From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mtiwmhc12.worldnet.att.net ([204.127.131.116]) by bombadil.infradead.org with esmtp (Exim 4.68 #1 (Red Hat Linux)) id 1KbkGr-0003Yq-Kr for linux-mtd@lists.infradead.org; Fri, 05 Sep 2008 22:59:17 +0000 From: Jeff Sutherland To: Vernon Sauder Subject: Re: mtd concatenate support for pxa2xx-flash mapping driver Date: Fri, 5 Sep 2008 18:59:02 -0400 References: <200809041359.33626.jeffs@fomsystems.com> <48C0B79E.2020909@gmail.com> In-Reply-To: <48C0B79E.2020909@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200809051859.03206.jeffs@fomsystems.com> Cc: linux-mtd@lists.infradead.org List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Friday 05 September 2008, apparently Vernon Sauder wrote: > Jeff Sutherland wrote: > > I'm working currently in 2.6.25. Is anyone working on mtd concatenation > > support for this particular driver? $customer selected 64Mbyte NOR flash > > chip for root filesystem, but in reality it is two 32Mb die in one > > package, and I need one mtd partition > 32mb after taking care of > > bootloader, environment, and kernel. > > Is this the P30 strataflash 48F4400 variety? I created a custom patch > for that but converted to using the phymap-flash support from 2.6.25.1 > and backported it to 2.6.24.4. I use the command line parsing. I got this sorted out today. It turns out the physmap-flash driver will fail silently if you specify a command line partition that spans twom chips. In other words, I had specified in the kernel config the total size of installed flash, but had left unchecked config_mtd_concat. On startup physmap-flash would correctly identify both chips inside the 48F4400 part, and even let me make a partition (passed in from the kernel command line) that spanned the boundary between chips. But jffs2 had a cow with the smaller erase blocks at the 32m boundary. Simply turning on CONFIG_MTD_CONCAT solved all the problems, sigh... Regards, -Jeff