From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from ns.tz.ru ([194.149.234.1]) by canuck.infradead.org with esmtp (Exim 4.63 #1 (Red Hat Linux)) id 1HTvxq-0007xL-Kv for linux-mtd@lists.infradead.org; Wed, 21 Mar 2007 04:14:35 -0400 Date: Wed, 21 Mar 2007 11:14:28 +0300 From: Igor Marnat Message-ID: <1055017593.20070321111428@rambler.ru> To: Thomas Gleixner Subject: Re: jffs2_gcd_mtd0 invoked oom-killer In-Reply-To: <1174323006.13341.504.camel@localhost.localdomain> References: <16125243125.20070312152801@rambler.ru> <1173705053.5493.28.camel@sauron> <41978953.20070319084552@rambler.ru> <1174288746.13341.490.camel@localhost.localdomain> <9323738734.20070319144832@rambler.ru> <1174323006.13341.504.camel@localhost.localdomain> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: linux-mtd@lists.infradead.org, David Woodhouse Reply-To: Igor Marnat List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Hello Thomas, TG> Interrestingly enough are the node CRCs of some of the nodes intact, so TG> the corruption must have happened before writing to flash. Loop "while (valid_ref)" (those that become infinite) in jffs2_get_inode_nodes also contains crc checking of nodes. They are checked in order of proceeding. It doesn't complain about nodes' crc, so it seems that crc are ok. So I don't think that crc checking of chain of nodes helps here. -- Best regards, Igor Marnat mailto:marny@rambler.ru