From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [200.73.64.66] (helo=apolo.virtuabyte.com) by canuck.infradead.org with esmtp (Exim 4.33 #1 (Red Hat Linux)) id 1Bvo09-0000Zv-42 for linux-mtd@lists.infradead.org; Fri, 13 Aug 2004 22:10:34 -0400 Received: from cm-vina3-148-69.cm.vtr.net ([200.104.148.69]) by apolo.virtuabyte.com (apolo.virtuabyte.com) with ASMTP id CQA74038 for ; Fri, 13 Aug 2004 22:10:30 -0400 From: Carlos Manuel Duclos Vergara To: linux-mtd@lists.infradead.org Date: Fri, 13 Aug 2004 22:19:38 +0200 References: <200408131855.16376.lucasvr@gobolinux.org> In-Reply-To: <200408131855.16376.lucasvr@gobolinux.org> MIME-Version: 1.0 Content-Disposition: inline Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <200408132219.39195.carlos@embedded.cl> Subject: Re: CRC errors and DMA List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , > Hi, > Hi, > What happens when the processor tryies to read data from memory while the > bus is being used for DMA transfers by another device? Is it possible that it depends on the processor, but that situation shouldn't happen... there must be some kind of bus master or some way to manage the bus (by using a token, by using a master-slave bus, ...). In most cases the DMA controller takes control of the bus during the DMA transfer, so the rest of the devices cannot use it > this kind of situation turns into CRC errors when trying to read data from > a flash device (JFFS2)? > Given that this problem can happen, is there a way to make JFFS2, or the > MTD subsystem to try to execute the operation again? > _IF_ you have that situation it is better to check your hardware or use another one, having a DMA controller that cannot take control of the bus to do its work is not a good situation... -- Carlos Manuel Duclos Vergara carlos@embedded.cl