summaryrefslogtreecommitdiffstats
path: root/fs/btrfs/file.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/btrfs/file.c')
-rw-r--r--fs/btrfs/file.c165
1 files changed, 130 insertions, 35 deletions
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 4205ba752d4..a005fe2c072 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -309,10 +309,6 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
ret = PTR_ERR(inode_root);
goto cleanup;
}
- if (btrfs_root_refs(&inode_root->root_item) == 0) {
- ret = -ENOENT;
- goto cleanup;
- }
key.objectid = defrag->ino;
btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
@@ -1317,6 +1313,56 @@ fail:
}
+static noinline int check_can_nocow(struct inode *inode, loff_t pos,
+ size_t *write_bytes)
+{
+ struct btrfs_trans_handle *trans;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_ordered_extent *ordered;
+ u64 lockstart, lockend;
+ u64 num_bytes;
+ int ret;
+
+ lockstart = round_down(pos, root->sectorsize);
+ lockend = lockstart + round_up(*write_bytes, root->sectorsize) - 1;
+
+ while (1) {
+ lock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+ ordered = btrfs_lookup_ordered_range(inode, lockstart,
+ lockend - lockstart + 1);
+ if (!ordered) {
+ break;
+ }
+ unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+ btrfs_start_ordered_extent(inode, ordered, 1);
+ btrfs_put_ordered_extent(ordered);
+ }
+
+ trans = btrfs_join_transaction(root);
+ if (IS_ERR(trans)) {
+ unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+ return PTR_ERR(trans);
+ }
+
+ num_bytes = lockend - lockstart + 1;
+ ret = can_nocow_extent(trans, inode, lockstart, &num_bytes, NULL, NULL,
+ NULL);
+ btrfs_end_transaction(trans, root);
+ if (ret <= 0) {
+ ret = 0;
+ } else {
+ clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
+ EXTENT_DIRTY | EXTENT_DELALLOC |
+ EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, 0, 0,
+ NULL, GFP_NOFS);
+ *write_bytes = min_t(size_t, *write_bytes, num_bytes);
+ }
+
+ unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend);
+
+ return ret;
+}
+
static noinline ssize_t __btrfs_buffered_write(struct file *file,
struct iov_iter *i,
loff_t pos)
@@ -1324,10 +1370,12 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
struct inode *inode = file_inode(file);
struct btrfs_root *root = BTRFS_I(inode)->root;
struct page **pages = NULL;
+ u64 release_bytes = 0;
unsigned long first_index;
size_t num_written = 0;
int nrptrs;
int ret = 0;
+ bool only_release_metadata = false;
bool force_page_uptodate = false;
nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) /
@@ -1348,6 +1396,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
offset);
size_t num_pages = (write_bytes + offset +
PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+ size_t reserve_bytes;
size_t dirty_pages;
size_t copied;
@@ -1362,11 +1411,41 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
break;
}
- ret = btrfs_delalloc_reserve_space(inode,
- num_pages << PAGE_CACHE_SHIFT);
+ reserve_bytes = num_pages << PAGE_CACHE_SHIFT;
+ ret = btrfs_check_data_free_space(inode, reserve_bytes);
+ if (ret == -ENOSPC &&
+ (BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW |
+ BTRFS_INODE_PREALLOC))) {
+ ret = check_can_nocow(inode, pos, &write_bytes);
+ if (ret > 0) {
+ only_release_metadata = true;
+ /*
+ * our prealloc extent may be smaller than
+ * write_bytes, so scale down.
+ */
+ num_pages = (write_bytes + offset +
+ PAGE_CACHE_SIZE - 1) >>
+ PAGE_CACHE_SHIFT;
+ reserve_bytes = num_pages << PAGE_CACHE_SHIFT;
+ ret = 0;
+ } else {
+ ret = -ENOSPC;
+ }
+ }
+
if (ret)
break;
+ ret = btrfs_delalloc_reserve_metadata(inode, reserve_bytes);
+ if (ret) {
+ if (!only_release_metadata)
+ btrfs_free_reserved_data_space(inode,
+ reserve_bytes);
+ break;
+ }
+
+ release_bytes = reserve_bytes;
+
/*
* This is going to setup the pages array with the number of
* pages we want, so we don't really need to worry about the
@@ -1375,11 +1454,8 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
ret = prepare_pages(root, file, pages, num_pages,
pos, first_index, write_bytes,
force_page_uptodate);
- if (ret) {
- btrfs_delalloc_release_space(inode,
- num_pages << PAGE_CACHE_SHIFT);
+ if (ret)
break;
- }
copied = btrfs_copy_from_user(pos, num_pages,
write_bytes, pages, i);
@@ -1409,30 +1485,46 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
* managed to copy.
*/
if (num_pages > dirty_pages) {
+ release_bytes = (num_pages - dirty_pages) <<
+ PAGE_CACHE_SHIFT;
if (copied > 0) {
spin_lock(&BTRFS_I(inode)->lock);
BTRFS_I(inode)->outstanding_extents++;
spin_unlock(&BTRFS_I(inode)->lock);
}
- btrfs_delalloc_release_space(inode,
- (num_pages - dirty_pages) <<
- PAGE_CACHE_SHIFT);
+ if (only_release_metadata)
+ btrfs_delalloc_release_metadata(inode,
+ release_bytes);
+ else
+ btrfs_delalloc_release_space(inode,
+ release_bytes);
}
+ release_bytes = dirty_pages << PAGE_CACHE_SHIFT;
if (copied > 0) {
ret = btrfs_dirty_pages(root, inode, pages,
dirty_pages, pos, copied,
NULL);
if (ret) {
- btrfs_delalloc_release_space(inode,
- dirty_pages << PAGE_CACHE_SHIFT);
btrfs_drop_pages(pages, num_pages);
break;
}
}
+ release_bytes = 0;
btrfs_drop_pages(pages, num_pages);
+ if (only_release_metadata && copied > 0) {
+ u64 lockstart = round_down(pos, root->sectorsize);
+ u64 lockend = lockstart +
+ (dirty_pages << PAGE_CACHE_SHIFT) - 1;
+
+ set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart,
+ lockend, EXTENT_NORESERVE, NULL,
+ NULL, GFP_NOFS);
+ only_release_metadata = false;
+ }
+
cond_resched();
balance_dirty_pages_ratelimited(inode->i_mapping);
@@ -1445,6 +1537,13 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
kfree(pages);
+ if (release_bytes) {
+ if (only_release_metadata)
+ btrfs_delalloc_release_metadata(inode, release_bytes);
+ else
+ btrfs_delalloc_release_space(inode, release_bytes);
+ }
+
return num_written ? num_written : ret;
}
@@ -2175,12 +2274,6 @@ static long btrfs_fallocate(struct file *file, int mode,
goto out_reserve_fail;
}
- /*
- * wait for ordered IO before we have any locks. We'll loop again
- * below with the locks held.
- */
- btrfs_wait_ordered_range(inode, alloc_start, alloc_end - alloc_start);
-
mutex_lock(&inode->i_mutex);
ret = inode_newsize_ok(inode, alloc_end);
if (ret)
@@ -2191,8 +2284,23 @@ static long btrfs_fallocate(struct file *file, int mode,
alloc_start);
if (ret)
goto out;
+ } else {
+ /*
+ * If we are fallocating from the end of the file onward we
+ * need to zero out the end of the page if i_size lands in the
+ * middle of a page.
+ */
+ ret = btrfs_truncate_page(inode, inode->i_size, 0, 0);
+ if (ret)
+ goto out;
}
+ /*
+ * wait for ordered IO before we have any locks. We'll loop again
+ * below with the locks held.
+ */
+ btrfs_wait_ordered_range(inode, alloc_start, alloc_end - alloc_start);
+
locked_end = alloc_end - 1;
while (1) {
struct btrfs_ordered_extent *ordered;
@@ -2425,20 +2533,7 @@ static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int whence)
}
}
- if (offset < 0 && !(file->f_mode & FMODE_UNSIGNED_OFFSET)) {
- offset = -EINVAL;
- goto out;
- }
- if (offset > inode->i_sb->s_maxbytes) {
- offset = -EINVAL;
- goto out;
- }
-
- /* Special lock needed here? */
- if (offset != file->f_pos) {
- file->f_pos = offset;
- file->f_version = 0;
- }
+ offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
out:
mutex_unlock(&inode->i_mutex);
return offset;
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-12-11 18:24:16 +0000