TWxSoftware/twx-linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

btrfs: simplify subpage handling of write_one_eb()

Browse Source 
Currently inside write_one_eb() we have two different ways of handling
subpage and regular metadata.

The differences are:

- Extra offset/length calculation when adding the folio range to bio for
  subpage cases
- Only decrease wbc->nr_to_write if the whole page is no longer dirty
  for subpage cases
- Use subpage helper for subpage cases

Merge the tow ways into a shared one:

- Always calculate the to-be-queued range
  So that bio_add_folio() can use the same calculated resulted length
  and offset for both cases.

- Use btrfs_meta_folio_clear_dirty() and
  btrfs_meta_folio_set_writeback() helpers
  This will cover both cases.

- Only decrease wbc->nr_to_write if the folio is no longer dirty
  Since we have the folio locked, no one else can modify the folio dirty
  flags (set_extent_buffer_dirty() will also lock the folio for subpage
  cases).

  Thus after our btrfs_meta_folio_clear_dirty() call, if the whole folio
  is no longer dirty, we're submitting the last dirty eb of the folio,
  and can decrease wbc->nr_to_write properly.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This commit is contained in:
Qu Wenruo
2025-01-29 14:05:26 +10:30
committed by David Sterba
parent 7895817b31
commit 6c6201278e
1 changed files with 13 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fs/btrfs/extent_io.c
+13 -29
View File
@@ -1786,6 +1786,7 @@ static noinline_for_stack void write_one_eb(struct extent_buffer *eb,
{
struct btrfs_fs_info *fs_info = eb->fs_info;
struct btrfs_bio *bbio;
const int num_folios = num_extent_folios(eb);
prepare_eb_write(eb);
@@ -1797,38 +1798,21 @@ static noinline_for_stack void write_one_eb(struct extent_buffer *eb,
wbc_init_bio(wbc, &bbio->bio);
bbio->inode = BTRFS_I(eb->fs_info->btree_inode);
bbio->file_offset = eb->start;
if (btrfs_meta_is_subpage(fs_info)) {
struct folio *folio = eb->folios[0];
bool ret;
for (int i = 0; i < num_folios; i++) {
struct folio *folio = eb->folios[i];
u64 range_start = max_t(u64, eb->start, folio_pos(folio));
u32 range_len = min_t(u64, folio_pos(folio) + folio_size(folio),
eb->start + eb->len) - range_start;
folio_lock(folio);
btrfs_subpage_set_writeback(fs_info, folio, eb->start, eb->len);
if (btrfs_subpage_clear_and_test_dirty(fs_info, folio, eb->start,
eb->len)) {
folio_clear_dirty_for_io(folio);
wbc->nr_to_write--;
}
ret = bio_add_folio(&bbio->bio, folio, eb->len,
eb->start - folio_pos(folio));
ASSERT(ret);
wbc_account_cgroup_owner(wbc, folio, eb->len);
folio_unlock(folio);
} else {
int num_folios = num_extent_folios(eb);
for (int i = 0; i < num_folios; i++) {
struct folio *folio = eb->folios[i];
bool ret;
folio_lock(folio);
folio_clear_dirty_for_io(folio);
folio_start_writeback(folio);
ret = bio_add_folio(&bbio->bio, folio, eb->folio_size, 0);
ASSERT(ret);
wbc_account_cgroup_owner(wbc, folio, eb->folio_size);
btrfs_meta_folio_clear_dirty(fs_info, folio, eb->start, eb->len);
btrfs_meta_folio_set_writeback(fs_info, folio, eb->start, eb->len);
if (!folio_test_dirty(folio))
wbc->nr_to_write -= folio_nr_pages(folio);
folio_unlock(folio);
}
bio_add_folio_nofail(&bbio->bio, folio, range_len,
offset_in_folio(folio, range_start));
wbc_account_cgroup_owner(wbc, folio, range_len);
folio_unlock(folio);
}
btrfs_submit_bbio(bbio, 0);
}