summaryrefslogtreecommitdiff
path: root/fs/btrfs/extent_io.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/btrfs/extent_io.c')
-rw-r--r--fs/btrfs/extent_io.c312
1 files changed, 244 insertions, 68 deletions
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 5e76a474cb7e..4f9893243dae 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -10,6 +10,8 @@
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/pagevec.h>
+#include <linux/prefetch.h>
+#include <linux/cleancache.h>
#include "extent_io.h"
#include "extent_map.h"
#include "compat.h"
@@ -690,6 +692,15 @@ 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.
@@ -940,10 +951,10 @@ static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
}
int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
+ struct extent_state **cached_state, gfp_t mask)
{
- return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
- NULL, mask);
+ return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0,
+ NULL, cached_state, mask);
}
static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
@@ -1012,8 +1023,7 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
mask);
}
-int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end,
- gfp_t mask)
+int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
{
return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL,
mask);
@@ -1433,12 +1443,13 @@ int extent_clear_unlock_delalloc(struct inode *inode,
*/
u64 count_range_bits(struct extent_io_tree *tree,
u64 *start, u64 search_end, u64 max_bytes,
- unsigned long bits)
+ unsigned long bits, int contig)
{
struct rb_node *node;
struct extent_state *state;
u64 cur_start = *start;
u64 total_bytes = 0;
+ u64 last = 0;
int found = 0;
if (search_end <= cur_start) {
@@ -1463,7 +1474,9 @@ u64 count_range_bits(struct extent_io_tree *tree,
state = rb_entry(node, struct extent_state, rb_node);
if (state->start > search_end)
break;
- if (state->end >= cur_start && (state->state & bits)) {
+ if (contig && found && state->start > last + 1)
+ break;
+ if (state->end >= cur_start && (state->state & bits) == bits) {
total_bytes += min(search_end, state->end) + 1 -
max(cur_start, state->start);
if (total_bytes >= max_bytes)
@@ -1472,6 +1485,9 @@ u64 count_range_bits(struct extent_io_tree *tree,
*start = state->start;
found = 1;
}
+ last = state->end;
+ } else if (contig && found) {
+ break;
}
node = rb_next(node);
if (!node)
@@ -1729,6 +1745,9 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
do {
struct page *page = bvec->bv_page;
+ struct extent_state *cached = NULL;
+ struct extent_state *state;
+
tree = &BTRFS_I(page->mapping->host)->io_tree;
start = ((u64)page->index << PAGE_CACHE_SHIFT) +
@@ -1743,9 +1762,20 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
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);
+ }
+ spin_unlock(&tree->lock);
+
if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
ret = tree->ops->readpage_end_io_hook(page, start, end,
- NULL);
+ state);
if (ret)
uptodate = 0;
}
@@ -1758,15 +1788,16 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
test_bit(BIO_UPTODATE, &bio->bi_flags);
if (err)
uptodate = 0;
+ uncache_state(&cached);
continue;
}
}
if (uptodate) {
- set_extent_uptodate(tree, start, end,
+ set_extent_uptodate(tree, start, end, &cached,
GFP_ATOMIC);
}
- unlock_extent(tree, start, end, GFP_ATOMIC);
+ unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
if (whole_page) {
if (uptodate) {
@@ -1805,6 +1836,7 @@ static void end_bio_extent_preparewrite(struct bio *bio, int err)
do {
struct page *page = bvec->bv_page;
+ struct extent_state *cached = NULL;
tree = &BTRFS_I(page->mapping->host)->io_tree;
start = ((u64)page->index << PAGE_CACHE_SHIFT) +
@@ -1815,13 +1847,14 @@ static void end_bio_extent_preparewrite(struct bio *bio, int err)
prefetchw(&bvec->bv_page->flags);
if (uptodate) {
- set_extent_uptodate(tree, start, end, GFP_ATOMIC);
+ set_extent_uptodate(tree, start, end, &cached,
+ GFP_ATOMIC);
} else {
ClearPageUptodate(page);
SetPageError(page);
}
- unlock_extent(tree, start, end, GFP_ATOMIC);
+ unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
} while (bvec >= bio->bi_io_vec);
@@ -1946,6 +1979,7 @@ void set_page_extent_mapped(struct page *page)
static void set_page_extent_head(struct page *page, unsigned long len)
{
+ WARN_ON(!PagePrivate(page));
set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2);
}
@@ -1983,6 +2017,13 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
set_page_extent_mapped(page);
+ if (!PageUptodate(page)) {
+ if (cleancache_get_page(page) == 0) {
+ BUG_ON(blocksize != PAGE_SIZE);
+ goto out;
+ }
+ }
+
end = page_end;
while (1) {
lock_extent(tree, start, end, GFP_NOFS);
@@ -2009,14 +2050,17 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
while (cur <= end) {
if (cur >= last_byte) {
char *userpage;
+ struct extent_state *cached = NULL;
+
iosize = PAGE_CACHE_SIZE - page_offset;
userpage = kmap_atomic(page, KM_USER0);
memset(userpage + page_offset, 0, iosize);
flush_dcache_page(page);
kunmap_atomic(userpage, KM_USER0);
set_extent_uptodate(tree, cur, cur + iosize - 1,
- GFP_NOFS);
- unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+ &cached, GFP_NOFS);
+ unlock_extent_cached(tree, cur, cur + iosize - 1,
+ &cached, GFP_NOFS);
break;
}
em = get_extent(inode, page, page_offset, cur,
@@ -2056,14 +2100,17 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
/* we've found a hole, just zero and go on */
if (block_start == EXTENT_MAP_HOLE) {
char *userpage;
+ struct extent_state *cached = NULL;
+
userpage = kmap_atomic(page, KM_USER0);
memset(userpage + page_offset, 0, iosize);
flush_dcache_page(page);
kunmap_atomic(userpage, KM_USER0);
set_extent_uptodate(tree, cur, cur + iosize - 1,
- GFP_NOFS);
- unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+ &cached, GFP_NOFS);
+ unlock_extent_cached(tree, cur, cur + iosize - 1,
+ &cached, GFP_NOFS);
cur = cur + iosize;
page_offset += iosize;
continue;
@@ -2110,6 +2157,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
cur = cur + iosize;
page_offset += iosize;
}
+out:
if (!nr) {
if (!PageError(page))
SetPageUptodate(page);
@@ -2181,10 +2229,12 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
unsigned long nr_written = 0;
if (wbc->sync_mode == WB_SYNC_ALL)
- write_flags = WRITE_SYNC_PLUG;
+ write_flags = WRITE_SYNC;
else
write_flags = WRITE;
+ trace___extent_writepage(page, inode, wbc);
+
WARN_ON(!PageLocked(page));
pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
if (page->index > end_index ||
@@ -2641,7 +2691,7 @@ int extent_readpages(struct extent_io_tree *tree,
prefetchw(&page->flags);
list_del(&page->lru);
if (!add_to_page_cache_lru(page, mapping,
- page->index, GFP_KERNEL)) {
+ page->index, GFP_NOFS)) {
__extent_read_full_page(tree, page, get_extent,
&bio, 0, &bio_flags);
}
@@ -2780,9 +2830,12 @@ int extent_prepare_write(struct extent_io_tree *tree,
iocount++;
block_start = block_start + iosize;
} else {
- set_extent_uptodate(tree, block_start, cur_end,
+ struct extent_state *cached = NULL;
+
+ set_extent_uptodate(tree, block_start, cur_end, &cached,
GFP_NOFS);
- unlock_extent(tree, block_start, cur_end, GFP_NOFS);
+ unlock_extent_cached(tree, block_start, cur_end,
+ &cached, GFP_NOFS);
block_start = cur_end + 1;
}
page_offset = block_start & (PAGE_CACHE_SIZE - 1);
@@ -2821,9 +2874,17 @@ int try_release_extent_state(struct extent_map_tree *map,
* at this point we can safely clear everything except the
* locked bit and the nodatasum bit
*/
- clear_extent_bit(tree, start, end,
+ ret = clear_extent_bit(tree, start, end,
~(EXTENT_LOCKED | EXTENT_NODATASUM),
0, 0, NULL, mask);
+
+ /* if clear_extent_bit failed for enomem reasons,
+ * we can't allow the release to continue.
+ */
+ if (ret < 0)
+ ret = 0;
+ else
+ ret = 1;
}
return ret;
}
@@ -2903,6 +2964,46 @@ out:
return sector;
}
+/*
+ * helper function for fiemap, which doesn't want to see any holes.
+ * This maps until we find something past 'last'
+ */
+static struct extent_map *get_extent_skip_holes(struct inode *inode,
+ u64 offset,
+ u64 last,
+ get_extent_t *get_extent)
+{
+ u64 sectorsize = BTRFS_I(inode)->root->sectorsize;
+ struct extent_map *em;
+ u64 len;
+
+ if (offset >= last)
+ return NULL;
+
+ while(1) {
+ len = last - offset;
+ if (len == 0)
+ break;
+ len = (len + sectorsize - 1) & ~(sectorsize - 1);
+ em = get_extent(inode, NULL, 0, offset, len, 0);
+ if (!em || IS_ERR(em))
+ return em;
+
+ /* if this isn't a hole return it */
+ if (!test_bit(EXTENT_FLAG_VACANCY, &em->flags) &&
+ em->block_start != EXTENT_MAP_HOLE) {
+ return em;
+ }
+
+ /* this is a hole, advance to the next extent */
+ offset = extent_map_end(em);
+ free_extent_map(em);
+ if (offset >= last)
+ break;
+ }
+ return NULL;
+}
+
int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
__u64 start, __u64 len, get_extent_t *get_extent)
{
@@ -2912,16 +3013,19 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
u32 flags = 0;
u32 found_type;
u64 last;
+ u64 last_for_get_extent = 0;
u64 disko = 0;
+ u64 isize = i_size_read(inode);
struct btrfs_key found_key;
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, em_len = 0;
+ u64 em_start = 0;
+ u64 em_len = 0;
+ u64 em_end = 0;
unsigned long emflags;
- int hole = 0;
if (len == 0)
return -EINVAL;
@@ -2931,6 +3035,10 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
return -ENOMEM;
path->leave_spinning = 1;
+ /*
+ * lookup the last file extent. We're not using i_size here
+ * because there might be preallocation past i_size
+ */
ret = btrfs_lookup_file_extent(NULL, BTRFS_I(inode)->root,
path, inode->i_ino, -1, 0);
if (ret < 0) {
@@ -2944,18 +3052,38 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]);
found_type = btrfs_key_type(&found_key);
- /* No extents, just return */
+ /* No extents, but there might be delalloc bits */
if (found_key.objectid != inode->i_ino ||
found_type != BTRFS_EXTENT_DATA_KEY) {
- btrfs_free_path(path);
- return 0;
+ /* have to trust i_size as the end */
+ last = (u64)-1;
+ last_for_get_extent = isize;
+ } else {
+ /*
+ * remember the start of the last extent. There are a
+ * bunch of different factors that go into the length of the
+ * extent, so its much less complex to remember where it started
+ */
+ last = found_key.offset;
+ last_for_get_extent = last + 1;
}
- last = found_key.offset;
btrfs_free_path(path);
+ /*
+ * we might have some extents allocated but more delalloc past those
+ * extents. so, we trust isize unless the start of the last extent is
+ * beyond isize
+ */
+ if (last < isize) {
+ last = (u64)-1;
+ last_for_get_extent = isize;
+ }
+
lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0,
&cached_state, GFP_NOFS);
- em = get_extent(inode, NULL, 0, off, max - off, 0);
+
+ em = get_extent_skip_holes(inode, off, last_for_get_extent,
+ get_extent);
if (!em)
goto out;
if (IS_ERR(em)) {
@@ -2964,22 +3092,38 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
}
while (!end) {
- hole = 0;
- off = em->start + em->len;
- if (off >= max)
- end = 1;
+ u64 offset_in_extent;
- if (em->block_start == EXTENT_MAP_HOLE) {
- hole = 1;
- goto next;
- }
+ /* break if the extent we found is outside the range */
+ if (em->start >= max || extent_map_end(em) < off)
+ break;
- em_start = em->start;
- em_len = em->len;
+ /*
+ * get_extent may return an extent that starts before our
+ * requested range. We have to make sure the ranges
+ * we return to fiemap always move forward and don't
+ * overlap, so adjust the offsets here
+ */
+ em_start = max(em->start, off);
+ /*
+ * record the offset from the start of the extent
+ * for adjusting the disk offset below
+ */
+ 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;
+ /*
+ * bump off for our next call to get_extent
+ */
+ off = extent_map_end(em);
+ if (off >= max)
+ end = 1;
+
if (em->block_start == EXTENT_MAP_LAST_BYTE) {
end = 1;
flags |= FIEMAP_EXTENT_LAST;
@@ -2990,42 +3134,34 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
flags |= (FIEMAP_EXTENT_DELALLOC |
FIEMAP_EXTENT_UNKNOWN);
} else {
- disko = em->block_start;
+ disko = em->block_start + offset_in_extent;
}
if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
flags |= FIEMAP_EXTENT_ENCODED;
-next:
- emflags = em->flags;
free_extent_map(em);
em = NULL;
- if (!end) {
- em = get_extent(inode, NULL, 0, off, max - off, 0);
- if (!em)
- goto out;
- if (IS_ERR(em)) {
- ret = PTR_ERR(em);
- goto out;
- }
- emflags = em->flags;
- }
-
- if (test_bit(EXTENT_FLAG_VACANCY, &emflags)) {
+ if ((em_start >= last) || em_len == (u64)-1 ||
+ (last == (u64)-1 && isize <= em_end)) {
flags |= FIEMAP_EXTENT_LAST;
end = 1;
}
- if (em_start == last) {
+ /* now scan forward to see if this is really the last extent. */
+ em = get_extent_skip_holes(inode, off, last_for_get_extent,
+ get_extent);
+ if (IS_ERR(em)) {
+ ret = PTR_ERR(em);
+ goto out;
+ }
+ if (!em) {
flags |= FIEMAP_EXTENT_LAST;
end = 1;
}
-
- if (!hole) {
- ret = fiemap_fill_next_extent(fieinfo, em_start, disko,
- em_len, flags);
- if (ret)
- goto out_free;
- }
+ ret = fiemap_fill_next_extent(fieinfo, em_start, disko,
+ em_len, flags);
+ if (ret)
+ goto out_free;
}
out_free:
free_extent_map(em);
@@ -3194,7 +3330,13 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
}
if (!PageUptodate(p))
uptodate = 0;
- unlock_page(p);
+
+ /*
+ * see below about how we avoid a nasty race with release page
+ * and why we unlock later
+ */
+ if (i != 0)
+ unlock_page(p);
}
if (uptodate)
set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
@@ -3218,9 +3360,26 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
atomic_inc(&eb->refs);
spin_unlock(&tree->buffer_lock);
radix_tree_preload_end();
+
+ /*
+ * there is a race where release page may have
+ * tried to find this extent buffer in the radix
+ * but failed. It will tell the VM it is safe to
+ * reclaim the, and it will clear the page private bit.
+ * We must make sure to set the page private bit properly
+ * after the extent buffer is in the radix tree so
+ * it doesn't get lost
+ */
+ set_page_extent_mapped(eb->first_page);
+ set_page_extent_head(eb->first_page, eb->len);
+ if (!page0)
+ unlock_page(eb->first_page);
return eb;
free_eb:
+ if (eb->first_page && !page0)
+ unlock_page(eb->first_page);
+
if (!atomic_dec_and_test(&eb->refs))
return exists;
btrfs_release_extent_buffer(eb);
@@ -3271,10 +3430,11 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree,
continue;
lock_page(page);
+ WARN_ON(!PagePrivate(page));
+
+ set_page_extent_mapped(page);
if (i == 0)
set_page_extent_head(page, eb->len);
- else
- set_page_private(page, EXTENT_PAGE_PRIVATE);
clear_page_dirty_for_io(page);
spin_lock_irq(&page->mapping->tree_lock);
@@ -3341,7 +3501,7 @@ int set_extent_buffer_uptodate(struct extent_io_tree *tree,
num_pages = num_extent_pages(eb->start, eb->len);
set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
- GFP_NOFS);
+ NULL, GFP_NOFS);
for (i = 0; i < num_pages; i++) {
page = extent_buffer_page(eb, i);
if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
@@ -3464,6 +3624,13 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
for (i = start_i; i < num_pages; i++) {
page = extent_buffer_page(eb, i);
+
+ WARN_ON(!PagePrivate(page));
+
+ set_page_extent_mapped(page);
+ if (i == 0)
+ set_page_extent_head(page, eb->len);
+
if (inc_all_pages)
page_cache_get(page);
if (!PageUptodate(page)) {
@@ -3569,6 +3736,7 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
"wanted %lu %lu\n", (unsigned long long)eb->start,
eb->len, start, min_len);
WARN_ON(1);
+ return -EINVAL;
}
p = extent_buffer_page(eb, i);
@@ -3761,6 +3929,12 @@ static void move_pages(struct page *dst_page, struct page *src_page,
kunmap_atomic(dst_kaddr, KM_USER0);
}
+static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len)
+{
+ unsigned long distance = (src > dst) ? src - dst : dst - src;
+ return distance < len;
+}
+
static void copy_pages(struct page *dst_page, struct page *src_page,
unsigned long dst_off, unsigned long src_off,
unsigned long len)
@@ -3768,10 +3942,12 @@ static void copy_pages(struct page *dst_page, struct page *src_page,
char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
char *src_kaddr;
- if (dst_page != src_page)
+ if (dst_page != src_page) {
src_kaddr = kmap_atomic(src_page, KM_USER1);
- else
+ } else {
src_kaddr = dst_kaddr;
+ BUG_ON(areas_overlap(src_off, dst_off, len));
+ }
memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
kunmap_atomic(dst_kaddr, KM_USER0);
@@ -3846,7 +4022,7 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
"len %lu len %lu\n", dst_offset, len, dst->len);
BUG_ON(1);
}
- if (dst_offset < src_offset) {
+ if (!areas_overlap(src_offset, dst_offset, len)) {
memcpy_extent_buffer(dst, dst_offset, src_offset, len);
return;
}
generated by cgit v1.2.3 (git 2.39.0) at 2024-12-03 01:03:37 +0300