From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755035AbaIDSIY (ORCPT ); Thu, 4 Sep 2014 14:08:24 -0400 Received: from mail-wi0-f174.google.com ([209.85.212.174]:50771 "EHLO mail-wi0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755006AbaIDSIW (ORCPT ); Thu, 4 Sep 2014 14:08:22 -0400 Message-ID: <5408AA8C.7040100@linaro.org> Date: Thu, 04 Sep 2014 19:08:12 +0100 From: Srinivas Kandagatla User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0 MIME-Version: 1.0 To: Pramod Gurav , linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, kernel@stlinux.com CC: Srinivas Kandagatla , Srinivas Kandagatla , Maxime Coquelin , Patrice Chotard , Linus Walleij , Bjorn Andersson Subject: Re: [PATCH] pinctrl: st: Add remove function and remove gpio_chip on failure References: <1409409935-8775-1-git-send-email-pramod.gurav@smartplayin.com> In-Reply-To: <1409409935-8775-1-git-send-email-pramod.gurav@smartplayin.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Pramod, sorry for delay in reply as I was travelling, still in Jet lag. > Signed-off-by: Pramod Gurav > --- > drivers/pinctrl/pinctrl-st.c | 25 +++++++++++++++++++++++++ > 1 files changed, 25 insertions(+), 0 deletions(-) > > diff --git a/drivers/pinctrl/pinctrl-st.c b/drivers/pinctrl/pinctrl-st.c > index 5475374..9296845 100644 > --- a/drivers/pinctrl/pinctrl-st.c > +++ b/drivers/pinctrl/pinctrl-st.c > @@ -1517,6 +1517,7 @@ static int st_gpiolib_register_bank(struct st_pinctrl *info, > 0, handle_simple_irq, > IRQ_TYPE_LEVEL_LOW); > if (err) { > + gpiochip_remove(&bank->gpio_chip); This change-set looks good. IMO, you can send a patch for this change set. > dev_info(dev, "could not add irqchip\n"); > return err; > } > @@ -1685,6 +1686,29 @@ static int st_pctl_probe(struct platform_device *pdev) > return 0; > } > > +static int st_pctl_remove(struct platform_device *pdev) > +{ Ideally this driver will not be removed, as other drivers depend on this, even the serial. so I see no big achievement in adding the remove functionality, as this is going to be a dead code and would never be tested. > + struct st_pinctrl *info = platform_get_drvdata(pdev); > + struct device_node *np = pdev->dev.of_node; > + struct device_node *child; > + struct gpio_chip gpio_chip; > + int bank = 0; > + > + if (info->nbanks) { > + for_each_child_of_node(np, child) { > + if (of_property_read_bool(child, "gpio-controller")) { We should not re-parse the DT nodes once we are done with it in the probe. > + gpio_chip = info->banks[bank].gpio_chip; > + gpiochip_remove(&gpio_chip); > + bank++; > + } > + } > + } > + I think the logic is very simple: while (nbanks--) gpiochip_remove(&info->banks[bank++].gpio_chip)) thanks, srini > + pinctrl_unregister(info->pctl); > + > + return 0; > +} > + > static struct platform_driver st_pctl_driver = { > .driver = { > .name = "st-pinctrl", > @@ -1692,6 +1716,7 @@ static struct platform_driver st_pctl_driver = { > .of_match_table = st_pctl_of_match, > }, > .probe = st_pctl_probe, > + .remove = st_pctl_remove, > }; > > static int __init st_pctl_init(void) >