From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934130AbeD0PAv (ORCPT ); Fri, 27 Apr 2018 11:00:51 -0400 Received: from mail-qk0-f171.google.com ([209.85.220.171]:37960 "EHLO mail-qk0-f171.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933698AbeD0OFq (ORCPT ); Fri, 27 Apr 2018 10:05:46 -0400 X-Google-Smtp-Source: AB8JxZqgxv1tosq5XPJ9crb/Xt4b97aaYbh+jYJn3nvD0Hnvzz6tqKP+B9ex5ELBs7fafVxA79P63A== Date: Fri, 27 Apr 2018 10:05:42 -0400 From: Vivek Unune To: =?utf-8?B?UmFmYcWCIE1pxYJlY2tp?= Cc: Florian Fainelli , Hauke Mehrtens , Jon Mason , bcm-kernel-feedback-list , Rob Herring , Mark Rutland , Russell King , "linux-arm-kernel@lists.infradead.org" , devicetree@vger.kernel.org, Linux Kernel Mailing List Subject: Re: [PATCH] ARM: BCM5301X: Fix NAND ECC parameters for Linksys Panamera Message-ID: <20180427140542.c3t5taztkciimfny@osboxes> References: <20180310171244.70299-1-npcomplete13@gmail.com> <20180311100313.znonhesgh462rz65@osboxes> <2ef6e4fc-cce7-372e-e35f-578fd6ada9ca@gmail.com> <20180313010210.ws4dugy3rrpqwzmk@osboxes> <00df9ee9-2bd2-ad2b-fdf8-682efdb9a4d1@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <00df9ee9-2bd2-ad2b-fdf8-682efdb9a4d1@gmail.com> User-Agent: NeoMutt/20170113 (1.7.2) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Apr 25, 2018 at 11:02:04AM +0200, Rafał Miłecki wrote: > > [ 13.530863] ubi0 warning: 0xc037c9e4: error -74 (ECC error) while reading 64 bytes from PEB 115:0, read only 64 bytes, retry > > [ 13.542763] ubi0 warning: 0xc037c9e4: error -74 (ECC error) while reading 64 bytes from PEB 115:0, read only 64 bytes, retry > > [ 13.554659] ubi0 warning: 0xc037c9e4: error -74 (ECC error) while reading 64 bytes from PEB 115:0, read only 64 bytes, retry > > [ 13.566548] ubi0 error: 0xc037ca08: error -74 (ECC error) while reading 64 bytes from PEB 115:0, read 64 bytes > > [ 13.576574] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.14.25 #0 > > [ 13.582589] Hardware name: BCM5301X > > [ 13.586078] Backtrace: > > [ 13.588526] Function entered at [] from [] > > [ 13.594368] r7:00000000 r6:60000013 r5:00000000 r4:c081d988 > > [ 13.600032] Function entered at [] from [] > > [ 13.605872] Function entered at [] from [] > > [ 13.611714] r7:00000000 r6:00000040 r5:c7954000 r4:ffffffb6 > > [ 13.617378] Function entered at [] from [] > > [ 13.623223] r10:00000000 r9:00000040 r8:00000073 r7:c7954000 r6:c7979000 r5:c7036180 > > [ 13.631071] r4:c7b72900 > > [ 13.633604] Function entered at [] from [] > > [ 13.639447] r10:00000000 r9:c7979800 r8:c7954000 r7:c7979000 r6:00000073 r5:c7036180 > > [ 13.647286] r4:c7b72900 > > [ 13.649819] Function entered at [] from [] > > [ 13.655663] r10:c7954040 r9:00000840 r8:c702e400 r7:00000000 r6:c702e400 r5:c7954000 > > [ 13.663510] r4:fffff800 > > [ 13.666042] Function entered at [] from [] > > [ 13.671886] r10:c0813a88 r9:c0651c18 r8:c0651bec r7:c084ff48 r6:c081e440 r5:c7823ea4 > > [ 13.679725] r4:c702e400 > > [ 13.682258] Function entered at [] from [] > > [ 13.688103] r10:c072e83c r9:c07366b8 r8:c0717954 r7:c072e838 r6:c081e440 r5:00000000 > > [ 13.695949] r4:c0803c08 > > [ 13.698482] Function entered at [] from [] > > [ 13.704327] r9:c07366b8 r8:c0700594 r7:c072e838 r6:c081e440 r5:c081e440 r4:00000008 > > [ 13.712087] Function entered at [] from [] > > [ 13.717930] r10:00000000 r9:00000000 r8:00000000 r7:00000000 r6:00000000 r5:c0506acc > > [ 13.725769] r4:00000000 > > [ 13.728303] Function entered at [] from [] > > [ 13.734144] r5:c0506acc r4:00000000 > > This is expected if you previously were using firmware programming flash > in non-BCH8 mode. The main reason for what you're seeing is that when > flashing image with CFE it doesn't erase all NAND pages. > > If previously you were using firmware programming flash in BCH1 mode, it > means that after flashing firmware using BCH8 you'll end up with mess in > ECC. > > Part of NAND flash programmed by CFE will use BCH8. > Part of NAND flash that remained from previous fw will use BCH1. > Linux will fail to read part of NAND flash from prev installation. > > The cleanest solution: erase all NAND content of firmware partition > before flashing firmware using BCH8. > > The easiest solution: ignore all these "error -74 (ECC error) while > reading" errors (they may last for few minutes!). I believe UBI should > simply recreate all pages it couldn't access and once you wait long > enough, all your NAND flash will be using BCH8. Let me build the firmware with BCH-8 and wait it out. I'll report back my results. Thanks, Vivek