Btrfs: allocate the free space by the existed max extent size when ENOSPC
By the current code, if the requested size is very large, and all the extents
in the free space cache are small, we will waste lots of the cpu time to cut
the requested size in half and search the cache again and again until it gets
down to the size the allocator can return. In fact, we can know the max extent
size in the cache after the first search, so we needn't cut the size in half
repeatedly, and just use the max extent size directly. This way can save
lots of cpu time and make the performance grow up when there are only fragments
in the free space cache.
According to my test, if there are only 4KB free space extents in the fs,
and the total size of those extents are 256MB, we can reduce the execute
time of the following test from 5.4s to 1.4s.
dd if=/dev/zero of=<testfile> bs=1MB count=1 oflag=sync
Changelog v2 -> v3:
- fix the problem that we skip the block group with the space which is
less than we need.
Changelog v1 -> v2:
- address the problem that we return a wrong start position when searching
the free space in a bitmap.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index ef3bea7..4f419ba 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -1433,13 +1433,19 @@
ctl->free_space += bytes;
}
+/*
+ * If we can not find suitable extent, we will use bytes to record
+ * the size of the max extent.
+ */
static int search_bitmap(struct btrfs_free_space_ctl *ctl,
struct btrfs_free_space *bitmap_info, u64 *offset,
u64 *bytes)
{
unsigned long found_bits = 0;
+ unsigned long max_bits = 0;
unsigned long bits, i;
unsigned long next_zero;
+ unsigned long extent_bits;
i = offset_to_bit(bitmap_info->offset, ctl->unit,
max_t(u64, *offset, bitmap_info->offset));
@@ -1448,9 +1454,12 @@
for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) {
next_zero = find_next_zero_bit(bitmap_info->bitmap,
BITS_PER_BITMAP, i);
- if ((next_zero - i) >= bits) {
- found_bits = next_zero - i;
+ extent_bits = next_zero - i;
+ if (extent_bits >= bits) {
+ found_bits = extent_bits;
break;
+ } else if (extent_bits > max_bits) {
+ max_bits = extent_bits;
}
i = next_zero;
}
@@ -1461,38 +1470,41 @@
return 0;
}
+ *bytes = (u64)(max_bits) * ctl->unit;
return -1;
}
+/* Cache the size of the max extent in bytes */
static struct btrfs_free_space *
find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
- unsigned long align)
+ unsigned long align, u64 *max_extent_size)
{
struct btrfs_free_space *entry;
struct rb_node *node;
- u64 ctl_off;
u64 tmp;
u64 align_off;
int ret;
if (!ctl->free_space_offset.rb_node)
- return NULL;
+ goto out;
entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
if (!entry)
- return NULL;
+ goto out;
for (node = &entry->offset_index; node; node = rb_next(node)) {
entry = rb_entry(node, struct btrfs_free_space, offset_index);
- if (entry->bytes < *bytes)
+ if (entry->bytes < *bytes) {
+ if (entry->bytes > *max_extent_size)
+ *max_extent_size = entry->bytes;
continue;
+ }
/* make sure the space returned is big enough
* to match our requested alignment
*/
if (*bytes >= align) {
- ctl_off = entry->offset - ctl->start;
- tmp = ctl_off + align - 1;;
+ tmp = entry->offset - ctl->start + align - 1;
do_div(tmp, align);
tmp = tmp * align + ctl->start;
align_off = tmp - entry->offset;
@@ -1501,14 +1513,22 @@
tmp = entry->offset;
}
- if (entry->bytes < *bytes + align_off)
+ if (entry->bytes < *bytes + align_off) {
+ if (entry->bytes > *max_extent_size)
+ *max_extent_size = entry->bytes;
continue;
+ }
if (entry->bitmap) {
- ret = search_bitmap(ctl, entry, &tmp, bytes);
+ u64 size = *bytes;
+
+ ret = search_bitmap(ctl, entry, &tmp, &size);
if (!ret) {
*offset = tmp;
+ *bytes = size;
return entry;
+ } else if (size > *max_extent_size) {
+ *max_extent_size = size;
}
continue;
}
@@ -1517,7 +1537,7 @@
*bytes = entry->bytes - align_off;
return entry;
}
-
+out:
return NULL;
}
@@ -2118,7 +2138,8 @@
}
u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
- u64 offset, u64 bytes, u64 empty_size)
+ u64 offset, u64 bytes, u64 empty_size,
+ u64 *max_extent_size)
{
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
struct btrfs_free_space *entry = NULL;
@@ -2129,7 +2150,7 @@
spin_lock(&ctl->tree_lock);
entry = find_free_space(ctl, &offset, &bytes_search,
- block_group->full_stripe_len);
+ block_group->full_stripe_len, max_extent_size);
if (!entry)
goto out;
@@ -2139,7 +2160,6 @@
if (!entry->bytes)
free_bitmap(ctl, entry);
} else {
-
unlink_free_space(ctl, entry);
align_gap_len = offset - entry->offset;
align_gap = entry->offset;
@@ -2153,7 +2173,6 @@
else
link_free_space(ctl, entry);
}
-
out:
spin_unlock(&ctl->tree_lock);
@@ -2208,7 +2227,8 @@
static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
struct btrfs_free_cluster *cluster,
struct btrfs_free_space *entry,
- u64 bytes, u64 min_start)
+ u64 bytes, u64 min_start,
+ u64 *max_extent_size)
{
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
int err;
@@ -2220,8 +2240,11 @@
search_bytes = bytes;
err = search_bitmap(ctl, entry, &search_start, &search_bytes);
- if (err)
+ if (err) {
+ if (search_bytes > *max_extent_size)
+ *max_extent_size = search_bytes;
return 0;
+ }
ret = search_start;
__bitmap_clear_bits(ctl, entry, ret, bytes);
@@ -2236,7 +2259,7 @@
*/
u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
struct btrfs_free_cluster *cluster, u64 bytes,
- u64 min_start)
+ u64 min_start, u64 *max_extent_size)
{
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
struct btrfs_free_space *entry = NULL;
@@ -2256,6 +2279,9 @@
entry = rb_entry(node, struct btrfs_free_space, offset_index);
while(1) {
+ if (entry->bytes < bytes && entry->bytes > *max_extent_size)
+ *max_extent_size = entry->bytes;
+
if (entry->bytes < bytes ||
(!entry->bitmap && entry->offset < min_start)) {
node = rb_next(&entry->offset_index);
@@ -2269,7 +2295,8 @@
if (entry->bitmap) {
ret = btrfs_alloc_from_bitmap(block_group,
cluster, entry, bytes,
- cluster->window_start);
+ cluster->window_start,
+ max_extent_size);
if (ret == 0) {
node = rb_next(&entry->offset_index);
if (!node)