From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Mon, 7 Feb 2005 21:15:47 +0100 From: Andi Kleen Subject: Re: [patch 19/24] TASK_SIZE is variable. Message-ID: <20050207201547.GA6790@wotan.suse.de> References: <20050205090619.C30866@flint.arm.linux.org.uk> <20050205154448.1cc89ee9.davem@davemloft.net> <20050206105047.GB6819@wotan.suse.de> <20050206131947.6524ece9.davem@davemloft.net> <20050206213148.GI18245@wotan.suse.de> <20050206133110.528ce713.davem@davemloft.net> <20050206215020.GJ18245@wotan.suse.de> <20050206143110.59dc039f.davem@davemloft.net> <20050207081106.GA16035@wotan.suse.de> <20050207112837.5594c314.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20050207112837.5594c314.davem@davemloft.net> To: "David S. Miller" Cc: Andi Kleen , rmk@arm.linux.org.uk, torvalds@osdl.org, akpm@osdl.org, dwmw2@infradead.org, linux-arch@vger.kernel.org List-ID: On Mon, Feb 07, 2005 at 11:28:37AM -0800, David S. Miller wrote: > On Mon, 7 Feb 2005 09:11:06 +0100 > Andi Kleen wrote: > > > Also BTW your proposed check wouldn't catch all cases anyways, > > it would only handle the case where the access_ok() check is > > done inside the KERNEL_DS. > > There must be a verify area done (either via explicit call or > via the user access macros which do not have the "__" prefix) > for each range of "userspace" accesses done through the uaccess.h accessors. Yes, but it is done outside KERNEL_DS (otherwise it is a security hole) And then later in KERNEL_DS there is no verify_area. So this change wouldn't help you at all because it would rarely if ever catch any bugs. -Andi