From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lars-Peter Clausen Subject: Re: [PATCH 19/19 v2] mfd/ti_am335x_tscadc: add private lock/unlock function for regmap read/write Date: Wed, 29 May 2013 13:25:39 +0200 Message-ID: <51A5E5B3.80201@metafoo.de> References: <1369681926-22185-1-git-send-email-bigeasy@linutronix.de> <1369681926-22185-20-git-send-email-bigeasy@linutronix.de> <20130529084642.GA18273@linutronix.de> <20130529111253.GR3660@sirena.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20130529111253.GR3660-GFdadSzt00ze9xe1eoZjHA@public.gmane.org> Sender: linux-iio-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Mark Brown , Sebastian Andrzej Siewior Cc: linux-input-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-iio-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Samuel Ortiz , Jonathan Cameron , Dmitry Torokhov , Felipe Balbi List-Id: linux-input@vger.kernel.org On 05/29/2013 01:12 PM, Mark Brown wrote: > On Wed, May 29, 2013 at 10:46:42AM +0200, Sebastian Andrzej Siewior wrote: >> Without this, devm_regmap_init_mmio() creates & uses a simple >> spin_lock() and this should be enough. Within the probe function the >> registers are read and written in process context. Later they are >> accessed from the ISR and lockdep complains because now the lock is >> taken suddenly with IRQs enabled. Currently I don't see any other way to >> keep lockdep quiet than doing this. > > This is not a good place to make this change, there's nothing specific > to the device here and in actual fact there's already a change in the > works from Lars-Peter Clausen converting the spinlock to always use > spin_lock_irqsave() so it should be resolved soon. The thread was on > lkml in the past few days. Yes, this patch shouldn't be necessary in next/master anymore. For the record the patch can be found here: https://patchwork.kernel.org/patch/2609721/ - Lars