From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754883Ab3FTHoI (ORCPT ); Thu, 20 Jun 2013 03:44:08 -0400 Received: from www.linutronix.de ([62.245.132.108]:59234 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753613Ab3FTHoF (ORCPT ); Thu, 20 Jun 2013 03:44:05 -0400 Date: Thu, 20 Jun 2013 09:44:05 +0200 (CEST) From: Thomas Gleixner To: Chen Gang cc: "linux-kernel@vger.kernel.org" Subject: Re: [PATCH] kernel/itimer.c: for return value, using -EINVAL instead of -EFAULT In-Reply-To: <51C2ACB8.6050701@asianux.com> Message-ID: References: <51C29FBC.1090507@asianux.com> <51C2ACB8.6050701@asianux.com> User-Agent: Alpine 2.02 (DEB 1266 2009-07-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Linutronix-Spam-Score: -1.0 X-Linutronix-Spam-Level: - X-Linutronix-Spam-Status: No , -1.0 points, 5.0 required, ALL_TRUSTED=-1,SHORTCIRCUIT=-0.0001 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 20 Jun 2013, Chen Gang wrote: > On 06/20/2013 02:59 PM, Thomas Gleixner wrote: > > On Thu, 20 Jun 2013, Chen Gang wrote: > > > >> > For the system call getitimer(), if the parameter 'value' is NULL, need > >> > return -EINVAL, not -EFAULT. > > Care to explain why? Because you are feeling so? > > > > I am not feeling so, the original implementation really just checks the parameter 'value', if it is invalid, need return, is it incorrect ?? > > > > I recommend reading the man page of getitimer: > > > > ERRORS > > EFAULT new_value, old_value, or curr_value is not valid a pointer. > > > > And NULL is definitely NOT a valid pointer. > > > > The Posix spec does not specify an explicit error value for this > > syscall, but the general policy is: > > > > [EFAULT] > > Bad address. The system detected an invalid address in attempting > > to use an argument of a call. The reliable detection of this error > > cannot be guaranteed, and when not detected may result in the > > generation of a signal, indicating an address violation, which is > > sent to the process. > > > > And we made use of this, which is correct and makes sense. > > > > Returning EINVAL makes no sense at all, because EINVAL _IS_ a > > specified error code for this syscall: > > > > [EINVAL] > > The which argument is not recognized. > > That means we need not check the parameter 'value' out side of copy_to_user(). We could do that, but that makes no sense. If we can detect it before copy_to_user() we can return the exactly same return value which we would return via copy_to_user(). That avoids to take a trap and run through the fixup code. Thanks, tglx _