALLOC, 1);
+ EXTENT_DELALLOC, 1, cached_state);
if (!ret) {
- unlock_extent(tree, delalloc_start, delalloc_end, GFP_NOFS);
+ unlock_extent_cached(tree, delalloc_start, delalloc_end,
+ &cached_state, GFP_NOFS);
__unlock_for_delalloc(inode, locked_page,
delalloc_start, delalloc_end);
cond_resched();
goto again;
}
+ free_extent_state(cached_state);
*start = delalloc_start;
*end = delalloc_end;
out_failed:
int extent_clear_unlock_delalloc(struct inode *inode,
struct extent_io_tree *tree,
u64 start, u64 end, struct page *locked_page,
- int unlock_pages,
- int clear_unlock,
- int clear_delalloc, int clear_dirty,
- int set_writeback,
- int end_writeback)
+ unsigned long op)
{
int ret;
struct page *pages[16];
int i;
int clear_bits = 0;
- if (clear_unlock)
+ if (op & EXTENT_CLEAR_UNLOCK)
clear_bits |= EXTENT_LOCKED;
- if (clear_dirty)
+ if (op & EXTENT_CLEAR_DIRTY)
clear_bits |= EXTENT_DIRTY;
- if (clear_delalloc)
+ if (op & EXTENT_CLEAR_DELALLOC)
clear_bits |= EXTENT_DELALLOC;
- clear_extent_bit(tree, start, end, clear_bits, 1, 0, GFP_NOFS);
- if (!(unlock_pages || clear_dirty || set_writeback || end_writeback))
+ 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 |
+ EXTENT_SET_PRIVATE2)))
return 0;
while (nr_pages > 0) {
min_t(unsigned long,
nr_pages, ARRAY_SIZE(pages)), pages);
for (i = 0; i < ret; i++) {
+
+ if (op & EXTENT_SET_PRIVATE2)
+ SetPagePrivate2(pages[i]);
+
if (pages[i] == locked_page) {
page_cache_release(pages[i]);
continue;
}
- if (clear_dirty)
+ if (op & EXTENT_CLEAR_DIRTY)
clear_page_dirty_for_io(pages[i]);
- if (set_writeback)
+ if (op & EXTENT_SET_WRITEBACK)
set_page_writeback(pages[i]);
- if (end_writeback)
+ if (op & EXTENT_END_WRITEBACK)
end_page_writeback(pages[i]);
- if (unlock_pages)
+ if (op & EXTENT_CLEAR_UNLOCK_PAGE)
unlock_page(pages[i]);
page_cache_release(pages[i]);
}
* range is found set.
*/
int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
- int bits, int filled)
+ int bits, int filled, struct extent_state *cached)
{
struct extent_state *state = NULL;
struct rb_node *node;
int bitset = 0;
spin_lock(&tree->lock);
- node = tree_search(tree, start);
+ if (cached && cached->tree && cached->start == start)
+ node = &cached->rb_node;
+ else
+ node = tree_search(tree, start);
while (node && start <= end) {
state = rb_entry(node, struct extent_state, rb_node);
bitset = 0;
break;
}
+
+ if (state->end == (u64)-1)
+ break;
+
start = state->end + 1;
if (start > end)
break;
{
u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
u64 end = start + PAGE_CACHE_SIZE - 1;
- if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1))
+ if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL))
SetPageUptodate(page);
return 0;
}
{
u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
u64 end = start + PAGE_CACHE_SIZE - 1;
- if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0))
+ if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0, NULL))
unlock_page(page);
return 0;
}
static int check_page_writeback(struct extent_io_tree *tree,
struct page *page)
{
- u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
- u64 end = start + PAGE_CACHE_SIZE - 1;
- if (!test_range_bit(tree, start, end, EXTENT_WRITEBACK, 0))
- end_page_writeback(page);
+ end_page_writeback(page);
return 0;
}
}
if (!uptodate) {
- clear_extent_uptodate(tree, start, end, GFP_ATOMIC);
+ clear_extent_uptodate(tree, start, end, NULL, GFP_NOFS);
ClearPageUptodate(page);
SetPageError(page);
}
- clear_extent_writeback(tree, start, end, GFP_ATOMIC);
-
if (whole_page)
end_page_writeback(page);
else
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;
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) {
}
check_page_locked(tree, page);
}
- } while (bvec >= bio->bi_io_vec);
+ } while (bvec <= bvec_end);
bio_put(bio);
}
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))
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;
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;
continue;
}
/* the get_extent function already copied into the page */
- if (test_range_bit(tree, cur, cur_end, EXTENT_UPTODATE, 1)) {
+ if (test_range_bit(tree, cur, cur_end,
+ EXTENT_UPTODATE, 1, NULL)) {
check_page_uptodate(tree, page);
unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
cur = cur + iosize;
u64 iosize;
u64 unlock_start;
sector_t sector;
+ struct extent_state *cached_state = NULL;
struct extent_map *em;
struct block_device *bdev;
int ret;
delalloc_end = 0;
page_started = 0;
if (!epd->extent_locked) {
- u64 delalloc_to_write;
+ u64 delalloc_to_write = 0;
/*
* make sure the wbc mapping index is at least updated
* to this page.
tree->ops->fill_delalloc(inode, page, delalloc_start,
delalloc_end, &page_started,
&nr_written);
- delalloc_to_write = (delalloc_end -
- max_t(u64, page_offset(page),
- delalloc_start) + 1) >>
- PAGE_CACHE_SHIFT;
- if (wbc->nr_to_write < delalloc_to_write) {
- wbc->nr_to_write = min_t(long, 8192,
- delalloc_to_write);
- }
+ /*
+ * delalloc_end is already one less than the total
+ * length, so we don't subtract one from
+ * PAGE_CACHE_SIZE
+ */
+ delalloc_to_write += (delalloc_end - delalloc_start +
+ PAGE_CACHE_SIZE) >>
+ PAGE_CACHE_SHIFT;
delalloc_start = delalloc_end + 1;
}
+ if (wbc->nr_to_write < delalloc_to_write) {
+ int thresh = 8192;
+
+ if (delalloc_to_write < thresh * 2)
+ thresh = delalloc_to_write;
+ wbc->nr_to_write = min_t(u64, delalloc_to_write,
+ thresh);
+ }
/* did the fill delalloc function already unlock and start
* the IO?
goto done_unlocked;
}
}
- lock_extent(tree, start, page_end, GFP_NOFS);
-
- unlock_start = start;
-
if (tree->ops && tree->ops->writepage_start_hook) {
ret = tree->ops->writepage_start_hook(page, start,
page_end);
if (ret == -EAGAIN) {
- unlock_extent(tree, start, page_end, GFP_NOFS);
redirty_page_for_writepage(wbc, page);
update_nr_written(page, wbc, nr_written);
unlock_page(page);
update_nr_written(page, wbc, nr_written + 1);
end = page_end;
- if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0))
- printk(KERN_ERR "btrfs delalloc bits after lock_extent\n");
-
if (last_byte <= start) {
- clear_extent_dirty(tree, start, page_end, GFP_NOFS);
- unlock_extent(tree, start, page_end, GFP_NOFS);
if (tree->ops && tree->ops->writepage_end_io_hook)
tree->ops->writepage_end_io_hook(page, start,
page_end, NULL, 1);
goto done;
}
- set_extent_uptodate(tree, start, page_end, GFP_NOFS);
blocksize = inode->i_sb->s_blocksize;
while (cur <= end) {
if (cur >= last_byte) {
- clear_extent_dirty(tree, cur, page_end, GFP_NOFS);
- unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
if (tree->ops && tree->ops->writepage_end_io_hook)
tree->ops->writepage_end_io_hook(page, cur,
page_end, NULL, 1);
*/
if (compressed || block_start == EXTENT_MAP_HOLE ||
block_start == EXTENT_MAP_INLINE) {
- clear_extent_dirty(tree, cur,
- cur + iosize - 1, GFP_NOFS);
-
- unlock_extent(tree, unlock_start, cur + iosize - 1,
- GFP_NOFS);
-
/*
* end_io notification does not happen here for
* compressed extents
}
/* leave this out until we have a page_mkwrite call */
if (0 && !test_range_bit(tree, cur, cur + iosize - 1,
- EXTENT_DIRTY, 0)) {
+ EXTENT_DIRTY, 0, NULL)) {
cur = cur + iosize;
pg_offset += iosize;
continue;
}
- clear_extent_dirty(tree, cur, cur + iosize - 1, GFP_NOFS);
if (tree->ops && tree->ops->writepage_io_hook) {
ret = tree->ops->writepage_io_hook(page, cur,
cur + iosize - 1);
set_page_writeback(page);
end_page_writeback(page);
}
- if (unlock_start <= page_end)
- unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
unlock_page(page);
done_unlocked:
+ /* drop our reference on any cached states */
+ free_extent_state(cached_state);
return 0;
}
{
int ret = 0;
int done = 0;
+ int nr_to_write_done = 0;
struct pagevec pvec;
int nr_pages;
pgoff_t index;
scanned = 1;
}
retry:
- while (!done && (index <= end) &&
+ while (!done && !nr_to_write_done && (index <= end) &&
(nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
PAGECACHE_TAG_DIRTY, min(end - index,
(pgoff_t)PAGEVEC_SIZE-1) + 1))) {
unlock_page(page);
ret = 0;
}
- if (ret || wbc->nr_to_write <= 0)
+ if (ret)
done = 1;
+
+ /*
+ * the filesystem may choose to bump up nr_to_write.
+ * We have to make sure to honor the new nr_to_write
+ * at any time
+ */
+ nr_to_write_done = wbc->nr_to_write <= 0;
}
pagevec_release(&pvec);
cond_resched();
{
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);
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;
if (start > end)
return 0;
- lock_extent(tree, start, end, GFP_NOFS);
- wait_on_extent_writeback(tree, start, end);
+ 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,
- 1, 1, GFP_NOFS);
+ EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
+ EXTENT_DO_ACCOUNTING,
+ 1, 1, &cached_state, GFP_NOFS);
return 0;
}
!isnew && !PageUptodate(page) &&
(block_off_end > to || block_off_start < from) &&
!test_range_bit(tree, block_start, cur_end,
- EXTENT_UPTODATE, 1)) {
+ EXTENT_UPTODATE, 1, NULL)) {
u64 sector;
u64 extent_offset = block_start - em->start;
size_t iosize;
*/
set_extent_bit(tree, block_start,
block_start + iosize - 1,
- EXTENT_LOCKED, 0, NULL, GFP_NOFS);
+ EXTENT_LOCKED, 0, NULL, NULL, GFP_NOFS);
ret = submit_extent_page(READ, tree, page,
sector, iosize, page_offset, em->bdev,
NULL, 1,
int ret = 1;
if (test_range_bit(tree, start, end,
- EXTENT_IOBITS | EXTENT_ORDERED, 0))
+ EXTENT_IOBITS, 0, NULL))
ret = 0;
else {
if ((mask & GFP_NOFS) == GFP_NOFS)
mask = GFP_NOFS;
- clear_extent_bit(tree, start, end, EXTENT_UPTODATE,
- 1, 1, mask);
+ /*
+ * at this point we can safely clear everything except the
+ * locked bit and the nodatasum bit
+ */
+ clear_extent_bit(tree, start, end,
+ ~(EXTENT_LOCKED | EXTENT_NODATASUM),
+ 0, 0, NULL, mask);
}
return ret;
}
u64 len;
while (start <= end) {
len = end - start + 1;
- spin_lock(&map->lock);
+ write_lock(&map->lock);
em = lookup_extent_mapping(map, start, len);
if (!em || IS_ERR(em)) {
- spin_unlock(&map->lock);
+ write_unlock(&map->lock);
break;
}
if (test_bit(EXTENT_FLAG_PINNED, &em->flags) ||
em->start != start) {
- spin_unlock(&map->lock);
+ write_unlock(&map->lock);
free_extent_map(em);
break;
}
if (!test_range_bit(tree, em->start,
extent_map_end(em) - 1,
- EXTENT_LOCKED | EXTENT_WRITEBACK |
- EXTENT_ORDERED,
- 0)) {
+ EXTENT_LOCKED | EXTENT_WRITEBACK,
+ 0, NULL)) {
remove_extent_mapping(map, em);
/* once for the rb tree */
free_extent_map(em);
}
start = extent_map_end(em);
- spin_unlock(&map->lock);
+ write_unlock(&map->lock);
/* once for us */
free_extent_map(em);
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;
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;
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;
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;
}
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:
}
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;
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)
int uptodate;
unsigned long index;
- ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1);
+ ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL);
if (ret)
return 1;
while (start <= end) {
}
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;
return 1;
ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
- EXTENT_UPTODATE, 1);
+ EXTENT_UPTODATE, 1, cached_state);
if (ret)
return ret;
return 0;
if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
- EXTENT_UPTODATE, 1)) {
+ EXTENT_UPTODATE, 1, NULL)) {
return 0;
}
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff