[PATCH] ext3: use generic_open_file to fix possible preemption bugs

Ext3 is currently using a duplicate version of generic_open_file, and this
should be fixed, before it gets out of sync.

In fact, it *has* got out of sync.

Apart some cosmetic changes (which are not a problem), note that it
directly reads inode->i_size, while the generic version uses i_size_read().

I'm not sure if this bug can actually be triggered, but here follows a
possible scenario: when

- a file is brought into inode cache
- that file is either truncated or extended (i.e.  inode->i_size is changed)
  and after this it changes it's "more than 2Gb" property,
- and process B opens that file in the meanwhile,

I suppose the size check can be hurted by a race condition, doesn't it?

Similar checks should probably be done for other FS's. Also, ext2 does not
have this problem - someone fixed this problem in ext2 but not ext3? How?
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade_spam@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

fs/ext3/file.c

@@ -43,20 +43,6 @@ static int ext3_release_file (struct inode * inode, struct file * filp)
 	return 0;
 }
 
-/*
- * Called when an inode is about to be opened.
- * We use this to disallow opening RW large files on 32bit systems if
- * the caller didn't specify O_LARGEFILE.  On 64bit systems we force
- * on this flag in sys_open.
- */
-static int ext3_open_file (struct inode *inode, struct file *filp)
-{
-	if (!(filp->f_flags & O_LARGEFILE) &&
-	    inode->i_size > 0x7FFFFFFFLL)
-		return -EFBIG;
-	return 0;
-}
-
 static ssize_t
 ext3_file_write(struct kiocb *iocb, const char __user *buf, size_t count, loff_t pos)
 {
@@ -125,7 +111,7 @@ struct file_operations ext3_file_operations = {
 	.writev		= generic_file_writev,
 	.ioctl		= ext3_ioctl,
 	.mmap		= generic_file_mmap,
-	.open		= ext3_open_file,
+	.open		= generic_file_open,
 	.release	= ext3_release_file,
 	.fsync		= ext3_sync_file,
 	.sendfile	= generic_file_sendfile,