From mboxrd@z Thu Jan 1 00:00:00 1970 From: Russell King - ARM Linux Subject: Re: [PATCH] tty/serial: digicolor: Fix bad usage of IS_ERR_VALUE Date: Thu, 11 Feb 2016 10:46:50 +0000 Message-ID: <20160211104650.GU10826@n2100.arm.linux.org.uk> References: <1455030539-10798-1-git-send-email-linux@roeck-us.net> <4571887.rTPndCnv20@wuerfel> <56BAA27A.9050605@roeck-us.net> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: <56BAA27A.9050605-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org> Sender: kernel-testers-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Guenter Roeck Cc: Arnd Bergmann , linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, Baruch Siach , Greg Kroah-Hartman , linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Andrzej Hajda , linux-serial-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Jiri Slaby , kernel-testers-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-serial@vger.kernel.org On Tue, Feb 09, 2016 at 06:37:46PM -0800, Guenter Roeck wrote: > On 02/09/2016 07:26 AM, Arnd Bergmann wrote: > >On Tuesday 09 February 2016 07:08:59 Guenter Roeck wrote: > >>IS_ERR_VALUE() assumes that its parameter is an unsigned long. > >>It can not be used to check if an unsigned int reflects an error. > >>Doing so can result in the following build warning. > >> > >>drivers/tty/serial/digicolor-usart.c: In function =E2=80=98digicolo= r_uart_probe=E2=80=99: > >>include/linux/err.h:21:38: warning: > >> comparison is always false due to limited range of data ty= pe > >>drivers/tty/serial/digicolor-usart.c:485:6: note: > >> in expansion of macro =E2=80=98IS_ERR_VALUE=E2=80=99 > >> > >>If that warning is seen, an error return from platform_get_irq() is= missed. > >> > >> > > > >The patch looks correct to me, but what compiler version and which k= ernel > >tree is it that triggered the warning? > > > >Andrzej Hajda just modified the definition of IS_ERR_VALUE(), and th= e > >changes are still under discussion, but I don't see that warning wit= h > >any of the versions. > > > I see it with gcc 5.1 and 5.2 (and W=3D1). I did not see / notice And= rzej's patch. >=20 > I agree that fixing the problem in IS_ERR_VALUE() is preferrable. I disagree. What happens if (eg) you decide to do this: u8 irq; irq =3D platform_get_irq(...); if (IS_ERR_VALUE(irq)) ... is that the fault of IS_ERR_VALUE() that it will never be true? No, it= 's a programming error in the caller, because the caller is using the wron= g type here - in fact, you can't do anything in IS_ERR_VALUE() to correct that. The same is true if you assign an error value to an unsigned int: the problem is in the caller, not in IS_ERR_VALUE(). What would be preferable is if there was some way IS_ERR_VALUE() could detect if it's used on something that isn't at least a short, and is signed, but I suspect coccinelle can do that much better than trying to work out some way to get the C compiler to check that. --=20 RMK's Patch system: http://www.arm.linux.org.uk/developer/patches/ =46TTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up according to speedtest.net.