diff options
Diffstat (limited to 'fs/btrfs/extent_io.c')
| -rw-r--r-- | fs/btrfs/extent_io.c | 2039 |
1 files changed, 1218 insertions, 821 deletions
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 1b319df29ee..a389820d158 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4,7 +4,6 @@ #include <linux/mm.h> #include <linux/pagemap.h> #include <linux/page-flags.h> -#include <linux/module.h> #include <linux/spinlock.h> #include <linux/blkdev.h> #include <linux/swap.h> @@ -14,23 +13,94 @@ #include <linux/cleancache.h> #include "extent_io.h" #include "extent_map.h" -#include "compat.h" #include "ctree.h" #include "btrfs_inode.h" #include "volumes.h" #include "check-integrity.h" #include "locking.h" #include "rcu-string.h" +#include "backref.h" static struct kmem_cache *extent_state_cache; static struct kmem_cache *extent_buffer_cache; +static struct bio_set *btrfs_bioset; +#ifdef CONFIG_BTRFS_DEBUG static LIST_HEAD(buffers); static LIST_HEAD(states); -#define LEAK_DEBUG 0 -#if LEAK_DEBUG static DEFINE_SPINLOCK(leak_lock); + +static inline +void btrfs_leak_debug_add(struct list_head *new, struct list_head *head) +{ + unsigned long flags; + + spin_lock_irqsave(&leak_lock, flags); + list_add(new, head); + spin_unlock_irqrestore(&leak_lock, flags); +} + +static inline +void btrfs_leak_debug_del(struct list_head *entry) +{ + unsigned long flags; + + spin_lock_irqsave(&leak_lock, flags); + list_del(entry); + spin_unlock_irqrestore(&leak_lock, flags); +} + +static inline +void btrfs_leak_debug_check(void) +{ + struct extent_state *state; + struct extent_buffer *eb; + + while (!list_empty(&states)) { + state = list_entry(states.next, struct extent_state, leak_list); + printk(KERN_ERR "BTRFS: state leak: start %llu end %llu " + "state %lu in tree %p refs %d\n", + state->start, 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(KERN_ERR "BTRFS: buffer leak start %llu len %lu " + "refs %d\n", + eb->start, eb->len, atomic_read(&eb->refs)); + list_del(&eb->leak_list); + kmem_cache_free(extent_buffer_cache, eb); + } +} + +#define btrfs_debug_check_extent_io_range(tree, start, end) \ + __btrfs_debug_check_extent_io_range(__func__, (tree), (start), (end)) +static inline void __btrfs_debug_check_extent_io_range(const char *caller, + struct extent_io_tree *tree, u64 start, u64 end) +{ + struct inode *inode; + u64 isize; + + if (!tree->mapping) + return; + + inode = tree->mapping->host; + isize = i_size_read(inode); + if (end >= PAGE_SIZE && (end % 2) == 0 && end != isize - 1) { + printk_ratelimited(KERN_DEBUG + "BTRFS: %s: ino %llu isize %llu odd range [%llu,%llu]\n", + caller, btrfs_ino(inode), isize, start, end); + } +} +#else +#define btrfs_leak_debug_add(new, head) do {} while (0) +#define btrfs_leak_debug_del(entry) do {} while (0) +#define btrfs_leak_debug_check() do {} while (0) +#define btrfs_debug_check_extent_io_range(c, s, e) do {} while (0) #endif #define BUFFER_LRU_MAX 64 @@ -60,6 +130,8 @@ static noinline void flush_write_bio(void *data); static inline struct btrfs_fs_info * tree_fs_info(struct extent_io_tree *tree) { + if (!tree->mapping) + return NULL; return btrfs_sb(tree->mapping->host->i_sb); } @@ -76,38 +148,34 @@ int __init extent_io_init(void) SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); if (!extent_buffer_cache) goto free_state_cache; + + btrfs_bioset = bioset_create(BIO_POOL_SIZE, + offsetof(struct btrfs_io_bio, bio)); + if (!btrfs_bioset) + goto free_buffer_cache; + + if (bioset_integrity_create(btrfs_bioset, BIO_POOL_SIZE)) + goto free_bioset; + return 0; +free_bioset: + bioset_free(btrfs_bioset); + btrfs_bioset = NULL; + +free_buffer_cache: + kmem_cache_destroy(extent_buffer_cache); + extent_buffer_cache = NULL; + free_state_cache: kmem_cache_destroy(extent_state_cache); + extent_state_cache = NULL; return -ENOMEM; } void extent_io_exit(void) { - struct extent_state *state; - struct extent_buffer *eb; - - while (!list_empty(&states)) { - state = list_entry(states.next, struct extent_state, leak_list); - 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(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); - } + btrfs_leak_debug_check(); /* * Make sure all delayed rcu free are flushed before we @@ -118,26 +186,23 @@ void extent_io_exit(void) kmem_cache_destroy(extent_state_cache); if (extent_buffer_cache) kmem_cache_destroy(extent_buffer_cache); + if (btrfs_bioset) + bioset_free(btrfs_bioset); } void extent_io_tree_init(struct extent_io_tree *tree, struct address_space *mapping) { tree->state = RB_ROOT; - INIT_RADIX_TREE(&tree->buffer, GFP_ATOMIC); tree->ops = NULL; tree->dirty_bytes = 0; spin_lock_init(&tree->lock); - spin_lock_init(&tree->buffer_lock); tree->mapping = mapping; } static struct extent_state *alloc_extent_state(gfp_t mask) { struct extent_state *state; -#if LEAK_DEBUG - unsigned long flags; -#endif state = kmem_cache_alloc(extent_state_cache, mask); if (!state) @@ -145,11 +210,7 @@ static struct extent_state *alloc_extent_state(gfp_t mask) state->state = 0; state->private = 0; state->tree = NULL; -#if LEAK_DEBUG - spin_lock_irqsave(&leak_lock, flags); - list_add(&state->leak_list, &states); - spin_unlock_irqrestore(&leak_lock, flags); -#endif + btrfs_leak_debug_add(&state->leak_list, &states); atomic_set(&state->refs, 1); init_waitqueue_head(&state->wq); trace_alloc_extent_state(state, mask, _RET_IP_); @@ -161,27 +222,31 @@ void free_extent_state(struct extent_state *state) if (!state) return; if (atomic_dec_and_test(&state->refs)) { -#if LEAK_DEBUG - unsigned long flags; -#endif WARN_ON(state->tree); -#if LEAK_DEBUG - spin_lock_irqsave(&leak_lock, flags); - list_del(&state->leak_list); - spin_unlock_irqrestore(&leak_lock, flags); -#endif + btrfs_leak_debug_del(&state->leak_list); trace_free_extent_state(state, _RET_IP_); kmem_cache_free(extent_state_cache, state); } } -static struct rb_node *tree_insert(struct rb_root *root, u64 offset, - struct rb_node *node) +static struct rb_node *tree_insert(struct rb_root *root, + struct rb_node *search_start, + u64 offset, + struct rb_node *node, + struct rb_node ***p_in, + struct rb_node **parent_in) { - struct rb_node **p = &root->rb_node; + struct rb_node **p; struct rb_node *parent = NULL; struct tree_entry *entry; + if (p_in && parent_in) { + p = *p_in; + parent = *parent_in; + goto do_insert; + } + + p = search_start ? &search_start : &root->rb_node; while (*p) { parent = *p; entry = rb_entry(parent, struct tree_entry, rb_node); @@ -194,35 +259,43 @@ static struct rb_node *tree_insert(struct rb_root *root, u64 offset, return parent; } +do_insert: rb_link_node(node, parent, p); rb_insert_color(node, root); return NULL; } static struct rb_node *__etree_search(struct extent_io_tree *tree, u64 offset, - struct rb_node **prev_ret, - struct rb_node **next_ret) + struct rb_node **prev_ret, + struct rb_node **next_ret, + struct rb_node ***p_ret, + struct rb_node **parent_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) { - entry = rb_entry(n, struct tree_entry, rb_node); - prev = n; + while (*n) { + prev = *n; + entry = rb_entry(prev, struct tree_entry, rb_node); prev_entry = entry; if (offset < entry->start) - n = n->rb_left; + n = &(*n)->rb_left; else if (offset > entry->end) - n = n->rb_right; + n = &(*n)->rb_right; else - return n; + return *n; } + if (p_ret) + *p_ret = n; + if (parent_ret) + *parent_ret = prev; + if (prev_ret) { orig_prev = prev; while (prev && offset > prev_entry->end) { @@ -244,18 +317,27 @@ static struct rb_node *__etree_search(struct extent_io_tree *tree, u64 offset, return NULL; } -static inline struct rb_node *tree_search(struct extent_io_tree *tree, - u64 offset) +static inline struct rb_node * +tree_search_for_insert(struct extent_io_tree *tree, + u64 offset, + struct rb_node ***p_ret, + struct rb_node **parent_ret) { struct rb_node *prev = NULL; struct rb_node *ret; - ret = __etree_search(tree, offset, &prev, NULL); + ret = __etree_search(tree, offset, &prev, NULL, p_ret, parent_ret); if (!ret) return prev; return ret; } +static inline struct rb_node *tree_search(struct extent_io_tree *tree, + u64 offset) +{ + return tree_search_for_insert(tree, offset, NULL, NULL); +} + static void merge_cb(struct extent_io_tree *tree, struct extent_state *new, struct extent_state *other) { @@ -309,21 +391,21 @@ static void merge_state(struct extent_io_tree *tree, } static void set_state_cb(struct extent_io_tree *tree, - struct extent_state *state, int *bits) + struct extent_state *state, unsigned long *bits) { if (tree->ops && tree->ops->set_bit_hook) tree->ops->set_bit_hook(tree->mapping->host, state, bits); } static void clear_state_cb(struct extent_io_tree *tree, - struct extent_state *state, int *bits) + struct extent_state *state, unsigned long *bits) { if (tree->ops && tree->ops->clear_bit_hook) tree->ops->clear_bit_hook(tree->mapping->host, state, bits); } static void set_state_bits(struct extent_io_tree *tree, - struct extent_state *state, int *bits); + struct extent_state *state, unsigned long *bits); /* * insert an extent_state struct into the tree. 'bits' are set on the @@ -337,27 +419,27 @@ static void set_state_bits(struct extent_io_tree *tree, */ static int insert_state(struct extent_io_tree *tree, struct extent_state *state, u64 start, u64 end, - int *bits) + struct rb_node ***p, + struct rb_node **parent, + unsigned long *bits) { struct rb_node *node; if (end < start) - WARN(1, KERN_ERR "btrfs end < start %llu %llu\n", - (unsigned long long)end, - (unsigned long long)start); + WARN(1, KERN_ERR "BTRFS: end < start %llu %llu\n", + end, start); state->start = start; state->end = end; set_state_bits(tree, state, bits); - node = tree_insert(&tree->state, end, &state->rb_node); + node = tree_insert(&tree->state, NULL, end, &state->rb_node, p, parent); if (node) { struct extent_state *found; found = rb_entry(node, struct extent_state, rb_node); - 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); + printk(KERN_ERR "BTRFS: found node %llu %llu on insert of " + "%llu %llu\n", + found->start, found->end, start, end); return -EEXIST; } state->tree = tree; @@ -398,7 +480,8 @@ static int split_state(struct extent_io_tree *tree, struct extent_state *orig, prealloc->state = orig->state; orig->start = split; - node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node); + node = tree_insert(&tree->state, &orig->rb_node, prealloc->end, + &prealloc->rb_node, NULL, NULL); if (node) { free_extent_state(prealloc); return -EEXIST; @@ -425,10 +508,10 @@ static struct extent_state *next_state(struct extent_state *state) */ static struct extent_state *clear_state_bit(struct extent_io_tree *tree, struct extent_state *state, - int *bits, int wake) + unsigned long *bits, int wake) { struct extent_state *next; - int bits_to_clear = *bits & ~EXTENT_CTLBITS; + unsigned long bits_to_clear = *bits & ~EXTENT_CTLBITS; if ((bits_to_clear & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) { u64 range = state->end - state->start + 1; @@ -464,7 +547,7 @@ alloc_extent_state_atomic(struct extent_state *prealloc) return prealloc; } -void extent_io_tree_panic(struct extent_io_tree *tree, int err) +static void extent_io_tree_panic(struct extent_io_tree *tree, int err) { btrfs_panic(tree_fs_info(tree), err, "Locking error: " "Extent tree was modified by another " @@ -484,7 +567,7 @@ void extent_io_tree_panic(struct extent_io_tree *tree, int err) * This takes the tree lock, and returns 0 on success and < 0 on error. */ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, - int bits, int wake, int delete, + unsigned long bits, int wake, int delete, struct extent_state **cached_state, gfp_t mask) { @@ -496,6 +579,11 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int err; int clear = 0; + btrfs_debug_check_extent_io_range(tree, start, end); + + if (bits & EXTENT_DELALLOC) + bits |= EXTENT_NORESERVE; + if (delete) bits |= ~EXTENT_CTLBITS; bits |= EXTENT_FIRST_DELALLOC; @@ -645,11 +733,14 @@ static void wait_on_state(struct extent_io_tree *tree, * The range [start, end] is inclusive. * The tree lock is taken by this function */ -void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits) +static void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, + unsigned long bits) { struct extent_state *state; struct rb_node *node; + btrfs_debug_check_extent_io_range(tree, start, end); + spin_lock(&tree->lock); again: while (1) { @@ -658,6 +749,7 @@ again: * our range starts */ node = tree_search(tree, start); +process_node: if (!node) break; @@ -678,7 +770,10 @@ again: if (start > end) break; - cond_resched_lock(&tree->lock); + if (!cond_resched_lock(&tree->lock)) { + node = rb_next(node); + goto process_node; + } } out: spin_unlock(&tree->lock); @@ -686,9 +781,9 @@ out: static void set_state_bits(struct extent_io_tree *tree, struct extent_state *state, - int *bits) + unsigned long *bits) { - int bits_to_set = *bits & ~EXTENT_CTLBITS; + unsigned long bits_to_set = *bits & ~EXTENT_CTLBITS; set_state_cb(tree, state, bits); if ((bits_to_set & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) { @@ -709,15 +804,6 @@ static void cache_state(struct extent_state *state, } } -static void uncache_state(struct extent_state **cached_ptr) -{ - if (cached_ptr && (*cached_ptr)) { - struct extent_state *state = *cached_ptr; - *cached_ptr = NULL; - free_extent_state(state); - } -} - /* * 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. @@ -731,16 +817,21 @@ static void uncache_state(struct extent_state **cached_ptr) static int __must_check __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) + unsigned long bits, unsigned long 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; + struct rb_node **p; + struct rb_node *parent; int err = 0; u64 last_start; u64 last_end; + btrfs_debug_check_extent_io_range(tree, start, end); + bits |= EXTENT_FIRST_DELALLOC; again: if (!prealloc && (mask & __GFP_WAIT)) { @@ -761,14 +852,16 @@ again: * this search will find all the extents that end after * our range starts. */ - node = tree_search(tree, start); + node = tree_search_for_insert(tree, start, &p, &parent); if (!node) { prealloc = alloc_extent_state_atomic(prealloc); BUG_ON(!prealloc); - err = insert_state(tree, prealloc, start, end, &bits); + err = insert_state(tree, prealloc, start, end, + &p, &parent, &bits); if (err) extent_io_tree_panic(tree, err); + cache_state(prealloc, cached_state); prealloc = NULL; goto out; } @@ -871,7 +964,7 @@ hit_next: * the later extent. */ err = insert_state(tree, prealloc, start, this_end, - &bits); + NULL, NULL, &bits); if (err) extent_io_tree_panic(tree, err); @@ -924,9 +1017,9 @@ search_again: goto again; } -int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int 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, + unsigned long bits, u64 * failed_start, + struct extent_state **cached_state, gfp_t mask) { return __set_extent_bit(tree, start, end, bits, 0, failed_start, cached_state, mask); @@ -951,16 +1044,20 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits, * boundary bits like LOCK. */ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, - int bits, int clear_bits, + unsigned long bits, unsigned long clear_bits, struct extent_state **cached_state, gfp_t mask) { struct extent_state *state; struct extent_state *prealloc = NULL; struct rb_node *node; + struct rb_node **p; + struct rb_node *parent; int err = 0; u64 last_start; u64 last_end; + btrfs_debug_check_extent_io_range(tree, start, end); + again: if (!prealloc && (mask & __GFP_WAIT)) { prealloc = alloc_extent_state(mask); @@ -982,17 +1079,19 @@ again: * this search will find all the extents that end after * our range starts. */ - node = tree_search(tree, start); + node = tree_search_for_insert(tree, start, &p, &parent); if (!node) { prealloc = alloc_extent_state_atomic(prealloc); if (!prealloc) { err = -ENOMEM; goto out; } - err = insert_state(tree, prealloc, start, end, &bits); - prealloc = NULL; + err = insert_state(tree, prealloc, start, end, + &p, &parent, &bits); if (err) extent_io_tree_panic(tree, err); + cache_state(prealloc, cached_state); + prealloc = NULL; goto out; } state = rb_entry(node, struct extent_state, rb_node); @@ -1085,7 +1184,7 @@ hit_next: * the later extent. */ err = insert_state(tree, prealloc, start, this_end, - &bits); + NULL, NULL, &bits); if (err) extent_io_tree_panic(tree, err); cache_state(prealloc, cached_state); @@ -1144,14 +1243,14 @@ int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, } int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - int bits, gfp_t mask) + unsigned long bits, gfp_t mask) { return set_extent_bit(tree, start, end, bits, NULL, NULL, mask); } int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - int bits, gfp_t mask) + unsigned long bits, gfp_t mask) { return clear_extent_bit(tree, start, end, bits, 0, 0, NULL, mask); } @@ -1190,7 +1289,7 @@ int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end, int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask) { - return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, + return set_extent_bit(tree, start, end, EXTENT_UPTODATE, NULL, cached_state, mask); } @@ -1206,7 +1305,7 @@ int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, * us if waiting is desired. */ int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - int bits, struct extent_state **cached_state) + unsigned long bits, struct extent_state **cached_state) { int err; u64 failed_start; @@ -1258,6 +1357,39 @@ int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end) GFP_NOFS); } +int extent_range_clear_dirty_for_io(struct inode *inode, 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(inode->i_mapping, index); + BUG_ON(!page); /* Pages should be in the extent_io_tree */ + clear_page_dirty_for_io(page); + page_cache_release(page); + index++; + } + return 0; +} + +int extent_range_redirty_for_io(struct inode *inode, 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(inode->i_mapping, index); + BUG_ON(!page); /* Pages should be in the extent_io_tree */ + account_page_redirty(page); + __set_page_dirty_nobuffers(page); + page_cache_release(page); + index++; + } + return 0; +} + /* * helper function to set both pages and extents in the tree writeback */ @@ -1281,8 +1413,9 @@ static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) * return it. tree->lock must be held. NULL will returned if * nothing was found after 'start' */ -struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree, - u64 start, int bits) +static struct extent_state * +find_first_extent_bit_state(struct extent_io_tree *tree, + u64 start, unsigned long bits) { struct rb_node *node; struct extent_state *state; @@ -1316,7 +1449,7 @@ out: * If nothing was found, 1 is returned. If found something, return 0. */ int find_first_extent_bit(struct extent_io_tree *tree, u64 start, - u64 *start_ret, u64 *end_ret, int bits, + u64 *start_ret, u64 *end_ret, unsigned long bits, struct extent_state **cached_state) { struct extent_state *state; @@ -1405,11 +1538,11 @@ static noinline u64 find_delalloc_range(struct extent_io_tree *tree, *end = state->end; cur_start = state->end + 1; node = rb_next(node); - if (!node) - break; total_bytes += state->end - state->start + 1; if (total_bytes >= max_bytes) break; + if (!node) + break; } out: spin_unlock(&tree->lock); @@ -1513,11 +1646,10 @@ done: * * 1 is returned if we find something, 0 if nothing was in the tree */ -static noinline u64 find_lock_delalloc_range(struct inode *inode, - struct extent_io_tree *tree, - struct page *locked_page, - u64 *start, u64 *end, - u64 max_bytes) +STATIC u64 find_lock_delalloc_range(struct inode *inode, + struct extent_io_tree *tree, + struct page *locked_page, u64 *start, + u64 *end, u64 max_bytes) { u64 delalloc_start; u64 delalloc_end; @@ -1536,7 +1668,7 @@ again: *start = delalloc_start; *end = delalloc_end; free_extent_state(cached_state); - return found; + return 0; } /* @@ -1549,10 +1681,9 @@ again: /* * 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) - delalloc_end = delalloc_start + PAGE_CACHE_SIZE - 1; + if (delalloc_end + 1 - delalloc_start > max_bytes) + delalloc_end = delalloc_start + max_bytes - 1; /* step two, lock all the pages after the page that has start */ ret = lock_delalloc_pages(inode, locked_page, @@ -1562,9 +1693,9 @@ again: * shortening the size of the delalloc range we're searching */ free_extent_state(cached_state); + cached_state = NULL; if (!loops) { - unsigned long offset = (*start) & (PAGE_CACHE_SIZE - 1); - max_bytes = PAGE_CACHE_SIZE - offset; + max_bytes = PAGE_CACHE_SIZE; loops = 1; goto again; } else { @@ -1595,31 +1726,21 @@ out_failed: return found; } -int extent_clear_unlock_delalloc(struct inode *inode, - struct extent_io_tree *tree, - u64 start, u64 end, struct page *locked_page, - unsigned long op) +int extent_clear_unlock_delalloc(struct inode *inode, u64 start, u64 end, + struct page *locked_page, + unsigned long clear_bits, + unsigned long page_ops) { + struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; int ret; struct page *pages[16]; unsigned long index = start >> PAGE_CACHE_SHIFT; unsigned long end_index = end >> PAGE_CACHE_SHIFT; unsigned long nr_pages = end_index - index + 1; int i; - int clear_bits = 0; - - if (op & EXTENT_CLEAR_UNLOCK) - clear_bits |= EXTENT_LOCKED; - if (op & EXTENT_CLEAR_DIRTY) - clear_bits |= EXTENT_DIRTY; - - if (op & EXTENT_CLEAR_DELALLOC) - clear_bits |= EXTENT_DELALLOC; 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))) + if (page_ops == 0) return 0; while (nr_pages > 0) { @@ -1628,20 +1749,20 @@ int extent_clear_unlock_delalloc(struct inode *inode, nr_pages, ARRAY_SIZE(pages)), pages); for (i = 0; i < ret; i++) { - if (op & EXTENT_SET_PRIVATE2) + if (page_ops & PAGE_SET_PRIVATE2) SetPagePrivate2(pages[i]); if (pages[i] == locked_page) { page_cache_release(pages[i]); continue; } - if (op & EXTENT_CLEAR_DIRTY) + if (page_ops & PAGE_CLEAR_DIRTY) clear_page_dirty_for_io(pages[i]); - if (op & EXTENT_SET_WRITEBACK) + if (page_ops & PAGE_SET_WRITEBACK) set_page_writeback(pages[i]); - if (op & EXTENT_END_WRITEBACK) + if (page_ops & PAGE_END_WRITEBACK) end_page_writeback(pages[i]); - if (op & EXTENT_CLEAR_UNLOCK_PAGE) + if (page_ops & PAGE_UNLOCK) unlock_page(pages[i]); page_cache_release(pages[i]); } @@ -1668,10 +1789,8 @@ u64 count_range_bits(struct extent_io_tree *tree, u64 last = 0; int found = 0; - if (search_end <= cur_start) { - WARN_ON(1); + if (WARN_ON(search_end <= cur_start)) return 0; - } spin_lock(&tree->lock); if (cur_start == 0 && bits == EXTENT_DIRTY) { @@ -1718,7 +1837,7 @@ out: * set the private field for a given byte offset in the tree. If there isn't * an extent_state there already, this does nothing. */ -int set_state_private(struct extent_io_tree *tree, u64 start, u64 private) +static int set_state_private(struct extent_io_tree *tree, u64 start, u64 private) { struct rb_node *node; struct extent_state *state; @@ -1779,7 +1898,7 @@ out: * range is found set. */ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end, - int bits, int filled, struct extent_state *cached) + unsigned long bits, int filled, struct extent_state *cached) { struct extent_state *state = NULL; struct rb_node *node; @@ -1834,35 +1953,13 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end, */ static void check_page_uptodate(struct extent_io_tree *tree, struct page *page) { - u64 start = (u64)page->index << PAGE_CACHE_SHIFT; + u64 start = page_offset(page); u64 end = start + PAGE_CACHE_SIZE - 1; if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL)) SetPageUptodate(page); } /* - * helper function to unlock a page if all the extents in the tree - * for that page are unlocked - */ -static void check_page_locked(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_LOCKED, 0, NULL)) - unlock_page(page); -} - -/* - * helper function to end page writeback if all the extents - * in the tree for that page are done with writeback - */ -static void check_page_writeback(struct extent_io_tree *tree, - struct page *page) -{ - end_page_writeback(page); -} - -/* * When IO fails, either with EIO or csum verification fails, we * try other mirrors that might have a good copy of the data. This * io_failure_record is used to record state as we go through all the @@ -1895,23 +1992,16 @@ static int free_io_failure(struct inode *inode, struct io_failure_record *rec, if (ret) err = ret; - if (did_repair) { - ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, - rec->start + rec->len - 1, - EXTENT_DAMAGED, GFP_NOFS); - if (ret && !err) - err = ret; - } + ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, + rec->start + rec->len - 1, + EXTENT_DAMAGED, GFP_NOFS); + if (ret && !err) + err = ret; kfree(rec); return err; } -static void repair_io_failure_callback(struct bio *bio, int err) -{ - complete(bio->bi_private); -} - /* * this bypasses the standard btrfs submit functions deliberately, as * the standard behavior is to write all copies in a raid setup. here we only @@ -1928,20 +2018,23 @@ int repair_io_failure(struct btrfs_fs_info *fs_info, u64 start, { struct bio *bio; struct btrfs_device *dev; - DECLARE_COMPLETION_ONSTACK(compl); u64 map_length = 0; u64 sector; struct btrfs_bio *bbio = NULL; + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; int ret; + ASSERT(!(fs_info->sb->s_flags & MS_RDONLY)); BUG_ON(!mirror_num); - bio = bio_alloc(GFP_NOFS, 1); + /* we can't repair anything in raid56 yet */ + if (btrfs_is_parity_mirror(map_tree, logical, length, mirror_num)) + return 0; + + bio = btrfs_io_bio_alloc(GFP_NOFS, 1); if (!bio) return -EIO; - bio->bi_private = &compl; - bio->bi_end_io = repair_io_failure_callback; - bio->bi_size = 0; + bio->bi_iter.bi_size = 0; map_length = length; ret = btrfs_map_block(fs_info, WRITE, logical, @@ -1952,7 +2045,7 @@ int repair_io_failure(struct btrfs_fs_info *fs_info, u64 start, } BUG_ON(mirror_num != bbio->mirror_num); sector = bbio->stripes[mirror_num-1].physical >> 9; - bio->bi_sector = sector; + bio->bi_iter.bi_sector = sector; dev = bbio->stripes[mirror_num-1].dev; kfree(bbio); if (!dev || !dev->bdev || !dev->writeable) { @@ -1960,20 +2053,19 @@ int repair_io_failure(struct btrfs_fs_info *fs_info, u64 start, return -EIO; } bio->bi_bdev = dev->bdev; - bio_add_page(bio, page, length, start-page_offset(page)); - btrfsic_submit_bio(WRITE_SYNC, bio); - wait_for_completion(&compl); + bio_add_page(bio, page, length, start - page_offset(page)); - if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) { + if (btrfsic_submit_bio_wait(WRITE_SYNC, bio)) { /* try to remap that extent elsewhere? */ bio_put(bio); btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_WRITE_ERRS); return -EIO; } - printk_ratelimited_in_rcu(KERN_INFO "btrfs read error corrected: ino %lu off %llu " - "(dev %s sector %llu)\n", page->mapping->host->i_ino, - start, rcu_str_deref(dev->name), sector); + printk_ratelimited_in_rcu(KERN_INFO + "BTRFS: read error corrected: ino %lu off %llu " + "(dev %s sector %llu)\n", page->mapping->host->i_ino, + start, rcu_str_deref(dev->name), sector); bio_put(bio); return 0; @@ -1986,6 +2078,9 @@ int repair_eb_io_failure(struct btrfs_root *root, struct extent_buffer *eb, unsigned long i, num_pages = num_extent_pages(eb->start, eb->len); int ret = 0; + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + for (i = 0; i < num_pages; i++) { struct page *p = extent_buffer_page(eb, i); ret = repair_io_failure(root->fs_info, start, PAGE_CACHE_SIZE, @@ -2007,12 +2102,12 @@ static int clean_io_failure(u64 start, struct page *page) u64 private; u64 private_failure; struct io_failure_record *failrec; - struct btrfs_fs_info *fs_info; + struct inode *inode = page->mapping->host; + struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info; struct extent_state *state; int num_copies; int did_repair = 0; int ret; - struct inode *inode = page->mapping->host; private = 0; ret = count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private, @@ -2035,6 +2130,8 @@ static int clean_io_failure(u64 start, struct page *page) did_repair = 1; goto out; } + if (fs_info->sb->s_flags & MS_RDONLY) + goto out; spin_lock(&BTRFS_I(inode)->io_tree.lock); state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree, @@ -2042,8 +2139,8 @@ static int clean_io_failure(u64 start, struct page *page) EXTENT_LOCKED); spin_unlock(&BTRFS_I(inode)->io_tree.lock); - if (state && state->start == failrec->start) { - fs_info = BTRFS_I(inode)->root->fs_info; + if (state && state->start <= failrec->start && + state->end >= failrec->start + failrec->len - 1) { num_copies = btrfs_num_copies(fs_info, failrec->logical, failrec->len); if (num_copies > 1) { @@ -2052,6 +2149,7 @@ static int clean_io_failure(u64 start, struct page *page) failrec->failed_mirror); did_repair = !ret; } + ret = 0; } out: @@ -2069,9 +2167,9 @@ out: * needed */ -static int bio_readpage_error(struct bio *failed_bio, struct page *page, - u64 start, u64 end, int failed_mirror, - struct extent_state *state) +static int bio_readpage_error(struct bio *failed_bio, u64 phy_offset, + struct page *page, u64 start, u64 end, + int failed_mirror) { struct io_failure_record *failrec = NULL; u64 private; @@ -2081,6 +2179,8 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page, struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; struct bio *bio; + struct btrfs_io_bio *btrfs_failed_bio; + struct btrfs_io_bio *btrfs_bio; int num_copies; int ret; int read_mode; @@ -2107,7 +2207,7 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page, return -EIO; } - if (em->start > start || em->start + em->len < start) { + if (em->start > start || em->start + em->len <= start) { free_extent_map(em); em = NULL; } @@ -2164,23 +2264,12 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page, * all the retry and error correction code that follows. no * matter what the error is, it is very likely to persist. */ - pr_debug("bio_readpage_error: cannot repair, num_copies == 1. " - "state=%p, num_copies=%d, next_mirror %d, " - "failed_mirror %d\n", state, num_copies, - failrec->this_mirror, failed_mirror); + pr_debug("bio_readpage_error: cannot repair, num_copies=%d, next_mirror %d, failed_mirror %d\n", + num_copies, failrec->this_mirror, failed_mirror); free_io_failure(inode, failrec, 0); return -EIO; } - if (!state) { - spin_lock(&tree->lock); - state = find_first_extent_bit_state(tree, failrec->start, - EXTENT_LOCKED); - if (state && state->start != failrec->start) - state = NULL; - spin_unlock(&tree->lock); - } - /* * there are two premises: * a) deliver good data to the caller @@ -2217,24 +2306,35 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page, read_mode = READ_SYNC; } - if (!state || failrec->this_mirror > num_copies) { - pr_debug("bio_readpage_error: (fail) state=%p, num_copies=%d, " - "next_mirror %d, failed_mirror %d\n", state, + if (failrec->this_mirror > num_copies) { + pr_debug("bio_readpage_error: (fail) num_copies=%d, next_mirror %d, failed_mirror %d\n", num_copies, failrec->this_mirror, failed_mirror); free_io_failure(inode, failrec, 0); return -EIO; } - bio = bio_alloc(GFP_NOFS, 1); + bio = btrfs_io_bio_alloc(GFP_NOFS, 1); if (!bio) { free_io_failure(inode, failrec, 0); return -EIO; } - bio->bi_private = state; bio->bi_end_io = failed_bio->bi_end_io; - bio->bi_sector = failrec->logical >> 9; + bio->bi_iter.bi_sector = failrec->logical >> 9; bio->bi_bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; - bio->bi_size = 0; + bio->bi_iter.bi_size = 0; + + btrfs_failed_bio = btrfs_io_bio(failed_bio); + if (btrfs_failed_bio->csum) { + struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info; + u16 csum_size = btrfs_super_csum_size(fs_info->super_copy); + + btrfs_bio = btrfs_io_bio(bio); + btrfs_bio->csum = btrfs_bio->csum_inline; + phy_offset >>= inode->i_sb->s_blocksize_bits; + phy_offset *= csum_size; + memcpy(btrfs_bio->csum, btrfs_failed_bio->csum + phy_offset, + csum_size); + } bio_add_page(bio, page, failrec->len, start - page_offset(page)); @@ -2254,7 +2354,7 @@ int end_extent_writepage(struct page *page, int err, u64 start, u64 end) { int uptodate = (err == 0); struct extent_io_tree *tree; - int ret; + int ret = 0; tree = &BTRFS_I(page->mapping->host)->io_tree; @@ -2268,6 +2368,8 @@ int end_extent_writepage(struct page *page, int err, u64 start, u64 end) if (!uptodate) { ClearPageUptodate(page); SetPageError(page); + ret = ret < 0 ? ret : -EIO; + mapping_set_error(page->mapping, ret); } return 0; } @@ -2283,40 +2385,55 @@ int end_extent_writepage(struct page *page, int err, u64 start, u64 end) */ static void end_bio_extent_writepage(struct bio *bio, int err) { - struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; - struct extent_io_tree *tree; + struct bio_vec *bvec; u64 start; u64 end; - int whole_page; + int i; - do { + bio_for_each_segment_all(bvec, bio, i) { struct page *page = bvec->bv_page; - tree = &BTRFS_I(page->mapping->host)->io_tree; - - start = ((u64)page->index << PAGE_CACHE_SHIFT) + - bvec->bv_offset; - end = start + bvec->bv_len - 1; - if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE) - whole_page = 1; - else - whole_page = 0; + /* We always issue full-page reads, but if some block + * in a page fails to read, blk_update_request() will + * advance bv_offset and adjust bv_len to compensate. + * Print a warning for nonzero offsets, and an error + * if they don't add up to a full page. */ + if (bvec->bv_offset || bvec->bv_len != PAGE_CACHE_SIZE) { + if (bvec->bv_offset + bvec->bv_len != PAGE_CACHE_SIZE) + btrfs_err(BTRFS_I(page->mapping->host)->root->fs_info, + "partial page write in btrfs with offset %u and length %u", + bvec->bv_offset, bvec->bv_len); + else + btrfs_info(BTRFS_I(page->mapping->host)->root->fs_info, + "incomplete page write in btrfs with offset %u and " + "length %u", + bvec->bv_offset, bvec->bv_len); + } - if (--bvec >= bio->bi_io_vec) - prefetchw(&bvec->bv_page->flags); + start = page_offset(page); + end = start + bvec->bv_offset + bvec->bv_len - 1; if (end_extent_writepage(page, err, start, end)) continue; - if (whole_page) - end_page_writeback(page); - else - check_page_writeback(tree, page); - } while (bvec >= bio->bi_io_vec); + end_page_writeback(page); + } bio_put(bio); } +static void +endio_readpage_release_extent(struct extent_io_tree *tree, u64 start, u64 len, + int uptodate) +{ + struct extent_state *cached = NULL; + u64 end = start + len - 1; + + if (uptodate && tree->track_uptodate) + set_extent_uptodate(tree, start, end, &cached, GFP_ATOMIC); + unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC); +} + /* * after a readpage IO is done, we need to: * clear the uptodate bits on error @@ -2330,68 +2447,74 @@ static void end_bio_extent_writepage(struct bio *bio, int err) */ static void end_bio_extent_readpage(struct bio *bio, int err) { + struct bio_vec *bvec; int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); - struct bio_vec *bvec_end = bio->bi_io_vec + bio->bi_vcnt - 1; - struct bio_vec *bvec = bio->bi_io_vec; + struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); struct extent_io_tree *tree; + u64 offset = 0; u64 start; u64 end; - int whole_page; + u64 len; + u64 extent_start = 0; + u64 extent_len = 0; int mirror; int ret; + int i; if (err) uptodate = 0; - do { + bio_for_each_segment_all(bvec, bio, i) { struct page *page = bvec->bv_page; - struct extent_state *cached = NULL; - struct extent_state *state; + struct inode *inode = page->mapping->host; pr_debug("end_bio_extent_readpage: bi_sector=%llu, err=%d, " - "mirror=%ld\n", (u64)bio->bi_sector, err, - (long int)bio->bi_bdev); - tree = &BTRFS_I(page->mapping->host)->io_tree; - - start = ((u64)page->index << PAGE_CACHE_SHIFT) + - bvec->bv_offset; - end = start + bvec->bv_len - 1; - - if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE) - whole_page = 1; - else - whole_page = 0; - - if (++bvec <= bvec_end) - prefetchw(&bvec->bv_page->flags); - - spin_lock(&tree->lock); - state = find_first_extent_bit_state(tree, start, EXTENT_LOCKED); - if (state && state->start == start) { - /* - * take a reference on the state, unlock will drop - * the ref - */ - cache_state(state, &cached); + "mirror=%lu\n", (u64)bio->bi_iter.bi_sector, err, + io_bio->mirror_num); + tree = &BTRFS_I(inode)->io_tree; + + /* We always issue full-page reads, but if some block + * in a page fails to read, blk_update_request() will + * advance bv_offset and adjust bv_len to compensate. + * Print a warning for nonzero offsets, and an error + * if they don't add up to a full page. */ + if (bvec->bv_offset || bvec->bv_len != PAGE_CACHE_SIZE) { + if (bvec->bv_offset + bvec->bv_len != PAGE_CACHE_SIZE) + btrfs_err(BTRFS_I(page->mapping->host)->root->fs_info, + "partial page read in btrfs with offset %u and length %u", + bvec->bv_offset, bvec->bv_len); + else + btrfs_info(BTRFS_I(page->mapping->host)->root->fs_info, + "incomplete page read in btrfs with offset %u and " + "length %u", + bvec->bv_offset, bvec->bv_len); } - spin_unlock(&tree->lock); - mirror = (int)(unsigned long)bio->bi_bdev; - if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) { - ret = tree->ops->readpage_end_io_hook(page, start, end, - state, mirror); + start = page_offset(page); + end = start + bvec->bv_offset + bvec->bv_len - 1; + len = bvec->bv_len; + + mirror = io_bio->mirror_num; + if (likely(uptodate && tree->ops && + tree->ops->readpage_end_io_hook)) { + ret = tree->ops->readpage_end_io_hook(io_bio, offset, + page, start, end, + mirror); if (ret) uptodate = 0; else clean_io_failure(start, page); } - if (!uptodate && tree->ops && tree->ops->readpage_io_failed_hook) { + if (likely(uptodate)) + goto readpage_ok; + + if (tree->ops && tree->ops->readpage_io_failed_hook) { ret = tree->ops->readpage_io_failed_hook(page, mirror); if (!ret && !err && test_bit(BIO_UPTODATE, &bio->bi_flags)) uptodate = 1; - } else if (!uptodate) { + } else { /* * The generic bio_readpage_error handles errors the * following way: If possible, new read requests are @@ -2402,66 +2525,119 @@ static void end_bio_extent_readpage(struct bio *bio, int err) * can't handle the error it will return -EIO and we * remain responsible for that page. */ - ret = bio_readpage_error(bio, page, start, end, mirror, NULL); + ret = bio_readpage_error(bio, offset, page, start, end, + mirror); if (ret == 0) { uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); if (err) uptodate = 0; - uncache_state(&cached); continue; } } - - if (uptodate && tree->track_uptodate) { - set_extent_uptodate(tree, start, end, &cached, - GFP_ATOMIC); +readpage_ok: + if (likely(uptodate)) { + loff_t i_size = i_size_read(inode); + pgoff_t end_index = i_size >> PAGE_CACHE_SHIFT; + unsigned offset; + + /* Zero out the end if this page straddles i_size */ + offset = i_size & (PAGE_CACHE_SIZE-1); + if (page->index == end_index && offset) + zero_user_segment(page, offset, PAGE_CACHE_SIZE); + SetPageUptodate(page); + } else { + ClearPageUptodate(page); + SetPageError(page); } - unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC); - - if (whole_page) { - if (uptodate) { - SetPageUptodate(page); - } else { - ClearPageUptodate(page); - SetPageError(page); + unlock_page(page); + offset += len; + + if (unlikely(!uptodate)) { + if (extent_len) { + endio_readpage_release_extent(tree, + extent_start, + extent_len, 1); + extent_start = 0; + extent_len = 0; } - unlock_page(page); + endio_readpage_release_extent(tree, start, + end - start + 1, 0); + } else if (!extent_len) { + extent_start = start; + extent_len = end + 1 - start; + } else if (extent_start + extent_len == start) { + extent_len += end + 1 - start; } else { - if (uptodate) { - check_page_uptodate(tree, page); - } else { - ClearPageUptodate(page); - SetPageError(page); - } - check_page_locked(tree, page); + endio_readpage_release_extent(tree, extent_start, + extent_len, uptodate); + extent_start = start; + extent_len = end + 1 - start; } - } while (bvec <= bvec_end); + } + if (extent_len) + endio_readpage_release_extent(tree, extent_start, extent_len, + uptodate); + if (io_bio->end_io) + io_bio->end_io(io_bio, err); bio_put(bio); } +/* + * this allocates from the btrfs_bioset. We're returning a bio right now + * but you can call btrfs_io_bio for the appropriate container_of magic + */ struct bio * btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, gfp_t gfp_flags) { + struct btrfs_io_bio *btrfs_bio; struct bio *bio; - bio = bio_alloc(gfp_flags, nr_vecs); + bio = bio_alloc_bioset(gfp_flags, nr_vecs, btrfs_bioset); if (bio == NULL && (current->flags & PF_MEMALLOC)) { - while (!bio && (nr_vecs /= 2)) - bio = bio_alloc(gfp_flags, nr_vecs); + while (!bio && (nr_vecs /= 2)) { + bio = bio_alloc_bioset(gfp_flags, + nr_vecs, btrfs_bioset); + } } if (bio) { - bio->bi_size = 0; bio->bi_bdev = bdev; - bio->bi_sector = first_sector; + bio->bi_iter.bi_sector = first_sector; + btrfs_bio = btrfs_io_bio(bio); + btrfs_bio->csum = NULL; + btrfs_bio->csum_allocated = NULL; + btrfs_bio->end_io = NULL; } return bio; } +struct bio *btrfs_bio_clone(struct bio *bio, gfp_t gfp_mask) +{ + return bio_clone_bioset(bio, gfp_mask, btrfs_bioset); +} + + +/* this also allocates from the btrfs_bioset */ +struct bio *btrfs_io_bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs) +{ + struct btrfs_io_bio *btrfs_bio; + struct bio *bio; + + bio = bio_alloc_bioset(gfp_mask, nr_iovecs, btrfs_bioset); + if (bio) { + btrfs_bio = btrfs_io_bio(bio); + btrfs_bio->csum = NULL; + btrfs_bio->csum_allocated = NULL; + btrfs_bio->end_io = NULL; + } + return bio; +} + + static int __must_check submit_one_bio(int rw, struct bio *bio, int mirror_num, unsigned long bio_flags) { @@ -2471,7 +2647,7 @@ static int __must_check submit_one_bio(int rw, struct bio *bio, struct extent_io_tree *tree = bio->bi_private; u64 start; - start = ((u64)page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset; + start = page_offset(page) + bvec->bv_offset; bio->bi_private = NULL; @@ -2489,13 +2665,13 @@ static int __must_check submit_one_bio(int rw, struct bio *bio, return ret; } -static int merge_bio(struct extent_io_tree *tree, struct page *page, +static int merge_bio(int rw, struct extent_io_tree *tree, struct page *page, unsigned long offset, size_t size, struct bio *bio, unsigned long bio_flags) { int ret = 0; if (tree->ops && tree->ops->merge_bio_hook) - ret = tree->ops->merge_bio_hook(page, offset, size, bio, + ret = tree->ops->merge_bio_hook(rw, page, offset, size, bio, bio_flags); BUG_ON(ret < 0); return ret; @@ -2524,13 +2700,12 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, if (bio_ret && *bio_ret) { bio = *bio_ret; if (old_compressed) - contig = bio->bi_sector == sector; + contig = bio->bi_iter.bi_sector == sector; else - contig = bio->bi_sector + (bio->bi_size >> 9) == - sector; + contig = bio_end_sector(bio) == sector; if (prev_bio_flags != bio_flags || !contig || - merge_bio(tree, page, offset, page_size, bio, bio_flags) || + merge_bio(rw, tree, page, offset, page_size, bio, bio_flags) || bio_add_page(bio, page, page_size, offset) < page_size) { ret = submit_one_bio(rw, bio, mirror_num, prev_bio_flags); @@ -2562,7 +2737,8 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, return ret; } -void attach_extent_buffer_page(struct extent_buffer *eb, struct page *page) +static void attach_extent_buffer_page(struct extent_buffer *eb, + struct page *page) { if (!PagePrivate(page)) { SetPagePrivate(page); @@ -2582,20 +2758,48 @@ void set_page_extent_mapped(struct page *page) } } +static struct extent_map * +__get_extent_map(struct inode *inode, struct page *page, size_t pg_offset, + u64 start, u64 len, get_extent_t *get_extent, + struct extent_map **em_cached) +{ + struct extent_map *em; + + if (em_cached && *em_cached) { + em = *em_cached; + if (extent_map_in_tree(em) && start >= em->start && + start < extent_map_end(em)) { + atomic_inc(&em->refs); + return em; + } + + free_extent_map(em); + *em_cached = NULL; + } + + em = get_extent(inode, page, pg_offset, start, len, 0); + if (em_cached && !IS_ERR_OR_NULL(em)) { + BUG_ON(*em_cached); + atomic_inc(&em->refs); + *em_cached = em; + } + return em; +} /* * basic readpage implementation. Locked extent state structs are inserted * into the tree that are removed when the IO is done (by the end_io * handlers) * XXX JDM: This needs looking at to ensure proper page locking */ -static int __extent_read_full_page(struct extent_io_tree *tree, - struct page *page, - get_extent_t *get_extent, - struct bio **bio, int mirror_num, - unsigned long *bio_flags) +static int __do_readpage(struct extent_io_tree *tree, + struct page *page, + get_extent_t *get_extent, + struct extent_map **em_cached, + struct bio **bio, int mirror_num, + unsigned long *bio_flags, int rw) { struct inode *inode = page->mapping->host; - u64 start = (u64)page->index << PAGE_CACHE_SHIFT; + u64 start = page_offset(page); u64 page_end = start + PAGE_CACHE_SIZE - 1; u64 end; u64 cur = start; @@ -2606,35 +2810,26 @@ 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; + int parent_locked = *bio_flags & EXTENT_BIO_PARENT_LOCKED; size_t pg_offset = 0; size_t iosize; size_t disk_io_size; size_t blocksize = inode->i_sb->s_blocksize; - unsigned long this_bio_flag = 0; + unsigned long this_bio_flag = *bio_flags & EXTENT_BIO_PARENT_LOCKED; set_page_extent_mapped(page); + end = page_end; if (!PageUptodate(page)) { if (cleancache_get_page(page) == 0) { BUG_ON(blocksize != PAGE_SIZE); + unlock_extent(tree, start, end); goto out; } } - end = page_end; - while (1) { - lock_extent(tree, start, end); - ordered = btrfs_lookup_ordered_extent(inode, start); - if (!ordered) - break; - unlock_extent(tree, start, end); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - } - if (page->index == last_byte >> PAGE_CACHE_SHIFT) { char *userpage; size_t zero_offset = last_byte & (PAGE_CACHE_SIZE - 1); @@ -2648,6 +2843,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree, } } while (cur <= end) { + unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1; + if (cur >= last_byte) { char *userpage; struct extent_state *cached = NULL; @@ -2659,15 +2856,18 @@ static int __extent_read_full_page(struct extent_io_tree *tree, kunmap_atomic(userpage); set_extent_uptodate(tree, cur, cur + iosize - 1, &cached, GFP_NOFS); - unlock_extent_cached(tree, cur, cur + iosize - 1, - &cached, GFP_NOFS); + if (!parent_locked) + unlock_extent_cached(tree, cur, + cur + iosize - 1, + &cached, GFP_NOFS); break; } - em = get_extent(inode, page, pg_offset, cur, - end - cur + 1, 0); + em = __get_extent_map(inode, page, pg_offset, cur, + end - cur + 1, get_extent, em_cached); if (IS_ERR_OR_NULL(em)) { SetPageError(page); - unlock_extent(tree, cur, end); + if (!parent_locked) + unlock_extent(tree, cur, end); break; } extent_offset = cur - em->start; @@ -2675,14 +2875,14 @@ static int __extent_read_full_page(struct extent_io_tree *tree, BUG_ON(end < cur); if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { - this_bio_flag = EXTENT_BIO_COMPRESSED; + this_bio_flag |= EXTENT_BIO_COMPRESSED; extent_set_compress_type(&this_bio_flag, em->compress_type); } iosize = min(extent_map_end(em) - cur, end - cur + 1); cur_end = min(extent_map_end(em) - 1, end); - iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1); + iosize = ALIGN(iosize, blocksize); if (this_bio_flag & EXTENT_BIO_COMPRESSED) { disk_io_size = em->block_len; sector = em->block_start >> 9; @@ -2719,7 +2919,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree, if (test_range_bit(tree, cur, cur_end, EXTENT_UPTODATE, 1, NULL)) { check_page_uptodate(tree, page); - unlock_extent(tree, cur, cur + iosize - 1); + if (!parent_locked) + unlock_extent(tree, cur, cur + iosize - 1); cur = cur + iosize; pg_offset += iosize; continue; @@ -2729,34 +2930,27 @@ static int __extent_read_full_page(struct extent_io_tree *tree, */ if (block_start == EXTENT_MAP_INLINE) { SetPageError(page); - unlock_extent(tree, cur, cur + iosize - 1); + if (!parent_locked) + unlock_extent(tree, cur, cur + iosize - 1); cur = cur + iosize; pg_offset += iosize; continue; } - ret = 0; - if (tree->ops && tree->ops->readpage_io_hook) { - ret = tree->ops->readpage_io_hook(page, cur, - cur + iosize - 1); - } - if (!ret) { - unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1; - pnr -= page->index; - ret = submit_extent_page(READ, tree, page, + pnr -= page->index; + ret = submit_extent_page(rw, tree, page, sector, disk_io_size, pg_offset, bdev, bio, pnr, end_bio_extent_readpage, mirror_num, *bio_flags, this_bio_flag); - if (!ret) { - nr++; - *bio_flags = this_bio_flag; - } - } - if (ret) { + if (!ret) { + nr++; + *bio_flags = this_bio_flag; + } else { SetPageError(page); - unlock_extent(tree, cur, cur + iosize - 1); + if (!parent_locked) + unlock_extent(tree, cur, cur + iosize - 1); } cur = cur + iosize; pg_offset += iosize; @@ -2770,6 +2964,104 @@ out: return 0; } +static inline void __do_contiguous_readpages(struct extent_io_tree *tree, + struct page *pages[], int nr_pages, + u64 start, u64 end, + get_extent_t *get_extent, + struct extent_map **em_cached, + struct bio **bio, int mirror_num, + unsigned long *bio_flags, int rw) +{ + struct inode *inode; + struct btrfs_ordered_extent *ordered; + int index; + + inode = pages[0]->mapping->host; + while (1) { + lock_extent(tree, start, end); + ordered = btrfs_lookup_ordered_range(inode, start, + end - start + 1); + if (!ordered) + break; + unlock_extent(tree, start, end); + btrfs_start_ordered_extent(inode, ordered, 1); + btrfs_put_ordered_extent(ordered); + } + + for (index = 0; index < nr_pages; index++) { + __do_readpage(tree, pages[index], get_extent, em_cached, bio, + mirror_num, bio_flags, rw); + page_cache_release(pages[index]); + } +} + +static void __extent_readpages(struct extent_io_tree *tree, + struct page *pages[], + int nr_pages, get_extent_t *get_extent, + struct extent_map **em_cached, + struct bio **bio, int mirror_num, + unsigned long *bio_flags, int rw) +{ + u64 start = 0; + u64 end = 0; + u64 page_start; + int index; + int first_index = 0; + + for (index = 0; index < nr_pages; index++) { + page_start = page_offset(pages[index]); + if (!end) { + start = page_start; + end = start + PAGE_CACHE_SIZE - 1; + first_index = index; + } else if (end + 1 == page_start) { + end += PAGE_CACHE_SIZE; + } else { + __do_contiguous_readpages(tree, &pages[first_index], + index - first_index, start, + end, get_extent, em_cached, + bio, mirror_num, bio_flags, + rw); + start = page_start; + end = start + PAGE_CACHE_SIZE - 1; + first_index = index; + } + } + + if (end) + __do_contiguous_readpages(tree, &pages[first_index], + index - first_index, start, + end, get_extent, em_cached, bio, + mirror_num, bio_flags, rw); +} + +static int __extent_read_full_page(struct extent_io_tree *tree, + struct page *page, + get_extent_t *get_extent, + struct bio **bio, int mirror_num, + unsigned long *bio_flags, int rw) +{ + struct inode *inode = page->mapping->host; + struct btrfs_ordered_extent *ordered; + u64 start = page_offset(page); + u64 end = start + PAGE_CACHE_SIZE - 1; + int ret; + + while (1) { + lock_extent(tree, start, end); + ordered = btrfs_lookup_ordered_extent(inode, start); + if (!ordered) + break; + unlock_extent(tree, start, end); + btrfs_start_ordered_extent(inode, ordered, 1); + btrfs_put_ordered_extent(ordered); + } + + ret = __do_readpage(tree, page, get_extent, NULL, bio, mirror_num, + bio_flags, rw); + return ret; +} + int extent_read_full_page(struct extent_io_tree *tree, struct page *page, get_extent_t *get_extent, int mirror_num) { @@ -2778,7 +3070,21 @@ int extent_read_full_page(struct extent_io_tree *tree, struct page *page, int ret; ret = __extent_read_full_page(tree, page, get_extent, &bio, mirror_num, - &bio_flags); + &bio_flags, READ); + if (bio) + ret = submit_one_bio(READ, bio, mirror_num, bio_flags); + return ret; +} + +int extent_read_full_page_nolock(struct extent_io_tree *tree, struct page *page, + get_extent_t *get_extent, int mirror_num) +{ + struct bio *bio = NULL; + unsigned long bio_flags = EXTENT_BIO_PARENT_LOCKED; + int ret; + + ret = __do_readpage(tree, page, get_extent, NULL, &bio, mirror_num, + &bio_flags, READ); if (bio) ret = submit_one_bio(READ, bio, mirror_num, bio_flags); return ret; @@ -2795,143 +3101,130 @@ static noinline void update_nr_written(struct page *page, } /* - * the writepage semantics are similar to regular writepage. extent - * records are inserted to lock ranges in the tree, and as dirty areas - * are found, they are marked writeback. Then the lock bits are removed - * and the end_io handler clears the writeback ranges + * helper for __extent_writepage, doing all of the delayed allocation setup. + * + * This returns 1 if our fill_delalloc function did all the work required + * to write the page (copy into inline extent). In this case the IO has + * been started and the page is already unlocked. + * + * This returns 0 if all went well (page still locked) + * This returns < 0 if there were errors (page still locked) */ -static int __extent_writepage(struct page *page, struct writeback_control *wbc, - void *data) +static noinline_for_stack int writepage_delalloc(struct inode *inode, + struct page *page, struct writeback_control *wbc, + struct extent_page_data *epd, + u64 delalloc_start, + unsigned long *nr_written) { - struct inode *inode = page->mapping->host; - struct extent_page_data *epd = data; struct extent_io_tree *tree = epd->tree; - u64 start = (u64)page->index << PAGE_CACHE_SHIFT; - u64 delalloc_start; + u64 page_end = delalloc_start + PAGE_CACHE_SIZE - 1; + u64 nr_delalloc; + u64 delalloc_to_write = 0; + u64 delalloc_end = 0; + int ret; + int page_started = 0; + + if (epd->extent_locked || !tree->ops || !tree->ops->fill_delalloc) + return 0; + + while (delalloc_end < page_end) { + nr_delalloc = find_lock_delalloc_range(inode, tree, + page, + &delalloc_start, + &delalloc_end, + 128 * 1024 * 1024); + if (nr_delalloc == 0) { + delalloc_start = delalloc_end + 1; + continue; + } + ret = tree->ops->fill_delalloc(inode, page, + delalloc_start, + delalloc_end, + &page_started, + nr_written); + /* File system has been set read-only */ + if (ret) { + SetPageError(page); + /* fill_delalloc should be return < 0 for error + * but just in case, we use > 0 here meaning the + * IO is started, so we don't want to return > 0 + * unless things are going well. + */ + ret = ret < 0 ? ret : -EIO; + goto done; + } + /* + * 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) { + /* + * 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; + return 1; + } + + ret = 0; + +done: + return ret; +} + +/* + * helper for __extent_writepage. This calls the writepage start hooks, + * and does the loop to map the page into extents and bios. + * + * We return 1 if the IO is started and the page is unlocked, + * 0 if all went well (page still locked) + * < 0 if there were errors (page still locked) + */ +static noinline_for_stack int __extent_writepage_io(struct inode *inode, + struct page *page, + struct writeback_control *wbc, + struct extent_page_data *epd, + loff_t i_size, + unsigned long nr_written, + int write_flags, int *nr_ret) +{ + struct extent_io_tree *tree = epd->tree; + u64 start = page_offset(page); u64 page_end = start + PAGE_CACHE_SIZE - 1; u64 end; u64 cur = start; u64 extent_offset; - u64 last_byte = i_size_read(inode); u64 block_start; u64 iosize; sector_t sector; struct extent_state *cached_state = NULL; struct extent_map *em; struct block_device *bdev; - int ret; - int nr = 0; size_t pg_offset = 0; size_t blocksize; - loff_t i_size = i_size_read(inode); - unsigned long end_index = i_size >> PAGE_CACHE_SHIFT; - u64 nr_delalloc; - u64 delalloc_end; - int page_started; - int compressed; - int write_flags; - unsigned long nr_written = 0; - bool fill_delalloc = true; - - if (wbc->sync_mode == WB_SYNC_ALL) - write_flags = WRITE_SYNC; - else - write_flags = WRITE; - - trace___extent_writepage(page, inode, wbc); - - WARN_ON(!PageLocked(page)); - - ClearPageError(page); - - pg_offset = i_size & (PAGE_CACHE_SIZE - 1); - if (page->index > end_index || - (page->index == end_index && !pg_offset)) { - page->mapping->a_ops->invalidatepage(page, 0); - unlock_page(page); - return 0; - } - - if (page->index == end_index) { - char *userpage; - - userpage = kmap_atomic(page); - memset(userpage + pg_offset, 0, - PAGE_CACHE_SIZE - pg_offset); - kunmap_atomic(userpage); - flush_dcache_page(page); - } - pg_offset = 0; - - set_page_extent_mapped(page); - - if (!tree->ops || !tree->ops->fill_delalloc) - fill_delalloc = false; - - delalloc_start = start; - delalloc_end = 0; - page_started = 0; - if (!epd->extent_locked && fill_delalloc) { - 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, - &delalloc_end, - 128 * 1024 * 1024); - if (nr_delalloc == 0) { - delalloc_start = delalloc_end + 1; - continue; - } - ret = tree->ops->fill_delalloc(inode, page, - delalloc_start, - delalloc_end, - &page_started, - &nr_written); - /* File system has been set read-only */ - if (ret) { - SetPageError(page); - goto done; - } - /* - * 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); - } + int ret = 0; + int nr = 0; + bool compressed; - /* did the fill delalloc function already unlock and start - * the IO? - */ - if (page_started) { - ret = 0; - /* - * 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; - } - } if (tree->ops && tree->ops->writepage_start_hook) { ret = tree->ops->writepage_start_hook(page, start, page_end); @@ -2941,9 +3234,10 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, wbc->pages_skipped++; else redirty_page_for_writepage(wbc, page); + update_nr_written(page, wbc, nr_written); unlock_page(page); - ret = 0; + ret = 1; goto done_unlocked; } } @@ -2955,7 +3249,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, update_nr_written(page, wbc, nr_written + 1); end = page_end; - if (last_byte <= start) { + if (i_size <= start) { if (tree->ops && tree->ops->writepage_end_io_hook) tree->ops->writepage_end_io_hook(page, start, page_end, NULL, 1); @@ -2965,7 +3259,8 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, blocksize = inode->i_sb->s_blocksize; while (cur <= end) { - if (cur >= last_byte) { + u64 em_end; + if (cur >= i_size) { if (tree->ops && tree->ops->writepage_end_io_hook) tree->ops->writepage_end_io_hook(page, cur, page_end, NULL, 1); @@ -2975,14 +3270,16 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, end - cur + 1, 1); if (IS_ERR_OR_NULL(em)) { SetPageError(page); + ret = PTR_ERR_OR_ZERO(em); break; } extent_offset = cur - em->start; - BUG_ON(extent_map_end(em) <= cur); + em_end = extent_map_end(em); + BUG_ON(em_end <= cur); BUG_ON(end < cur); - iosize = min(extent_map_end(em) - cur, end - cur + 1); - iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1); + iosize = min(em_end - cur, end - cur + 1); + iosize = ALIGN(iosize, blocksize); sector = (em->block_start + extent_offset) >> 9; bdev = em->bdev; block_start = em->block_start; @@ -3017,13 +3314,6 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, pg_offset += iosize; continue; } - /* leave this out until we have a page_mkwrite call */ - if (0 && !test_range_bit(tree, cur, cur + iosize - 1, - EXTENT_DIRTY, 0, NULL)) { - cur = cur + iosize; - pg_offset += iosize; - continue; - } if (tree->ops && tree->ops->writepage_io_hook) { ret = tree->ops->writepage_io_hook(page, cur, @@ -3034,14 +3324,13 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, if (ret) { SetPageError(page); } else { - unsigned long max_nr = end_index + 1; + unsigned long max_nr = (i_size >> PAGE_CACHE_SHIFT) + 1; set_range_writeback(tree, cur, cur + iosize - 1); if (!PageWriteback(page)) { - printk(KERN_ERR "btrfs warning page %lu not " - "writeback, cur %llu end %llu\n", - page->index, (unsigned long long)cur, - (unsigned long long)end); + btrfs_err(BTRFS_I(inode)->root->fs_info, + "page %lu not writeback, cur %llu end %llu", + page->index, cur, end); } ret = submit_extent_page(write_flags, tree, page, @@ -3057,17 +3346,94 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, nr++; } done: + *nr_ret = nr; + +done_unlocked: + + /* drop our reference on any cached states */ + free_extent_state(cached_state); + return ret; +} + +/* + * the writepage semantics are similar to regular writepage. extent + * records are inserted to lock ranges in the tree, and as dirty areas + * are found, they are marked writeback. Then the lock bits are removed + * and the end_io handler clears the writeback ranges + */ +static int __extent_writepage(struct page *page, struct writeback_control *wbc, + void *data) +{ + struct inode *inode = page->mapping->host; + struct extent_page_data *epd = data; + u64 start = page_offset(page); + u64 page_end = start + PAGE_CACHE_SIZE - 1; + int ret; + int nr = 0; + size_t pg_offset = 0; + loff_t i_size = i_size_read(inode); + unsigned long end_index = i_size >> PAGE_CACHE_SHIFT; + int write_flags; + unsigned long nr_written = 0; + + if (wbc->sync_mode == WB_SYNC_ALL) + write_flags = WRITE_SYNC; + else + write_flags = WRITE; + + trace___extent_writepage(page, inode, wbc); + + WARN_ON(!PageLocked(page)); + + ClearPageError(page); + + pg_offset = i_size & (PAGE_CACHE_SIZE - 1); + if (page->index > end_index || + (page->index == end_index && !pg_offset)) { + page->mapping->a_ops->invalidatepage(page, 0, PAGE_CACHE_SIZE); + unlock_page(page); + return 0; + } + + if (page->index == end_index) { + char *userpage; + + userpage = kmap_atomic(page); + memset(userpage + pg_offset, 0, + PAGE_CACHE_SIZE - pg_offset); + kunmap_atomic(userpage); + flush_dcache_page(page); + } + + pg_offset = 0; + + set_page_extent_mapped(page); + + ret = writepage_delalloc(inode, page, wbc, epd, start, &nr_written); + if (ret == 1) + goto done_unlocked; + if (ret) + goto done; + + ret = __extent_writepage_io(inode, page, wbc, epd, + i_size, nr_written, write_flags, &nr); + if (ret == 1) + goto done_unlocked; + +done: if (nr == 0) { /* make sure the mapping tag for page dirty gets cleared */ set_page_writeback(page); end_page_writeback(page); } + if (PageError(page)) { + ret = ret < 0 ? ret : -EIO; + end_extent_writepage(page, ret, start, page_end); + } unlock_page(page); + return ret; done_unlocked: - - /* drop our reference on any cached states */ - free_extent_state(cached_state); return 0; } @@ -3077,15 +3443,16 @@ static int eb_wait(void *word) return 0; } -static void wait_on_extent_buffer_writeback(struct extent_buffer *eb) +void wait_on_extent_buffer_writeback(struct extent_buffer *eb) { wait_on_bit(&eb->bflags, EXTENT_BUFFER_WRITEBACK, eb_wait, TASK_UNINTERRUPTIBLE); } -static int lock_extent_buffer_for_io(struct extent_buffer *eb, - struct btrfs_fs_info *fs_info, - struct extent_page_data *epd) +static noinline_for_stack int +lock_extent_buffer_for_io(struct extent_buffer *eb, + struct btrfs_fs_info *fs_info, + struct extent_page_data *epd) { unsigned long i, num_pages; int flush = 0; @@ -3124,12 +3491,9 @@ static int lock_extent_buffer_for_io(struct extent_buffer *eb, set_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags); spin_unlock(&eb->refs_lock); btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN); - spin_lock(&fs_info->delalloc_lock); - if (fs_info->dirty_metadata_bytes >= eb->len) - fs_info->dirty_metadata_bytes -= eb->len; - else - WARN_ON(1); - spin_unlock(&fs_info->delalloc_lock); + __percpu_counter_add(&fs_info->dirty_metadata_bytes, + -eb->len, + fs_info->dirty_metadata_batch); ret = 1; } else { spin_unlock(&eb->refs_lock); @@ -3159,26 +3523,24 @@ static int lock_extent_buffer_for_io(struct extent_buffer *eb, static void end_extent_buffer_writeback(struct extent_buffer *eb) { clear_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); wake_up_bit(&eb->bflags, EXTENT_BUFFER_WRITEBACK); } static void end_bio_extent_buffer_writepage(struct bio *bio, int err) { - int uptodate = err == 0; - struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; + struct bio_vec *bvec; struct extent_buffer *eb; - int done; + int i, done; - do { + bio_for_each_segment_all(bvec, bio, i) { struct page *page = bvec->bv_page; - bvec--; eb = (struct extent_buffer *)page->private; BUG_ON(!eb); done = atomic_dec_and_test(&eb->io_pages); - if (!uptodate || test_bit(EXTENT_BUFFER_IOERR, &eb->bflags)) { + if (err || test_bit(EXTENT_BUFFER_IOERR, &eb->bflags)) { set_bit(EXTENT_BUFFER_IOERR, &eb->bflags); ClearPageUptodate(page); SetPageError(page); @@ -3190,22 +3552,22 @@ static void end_bio_extent_buffer_writepage(struct bio *bio, int err) continue; end_extent_buffer_writeback(eb); - } while (bvec >= bio->bi_io_vec); + } bio_put(bio); - } -static int write_one_eb(struct extent_buffer *eb, +static noinline_for_stack int write_one_eb(struct extent_buffer *eb, struct btrfs_fs_info *fs_info, struct writeback_control *wbc, struct extent_page_data *epd) { struct block_device *bdev = fs_info->fs_devices->latest_bdev; + struct extent_io_tree *tree = &BTRFS_I(fs_info->btree_inode)->io_tree; u64 offset = eb->start; unsigned long i, num_pages; unsigned long bio_flags = 0; - int rw = (epd->sync_io ? WRITE_SYNC : WRITE); + int rw = (epd->sync_io ? WRITE_SYNC : WRITE) | REQ_META; int ret = 0; clear_bit(EXTENT_BUFFER_IOERR, &eb->bflags); @@ -3219,7 +3581,7 @@ static int write_one_eb(struct extent_buffer *eb, clear_page_dirty_for_io(p); set_page_writeback(p); - ret = submit_extent_page(rw, eb->tree, p, offset >> 9, + ret = submit_extent_page(rw, tree, p, offset >> 9, PAGE_CACHE_SIZE, 0, bdev, &epd->bio, -1, end_bio_extent_buffer_writepage, 0, epd->bio_flags, bio_flags); @@ -3316,9 +3678,8 @@ retry: * but no sense in crashing the users box for something * we can survive anyway. */ - if (!eb) { + if (WARN_ON(!eb)) { spin_unlock(&mapping->private_lock); - WARN_ON(1); continue; } @@ -3394,6 +3755,7 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, struct inode *inode = mapping->host; int ret = 0; int done = 0; + int err = 0; int nr_to_write_done = 0; struct pagevec pvec; int nr_pages; @@ -3446,15 +3808,9 @@ retry: * swizzled back from swapper_space to tmpfs file * mapping */ - if (tree->ops && - tree->ops->write_cache_pages_lock_hook) { - tree->ops->write_cache_pages_lock_hook(page, - data, flush_fn); - } else { - if (!trylock_page(page)) { - flush_fn(data); - lock_page(page); - } + if (!trylock_page(page)) { + flush_fn(data); + lock_page(page); } if (unlikely(page->mapping != mapping)) { @@ -3486,8 +3842,8 @@ retry: unlock_page(page); ret = 0; } - if (ret) - done = 1; + if (!err && ret < 0) + err = ret; /* * the filesystem may choose to bump up nr_to_write. @@ -3499,7 +3855,7 @@ retry: pagevec_release(&pvec); cond_resched(); } - if (!scanned && !done) { + if (!scanned && !done && !err) { /* * We hit the last page and there is more work to be done: wrap * back to the start of the file @@ -3509,7 +3865,7 @@ retry: goto retry; } btrfs_add_delayed_iput(inode); - return ret; + return err; } static void flush_epd_write_bio(struct extent_page_data *epd) @@ -3629,7 +3985,7 @@ int extent_readpages(struct extent_io_tree *tree, unsigned long bio_flags = 0; struct page *pagepool[16]; struct page *page; - int i = 0; + struct extent_map *em_cached = NULL; int nr = 0; for (page_idx = 0; page_idx < nr_pages; page_idx++) { @@ -3646,18 +4002,16 @@ int extent_readpages(struct extent_io_tree *tree, pagepool[nr++] = page; if (nr < ARRAY_SIZE(pagepool)) continue; - for (i = 0; i < nr; i++) { - __extent_read_full_page(tree, pagepool[i], get_extent, - &bio, 0, &bio_flags); - page_cache_release(pagepool[i]); - } + __extent_readpages(tree, pagepool, nr, get_extent, &em_cached, + &bio, 0, &bio_flags, READ); nr = 0; } - for (i = 0; i < nr; i++) { - __extent_read_full_page(tree, pagepool[i], get_extent, - &bio, 0, &bio_flags); - page_cache_release(pagepool[i]); - } + if (nr) + __extent_readpages(tree, pagepool, nr, get_extent, &em_cached, + &bio, 0, &bio_flags, READ); + + if (em_cached) + free_extent_map(em_cached); BUG_ON(!list_empty(pages)); if (bio) @@ -3674,11 +4028,11 @@ 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 start = page_offset(page); u64 end = start + PAGE_CACHE_SIZE - 1; size_t blocksize = page->mapping->host->i_sb->s_blocksize; - start += (offset + blocksize - 1) & ~(blocksize - 1); + start += ALIGN(offset, blocksize); if (start > end) return 0; @@ -3696,11 +4050,11 @@ int extent_invalidatepage(struct extent_io_tree *tree, * are locked or under IO and drops the related state bits if it is safe * to drop the page. */ -int try_release_extent_state(struct extent_map_tree *map, - struct extent_io_tree *tree, struct page *page, - gfp_t mask) +static int try_release_extent_state(struct extent_map_tree *map, + struct extent_io_tree *tree, + struct page *page, gfp_t mask) { - u64 start = (u64)page->index << PAGE_CACHE_SHIFT; + u64 start = page_offset(page); u64 end = start + PAGE_CACHE_SIZE - 1; int ret = 1; @@ -3739,7 +4093,7 @@ int try_release_extent_mapping(struct extent_map_tree *map, gfp_t mask) { struct extent_map *em; - u64 start = (u64)page->index << PAGE_CACHE_SHIFT; + u64 start = page_offset(page); u64 end = start + PAGE_CACHE_SIZE - 1; if ((mask & __GFP_WAIT) && @@ -3793,11 +4147,11 @@ static struct extent_map *get_extent_skip_holes(struct inode *inode, if (offset >= last) return NULL; - while(1) { + while (1) { len = last - offset; if (len == 0) break; - len = (len + sectorsize - 1) & ~(sectorsize - 1); + len = ALIGN(len, sectorsize); em = get_extent(inode, NULL, 0, offset, len, 0); if (IS_ERR_OR_NULL(em)) return em; @@ -3817,6 +4171,19 @@ static struct extent_map *get_extent_skip_holes(struct inode *inode, return NULL; } +static noinline int count_ext_ref(u64 inum, u64 offset, u64 root_id, void *ctx) +{ + unsigned long cnt = *((unsigned long *)ctx); + + cnt++; + *((unsigned long *)ctx) = cnt; + + /* Now we're sure that the extent is shared. */ + if (cnt > 1) + return 1; + return 0; +} + int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, __u64 start, __u64 len, get_extent_t *get_extent) { @@ -3833,12 +4200,10 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, struct extent_map *em = NULL; struct extent_state *cached_state = NULL; struct btrfs_path *path; - struct btrfs_file_extent_item *item; int end = 0; u64 em_start = 0; u64 em_len = 0; u64 em_end = 0; - unsigned long emflags; if (len == 0) return -EINVAL; @@ -3863,8 +4228,6 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, } WARN_ON(!ret); path->slots[0]--; - item = btrfs_item_ptr(path->nodes[0], path->slots[0], - struct btrfs_file_extent_item); btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]); found_type = btrfs_key_type(&found_key); @@ -3883,7 +4246,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, last = found_key.offset; last_for_get_extent = last + 1; } - btrfs_free_path(path); + btrfs_release_path(path); /* * we might have some extents allocated but more delalloc past those @@ -3895,7 +4258,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, last_for_get_extent = isize; } - lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0, + lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len - 1, 0, &cached_state); em = get_extent_skip_holes(inode, start, last_for_get_extent, @@ -3908,7 +4271,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, } while (!end) { - u64 offset_in_extent; + u64 offset_in_extent = 0; /* break if the extent we found is outside the range */ if (em->start >= max || extent_map_end(em) < off) @@ -3924,12 +4287,14 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, /* * record the offset from the start of the extent - * for adjusting the disk offset below + * for adjusting the disk offset below. Only do this if the + * extent isn't compressed since our in ram offset may be past + * what we have actually allocated on disk. */ - offset_in_extent = em_start - em->start; + if (!test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) + offset_in_extent = em_start - em->start; em_end = extent_map_end(em); em_len = em_end - em_start; - emflags = em->flags; disko = 0; flags = 0; @@ -3950,7 +4315,24 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, flags |= (FIEMAP_EXTENT_DELALLOC | FIEMAP_EXTENT_UNKNOWN); } else { + unsigned long ref_cnt = 0; + disko = em->block_start + offset_in_extent; + + /* + * As btrfs supports shared space, this information + * can be exported to userspace tools via + * flag FIEMAP_EXTENT_SHARED. + */ + ret = iterate_inodes_from_logical( + em->block_start, + BTRFS_I(inode)->root->fs_info, + path, count_ext_ref, &ref_cnt); + if (ret < 0 && ret != -ENOENT) + goto out_free; + + if (ref_cnt > 1) + flags |= FIEMAP_EXTENT_SHARED; } if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) flags |= FIEMAP_EXTENT_ENCODED; @@ -3982,40 +4364,100 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, out_free: free_extent_map(em); out: - unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, start + len, + btrfs_free_path(path); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, start + len - 1, &cached_state, GFP_NOFS); return ret; } static void __free_extent_buffer(struct extent_buffer *eb) { -#if LEAK_DEBUG - unsigned long flags; - spin_lock_irqsave(&leak_lock, flags); - list_del(&eb->leak_list); - spin_unlock_irqrestore(&leak_lock, flags); -#endif - if (eb->pages && eb->pages != eb->inline_pages) - kfree(eb->pages); + btrfs_leak_debug_del(&eb->leak_list); kmem_cache_free(extent_buffer_cache, eb); } -static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, - u64 start, - unsigned long len, - gfp_t mask) +int extent_buffer_under_io(struct extent_buffer *eb) +{ + return (atomic_read(&eb->io_pages) || + test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags) || + test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); +} + +/* + * Helper for releasing extent buffer page. + */ +static void btrfs_release_extent_buffer_page(struct extent_buffer *eb, + unsigned long start_idx) +{ + unsigned long index; + unsigned long num_pages; + struct page *page; + int mapped = !test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags); + + BUG_ON(extent_buffer_under_io(eb)); + + num_pages = num_extent_pages(eb->start, eb->len); + index = start_idx + num_pages; + if (start_idx >= index) + return; + + do { + index--; + page = extent_buffer_page(eb, index); + if (page && mapped) { + spin_lock(&page->mapping->private_lock); + /* + * We do this since we'll remove the pages after we've + * removed the eb from the radix tree, so we could race + * and have this page now attached to the new eb. So + * only clear page_private if it's still connected to + * this eb. + */ + if (PagePrivate(page) && + page->private == (unsigned long)eb) { + BUG_ON(test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); + BUG_ON(PageDirty(page)); + BUG_ON(PageWriteback(page)); + /* + * We need to make sure we haven't be attached + * to a new eb. + */ + ClearPagePrivate(page); + set_page_private(page, 0); + /* One for the page private */ + page_cache_release(page); + } + spin_unlock(&page->mapping->private_lock); + + } + if (page) { + /* One for when we alloced the page */ + page_cache_release(page); + } + } while (index != start_idx); +} + +/* + * Helper for releasing the extent buffer. + */ +static inline void btrfs_release_extent_buffer(struct extent_buffer *eb) +{ + btrfs_release_extent_buffer_page(eb, 0); + __free_extent_buffer(eb); +} + +static struct extent_buffer * +__alloc_extent_buffer(struct btrfs_fs_info *fs_info, u64 start, + unsigned long len, gfp_t mask) { struct extent_buffer *eb = NULL; -#if LEAK_DEBUG - unsigned long flags; -#endif eb = kmem_cache_zalloc(extent_buffer_cache, mask); if (eb == NULL) return NULL; eb->start = start; eb->len = len; - eb->tree = tree; + eb->fs_info = fs_info; eb->bflags = 0; rwlock_init(&eb->lock); atomic_set(&eb->write_locks, 0); @@ -4028,28 +4470,18 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, init_waitqueue_head(&eb->write_lock_wq); init_waitqueue_head(&eb->read_lock_wq); -#if LEAK_DEBUG - spin_lock_irqsave(&leak_lock, flags); - list_add(&eb->leak_list, &buffers); - spin_unlock_irqrestore(&leak_lock, flags); -#endif + btrfs_leak_debug_add(&eb->leak_list, &buffers); + spin_lock_init(&eb->refs_lock); atomic_set(&eb->refs, 1); atomic_set(&eb->io_pages, 0); - if (len > MAX_INLINE_EXTENT_BUFFER_SIZE) { - struct page **pages; - int num_pages = (len + PAGE_CACHE_SIZE - 1) >> - PAGE_CACHE_SHIFT; - pages = kzalloc(num_pages, mask); - if (!pages) { - __free_extent_buffer(eb); - return NULL; - } - eb->pages = pages; - } else { - eb->pages = eb->inline_pages; - } + /* + * Sanity checks, currently the maximum is 64k covered by 16x 4k pages + */ + BUILD_BUG_ON(BTRFS_MAX_METADATA_BLOCKSIZE + > MAX_INLINE_EXTENT_BUFFER_SIZE); + BUG_ON(len > MAX_INLINE_EXTENT_BUFFER_SIZE); return eb; } @@ -4061,13 +4493,16 @@ struct extent_buffer *btrfs_clone_extent_buffer(struct extent_buffer *src) struct extent_buffer *new; unsigned long num_pages = num_extent_pages(src->start, src->len); - new = __alloc_extent_buffer(NULL, src->start, src->len, GFP_ATOMIC); + new = __alloc_extent_buffer(NULL, src->start, src->len, GFP_NOFS); if (new == NULL) return NULL; for (i = 0; i < num_pages; i++) { - p = alloc_page(GFP_ATOMIC); - BUG_ON(!p); + p = alloc_page(GFP_NOFS); + if (!p) { + btrfs_release_extent_buffer(new); + return NULL; + } attach_extent_buffer_page(new, p); WARN_ON(PageDirty(p)); SetPageUptodate(p); @@ -4087,12 +4522,12 @@ struct extent_buffer *alloc_dummy_extent_buffer(u64 start, unsigned long len) unsigned long num_pages = num_extent_pages(0, len); unsigned long i; - eb = __alloc_extent_buffer(NULL, start, len, GFP_ATOMIC); + eb = __alloc_extent_buffer(NULL, start, len, GFP_NOFS); if (!eb) return NULL; for (i = 0; i < num_pages; i++) { - eb->pages[i] = alloc_page(GFP_ATOMIC); + eb->pages[i] = alloc_page(GFP_NOFS); if (!eb->pages[i]) goto err; } @@ -4108,78 +4543,9 @@ err: return NULL; } -static int extent_buffer_under_io(struct extent_buffer *eb) -{ - return (atomic_read(&eb->io_pages) || - test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags) || - test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); -} - -/* - * Helper for releasing extent buffer page. - */ -static void btrfs_release_extent_buffer_page(struct extent_buffer *eb, - unsigned long start_idx) -{ - unsigned long index; - unsigned long num_pages; - struct page *page; - int mapped = !test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags); - - BUG_ON(extent_buffer_under_io(eb)); - - num_pages = num_extent_pages(eb->start, eb->len); - index = start_idx + num_pages; - if (start_idx >= index) - return; - - do { - index--; - page = extent_buffer_page(eb, index); - if (page && mapped) { - spin_lock(&page->mapping->private_lock); - /* - * We do this since we'll remove the pages after we've - * removed the eb from the radix tree, so we could race - * and have this page now attached to the new eb. So - * only clear page_private if it's still connected to - * this eb. - */ - if (PagePrivate(page) && - page->private == (unsigned long)eb) { - BUG_ON(test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); - BUG_ON(PageDirty(page)); - BUG_ON(PageWriteback(page)); - /* - * We need to make sure we haven't be attached - * to a new eb. - */ - ClearPagePrivate(page); - set_page_private(page, 0); - /* One for the page private */ - page_cache_release(page); - } - spin_unlock(&page->mapping->private_lock); - - } - if (page) { - /* One for when we alloced the page */ - page_cache_release(page); - } - } while (index != start_idx); -} - -/* - * Helper for releasing the extent buffer. - */ -static inline void btrfs_release_extent_buffer(struct extent_buffer *eb) -{ - btrfs_release_extent_buffer_page(eb, 0); - __free_extent_buffer(eb); -} - static void check_buffer_tree_ref(struct extent_buffer *eb) { + int refs; /* the ref bit is tricky. We have to make sure it is set * if we have the buffer dirty. Otherwise the * code to free a buffer can end up dropping a dirty @@ -4200,13 +4566,18 @@ static void check_buffer_tree_ref(struct extent_buffer *eb) * So bump the ref count first, then set the bit. If someone * beat us to it, drop the ref we added. */ + refs = atomic_read(&eb->refs); + if (refs >= 2 && test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) + return; + spin_lock(&eb->refs_lock); if (!test_and_set_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) atomic_inc(&eb->refs); spin_unlock(&eb->refs_lock); } -static void mark_extent_buffer_accessed(struct extent_buffer *eb) +static void mark_extent_buffer_accessed(struct extent_buffer *eb, + struct page *accessed) { unsigned long num_pages, i; @@ -4215,11 +4586,77 @@ static void mark_extent_buffer_accessed(struct extent_buffer *eb) num_pages = num_extent_pages(eb->start, eb->len); for (i = 0; i < num_pages; i++) { struct page *p = extent_buffer_page(eb, i); - mark_page_accessed(p); + if (p != accessed) + mark_page_accessed(p); + } +} + +struct extent_buffer *find_extent_buffer(struct btrfs_fs_info *fs_info, + u64 start) +{ + struct extent_buffer *eb; + + rcu_read_lock(); + eb = radix_tree_lookup(&fs_info->buffer_radix, + start >> PAGE_CACHE_SHIFT); + if (eb && atomic_inc_not_zero(&eb->refs)) { + rcu_read_unlock(); + mark_extent_buffer_accessed(eb, NULL); + return eb; } + rcu_read_unlock(); + + return NULL; } -struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, +#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS +struct extent_buffer *alloc_test_extent_buffer(struct btrfs_fs_info *fs_info, + u64 start, unsigned long len) +{ + struct extent_buffer *eb, *exists = NULL; + int ret; + + eb = find_extent_buffer(fs_info, start); + if (eb) + return eb; + eb = alloc_dummy_extent_buffer(start, len); + if (!eb) + return NULL; + eb->fs_info = fs_info; +again: + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + if (ret) + goto free_eb; + spin_lock(&fs_info->buffer_lock); + ret = radix_tree_insert(&fs_info->buffer_radix, + start >> PAGE_CACHE_SHIFT, eb); + spin_unlock(&fs_info->buffer_lock); + radix_tree_preload_end(); + if (ret == -EEXIST) { + exists = find_extent_buffer(fs_info, start); + if (exists) + goto free_eb; + else + goto again; + } + check_buffer_tree_ref(eb); + set_bit(EXTENT_BUFFER_IN_TREE, &eb->bflags); + + /* + * We will free dummy extent buffer's if they come into + * free_extent_buffer with a ref count of 2, but if we are using this we + * want the buffers to stay in memory until we're done with them, so + * bump the ref count again. + */ + atomic_inc(&eb->refs); + return eb; +free_eb: + btrfs_release_extent_buffer(eb); + return exists; +} +#endif + +struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, u64 start, unsigned long len) { unsigned long num_pages = num_extent_pages(start, len); @@ -4228,20 +4665,15 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, struct extent_buffer *eb; struct extent_buffer *exists = NULL; struct page *p; - struct address_space *mapping = tree->mapping; + struct address_space *mapping = fs_info->btree_inode->i_mapping; int uptodate = 1; int ret; - rcu_read_lock(); - eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT); - if (eb && atomic_inc_not_zero(&eb->refs)) { - rcu_read_unlock(); - mark_extent_buffer_accessed(eb); + eb = find_extent_buffer(fs_info, start); + if (eb) return eb; - } - rcu_read_unlock(); - eb = __alloc_extent_buffer(tree, start, len, GFP_NOFS); + eb = __alloc_extent_buffer(fs_info, start, len, GFP_NOFS); if (!eb) return NULL; @@ -4264,7 +4696,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, spin_unlock(&mapping->private_lock); unlock_page(p); page_cache_release(p); - mark_extent_buffer_accessed(exists); + mark_extent_buffer_accessed(exists, p); goto free_eb; } @@ -4279,7 +4711,6 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, attach_extent_buffer_page(eb, p); spin_unlock(&mapping->private_lock); WARN_ON(PageDirty(p)); - mark_page_accessed(p); eb->pages[i] = p; if (!PageUptodate(p)) uptodate = 0; @@ -4296,26 +4727,21 @@ again: if (ret) goto free_eb; - spin_lock(&tree->buffer_lock); - ret = radix_tree_insert(&tree->buffer, start >> PAGE_CACHE_SHIFT, eb); + spin_lock(&fs_info->buffer_lock); + ret = radix_tree_insert(&fs_info->buffer_radix, + start >> PAGE_CACHE_SHIFT, eb); + spin_unlock(&fs_info->buffer_lock); + radix_tree_preload_end(); if (ret == -EEXIST) { - exists = radix_tree_lookup(&tree->buffer, - start >> PAGE_CACHE_SHIFT); - if (!atomic_inc_not_zero(&exists->refs)) { - spin_unlock(&tree->buffer_lock); - radix_tree_preload_end(); - exists = NULL; + exists = find_extent_buffer(fs_info, start); + if (exists) + goto free_eb; + else goto again; - } - spin_unlock(&tree->buffer_lock); - radix_tree_preload_end(); - mark_extent_buffer_accessed(exists); - goto free_eb; } /* add one reference for the tree */ check_buffer_tree_ref(eb); - spin_unlock(&tree->buffer_lock); - radix_tree_preload_end(); + set_bit(EXTENT_BUFFER_IN_TREE, &eb->bflags); /* * there is a race where release page may have @@ -4346,23 +4772,6 @@ free_eb: return exists; } -struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len) -{ - struct extent_buffer *eb; - - rcu_read_lock(); - eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT); - if (eb && atomic_inc_not_zero(&eb->refs)) { - rcu_read_unlock(); - mark_extent_buffer_accessed(eb); - return eb; - } - rcu_read_unlock(); - - return NULL; -} - static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head) { struct extent_buffer *eb = @@ -4372,21 +4781,21 @@ static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head) } /* Expects to have eb->eb_lock already held */ -static int release_extent_buffer(struct extent_buffer *eb, gfp_t mask) +static int release_extent_buffer(struct extent_buffer *eb) { WARN_ON(atomic_read(&eb->refs) == 0); if (atomic_dec_and_test(&eb->refs)) { - if (test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags)) { - spin_unlock(&eb->refs_lock); - } else { - struct extent_io_tree *tree = eb->tree; + if (test_and_clear_bit(EXTENT_BUFFER_IN_TREE, &eb->bflags)) { + struct btrfs_fs_info *fs_info = eb->fs_info; spin_unlock(&eb->refs_lock); - spin_lock(&tree->buffer_lock); - radix_tree_delete(&tree->buffer, + spin_lock(&fs_info->buffer_lock); + radix_tree_delete(&fs_info->buffer_radix, eb->start >> PAGE_CACHE_SHIFT); - spin_unlock(&tree->buffer_lock); + spin_unlock(&fs_info->buffer_lock); + } else { + spin_unlock(&eb->refs_lock); } /* Should be safe to release our pages at this point */ @@ -4401,9 +4810,20 @@ static int release_extent_buffer(struct extent_buffer *eb, gfp_t mask) void free_extent_buffer(struct extent_buffer *eb) { + int refs; + int old; if (!eb) return; + while (1) { + refs = atomic_read(&eb->refs); + if (refs <= 3) + break; + old = atomic_cmpxchg(&eb->refs, refs, refs - 1); + if (old == refs) + return; + } + spin_lock(&eb->refs_lock); if (atomic_read(&eb->refs) == 2 && test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags)) @@ -4419,7 +4839,7 @@ void free_extent_buffer(struct extent_buffer *eb) * I know this is terrible, but it's temporary until we stop tracking * the uptodate bits and such for the extent buffers. */ - release_extent_buffer(eb, GFP_ATOMIC); + release_extent_buffer(eb); } void free_extent_buffer_stale(struct extent_buffer *eb) @@ -4433,7 +4853,7 @@ void free_extent_buffer_stale(struct extent_buffer *eb) if (atomic_read(&eb->refs) == 2 && !extent_buffer_under_io(eb) && test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) atomic_dec(&eb->refs); - release_extent_buffer(eb, GFP_NOFS); + release_extent_buffer(eb); } void clear_extent_buffer_dirty(struct extent_buffer *eb) @@ -4485,17 +4905,6 @@ int set_extent_buffer_dirty(struct extent_buffer *eb) return was_dirty; } -static int range_straddles_pages(u64 start, u64 len) -{ - if (len < PAGE_CACHE_SIZE) - return 1; - if (start & (PAGE_CACHE_SIZE - 1)) - return 1; - if ((start + len) & (PAGE_CACHE_SIZE - 1)) - return 1; - return 0; -} - int clear_extent_buffer_uptodate(struct extent_buffer *eb) { unsigned long i; @@ -4527,37 +4936,6 @@ int set_extent_buffer_uptodate(struct extent_buffer *eb) return 0; } -int extent_range_uptodate(struct extent_io_tree *tree, - u64 start, u64 end) -{ - struct page *page; - int ret; - int pg_uptodate = 1; - int uptodate; - unsigned long index; - - if (range_straddles_pages(start, end - start + 1)) { - ret = test_range_bit(tree, start, end, - EXTENT_UPTODATE, 1, NULL); - if (ret) - return 1; - } - while (start <= end) { - index = start >> PAGE_CACHE_SHIFT; - page = find_get_page(tree->mapping, index); - if (!page) - return 1; - uptodate = PageUptodate(page); - page_cache_release(page); - if (!uptodate) { - pg_uptodate = 0; - break; - } - start += PAGE_CACHE_SIZE; - } - return pg_uptodate; -} - int extent_buffer_uptodate(struct extent_buffer *eb) { return test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); @@ -4620,7 +4998,8 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, ClearPageError(page); err = __extent_read_full_page(tree, page, get_extent, &bio, - mirror_num, &bio_flags); + mirror_num, &bio_flags, + READ | REQ_META); if (err) ret = err; } else { @@ -4629,7 +5008,8 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, } if (bio) { - err = submit_one_bio(READ, bio, mirror_num, bio_flags); + err = submit_one_bio(READ | REQ_META, bio, mirror_num, + bio_flags); if (err) return err; } @@ -4672,7 +5052,7 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv, WARN_ON(start > eb->len); WARN_ON(start + len > eb->start + eb->len); - offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1); + offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1); while (len > 0) { page = extent_buffer_page(eb, i); @@ -4688,6 +5068,43 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv, } } +int read_extent_buffer_to_user(struct extent_buffer *eb, void __user *dstv, + unsigned long start, + unsigned long len) +{ + size_t cur; + size_t offset; + struct page *page; + char *kaddr; + char __user *dst = (char __user *)dstv; + size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1); + unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT; + int ret = 0; + + WARN_ON(start > eb->len); + WARN_ON(start + len > eb->start + eb->len); + + offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1); + + while (len > 0) { + page = extent_buffer_page(eb, i); + + cur = min(len, (PAGE_CACHE_SIZE - offset)); + kaddr = page_address(page); + if (copy_to_user(dst, kaddr + offset, cur)) { + ret = -EFAULT; + break; + } + + dst += cur; + len -= cur; + offset = 0; + i++; + } + + return ret; +} + int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, unsigned long min_len, char **map, unsigned long *map_start, @@ -4714,8 +5131,8 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, if (start + min_len > eb->len) { WARN(1, KERN_ERR "btrfs bad mapping eb start %llu len %lu, " - "wanted %lu %lu\n", (unsigned long long)eb->start, - eb->len, start, min_len); + "wanted %lu %lu\n", + eb->start, eb->len, start, min_len); return -EINVAL; } @@ -4742,7 +5159,7 @@ int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv, WARN_ON(start > eb->len); WARN_ON(start + len > eb->start + eb->len); - offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1); + offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1); while (len > 0) { page = extent_buffer_page(eb, i); @@ -4776,7 +5193,7 @@ void write_extent_buffer(struct extent_buffer *eb, const void *srcv, WARN_ON(start > eb->len); WARN_ON(start + len > eb->start + eb->len); - offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1); + offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1); while (len > 0) { page = extent_buffer_page(eb, i); @@ -4806,7 +5223,7 @@ void memset_extent_buffer(struct extent_buffer *eb, char c, WARN_ON(start > eb->len); WARN_ON(start + len > eb->start + eb->len); - offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1); + offset = (start_offset + start) & (PAGE_CACHE_SIZE - 1); while (len > 0) { page = extent_buffer_page(eb, i); @@ -4837,7 +5254,7 @@ void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src, WARN_ON(src->len != dst_len); offset = (start_offset + dst_offset) & - ((unsigned long)PAGE_CACHE_SIZE - 1); + (PAGE_CACHE_SIZE - 1); while (len > 0) { page = extent_buffer_page(dst, i); @@ -4855,23 +5272,6 @@ void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src, } } -static void move_pages(struct page *dst_page, struct page *src_page, - unsigned long dst_off, unsigned long src_off, - unsigned long len) -{ - char *dst_kaddr = page_address(dst_page); - if (dst_page == src_page) { - memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len); - } else { - char *src_kaddr = page_address(src_page); - char *p = dst_kaddr + dst_off + len; - char *s = src_kaddr + src_off + len; - - while (len--) - *--p = *--s; - } -} - static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len) { unsigned long distance = (src > dst) ? src - dst : dst - src; @@ -4911,21 +5311,21 @@ void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, unsigned long src_i; if (src_offset + len > dst->len) { - printk(KERN_ERR "btrfs memmove bogus src_offset %lu move " + 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(KERN_ERR "btrfs memmove bogus dst_offset %lu move " + 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) { dst_off_in_page = (start_offset + dst_offset) & - ((unsigned long)PAGE_CACHE_SIZE - 1); + (PAGE_CACHE_SIZE - 1); src_off_in_page = (start_offset + src_offset) & - ((unsigned long)PAGE_CACHE_SIZE - 1); + (PAGE_CACHE_SIZE - 1); dst_i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT; src_i = (start_offset + src_offset) >> PAGE_CACHE_SHIFT; @@ -4958,12 +5358,12 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, unsigned long src_i; if (src_offset + len > dst->len) { - printk(KERN_ERR "btrfs memmove bogus src_offset %lu move " + 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(KERN_ERR "btrfs memmove bogus dst_offset %lu move " + printk(KERN_ERR "BTRFS: memmove bogus dst_offset %lu move " "len %lu len %lu\n", dst_offset, len, dst->len); BUG_ON(1); } @@ -4976,13 +5376,13 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, src_i = (start_offset + src_end) >> PAGE_CACHE_SHIFT; dst_off_in_page = (start_offset + dst_end) & - ((unsigned long)PAGE_CACHE_SIZE - 1); + (PAGE_CACHE_SIZE - 1); src_off_in_page = (start_offset + src_end) & - ((unsigned long)PAGE_CACHE_SIZE - 1); + (PAGE_CACHE_SIZE - 1); cur = min_t(unsigned long, len, src_off_in_page + 1); cur = min(cur, dst_off_in_page + 1); - move_pages(extent_buffer_page(dst, dst_i), + copy_pages(extent_buffer_page(dst, dst_i), extent_buffer_page(dst, src_i), dst_off_in_page - cur + 1, src_off_in_page - cur + 1, cur); @@ -4993,7 +5393,7 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, } } -int try_release_extent_buffer(struct page *page, gfp_t mask) +int try_release_extent_buffer(struct page *page) { struct extent_buffer *eb; @@ -5023,9 +5423,6 @@ int try_release_extent_buffer(struct page *page, gfp_t mask) } spin_unlock(&page->mapping->private_lock); - if ((mask & GFP_NOFS) == GFP_NOFS) - mask = GFP_NOFS; - /* * If tree ref isn't set then we know the ref on this eb is a real ref, * so just return, this page will likely be freed soon anyway. @@ -5035,5 +5432,5 @@ int try_release_extent_buffer(struct page *page, gfp_t mask) return 0; } - return release_extent_buffer(eb, mask); + return release_extent_buffer(eb); } |