ext4: use pagevec_lookup_range_tag()
authorJan Kara <jack@suse.cz>
Thu, 16 Nov 2017 01:34:44 +0000 (17:34 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 16 Nov 2017 02:21:03 +0000 (18:21 -0800)
We want only pages from given range in ext4_writepages().  Use
pagevec_lookup_range_tag() instead of pagevec_lookup_tag() and remove
unnecessary code.

Link: http://lkml.kernel.org/r/20171009151359.31984-5-jack@suse.cz
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/ext4/inode.c

index 2633150e41b963f4fc449a02e68819a532018b67..3d0708c91c5a659d2aee04f61ed6ff314f4b901d 100644 (file)
@@ -2620,24 +2620,14 @@ static int mpage_prepare_extent_to_map(struct mpage_da_data *mpd)
        mpd->map.m_len = 0;
        mpd->next_page = index;
        while (index <= end) {
-               nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
-                             min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
+               nr_pages = pagevec_lookup_range_tag(&pvec, mapping, &index, end,
+                               tag, PAGEVEC_SIZE);
                if (nr_pages == 0)
                        goto out;
 
                for (i = 0; i < nr_pages; i++) {
                        struct page *page = pvec.pages[i];
 
-                       /*
-                        * At this point, the page may be truncated or
-                        * invalidated (changing page->mapping to NULL), or
-                        * even swizzled back from swapper_space to tmpfs file
-                        * mapping. However, page->index will not change
-                        * because we have a reference on the page.
-                        */
-                       if (page->index > end)
-                               goto out;
-
                        /*
                         * Accumulated enough dirty pages? This doesn't apply
                         * to WB_SYNC_ALL mode. For integrity sync we have to