From mboxrd@z Thu Jan 1 00:00:00 1970 From: jon-hunter@ti.com (Jon Hunter) Date: Mon, 15 Apr 2013 15:00:31 -0500 Subject: [PATCH 3/5] gpio/omap: Add DT support to GPIO driver In-Reply-To: <516C3076.3000107@wwwdotorg.org> References: <1329321854-24490-1-git-send-email-b-cousson@ti.com> <512D3AB1.1080202@wwwdotorg.org> <512D3EC2.6050408@ti.com> <20130302200524.D230F3E1571@localhost> <51391F41.5000303@ti.com> <514C79E1.4090106@wwwdotorg.org> <514CE0AB.6060207@ti.com> <515319D5.20105@wwwdotorg.org> <5155C902.7080207@wwwdotorg.org> <5165CB9D.1090202@wwwdotorg.org> <51671D7B.5060303@wwwdotorg.org> <51673D70.3010503@wwwdotorg.org> <516C3076.3000107@wwwdotor! g.org> Message-ID: <516C5C5F.4060106@ti.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 04/15/2013 11:53 AM, Stephen Warren wrote: > On 04/13/2013 07:35 PM, Javier Martinez Canillas wrote: > ... >> Is the following inlined patch [1] what you were thinking that would >> be the right approach? > ... >> diff --git a/drivers/gpio/gpio-omap.c b/drivers/gpio/gpio-omap.c > ... >> +static int omap_gpio_irq_request(struct irq_data *d) >> +{ >> + struct gpio_bank *bank = irq_data_get_irq_chip_data(d); >> + >> + return omap_gpio_request(&bank->chip, d->hwirq); > > If you want the GPIO usage to be known to the GPIO subsystem, then > wouldn't you call gpio_request() here rather than omap_gpio_request()? > The above code will certainly do enough so that the OMAP GPIO HW is > fully enabled as you need, but I thought the idea was to also prevent > some other code successfully running gpio_request() on that same GPIO? Also, although omap gpios default to being inputs, we should not assume that. So may be you should call gpio_request_one() here passing as flags GPIOF_IN to configure as an input. Cheers Jon