From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Howells <dhowells@redhat.com>
Subject: Using __user with compat_uptr_t
Date: Tue, 26 Apr 2005 10:06:12 +0100
Message-ID: <4872.1114506372@redhat.com>
References: <20050417033806.65a5786a.sfr@canb.auug.org.au>  <26687.1113576302@redhat.com>
Cc: akpm@osdl.org, Stephen Rothwell <sfr@canb.auug.org.au>,
	linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Return-path: <linux-kernel-owner+glk-linux-kernel=40m.gmane.org-S261412AbVDZJG4@vger.kernel.org>
In-Reply-To: <20050417033806.65a5786a.sfr@canb.auug.org.au> 
To: torvalds@osdl.org
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org


Hi Linus,

I've added a couple of structures for dealing with 32-bit -> 64-bit upgrade of
NFS4 mounts. They represent the NFS4 mount information provided by userspace
and they contain some pointers to further userspace data. Should these further
userspace pointers be labelled __user?

For example, is this right?:

	struct compat_nfs_string {
		compat_uint_t len;
		compat_uptr_t __user data;
	};

Or is this right?:

	struct compat_nfs_string {
		compat_uint_t len;
		compat_uptr_t data;
	};

Now it makes no difference to the compiler, but it might affect the checker
tool.

David