[PATCH 2/5] smb: client: discard post-EOF pagecache when extending a file via truncate

[Paulo Alcantara <pc@manguebit.org>]

cifs_setsize() relied on pagecache_isize_extended() to zero the tail of
the folio straddling the old EOF, but that helper is a no-op on CIFS
(i_blkbits is 14), so data dirtied past EOF through an mmap survived and
became visible once the file was extended.

Drop that call and lower the truncate_pagecache() boundary to the smaller
of the old and new sizes, which is block-size independent and reliably
discards the pagecache beyond the old EOF.

Fixes: c510edb9734a ("cifs: call pagecache_isize_extended() in cifs_setsize() when extending")
Reviewed-by: David Howells <dhowells@redhat.com>
Signed-off-by: Paulo Alcantara <pc@manguebit.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Namjae Jeon <linkinjeon@kernel.org>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Shyam Prasad N <sprasad@microsoft.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Bharath SM <bharathsm@microsoft.com>
Cc: stable@vger.kernel.org
---
 fs/smb/client/inode.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/fs/smb/client/inode.c b/fs/smb/client/inode.c
index 1fe0ef0a95db..d79a9d80ef34 100644
--- a/fs/smb/client/inode.c
+++ b/fs/smb/client/inode.c
@@ -3071,9 +3071,13 @@ void cifs_setsize(struct inode *inode, loff_t offset)
 		inode->i_blocks = blocks;
 	spin_unlock(&inode->i_lock);
 	inode_set_mtime_to_ts(inode, inode_set_ctime_current(inode));
-	if (offset > old_size)
-		pagecache_isize_extended(inode, old_size, offset);
-	truncate_pagecache(inode, offset);
+	/*
+	 * Drop the pagecache beyond the smaller of the old and new size so that
+	 * data written past the old EOF through an mmap is not exposed once the
+	 * file is extended.  pagecache_isize_extended() can't be used here as
+	 * it is a no-op when the block size is >= PAGE_SIZE, as on CIFS.
+	 */
+	truncate_pagecache(inode, min(offset, old_size));
 	netfs_wait_for_outstanding_io(inode);
 }
 
-- 
2.55.0