From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [RFC PATCH (resend)] block layer zero-copy: missing access_ok() check Date: Thu, 21 Mar 2013 17:33:13 -0400 (EDT) Message-ID: <20130321.173313.827411684773319604.davem@davemloft.net> References: <20130315171000.GA2342@Krystal> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Return-path: Received: from shards.monkeyblade.net ([149.20.54.216]:47336 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753461Ab3CUVdQ (ORCPT ); Thu, 21 Mar 2013 17:33:16 -0400 In-Reply-To: Sender: linux-arch-owner@vger.kernel.org List-ID: To: torvalds@linux-foundation.org Cc: mathieu.desnoyers@efficios.com, linux-arch@vger.kernel.org, axboe@kernel.dk, security@kernel.org, gregkh@linuxfoundation.org, viro@zeniv.linux.org.uk, npiggin@kernel.dk From: Linus Torvalds Date: Fri, 15 Mar 2013 10:21:41 -0700 > Adding linux-arch. Guys, can you check your architectures? ... >> * sparc: access_ok missing in get_user_pages_fast, >> -> no indication of any _PAGE_USER flag. Well, access_ok() returns 1 unconditionally on sparc64. I can add a call there if you want. :-) And that's because the user virtual address space is %100 segregated from the privileged one. There are no "user address ranges" vs. kernel ones. For example, we can and do give 64-bit processes the full 64-bit virtual address space for user mappings. The page tables, outside of swapper_pg_dir, only map user pages.