From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from 3.mo5.mail-out.ovh.net ([46.105.40.108] helo=mo5.mail-out.ovh.net) by merlin.infradead.org with esmtp (Exim 4.80.1 #2 (Red Hat Linux)) id 1UZSOW-0004T2-Jc for barebox@lists.infradead.org; Mon, 06 May 2013 20:52:25 +0000 Received: from mail414.ha.ovh.net (b6.ovh.net [213.186.33.56]) by mo5.mail-out.ovh.net (Postfix) with SMTP id 5F9F7FF9621 for ; Mon, 6 May 2013 22:52:02 +0200 (CEST) Date: Mon, 6 May 2013 22:47:39 +0200 From: Jean-Christophe PLAGNIOL-VILLARD Message-ID: <20130506204739.GD30509@game.jcrosoft.org> References: <1367852182-28870-1-git-send-email-dev@lynxeye.de> <20130506152725.GP13393@game.jcrosoft.org> <20130506195903.GJ32299@pengutronix.de> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <20130506195903.GJ32299@pengutronix.de> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "barebox" Errors-To: barebox-bounces+u.kleine-koenig=pengutronix.de@lists.infradead.org Subject: Re: [PATCH 1/2] tegra20: add pinctrl driver To: Sascha Hauer Cc: barebox@lists.infradead.org On 21:59 Mon 06 May , Sascha Hauer wrote: > On Mon, May 06, 2013 at 05:27:25PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote: > > > + > > > +static struct pinctrl_ops pinctrl_tegra20_ops = { > > > + .set_state = pinctrl_tegra20_set_state, > > > +}; > > > + > > > +static int pinctrl_tegra20_probe(struct device_d *dev) > > > +{ > > > + struct pinctrl_tegra20 *ctrl; > > > + int i, ret = 0; > > no need the init of ret > > > + > > > + ctrl = xzalloc(sizeof(*ctrl)); > > > + > > > + /* > > > + * Tegra pincontrol is split out into four independent memory ranges: > > > + * tristate control, function mux, pullup/down control, pad control > > > + * (from lowest to highest hardware address). > > > + * We are only interested in the first three for now. > > > + */ > > > + for (i = 0; i <= 2; i++) { > > > + ctrl->regs[i] = dev_request_mem_region(dev, i); > > The return values of dev_request_mem_region should be checked for new > drivers. I know 98% of the code currently does not do this and it was a > mistake. The above may fail if regions overlap or there are other bugs > in the devicetree. Having this silently fail means that the actual > register accesses do a NULL pointer deref later. why not simply put a BUG in the request_mem > > > > + } > > drop {} > > > > but use a specific name will simplify debug > > I don't think this should be mandatory. the issue is that a number can be easly endup to a hiden typo (no compile error) where a ttypo in the name will no compile. > > > > > and make the code easier > > > + > > > + ctrl->pinctrl.dev = dev; > > > + ctrl->pinctrl.ops = &pinctrl_tegra20_ops; > > > + > > > + ret = pinctrl_register(&ctrl->pinctrl); > > > + if (ret) > > > + free(ctrl); > > report an error message > > The core will answer with an error message here. There is no need to > litter the code with error messages which only add to the binary size. > The above may fail for a developer but really should not fail for a user. > A developer has enough information with the probe functions error code. pinctrl_register print no error message it's good to known why we failled no just that the probe return -Exxx and this will not impact the binary size if we put a dev_warn or dev_debug as we can drop them with the printk level _______________________________________________ barebox mailing list barebox@lists.infradead.org http://lists.infradead.org/mailman/listinfo/barebox