From mboxrd@z Thu Jan 1 00:00:00 1970 From: Casper Lyngesen Mogensen Subject: RE: OMAP 3630 random crashes Date: Tue, 1 Jul 2014 10:35:57 +0000 Message-ID: <1404210956640.67615@grundfos.com> References: <1403254242990.39117@grundfos.com> <20140620091431.GA28884@atomide.com> <1403272840480.35753@grundfos.com> <1403274413159.44797@grundfos.com> <1404135826925.10774@grundfos.com> ,<20140701075437.GM28884@atomide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from mail-db3lp0080.outbound.protection.outlook.com ([213.199.154.80]:8970 "EHLO emea01-db3-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751425AbaGAKf7 convert rfc822-to-8bit (ORCPT ); Tue, 1 Jul 2014 06:35:59 -0400 In-Reply-To: <20140701075437.GM28884@atomide.com> Content-Language: en-US Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Tony Lindgren , Michael Trimarchi Cc: Ash Charles , "linux-omap@vger.kernel.org" Hi You are right. I used a different bootloader for the newer kernels, because of another build system and did not think about it. The u-boot from github/gumstix sets the memory timings different than the u-boot directly from denx.de. I switched to the gumstix-uboot and i can now allocate all the memory in the system with all peripherals enabled, and i just get killed by the OOM a some point. But thanks for the help so far, even though it was not a kernel issue :-) Med venlig hilsen/Best regards Casper Mogensen ________________________________________ From: Tony Lindgren Sent: Tuesday, July 01, 2014 09:54 To: Michael Trimarchi Cc: Casper Lyngesen Mogensen; Ash Charles; linux-omap@vger.kernel.org Subject: Re: OMAP 3630 random crashes * Michael Trimarchi [140630 06:49]: > > On Mon, Jun 30, 2014 at 3:43 PM, Casper Lyngesen Mogensen > wrote: > > > > I got two new Overo AirStorm today, and they have same behavior. I tried disabling > > the SMSC9xxx again and booted from MMC. This seems to make the backtrace a > > little more consistent(Tried on both new and old). There is (almost) always something > > about Alignment trap, the crash log itself differs every time. Also i get to allocate a lot > > more memory before crash with these settings. > > What kernel version are you using? We should have v3.15 working fine, but PM features are not merged in util for v3.16. I would try dding calls to omap_sdrc_init to pdata-quirks.c the same way board-over.c is doing: omap_sdrc_init(mt46h32m32lf6_sdrc_params, mt46h32m32lf6_sdrc_params); If the memory timings are not correctly set by the bootloader that could explain the issues. Regards, Tony