From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: Regmap regression in next caused by of_iomap change Date: Thu, 08 Sep 2016 21:59:20 +0200 Message-ID: <4088076.eNNe5WgnxT@wuerfel> References: <20160908184111.vj3sb2ik6ydehkag@atomide.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20160908184111.vj3sb2ik6ydehkag@atomide.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=m.gmane.org@lists.infradead.org To: linux-arm-kernel@lists.infradead.org Cc: Tony Lindgren , Andrey Smirnov , linux-omap@vger.kernel.org, Lee Jones List-Id: linux-omap@vger.kernel.org On Thursday, September 8, 2016 11:41:12 AM CEST Tony Lindgren wrote: > Hi, > > Looks like commit 39de2c4275a9 ("mfd: syscon: Make use of of_iomap") > makes at least MMC PBIAS regulator stop working on omap3. > > That's probably because of_syscon_register uses the now unintialized > struct resource res to set syscon_config.max_register. Looks correct. Lee, please send a revert of that patch to Linus along with any other fixes you may have. > I have a strange deja vu feeling that we've been through this > before See commit ca668f0edfae ("mfd: syscon: Set regmap max_register > in of_syscon_register") > > Maybe we should add comments there instead? I wonder why gcc didn't detect this case. I usually do my test builds with -Wmaybe-uninitialized that should have caught it but didn't. Arnd