fs: Turn page_offset() into a wrapper around folio_pos()

This is far less efficient for the lagging filesystems which still
use page_offset(), but it removes an access to page->index.  It also
fixes a bug -- if any filesystem passed a tail page to page_offset(),
it would return garbage which might result in the filesystem choosing
to not writeback a dirty page.  There probably aren't any examples
of this, but I can't be certain.

Signed-off-by: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Link: https://lore.kernel.org/r/20250221203932.3588740-1-willy@infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
This commit is contained in:
Matthew Wilcox (Oracle) 2025-02-21 20:39:29 +00:00 committed by Christian Brauner
parent d1c735d44c
commit 12851bd921
No known key found for this signature in database
GPG Key ID: 91C61BC06578DCA2

View File

@ -1044,21 +1044,23 @@ static inline pgoff_t page_pgoff(const struct folio *folio,
return folio->index + folio_page_idx(folio, page);
}
/**
* folio_pos - Returns the byte position of this folio in its file.
* @folio: The folio.
*/
static inline loff_t folio_pos(const struct folio *folio)
{
return ((loff_t)folio->index) * PAGE_SIZE;
}
/*
* Return byte-offset into filesystem object for page.
*/
static inline loff_t page_offset(struct page *page)
{
return ((loff_t)page->index) << PAGE_SHIFT;
}
struct folio *folio = page_folio(page);
/**
* folio_pos - Returns the byte position of this folio in its file.
* @folio: The folio.
*/
static inline loff_t folio_pos(struct folio *folio)
{
return page_offset(&folio->page);
return folio_pos(folio) + folio_page_idx(folio, page) * PAGE_SIZE;
}
/*