From mboxrd@z Thu Jan 1 00:00:00 1970 From: Felipe Balbi Subject: Re: [RFC/PATCH 2/4] input: misc: twl4030: move to request_threaded_irq Date: Mon, 14 Dec 2009 13:40:09 +0200 Message-ID: <20091214114009.GI12526@nokia.com> References: <20091212183410.GF3092@sirena.org.uk> <1260786654-13294-3-git-send-email-felipe.balbi@nokia.com> Reply-To: felipe.balbi@nokia.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: "ext Shilimkar, Santosh" Cc: "Balbi Felipe (Nokia-D/Helsinki)" , "linux-kernel@vger.kernel.org" , Linux OMAP Mailing List , Tony Lindgren , "Koskinen Aaro (Nokia-D/Helsinki)" , David Brownell , Linux USB Mailing List , Anton Vorontsov , Grazvydas Ignotas , "Chikkature Rajashekar, Madhusudhan" , Greg Kroah-Hartman , Mark Brown , Samuel Ortiz List-Id: linux-omap@vger.kernel.org Hi, On Mon, Dec 14, 2009 at 12:31:11PM +0100, ext Shilimkar, Santosh wrote: >In whole of the series the ISR you have converted to threads using threaded_irq are very >small in size. They are like quick_change_handlers. So only advantage is the particular >interrupt is masked for bit longer than with you change. > >I might be wrong here but it might introduce the spurious IRQ's because >of bit of delay in the processing.What is the motive for this change ? >Are we using this API as it suppose to be ? the irq chip is connected through i2c and that mandate us to handle irqs in thread context. Now that we have kernel-wise api for that, we're just moving towards it instead of definint our own stuff and the lockdep shortcuts we had to put before. I boot tested these patches on a 3430-based board and it seems to be working fine. Did a few tests with usb only. I can see I get the usb presence irq correctly. -- balbi