From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Mundt Date: Fri, 18 May 2012 08:57:50 +0000 Subject: Re: [PATCH 05/29] regulator: use IS_ERR_OR_NULL() instead of open-coding Message-Id: <20120518085749.GE24355@linux-sh.org> List-Id: References: <1336057558-11031-1-git-send-email-g.liakhovetski@gmx.de> <1336057558-11031-6-git-send-email-g.liakhovetski@gmx.de> <20120503153239.GN3955@opensource.wolfsonmicro.com> In-Reply-To: <20120503153239.GN3955@opensource.wolfsonmicro.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Mark Brown Cc: Guennadi Liakhovetski , linux-mmc@vger.kernel.org, linux-sh@vger.kernel.org On Thu, May 03, 2012 at 04:32:39PM +0100, Mark Brown wrote: > On Thu, May 03, 2012 at 05:05:34PM +0200, Guennadi Liakhovetski wrote: > > > - if (regulator = NULL || IS_ERR(regulator)) > > + if (IS_ERR_OR_NULL(regulator)) > > The bigger question here is why we're accepting NULL in the first place. I'm not sure about the regulator case, but it's been useful to support passing NULL around in the clock framework case. There are plenty of cases where a struct clk is optional and if we fail to find the clock we just set clk to NULL and continue on without having to constantly check the value of the clk pointer, which helps considerably when you consider the number of clk_enable/disable() pairs some drivers have. Presumably the same applies for regulators?