[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20200507192933.GK89269@dtor-ws>
Date: Thu, 7 May 2020 12:29:33 -0700
From: Dmitry Torokhov <dmitry.torokhov@...il.com>
To: ChenTao <chentao107@...wei.com>
Cc: kstewart@...uxfoundation.org, tglx@...utronix.de,
gregkh@...uxfoundation.org, allison@...utok.net,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] input/misc/drv260x: Remove a useless comparison
Hi,
On Thu, May 07, 2020 at 01:16:17PM +0800, ChenTao wrote:
> Fix the following warning:
>
> 'mode' and 'library' are u32, they are never be negative,
> DRV260X_LRA_MODE and DRV260X_LIB_EMPTY are 0x00, the comparison
> is always false.
The fact that the symbolic names resolve to 0 is just a happenstance. We
should not be writing code based on concrete values in this case. I
wish we could tell the compiler to ignore this kind of cases, where we
want to test a range of values, and do not care about concrete numeric
bounds of said range.
>
> drivers/input/misc/drv260x.c:478:20: warning:
> comparison of unsigned expression < 0 is always false [-Wtype-limits]
> if (haptics->mode < DRV260X_LRA_MODE ||
> drivers/input/misc/drv260x.c:490:23: warning:
> comparison of unsigned expression < 0 is always false [-Wtype-limits]
> if (haptics->library < DRV260X_LIB_EMPTY ||
Are we planning on enabling this warning in kernel by default?
Thanks.
--
Dmitry
Powered by blists - more mailing lists