From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Holler Subject: Re: [PATCH v4 1/2] gpio/omap: don't create an IRQ mapping for every GPIO on DT Date: Sun, 28 Jul 2013 14:59:21 +0200 Message-ID: <51F515A9.9010005@ahsoftware.de> References: <1372433223-9053-1-git-send-email-javier.martinez@collabora.co.uk> <51F4F973.8000303@ahsoftware.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from h1446028.stratoserver.net ([85.214.92.142]:40815 "EHLO mail.ahsoftware.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753529Ab3G1NA6 (ORCPT ); Sun, 28 Jul 2013 09:00:58 -0400 Received: from eiche.ahsoftware (p57B21DD0.dip0.t-ipconnect.de [87.178.29.208]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by mail.ahsoftware.de (Postfix) with ESMTPSA id 693B0423C272 for ; Sun, 28 Jul 2013 15:00:56 +0200 (CEST) In-Reply-To: Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Linus Walleij Cc: ext Tony Lindgren , Grant Likely , "santosh.shilimkar@ti.com" , Kevin Hilman , Javier Martinez Canillas , Jon Hunter , Jean-Christophe PLAGNIOL-VILLARD , Enric Balletbo Serra , Linux-OMAP , Florian Vaussard , Aaro Koskinen , Balaji T K Am 28.07.2013 13:14, schrieb Linus Walleij: > On Sun, Jul 28, 2013 at 12:58 PM, Alexander Holler wrote: >> Am 28.06.2013 17:27, schrieb Javier Martinez Canillas: >> >>> When a GPIO is defined as an interrupt line using Device >>> Tree, a call to irq_create_of_mapping() is made that calls >>> irq_create_mapping(). So, is not necessary to do the mapping >>> for all OMAP GPIO lines and explicitly call irq_create_mapping() >>> on the driver probe() when booting with Device Tree. >>> >>> Add a custom IRQ domain .map function handler that will be >>> called by irq_create_mapping() to map the GPIO lines used as IRQ. >>> This also allows to execute needed setup code such as configuring >>> a GPIO as input and enabling the GPIO bank. >> >> >> This patch basically broke every usage of >> >> irq = gpio_to_irq(gpio); >> request[_threaded]_irq(irq, ...); >> >> because request[_threaded]_irq(irq, ...) now fails because of a missing >> irq_domain (no mapping => no domain). > > OK and I had ACKs from Santosh and even a Test-by tag from Enric > on that thing. > > What shall we do with this mess now? > > I was told that these patches really needed to be applied because > they were fixing a regression in v3.11, and now it seems they are > causing another regression. > > What I need to know is what is worst: having these three patches > there or reverting them? > > Or should I simply revert *all* the TI GPIO stuff merged for v3.11 > now that is seems like this is a can of worms? :-/ > > Tony, Kevin, Santosh, someone? What will make all happy? Hmm, maybe something which adds a mapping for a IRQ when gpio_to_irq() is called would help. I've just reverted the 3 patches on that topic because I was to lazy to find out how to add a mapping for an irq in gpio_to_irq(). And just in case, I've only seen a breakage with that cd-signal for MMC and nothing else. But I haven't really tested 3.11-rc much and only just added that cd-gpio to my DT which is the reason I've seen the problem. And my stripped down kernel doesn't seem to use a gpio-irq somewhere else. But nevertheless, I assume it is a problem. Regards, Alexander Holler