On 2/14/22 12:00, Matthew Wilcox (Oracle) wrote:
invalidate_inode_page() is the only caller of invalidate_complete_page()
and inlining it reveals that the first check is unnecessary (because we
hold the page locked, and we just retrieved the mapping from the page).
I just noticed this yesterday, when reviewing Rik's page poisoning fix.
I had a patch for it squirreled away, but I missed the point about
removing that extraneous mapping check. Glad you spotted it.
...
@@ -309,7 +288,10 @@ int invalidate_inode_page(struct page *page)
It would be nice to retain some of the original comments. May I suggest
this (it has an additional paragraph) for an updated version of comments
above invalidate_inode_page():
/*
* Safely invalidate one page from its pagecache mapping.
* It only drops clean, unused pages. The page must be locked.
*
* This function can be called at any time, and is not supposed to throw away
* dirty pages. But pages can be marked dirty at any time too, so use
* remove_mapping(), which safely discards clean, unused pages.
*
* Returns 1 if the page is successfully invalidated, otherwise 0.
*/
Also, as long as you're there, a newline after the mapping declaration
would bring this routine into compliance with that convention.
hmmm, now I wonder why this isn't a boolean function. And I think the
reason is that it's quite old.
Either way, looks good:
Reviewed-by: John Hubbard <jhubbard@xxxxxxxxxx>
thanks,
--
John Hubbard
NVIDIA
return 0;
if (page_mapped(page))
return 0;
- return invalidate_complete_page(mapping, page);
+ if (page_has_private(page) && !try_to_release_page(page, 0))
+ return 0;
+
+ return remove_mapping(mapping, page);
}
/**
@@ -584,7 +566,7 @@ void invalidate_mapping_pagevec(struct address_space *mapping,
}
/*
- * This is like invalidate_complete_page(), except it ignores the page's
+ * This is like invalidate_inode_page(), except it ignores the page's
* refcount. We do this because invalidate_inode_pages2() needs stronger
* invalidation guarantees, and cannot afford to leave pages behind because
* shrink_page_list() has a temp ref on them, or because they're transiently
