#include <linux/spinlock.h>
#include <linux/blkdev.h>
#include <linux/swap.h>
-#include <linux/version.h>
#include <linux/writeback.h>
#include <linux/pagevec.h>
#include "extent_io.h"
#include "ctree.h"
#include "btrfs_inode.h"
-/* temporary define until extent_map moves out of btrfs */
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
- unsigned long extra_flags,
- void (*ctor)(void *, struct kmem_cache *,
- unsigned long));
-
static struct kmem_cache *extent_state_cache;
static struct kmem_cache *extent_buffer_cache;
static LIST_HEAD(states);
#define LEAK_DEBUG 0
-#ifdef LEAK_DEBUG
-static spinlock_t leak_lock = SPIN_LOCK_UNLOCKED;
+#if LEAK_DEBUG
+static DEFINE_SPINLOCK(leak_lock);
#endif
#define BUFFER_LRU_MAX 64
/* tells writepage not to lock the state bits for this range
* it still does the unlocking
*/
- int extent_locked;
+ unsigned int extent_locked:1;
+
+ /* tells the submit_bio code to use a WRITE_SYNC */
+ unsigned int sync_io:1;
};
int __init extent_io_init(void)
{
- extent_state_cache = btrfs_cache_create("extent_state",
- sizeof(struct extent_state), 0,
- NULL);
+ extent_state_cache = kmem_cache_create("extent_state",
+ sizeof(struct extent_state), 0,
+ SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
if (!extent_state_cache)
return -ENOMEM;
- extent_buffer_cache = btrfs_cache_create("extent_buffers",
- sizeof(struct extent_buffer), 0,
- NULL);
+ extent_buffer_cache = kmem_cache_create("extent_buffers",
+ sizeof(struct extent_buffer), 0,
+ SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
if (!extent_buffer_cache)
goto free_state_cache;
return 0;
while (!list_empty(&states)) {
state = list_entry(states.next, struct extent_state, leak_list);
- printk("state leak: start %Lu end %Lu state %lu in tree %p refs %d\n", state->start, state->end, state->state, state->tree, atomic_read(&state->refs));
+ printk(KERN_ERR "btrfs state leak: start %llu end %llu "
+ "state %lu in tree %p refs %d\n",
+ (unsigned long long)state->start,
+ (unsigned long long)state->end,
+ state->state, state->tree, atomic_read(&state->refs));
list_del(&state->leak_list);
kmem_cache_free(extent_state_cache, state);
while (!list_empty(&buffers)) {
eb = list_entry(buffers.next, struct extent_buffer, leak_list);
- printk("buffer leak start %Lu len %lu refs %d\n", eb->start, eb->len, atomic_read(&eb->refs));
+ printk(KERN_ERR "btrfs buffer leak start %llu len %lu "
+ "refs %d\n", (unsigned long long)eb->start,
+ eb->len, atomic_read(&eb->refs));
list_del(&eb->leak_list);
kmem_cache_free(extent_buffer_cache, eb);
}
spin_lock_init(&tree->buffer_lock);
tree->mapping = mapping;
}
-EXPORT_SYMBOL(extent_io_tree_init);
static struct extent_state *alloc_extent_state(gfp_t mask)
{
struct extent_state *state;
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
unsigned long flags;
#endif
state->state = 0;
state->private = 0;
state->tree = NULL;
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
spin_lock_irqsave(&leak_lock, flags);
list_add(&state->leak_list, &states);
spin_unlock_irqrestore(&leak_lock, flags);
init_waitqueue_head(&state->wq);
return state;
}
-EXPORT_SYMBOL(alloc_extent_state);
static void free_extent_state(struct extent_state *state)
{
if (!state)
return;
if (atomic_dec_and_test(&state->refs)) {
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
unsigned long flags;
#endif
WARN_ON(state->tree);
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
spin_lock_irqsave(&leak_lock, flags);
list_del(&state->leak_list);
spin_unlock_irqrestore(&leak_lock, flags);
kmem_cache_free(extent_state_cache, state);
}
}
-EXPORT_SYMBOL(free_extent_state);
static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
struct rb_node *node)
{
- struct rb_node ** p = &root->rb_node;
- struct rb_node * parent = NULL;
+ struct rb_node **p = &root->rb_node;
+ struct rb_node *parent = NULL;
struct tree_entry *entry;
- while(*p) {
+ while (*p) {
parent = *p;
entry = rb_entry(parent, struct tree_entry, rb_node);
struct rb_node **next_ret)
{
struct rb_root *root = &tree->state;
- struct rb_node * n = root->rb_node;
+ struct rb_node *n = root->rb_node;
struct rb_node *prev = NULL;
struct rb_node *orig_prev = NULL;
struct tree_entry *entry;
struct tree_entry *prev_entry = NULL;
- while(n) {
+ while (n) {
entry = rb_entry(n, struct tree_entry, rb_node);
prev = n;
prev_entry = entry;
n = n->rb_left;
else if (offset > entry->end)
n = n->rb_right;
- else {
+ else
return n;
- }
}
if (prev_ret) {
orig_prev = prev;
- while(prev && offset > prev_entry->end) {
+ while (prev && offset > prev_entry->end) {
prev = rb_next(prev);
prev_entry = rb_entry(prev, struct tree_entry, rb_node);
}
if (next_ret) {
prev_entry = rb_entry(prev, struct tree_entry, rb_node);
- while(prev && offset < prev_entry->start) {
+ while (prev && offset < prev_entry->start) {
prev = rb_prev(prev);
prev_entry = rb_entry(prev, struct tree_entry, rb_node);
}
struct rb_node *ret;
ret = __etree_search(tree, offset, &prev, NULL);
- if (!ret) {
+ if (!ret)
return prev;
- }
return ret;
}
u64 offset, struct rb_node *node)
{
struct rb_root *root = &tree->buffer;
- struct rb_node ** p = &root->rb_node;
- struct rb_node * parent = NULL;
+ struct rb_node **p = &root->rb_node;
+ struct rb_node *parent = NULL;
struct extent_buffer *eb;
- while(*p) {
+ while (*p) {
parent = *p;
eb = rb_entry(parent, struct extent_buffer, rb_node);
u64 offset)
{
struct rb_root *root = &tree->buffer;
- struct rb_node * n = root->rb_node;
+ struct rb_node *n = root->rb_node;
struct extent_buffer *eb;
- while(n) {
+ while (n) {
eb = rb_entry(n, struct extent_buffer, rb_node);
if (offset < eb->start)
n = n->rb_left;
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,
+static int set_state_cb(struct extent_io_tree *tree,
struct extent_state *state,
unsigned long 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->start, state->end,
+ state->state, bits);
}
+
+ return 0;
}
static void clear_state_cb(struct extent_io_tree *tree,
struct extent_state *state,
unsigned long bits)
{
- if (tree->ops && tree->ops->set_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);
}
/*
int bits)
{
struct rb_node *node;
+ int ret;
if (end < start) {
- printk("end < start %Lu %Lu\n", end, start);
+ printk(KERN_ERR "btrfs end < start %llu %llu\n",
+ (unsigned long long)end,
+ (unsigned long long)start);
WARN_ON(1);
}
+ state->start = start;
+ state->end = end;
+ ret = set_state_cb(tree, state, bits);
+ if (ret)
+ return ret;
+
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;
node = tree_insert(&tree->state, end, &state->rb_node);
if (node) {
struct extent_state *found;
found = rb_entry(node, struct extent_state, rb_node);
- printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, start, end);
+ printk(KERN_ERR "btrfs found node %llu %llu on insert of "
+ "%llu %llu\n", (unsigned long long)found->start,
+ (unsigned long long)found->end,
+ (unsigned long long)start, (unsigned long long)end);
free_extent_state(state);
return -EEXIST;
}
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;
node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node);
if (node) {
- struct extent_state *found;
- found = rb_entry(node, struct extent_state, rb_node);
- printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, prealloc->start, prealloc->end);
free_extent_state(prealloc);
return -EEXIST;
}
struct extent_state *state, int bits, int wake,
int delete)
{
- int ret = state->state & bits;
+ int bits_to_clear = bits & ~EXTENT_DO_ACCOUNTING;
+ int ret = state->state & bits_to_clear;
if ((bits & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
u64 range = state->end - state->start + 1;
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) {
* 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;
}
spin_lock(&tree->lock);
+ if (cached_state) {
+ cached = *cached_state;
+ *cached_state = NULL;
+ cached_state = NULL;
+ if (cached && cached->tree && cached->start == start) {
+ atomic_dec(&cached->refs);
+ state = cached;
+ goto hit_next;
+ }
+ 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);
+ last_end = state->end;
/*
* | ---- desired range ---- |
if (err)
goto out;
if (state->end <= end) {
- start = state->end + 1;
- set |= clear_state_bit(tree, state, bits,
- wake, delete);
- } else {
- start = state->start;
+ set |= clear_state_bit(tree, state, bits, wake,
+ delete);
+ if (last_end == (u64)-1)
+ goto out;
+ start = last_end + 1;
}
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, delete);
+
prealloc = NULL;
goto out;
}
- start = state->end + 1;
+ 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, delete);
+ 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:
cond_resched();
goto again;
}
-EXPORT_SYMBOL(clear_extent_bit);
static int wait_on_state(struct extent_io_tree *tree,
struct extent_state *state)
spin_unlock(&tree->lock);
return 0;
}
-EXPORT_SYMBOL(wait_extent_bit);
-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 ret;
+
+ ret = set_state_cb(tree, state, bits);
+ if (ret)
+ return ret;
+
if ((bits & 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;
+
+ 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)
+
+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)
{
struct extent_state *state;
struct extent_state *prealloc = NULL;
struct rb_node *node;
int err = 0;
- int set;
u64 last_start;
u64 last_end;
+
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.
BUG_ON(err == -EEXIST);
goto out;
}
-
state = rb_entry(node, struct extent_state, rb_node);
+hit_next:
last_start = state->start;
last_end = state->end;
* Just lock what we found and keep going
*/
if (state->start == start && state->end <= end) {
- set = state->state & bits;
- if (set && exclusive) {
+ struct rb_node *next_node;
+ if (state->state & exclusive_bits) {
*failed_start = state->start;
err = -EEXIST;
goto out;
}
- set_state_bits(tree, state, bits);
- start = state->end + 1;
+
+ 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;
+ if (start < end && prealloc && !need_resched()) {
+ next_node = rb_next(node);
+ if (next_node) {
+ state = rb_entry(next_node, struct extent_state,
+ rb_node);
+ if (state->start == start)
+ goto hit_next;
+ }
+ }
goto search_again;
}
* 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);
- start = state->end + 1;
+ err = set_state_bits(tree, state, bits);
+ if (err)
+ goto out;
+ cache_state(state, cached_state);
merge_state(tree, state);
- } else {
- start = state->start;
+ if (last_end == (u64)-1)
+ goto out;
+ start = last_end + 1;
}
goto search_again;
}
if (end < last_start)
this_end = end;
else
- this_end = last_start -1;
+ this_end = last_start - 1;
err = insert_state(tree, prealloc, start, this_end,
bits);
- prealloc = NULL;
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;
cond_resched();
goto again;
}
-EXPORT_SYMBOL(set_extent_bit);
/* wrappers around set/clear extent bit */
int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
- mask);
+ NULL, mask);
}
-EXPORT_SYMBOL(set_extent_dirty);
-
-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);
-}
-EXPORT_SYMBOL(set_extent_ordered);
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);
}
-EXPORT_SYMBOL(set_extent_bits);
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);
}
-EXPORT_SYMBOL(clear_extent_bits);
int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
gfp_t mask)
{
return set_extent_bit(tree, start, end,
- EXTENT_DELALLOC | EXTENT_DIRTY,
- 0, NULL, mask);
+ EXTENT_DELALLOC | EXTENT_DIRTY | EXTENT_UPTODATE,
+ 0, NULL, NULL, mask);
}
-EXPORT_SYMBOL(set_extent_delalloc);
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);
-}
-EXPORT_SYMBOL(clear_extent_dirty);
-
-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);
}
-EXPORT_SYMBOL(clear_extent_ordered);
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);
}
-EXPORT_SYMBOL(set_extent_new);
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);
-}
-EXPORT_SYMBOL(set_extent_uptodate);
-
-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);
+ NULL, mask);
}
-static int set_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
+static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
+ u64 end, 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,
+ NULL, mask);
}
int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end)
{
return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
}
-EXPORT_SYMBOL(wait_on_extent_writeback);
/*
* 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;
}
-EXPORT_SYMBOL(lock_extent);
+
+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;
}
-EXPORT_SYMBOL(try_lock_extent);
+
+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);
}
-EXPORT_SYMBOL(unlock_extent);
/*
* helper function to set pages and extents in the tree dirty
page_cache_release(page);
index++;
}
- set_extent_dirty(tree, start, end, GFP_NOFS);
return 0;
}
-EXPORT_SYMBOL(set_range_dirty);
/*
* helper function to set both pages and extents in the tree writeback
page_cache_release(page);
index++;
}
- set_extent_writeback(tree, start, end, GFP_NOFS);
return 0;
}
* our range starts.
*/
node = tree_search(tree, start);
- if (!node) {
+ if (!node)
goto out;
- }
- while(1) {
+ while (1) {
state = rb_entry(node, struct extent_state, rb_node);
if (state->end >= start && (state->state & bits)) {
*start_ret = state->start;
spin_unlock(&tree->lock);
return ret;
}
-EXPORT_SYMBOL(find_first_extent_bit);
/* find the first state struct with 'bits' set after 'start', and
* return it. tree->lock must be held. NULL will returned if
* our range starts.
*/
node = tree_search(tree, start);
- if (!node) {
+ if (!node)
goto out;
- }
- while(1) {
+ while (1) {
state = rb_entry(node, struct extent_state, rb_node);
- if (state->end >= start && (state->state & bits)) {
+ if (state->end >= start && (state->state & bits))
return state;
- }
+
node = rb_next(node);
if (!node)
break;
out:
return NULL;
}
-EXPORT_SYMBOL(find_first_extent_bit_state);
/*
* find a contiguous range of bytes in the file marked as delalloc, not
goto out;
}
- while(1) {
+ while (1) {
state = rb_entry(node, struct extent_state, rb_node);
if (found && (state->start != cur_start ||
(state->state & EXTENT_BOUNDARY))) {
if (index == locked_page->index && end_index == index)
return 0;
- while(nr_pages > 0) {
+ while (nr_pages > 0) {
ret = find_get_pages_contig(inode->i_mapping, index,
min_t(unsigned long, nr_pages,
ARRAY_SIZE(pages)), pages);
/* skip the page at the start index */
nrpages = end_index - index + 1;
- while(nrpages > 0) {
+ while (nrpages > 0) {
ret = find_get_pages_contig(inode->i_mapping, index,
min_t(unsigned long,
nrpages, ARRAY_SIZE(pages)), pages);
u64 delalloc_start;
u64 delalloc_end;
u64 found;
+ struct extent_state *cached_state = NULL;
int ret;
int loops = 0;
* pages in order, so we can't process delalloc bytes before
* locked_page
*/
- if (delalloc_start < *start) {
+ if (delalloc_start < *start)
delalloc_start = *start;
- }
/*
* make sure to limit the number of pages we try to lock down
* if we're looping.
*/
- if (delalloc_end + 1 - delalloc_start > max_bytes && loops) {
+ if (delalloc_end + 1 - delalloc_start > max_bytes && loops)
delalloc_end = delalloc_start + PAGE_CACHE_SIZE - 1;
- }
+
/* step two, lock all the pages after the page that has start */
ret = lock_delalloc_pages(inode, locked_page,
delalloc_start, delalloc_end);
/* 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))
+ 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 |
+ EXTENT_SET_PRIVATE2)))
return 0;
- while(nr_pages > 0) {
+ while (nr_pages > 0) {
ret = find_get_pages_contig(inode->i_mapping, index,
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]);
}
}
return 0;
}
-EXPORT_SYMBOL(extent_clear_unlock_delalloc);
/*
* count the number of bytes in the tree that have a given bit(s)
int found = 0;
if (search_end <= cur_start) {
- printk("search_end %Lu start %Lu\n", search_end, cur_start);
WARN_ON(1);
return 0;
}
* our range starts.
*/
node = tree_search(tree, cur_start);
- if (!node) {
+ if (!node)
goto out;
- }
- while(1) {
+ while (1) {
state = rb_entry(node, struct extent_state, rb_node);
if (state->start > search_end)
break;
return total_bytes;
}
-#if 0
-/*
- * helper function to lock both pages and extents in the tree.
- * pages must be locked first.
- */
-static int lock_range(struct extent_io_tree *tree, u64 start, u64 end)
-{
- unsigned long index = start >> PAGE_CACHE_SHIFT;
- unsigned long end_index = end >> PAGE_CACHE_SHIFT;
- struct page *page;
- int err;
-
- while (index <= end_index) {
- page = grab_cache_page(tree->mapping, index);
- if (!page) {
- err = -ENOMEM;
- goto failed;
- }
- if (IS_ERR(page)) {
- err = PTR_ERR(page);
- goto failed;
- }
- index++;
- }
- lock_extent(tree, start, end, GFP_NOFS);
- return 0;
-
-failed:
- /*
- * we failed above in getting the page at 'index', so we undo here
- * up to but not including the page at 'index'
- */
- end_index = index;
- index = start >> PAGE_CACHE_SHIFT;
- while (index < end_index) {
- page = find_get_page(tree->mapping, index);
- unlock_page(page);
- page_cache_release(page);
- index++;
- }
- return err;
-}
-
-/*
- * helper function to unlock both pages and extents in the tree.
- */
-static int unlock_range(struct extent_io_tree *tree, u64 start, u64 end)
-{
- unsigned long index = start >> PAGE_CACHE_SHIFT;
- unsigned long end_index = end >> PAGE_CACHE_SHIFT;
- struct page *page;
-
- while (index <= end_index) {
- page = find_get_page(tree->mapping, index);
- unlock_page(page);
- page_cache_release(page);
- index++;
- }
- unlock_extent(tree, start, end, GFP_NOFS);
- return 0;
-}
-#endif
-
/*
* set the private field for a given byte offset in the tree. If there isn't
* an extent_state there already, this does nothing.
* 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;
spin_unlock(&tree->lock);
return bitset;
}
-EXPORT_SYMBOL(test_range_bit);
/*
* helper function to set a given page up to date if all the
{
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, GFP_NOFS);
ClearPageUptodate(page);
SetPageError(page);
}
- clear_extent_writeback(tree, start, end, GFP_ATOMIC);
-
if (whole_page)
end_page_writeback(page);
else
nr = bio_get_nr_vecs(bdev);
bio = extent_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH);
- if (!bio) {
- printk("failed to allocate bio nr %d\n", nr);
- }
bio_add_page(bio, page, page_size, offset);
bio->bi_end_io = end_io_func;
bio->bi_private = tree;
- if (bio_ret) {
+ if (bio_ret)
*bio_ret = bio;
- } else {
+ else
ret = submit_one_bio(rw, bio, mirror_num, bio_flags);
- }
return ret;
}
set_page_private(page, EXTENT_PAGE_PRIVATE);
}
}
-EXPORT_SYMBOL(set_page_extent_mapped);
static void set_page_extent_head(struct page *page, unsigned long len)
{
break;
}
extent_offset = cur - em->start;
- if (extent_map_end(em) <= cur) {
-printk("bad mapping em [%Lu %Lu] cur %Lu\n", em->start, extent_map_end(em), cur);
- }
BUG_ON(extent_map_end(em) <= cur);
- if (end < cur) {
-printk("2bad mapping end %Lu cur %Lu\n", end, cur);
- }
BUG_ON(end < cur);
if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
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;
submit_one_bio(READ, bio, 0, bio_flags);
return ret;
}
-EXPORT_SYMBOL(extent_read_full_page);
+
+static noinline void update_nr_written(struct page *page,
+ struct writeback_control *wbc,
+ unsigned long nr_written)
+{
+ wbc->nr_to_write -= nr_written;
+ if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
+ wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
+ page->mapping->writeback_index = page->index + nr_written;
+}
/*
* the writepage semantics are similar to regular writepage. extent
u64 iosize;
u64 unlock_start;
sector_t sector;
+ struct extent_state *cached_state = NULL;
struct extent_map *em;
struct block_device *bdev;
int ret;
u64 delalloc_end;
int page_started;
int compressed;
+ int write_flags;
unsigned long nr_written = 0;
+ if (wbc->sync_mode == WB_SYNC_ALL)
+ write_flags = WRITE_SYNC_PLUG;
+ else
+ write_flags = WRITE;
+
WARN_ON(!PageLocked(page));
pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
if (page->index > end_index ||
delalloc_end = 0;
page_started = 0;
if (!epd->extent_locked) {
- while(delalloc_end < page_end) {
+ u64 delalloc_to_write = 0;
+ /*
+ * make sure the wbc mapping index is at least updated
+ * to this page.
+ */
+ update_nr_written(page, wbc, 0);
+
+ while (delalloc_end < page_end) {
nr_delalloc = find_lock_delalloc_range(inode, tree,
page,
&delalloc_start,
tree->ops->fill_delalloc(inode, page, delalloc_start,
delalloc_end, &page_started,
&nr_written);
+ /*
+ * 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?
*/
if (page_started) {
ret = 0;
- goto update_nr_written;
+ /*
+ * we've unlocked the page, so we can't update
+ * the mapping's writeback index, just update
+ * nr_to_write.
+ */
+ wbc->nr_to_write -= nr_written;
+ 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);
ret = 0;
- goto update_nr_written;
+ goto done_unlocked;
}
}
- nr_written++;
+ /*
+ * we don't want to touch the inode after unlocking the page,
+ * so we update the mapping writeback index now
+ */
+ update_nr_written(page, wbc, nr_written + 1);
end = page_end;
- if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0)) {
- printk("found 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_range_writeback(tree, cur, cur + iosize - 1);
if (!PageWriteback(page)) {
- printk("warning page %lu not writeback, "
- "cur %llu end %llu\n", page->index,
- (unsigned long long)cur,
+ printk(KERN_ERR "btrfs warning page %lu not "
+ "writeback, cur %llu end %llu\n",
+ page->index, (unsigned long long)cur,
(unsigned long long)end);
}
- ret = submit_extent_page(WRITE, tree, page, sector,
- iosize, pg_offset, bdev,
- &epd->bio, max_nr,
+ ret = submit_extent_page(write_flags, tree, page,
+ sector, iosize, pg_offset,
+ bdev, &epd->bio, max_nr,
end_bio_extent_writepage,
0, 0, 0);
if (ret)
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);
-update_nr_written:
- wbc->nr_to_write -= nr_written;
- if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
- wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
- page->mapping->writeback_index = page->index + nr_written;
+done_unlocked:
+
+ /* drop our reference on any cached states */
+ free_extent_state(cached_state);
return 0;
}
writepage_t writepage, void *data,
void (*flush_fn)(void *))
{
- struct backing_dev_info *bdi = mapping->backing_dev_info;
int ret = 0;
int done = 0;
+ int nr_to_write_done = 0;
struct pagevec pvec;
int nr_pages;
pgoff_t index;
int scanned = 0;
int range_whole = 0;
- if (wbc->nonblocking && bdi_write_congested(bdi)) {
- wbc->encountered_congestion = 1;
- return 0;
- }
-
pagevec_init(&pvec, 0);
if (wbc->range_cyclic) {
index = mapping->writeback_index; /* Start from prev offset */
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))) {
+ PAGECACHE_TAG_DIRTY, min(end - index,
+ (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
unsigned i;
scanned = 1;
unlock_page(page);
ret = 0;
}
- if (ret || wbc->nr_to_write <= 0)
- done = 1;
- if (wbc->nonblocking && bdi_write_congested(bdi)) {
- wbc->encountered_congestion = 1;
+ 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();
return ret;
}
-static noinline void flush_write_bio(void *data)
+static void flush_epd_write_bio(struct extent_page_data *epd)
{
- struct extent_page_data *epd = data;
if (epd->bio) {
- submit_one_bio(WRITE, epd->bio, 0, 0);
+ if (epd->sync_io)
+ submit_one_bio(WRITE_SYNC, epd->bio, 0, 0);
+ else
+ submit_one_bio(WRITE, epd->bio, 0, 0);
epd->bio = NULL;
}
}
+static noinline void flush_write_bio(void *data)
+{
+ struct extent_page_data *epd = data;
+ flush_epd_write_bio(epd);
+}
+
int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
get_extent_t *get_extent,
struct writeback_control *wbc)
.tree = tree,
.get_extent = get_extent,
.extent_locked = 0,
+ .sync_io = wbc->sync_mode == WB_SYNC_ALL,
};
struct writeback_control wbc_writepages = {
.bdi = wbc->bdi,
- .sync_mode = WB_SYNC_NONE,
+ .sync_mode = wbc->sync_mode,
.older_than_this = NULL,
.nr_to_write = 64,
.range_start = page_offset(page) + PAGE_CACHE_SIZE,
.range_end = (loff_t)-1,
};
-
ret = __extent_writepage(page, wbc, &epd);
extent_write_cache_pages(tree, mapping, &wbc_writepages,
__extent_writepage, &epd, flush_write_bio);
- if (epd.bio) {
- submit_one_bio(WRITE, epd.bio, 0, 0);
- }
+ flush_epd_write_bio(&epd);
return ret;
}
-EXPORT_SYMBOL(extent_write_full_page);
int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
u64 start, u64 end, get_extent_t *get_extent,
.tree = tree,
.get_extent = get_extent,
.extent_locked = 1,
+ .sync_io = mode == WB_SYNC_ALL,
};
struct writeback_control wbc_writepages = {
.bdi = inode->i_mapping->backing_dev_info,
.range_end = end + 1,
};
- while(start <= end) {
+ while (start <= end) {
page = find_get_page(mapping, start >> PAGE_CACHE_SHIFT);
if (clear_page_dirty_for_io(page))
ret = __extent_writepage(page, &wbc_writepages, &epd);
start += PAGE_CACHE_SIZE;
}
- if (epd.bio)
- submit_one_bio(WRITE, epd.bio, 0, 0);
+ flush_epd_write_bio(&epd);
return ret;
}
-EXPORT_SYMBOL(extent_write_locked_range);
-
int extent_writepages(struct extent_io_tree *tree,
struct address_space *mapping,
.tree = tree,
.get_extent = get_extent,
.extent_locked = 0,
+ .sync_io = wbc->sync_mode == WB_SYNC_ALL,
};
ret = extent_write_cache_pages(tree, mapping, wbc,
__extent_writepage, &epd,
flush_write_bio);
- if (epd.bio) {
- submit_one_bio(WRITE, epd.bio, 0, 0);
- }
+ flush_epd_write_bio(&epd);
return ret;
}
-EXPORT_SYMBOL(extent_writepages);
int extent_readpages(struct extent_io_tree *tree,
struct address_space *mapping,
submit_one_bio(READ, bio, 0, bio_flags);
return 0;
}
-EXPORT_SYMBOL(extent_readpages);
/*
* basic invalidatepage code, this waits on any locked or writeback
u64 end = start + PAGE_CACHE_SIZE - 1;
size_t blocksize = page->mapping->host->i_sb->s_blocksize;
- start += (offset + blocksize -1) & ~(blocksize - 1);
+ start += (offset + blocksize - 1) & ~(blocksize - 1);
if (start > end)
return 0;
lock_extent(tree, start, end, GFP_NOFS);
- wait_on_extent_writeback(tree, start, end);
+ 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, NULL, GFP_NOFS);
return 0;
}
-EXPORT_SYMBOL(extent_invalidatepage);
/*
* simple commit_write call, set_range_dirty is used to mark both
}
return 0;
}
-EXPORT_SYMBOL(extent_commit_write);
int extent_prepare_write(struct extent_io_tree *tree,
struct inode *inode, struct page *page,
orig_block_start = block_start;
lock_extent(tree, page_start, page_end, GFP_NOFS);
- while(block_start <= block_end) {
+ while (block_start <= block_end) {
em = get_extent(inode, page, page_offset, block_start,
block_end - block_start + 1, 1);
- if (IS_ERR(em) || !em) {
+ if (IS_ERR(em) || !em)
goto err;
- }
+
cur_end = min(block_end, extent_map_end(em) - 1);
block_off_start = block_start & (PAGE_CACHE_SIZE - 1);
block_off_end = block_off_start + blocksize;
!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,
/* FIXME, zero out newly allocated blocks on error */
return err;
}
-EXPORT_SYMBOL(extent_prepare_write);
/*
* a helper for releasepage, this tests for areas of the page that
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;
}
-EXPORT_SYMBOL(try_release_extent_state);
/*
* a helper for releasepage. As long as there are no locked extents
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);
}
return try_release_extent_state(map, tree, page, mask);
}
-EXPORT_SYMBOL(try_release_extent_mapping);
sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
get_extent_t *get_extent)
return sector;
}
+int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+ __u64 start, __u64 len, get_extent_t *get_extent)
+{
+ int ret;
+ u64 off = start;
+ u64 max = start + len;
+ u32 flags = 0;
+ u64 disko = 0;
+ struct extent_map *em = NULL;
+ int end = 0;
+ u64 em_start = 0, em_len = 0;
+ unsigned long emflags;
+ ret = 0;
+
+ if (len == 0)
+ return -EINVAL;
+
+ lock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
+ GFP_NOFS);
+ em = get_extent(inode, NULL, 0, off, max - off, 0);
+ if (!em)
+ goto out;
+ if (IS_ERR(em)) {
+ ret = PTR_ERR(em);
+ goto out;
+ }
+ while (!end) {
+ off = em->start + em->len;
+ if (off >= max)
+ end = 1;
+
+ em_start = em->start;
+ em_len = em->len;
+
+ disko = 0;
+ flags = 0;
+
+ if (em->block_start == EXTENT_MAP_LAST_BYTE) {
+ end = 1;
+ flags |= FIEMAP_EXTENT_LAST;
+ } else if (em->block_start == EXTENT_MAP_HOLE) {
+ flags |= FIEMAP_EXTENT_UNWRITTEN;
+ } else if (em->block_start == EXTENT_MAP_INLINE) {
+ flags |= (FIEMAP_EXTENT_DATA_INLINE |
+ FIEMAP_EXTENT_NOT_ALIGNED);
+ } else if (em->block_start == EXTENT_MAP_DELALLOC) {
+ flags |= (FIEMAP_EXTENT_DELALLOC |
+ FIEMAP_EXTENT_UNKNOWN);
+ } else {
+ disko = em->block_start;
+ }
+ if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
+ flags |= FIEMAP_EXTENT_ENCODED;
+
+ emflags = em->flags;
+ free_extent_map(em);
+ em = NULL;
+
+ if (!end) {
+ em = get_extent(inode, NULL, 0, off, max - off, 0);
+ if (!em)
+ goto out;
+ if (IS_ERR(em)) {
+ ret = PTR_ERR(em);
+ goto out;
+ }
+ emflags = em->flags;
+ }
+ if (test_bit(EXTENT_FLAG_VACANCY, &emflags)) {
+ flags |= FIEMAP_EXTENT_LAST;
+ end = 1;
+ }
+
+ ret = fiemap_fill_next_extent(fieinfo, em_start, disko,
+ em_len, flags);
+ if (ret)
+ goto out_free;
+ }
+out_free:
+ free_extent_map(em);
+out:
+ unlock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
+ GFP_NOFS);
+ return ret;
+}
+
static inline struct page *extent_buffer_page(struct extent_buffer *eb,
unsigned long i)
{
gfp_t mask)
{
struct extent_buffer *eb = NULL;
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
unsigned long flags;
#endif
eb = kmem_cache_zalloc(extent_buffer_cache, mask);
eb->start = start;
eb->len = len;
- mutex_init(&eb->mutex);
-#ifdef LEAK_DEBUG
+ spin_lock_init(&eb->lock);
+ init_waitqueue_head(&eb->lock_wq);
+
+#if LEAK_DEBUG
spin_lock_irqsave(&leak_lock, flags);
list_add(&eb->leak_list, &buffers);
spin_unlock_irqrestore(&leak_lock, flags);
static void __free_extent_buffer(struct extent_buffer *eb)
{
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
unsigned long flags;
spin_lock_irqsave(&leak_lock, flags);
list_del(&eb->leak_list);
unlock_page(p);
}
if (uptodate)
- eb->flags |= EXTENT_UPTODATE;
- eb->flags |= EXTENT_BUFFER_FILLED;
+ set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
spin_lock(&tree->buffer_lock);
exists = buffer_tree_insert(tree, start, &eb->rb_node);
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:
__free_extent_buffer(eb);
return exists;
}
-EXPORT_SYMBOL(alloc_extent_buffer);
struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
u64 start, unsigned long len,
return eb;
}
-EXPORT_SYMBOL(find_extent_buffer);
void free_extent_buffer(struct extent_buffer *eb)
{
WARN_ON(1);
}
-EXPORT_SYMBOL(free_extent_buffer);
int clear_extent_buffer_dirty(struct extent_io_tree *tree,
struct extent_buffer *eb)
{
- int set;
unsigned long i;
unsigned long num_pages;
struct page *page;
- u64 start = eb->start;
- u64 end = start + eb->len - 1;
-
- set = clear_extent_dirty(tree, start, end, GFP_NOFS);
num_pages = num_extent_pages(eb->start, eb->len);
for (i = 0; i < num_pages; i++) {
page = extent_buffer_page(eb, i);
- if (!set && !PageDirty(page))
+ if (!PageDirty(page))
continue;
lock_page(page);
else
set_page_private(page, EXTENT_PAGE_PRIVATE);
- /*
- * if we're on the last page or the first page and the
- * block isn't aligned on a page boundary, do extra checks
- * to make sure we don't clean page that is partially dirty
- */
- if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
- ((i == num_pages - 1) &&
- ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) {
- start = (u64)page->index << PAGE_CACHE_SHIFT;
- end = start + PAGE_CACHE_SIZE - 1;
- if (test_range_bit(tree, start, end,
- EXTENT_DIRTY, 0)) {
- unlock_page(page);
- continue;
- }
- }
clear_page_dirty_for_io(page);
spin_lock_irq(&page->mapping->tree_lock);
if (!PageDirty(page)) {
}
return 0;
}
-EXPORT_SYMBOL(clear_extent_buffer_dirty);
int wait_on_extent_buffer_writeback(struct extent_io_tree *tree,
struct extent_buffer *eb)
return wait_on_extent_writeback(tree, eb->start,
eb->start + eb->len - 1);
}
-EXPORT_SYMBOL(wait_on_extent_buffer_writeback);
int set_extent_buffer_dirty(struct extent_io_tree *tree,
struct extent_buffer *eb)
{
unsigned long i;
unsigned long num_pages;
+ int was_dirty = 0;
+ was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags);
num_pages = num_extent_pages(eb->start, eb->len);
- for (i = 0; i < num_pages; i++) {
- struct page *page = extent_buffer_page(eb, i);
- /* writepage may need to do something special for the
- * first page, we have to make sure page->private is
- * properly set. releasepage may drop page->private
- * on us if the page isn't already dirty.
- */
- lock_page(page);
- if (i == 0) {
- set_page_extent_head(page, eb->len);
- } else if (PagePrivate(page) &&
- page->private != EXTENT_PAGE_PRIVATE) {
- set_page_extent_mapped(page);
- }
+ for (i = 0; i < num_pages; i++)
__set_page_dirty_nobuffers(extent_buffer_page(eb, i));
- set_extent_dirty(tree, page_offset(page),
- page_offset(page) + PAGE_CACHE_SIZE -1,
- GFP_NOFS);
- unlock_page(page);
- }
- return 0;
+ return was_dirty;
}
-EXPORT_SYMBOL(set_extent_buffer_dirty);
int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
struct extent_buffer *eb)
unsigned long num_pages;
num_pages = num_extent_pages(eb->start, eb->len);
- eb->flags &= ~EXTENT_UPTODATE;
+ clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
GFP_NOFS);
}
return 0;
}
-EXPORT_SYMBOL(set_extent_buffer_uptodate);
int extent_range_uptodate(struct extent_io_tree *tree,
u64 start, u64 end)
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) {
+ while (start <= end) {
index = start >> PAGE_CACHE_SHIFT;
page = find_get_page(tree->mapping, index);
uptodate = PageUptodate(page);
struct page *page;
int pg_uptodate = 1;
- if (eb->flags & EXTENT_UPTODATE)
+ if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
return 1;
ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
- EXTENT_UPTODATE, 1);
+ EXTENT_UPTODATE, 1, NULL);
if (ret)
return ret;
}
return pg_uptodate;
}
-EXPORT_SYMBOL(extent_buffer_uptodate);
int read_extent_buffer_pages(struct extent_io_tree *tree,
struct extent_buffer *eb,
struct bio *bio = NULL;
unsigned long bio_flags = 0;
- if (eb->flags & EXTENT_UPTODATE)
+ if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
return 0;
if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
- EXTENT_UPTODATE, 1)) {
+ EXTENT_UPTODATE, 1, NULL)) {
return 0;
}
lock_page(page);
}
locked_pages++;
- if (!PageUptodate(page)) {
+ if (!PageUptodate(page))
all_uptodate = 0;
- }
}
if (all_uptodate) {
if (start_i == 0)
- eb->flags |= EXTENT_UPTODATE;
- if (ret) {
- printk("all up to date but ret is %d\n", ret);
- }
+ set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
goto unlock_exit;
}
err = __extent_read_full_page(tree, page,
get_extent, &bio,
mirror_num, &bio_flags);
- if (err) {
+ if (err)
ret = err;
- printk("err %d from __extent_read_full_page\n", ret);
- }
} else {
unlock_page(page);
}
if (bio)
submit_one_bio(READ, bio, mirror_num, bio_flags);
- if (ret || !wait) {
- if (ret)
- printk("ret %d wait %d returning\n", ret, wait);
+ if (ret || !wait)
return ret;
- }
+
for (i = start_i; i < num_pages; i++) {
page = extent_buffer_page(eb, i);
wait_on_page_locked(page);
- if (!PageUptodate(page)) {
- printk("page not uptodate after wait_on_page_locked\n");
+ if (!PageUptodate(page))
ret = -EIO;
- }
}
+
if (!ret)
- eb->flags |= EXTENT_UPTODATE;
+ set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
return ret;
unlock_exit:
i = start_i;
- while(locked_pages > 0) {
+ while (locked_pages > 0) {
page = extent_buffer_page(eb, i);
i++;
unlock_page(page);
}
return ret;
}
-EXPORT_SYMBOL(read_extent_buffer_pages);
void read_extent_buffer(struct extent_buffer *eb, void *dstv,
unsigned long start,
offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
- while(len > 0) {
+ while (len > 0) {
page = extent_buffer_page(eb, i);
cur = min(len, (PAGE_CACHE_SIZE - offset));
i++;
}
}
-EXPORT_SYMBOL(read_extent_buffer);
int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
unsigned long min_len, char **token, char **map,
offset = 0;
*map_start = ((u64)i << PAGE_CACHE_SHIFT) - start_offset;
}
+
if (start + min_len > eb->len) {
-printk("bad mapping eb start %Lu len %lu, wanted %lu %lu\n", eb->start, eb->len, start, min_len);
+ printk(KERN_ERR "btrfs bad mapping eb start %llu len %lu, "
+ "wanted %lu %lu\n", (unsigned long long)eb->start,
+ eb->len, start, min_len);
WARN_ON(1);
}
*map_len = PAGE_CACHE_SIZE - offset;
return 0;
}
-EXPORT_SYMBOL(map_private_extent_buffer);
int map_extent_buffer(struct extent_buffer *eb, unsigned long start,
unsigned long min_len,
unmap_extent_buffer(eb, eb->map_token, km);
eb->map_token = NULL;
save = 1;
- WARN_ON(!mutex_is_locked(&eb->mutex));
}
err = map_private_extent_buffer(eb, start, min_len, token, map,
map_start, map_len, km);
}
return err;
}
-EXPORT_SYMBOL(map_extent_buffer);
void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km)
{
kunmap_atomic(token, km);
}
-EXPORT_SYMBOL(unmap_extent_buffer);
int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
unsigned long start,
offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
- while(len > 0) {
+ while (len > 0) {
page = extent_buffer_page(eb, i);
cur = min(len, (PAGE_CACHE_SIZE - offset));
}
return ret;
}
-EXPORT_SYMBOL(memcmp_extent_buffer);
void write_extent_buffer(struct extent_buffer *eb, const void *srcv,
unsigned long start, unsigned long len)
offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
- while(len > 0) {
+ while (len > 0) {
page = extent_buffer_page(eb, i);
WARN_ON(!PageUptodate(page));
i++;
}
}
-EXPORT_SYMBOL(write_extent_buffer);
void memset_extent_buffer(struct extent_buffer *eb, char c,
unsigned long start, unsigned long len)
offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
- while(len > 0) {
+ while (len > 0) {
page = extent_buffer_page(eb, i);
WARN_ON(!PageUptodate(page));
i++;
}
}
-EXPORT_SYMBOL(memset_extent_buffer);
void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
unsigned long dst_offset, unsigned long src_offset,
offset = (start_offset + dst_offset) &
((unsigned long)PAGE_CACHE_SIZE - 1);
- while(len > 0) {
+ while (len > 0) {
page = extent_buffer_page(dst, i);
WARN_ON(!PageUptodate(page));
i++;
}
}
-EXPORT_SYMBOL(copy_extent_buffer);
static void move_pages(struct page *dst_page, struct page *src_page,
unsigned long dst_off, unsigned long src_off,
unsigned long src_i;
if (src_offset + len > dst->len) {
- printk("memmove bogus src_offset %lu move len %lu len %lu\n",
- src_offset, len, dst->len);
+ printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
+ "len %lu dst len %lu\n", src_offset, len, dst->len);
BUG_ON(1);
}
if (dst_offset + len > dst->len) {
- printk("memmove bogus dst_offset %lu move len %lu len %lu\n",
- dst_offset, len, dst->len);
+ printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
+ "len %lu dst len %lu\n", dst_offset, len, dst->len);
BUG_ON(1);
}
- while(len > 0) {
+ while (len > 0) {
dst_off_in_page = (start_offset + dst_offset) &
((unsigned long)PAGE_CACHE_SIZE - 1);
src_off_in_page = (start_offset + src_offset) &
len -= cur;
}
}
-EXPORT_SYMBOL(memcpy_extent_buffer);
void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
unsigned long src_offset, unsigned long len)
unsigned long src_i;
if (src_offset + len > dst->len) {
- printk("memmove bogus src_offset %lu move len %lu len %lu\n",
- src_offset, len, dst->len);
+ printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
+ "len %lu len %lu\n", src_offset, len, dst->len);
BUG_ON(1);
}
if (dst_offset + len > dst->len) {
- printk("memmove bogus dst_offset %lu move len %lu len %lu\n",
- dst_offset, len, dst->len);
+ printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
+ "len %lu len %lu\n", dst_offset, len, dst->len);
BUG_ON(1);
}
if (dst_offset < src_offset) {
memcpy_extent_buffer(dst, dst_offset, src_offset, len);
return;
}
- while(len > 0) {
+ while (len > 0) {
dst_i = (start_offset + dst_end) >> PAGE_CACHE_SHIFT;
src_i = (start_offset + src_end) >> PAGE_CACHE_SHIFT;
len -= cur;
}
}
-EXPORT_SYMBOL(memmove_extent_buffer);
int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page)
{
ret = 0;
goto out;
}
+ if (test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
+ ret = 0;
+ goto out;
+ }
/* at this point we can safely release the extent buffer */
num_pages = num_extent_pages(eb->start, eb->len);
for (i = 0; i < num_pages; i++)
spin_unlock(&tree->buffer_lock);
return ret;
}
-EXPORT_SYMBOL(try_release_extent_buffer);