From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from arroyo.ext.ti.com (arroyo.ext.ti.com [192.94.94.40]) by yocto-www.yoctoproject.org (Postfix) with ESMTP id 815C4E0144E for ; Thu, 15 Aug 2013 12:06:03 -0700 (PDT) Received: from dlelxv90.itg.ti.com ([172.17.2.17]) by arroyo.ext.ti.com (8.13.7/8.13.7) with ESMTP id r7FJ61kP030623; Thu, 15 Aug 2013 14:06:01 -0500 Received: from DLEE70.ent.ti.com (dlee70.ent.ti.com [157.170.170.113]) by dlelxv90.itg.ti.com (8.14.3/8.13.8) with ESMTP id r7FJ60bl008083; Thu, 15 Aug 2013 14:06:01 -0500 Received: from dlelxv22.itg.ti.com (172.17.1.197) by DLEE70.ent.ti.com (157.170.170.113) with Microsoft SMTP Server id 14.2.342.3; Thu, 15 Aug 2013 14:06:01 -0500 Received: from localhost (gtudedge.gt.design.ti.com [158.218.102.158]) by dlelxv22.itg.ti.com (8.13.8/8.13.8) with ESMTP id r7FJ60L7015370; Thu, 15 Aug 2013 14:06:01 -0500 Date: Thu, 15 Aug 2013 15:06:00 -0400 From: Denys Dmytriyenko To: Chris Larson Message-ID: <20130815190600.GF2190@edge> References: MIME-Version: 1.0 In-Reply-To: User-Agent: Mutt/1.5.20 (2009-06-14) Cc: "meta-ti@yoctoproject.org" Subject: Re: Stumped. Can't get 2.6.37 TI kernel to boot when built with toolchain from denzil or Linaro gcc-linaro-arm-linux-gnueabihf-4.8-2013.07-1_linux/ X-BeenThere: meta-ti@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: Usage and development list for the meta-ti layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 15 Aug 2013 19:06:07 -0000 Content-Type: text/plain; charset="us-ascii" Content-Disposition: inline On Thu, Aug 15, 2013 at 09:25:38AM -0700, Chris Larson wrote: > On Thu, Aug 15, 2013 at 8:52 AM, Brian Hutchinson wrote: > > > I have a custom TI816x board that uses a modified kernel from the TI PSP. > > Kernel is of the 2.6.37 flavor. I'm trying to migrate to OE Core > > environment and I'm having trouble with getting the kernel to work. > > > > With the toolchain in denzil (gcc version 4.6.4 20120303), kernel will at > > least boot but will die with: > > > > Unable to handle kernel NULL pointer dereference at virtual address > > 00000002 > > pgd = c0004000 > > [00000002] *pgd=00000000 > > Internal error: Oops: 5 [#1] > > > > Any of the more modern toolchains (poky/meta-ti master or Linaro > > (gcc-linaro-arm-linux-gnueabihf-4.8-2013.07-1_linux) are used, the kernel > > won't even make a peep ... nothing is displayed. > > > > I haven't seen the panic, but I have seen a complete silent boot hang when > moving from 4.6 to 4.7 or 4.8 on my OMAP5 EVM board. The gcc-linaro > toolchain worked, as did a 2012.03 sourcery toolchain, but neither the > 2013.05 toolchain nor the internal built yocto toolchain from 1.5/master > resulted in a successful boot. I got past that issue by cherry picking > these two commits from a more recent kernel: > > 455bd4c430b0c0a361f38e8658a0d6cb469942b5 - ARM: 7668/1: fix memset-related > crashes caused by recent GCC (4.7.2) optimizations > 418df63adac56841ef6b0f1fcf435bc64d4ed177 - ARM: 7670/1: fix the memset fix > > See also > http://comments.gmane.org/gmane.comp.handhelds.openembedded.core/39864 > > Presumably you may need more than just that since you're also hitting a > panic with 4.6, but it may be worth trying those as a starting point. Good > luck! Yes, it is often a PITA to get old kernels build and work properly with new toolchains. E.g. you are aware of the alignment issue with 4.7, right? http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-kernel/linux/files/0001-arm-fix-builds-with-gcc-4.7.patch -- Denys