| Message ID | 20231214053035.1018876-1-willy@infradead.org |
|---|---|
| State | Awaiting Upstream |
| Headers | show |
| Series | ext4: Convert ext4_da_do_write_end() to take a folio | expand |
On Thu 14-12-23 05:30:35, Matthew Wilcox (Oracle) wrote: > There's nothing page-specific happening in ext4_da_do_write_end(); > it's merely used for its refcount & lock, both of which are folio > properties. Saves four calls to compound_head(). > > Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> Looks good. Feel free to add: Reviewed-by: Jan Kara <jack@suse.cz> Honza > --- > fs/ext4/inode.c | 19 ++++++++++--------- > 1 file changed, 10 insertions(+), 9 deletions(-) > > diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c > index 83ee4e0f46f4..216ad9bcca45 100644 > --- a/fs/ext4/inode.c > +++ b/fs/ext4/inode.c > @@ -2947,7 +2947,7 @@ static int ext4_da_should_update_i_disksize(struct folio *folio, > > static int ext4_da_do_write_end(struct address_space *mapping, > loff_t pos, unsigned len, unsigned copied, > - struct page *page) > + struct folio *folio) > { > struct inode *inode = mapping->host; > loff_t old_size = inode->i_size; > @@ -2958,12 +2958,13 @@ static int ext4_da_do_write_end(struct address_space *mapping, > * block_write_end() will mark the inode as dirty with I_DIRTY_PAGES > * flag, which all that's needed to trigger page writeback. > */ > - copied = block_write_end(NULL, mapping, pos, len, copied, page, NULL); > + copied = block_write_end(NULL, mapping, pos, len, copied, > + &folio->page, NULL); > new_i_size = pos + copied; > > /* > - * It's important to update i_size while still holding page lock, > - * because page writeout could otherwise come in and zero beyond > + * It's important to update i_size while still holding folio lock, > + * because folio writeout could otherwise come in and zero beyond > * i_size. > * > * Since we are holding inode lock, we are sure i_disksize <= > @@ -2981,14 +2982,14 @@ static int ext4_da_do_write_end(struct address_space *mapping, > > i_size_write(inode, new_i_size); > end = (new_i_size - 1) & (PAGE_SIZE - 1); > - if (copied && ext4_da_should_update_i_disksize(page_folio(page), end)) { > + if (copied && ext4_da_should_update_i_disksize(folio, end)) { > ext4_update_i_disksize(inode, new_i_size); > disksize_changed = true; > } > } > > - unlock_page(page); > - put_page(page); > + folio_unlock(folio); > + folio_put(folio); > > if (old_size < pos) > pagecache_isize_extended(inode, old_size, pos); > @@ -3027,10 +3028,10 @@ static int ext4_da_write_end(struct file *file, > return ext4_write_inline_data_end(inode, pos, len, copied, > folio); > > - if (unlikely(copied < len) && !PageUptodate(page)) > + if (unlikely(copied < len) && !folio_test_uptodate(folio)) > copied = 0; > > - return ext4_da_do_write_end(mapping, pos, len, copied, &folio->page); > + return ext4_da_do_write_end(mapping, pos, len, copied, folio); > } > > /* > -- > 2.42.0 > >
On Thu, 14 Dec 2023 05:30:35 +0000, Matthew Wilcox (Oracle) wrote: > There's nothing page-specific happening in ext4_da_do_write_end(); > it's merely used for its refcount & lock, both of which are folio > properties. Saves four calls to compound_head(). > > Applied, thanks! [1/1] ext4: Convert ext4_da_do_write_end() to take a folio commit: 05f240655f03fd7fcdbfc6129b6fb7dcc3f64e0e Best regards,
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 83ee4e0f46f4..216ad9bcca45 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2947,7 +2947,7 @@ static int ext4_da_should_update_i_disksize(struct folio *folio, static int ext4_da_do_write_end(struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, - struct page *page) + struct folio *folio) { struct inode *inode = mapping->host; loff_t old_size = inode->i_size; @@ -2958,12 +2958,13 @@ static int ext4_da_do_write_end(struct address_space *mapping, * block_write_end() will mark the inode as dirty with I_DIRTY_PAGES * flag, which all that's needed to trigger page writeback. */ - copied = block_write_end(NULL, mapping, pos, len, copied, page, NULL); + copied = block_write_end(NULL, mapping, pos, len, copied, + &folio->page, NULL); new_i_size = pos + copied; /* - * It's important to update i_size while still holding page lock, - * because page writeout could otherwise come in and zero beyond + * It's important to update i_size while still holding folio lock, + * because folio writeout could otherwise come in and zero beyond * i_size. * * Since we are holding inode lock, we are sure i_disksize <= @@ -2981,14 +2982,14 @@ static int ext4_da_do_write_end(struct address_space *mapping, i_size_write(inode, new_i_size); end = (new_i_size - 1) & (PAGE_SIZE - 1); - if (copied && ext4_da_should_update_i_disksize(page_folio(page), end)) { + if (copied && ext4_da_should_update_i_disksize(folio, end)) { ext4_update_i_disksize(inode, new_i_size); disksize_changed = true; } } - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); if (old_size < pos) pagecache_isize_extended(inode, old_size, pos); @@ -3027,10 +3028,10 @@ static int ext4_da_write_end(struct file *file, return ext4_write_inline_data_end(inode, pos, len, copied, folio); - if (unlikely(copied < len) && !PageUptodate(page)) + if (unlikely(copied < len) && !folio_test_uptodate(folio)) copied = 0; - return ext4_da_do_write_end(mapping, pos, len, copied, &folio->page); + return ext4_da_do_write_end(mapping, pos, len, copied, folio); } /*
There's nothing page-specific happening in ext4_da_do_write_end(); it's merely used for its refcount & lock, both of which are folio properties. Saves four calls to compound_head(). Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> --- fs/ext4/inode.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)