From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mika.eatserver.nl ([195.20.9.75]) by casper.infradead.org with esmtps (Exim 4.72 #1 (Red Hat Linux)) id 1Pfvzp-0000wL-70 for linux-mtd@lists.infradead.org; Thu, 20 Jan 2011 15:00:21 +0000 Message-ID: <4D384DE7.10602@aimvalley.nl> Date: Thu, 20 Jan 2011 15:59:51 +0100 From: Norbert van Bolhuis MIME-Version: 1.0 To: Holger Brunck Subject: Re: MTD concat support for linux and UBI References: <4D3845AE.2040002@keymile.com> In-Reply-To: <4D3845AE.2040002@keymile.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Stefan Roese , linux-mtd@lists.infradead.org, "agust@denx.de" , Detlev Zundel List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 01/20/11 15:24, Holger Brunck wrote: > Hi all, > I have got a general question to the MTD concat feature in Linux. Is it allowed > to concatenate NOR flashes with different writebuffersizes? Currently the code > allows this. > If so I see a problem for different sizes when dealing with UBI volumes > comprehending such flashes. Some days ago some patches were added to adjust the > min I/O size for UBI to the writebuffersize of the flash. If now two flashes > with different writebuffersizes are concatenated whats the way forward? I see > two solutions: > 1) Using the writebuffersize from the flash with the largest writebuffer for min > I/O size in UBI. > 2) Adapting the concat feature in Linux to forbid concatenating flashes with > different writebuffersizes. > You're right. MTD concat already checks for incompatibilites between the subdevices and I guess it should take into account writebuffer size too. --- N. van Bolhuis.