ramfs: fix nommu mmap with gaps in the page cache

[ Upstream commit 50b7d85 ]

ramfs needs to check that pages are both physically contiguous and
contiguous in the file.  If the page cache happens to have, eg, page A for
index 0 of the file, no page for index 1, and page A+1 for index 2, then
an mmap of the first two pages of the file will succeed when it should
fail.

Fixes: 642fb4d ("[PATCH] NOMMU: Provide shared-writable mmap support on ramfs")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: David Howells <dhowells@redhat.com>
Link: https://lkml.kernel.org/r/20200914122239.GO6583@casper.infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

fs/ramfs/file-nommu.c

@@ -224,7 +224,7 @@ static unsigned long ramfs_nommu_get_unmapped_area(struct file *file,
 	if (!pages)
 		goto out_free;
 
-	nr = find_get_pages(inode->i_mapping, &pgoff, lpages, pages);
+	nr = find_get_pages_contig(inode->i_mapping, pgoff, lpages, pages);
 	if (nr != lpages)
 		goto out_free_pages; /* leave if some pages were missing */