From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-eopbgr40139.outbound.protection.outlook.com ([40.107.4.139] helo=EUR03-DB5-obe.outbound.protection.outlook.com) by bombadil.infradead.org with esmtps (Exim 4.87 #1 (Red Hat Linux)) id 1cdF9G-0005Cy-7S for linux-mtd@lists.infradead.org; Mon, 13 Feb 2017 11:50:28 +0000 Subject: Re: UBI/UBIFS corruptions during random power-cuts To: Richard Weinberger , References: <0500de05-b3d7-6cbc-0236-5d4c83a5f00f@nod.at> <961abad6-7d46-4427-fb2b-97477a4a6a10@toradex.com> <2be57b2b-8fad-a987-8257-18f8ea024621@nod.at> <89cd8327-581b-4285-d9ae-138d800211cf@toradex.com> <5a769de2-31f1-f011-c6bc-e94427407751@nod.at> <306876cd-0855-b9ee-659d-2164cfc3b2c4@nod.at> <877dea3c-717d-9b1a-c01d-537f3b46b2de@toradex.com> <47db4d54-f8f5-a663-75b3-d2cd5fd0769f@toradex.com> <17f16834-e200-60c7-62d4-b083b00335ac@toradex.com> <24fa997a-ba80-efb7-231d-8d4fddcd954d@nod.at> <750f1307-de2d-8735-d407-ac92b001d009@toradex.com> CC: Artem Bityutskiy From: Bhuvanchandra DV Message-ID: Date: Mon, 13 Feb 2017 17:15:51 +0530 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset="windows-1252"; format=flowed Content-Transfer-Encoding: 7bit List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 02/13/2017 01:27 PM, Richard Weinberger wrote: > Bhuvan, > > Am 13.02.2017 um 04:34 schrieb Bhuvanchandra DV: >> On 02/10/2017 06:31 PM, Richard Weinberger wrote: >> >>> Bhuvan, >>> >>> Am 09.02.2017 um 05:59 schrieb Bhuvanchandra DV: >>>>> Just in case, you can trigger this also with a fresh filesystem? >>>>> Or did you always keep the same one? >>>> Yes, I triggered it with fresh filesystem. >>> Okay. Not good. >>> >>> How do you trigger this? What workload is running while the power-cut >>> happens? >> The issue is triggered when doing a continous external power-cut for every >> 25sec since power ON. After a week or two observed this issue. There is no >> additional work load on CPU during the test, it just boots to a lxde desktop. > So, the power-cut happens while the system is idle? Hmm. May be not, sometimes power-cut might happen during boot. Sometimes I observed journalctl logging just before power-cut. > >>> Did you try also kernel 4.9? >> I did ran similar testes on our downstream kernel 4.4 and 4.9 kernel few weeks >> back, observed similar issues. > Okay, so I didn't break it lately. Puh... ;) :) > >>> Does it also only happen then xattrs are used? >> Didn't check that. How exactly to check whether xattr is used ? > i.e. is a LSM such das SELinux being used? No, not used so far. -- Bhuvan > > Thanks, > //richard