From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [195.209.228.254] (helo=shelob.oktetlabs.ru) by canuck.infradead.org with esmtps (Exim 4.54 #1 (Red Hat Linux)) id 1Ebbl8-0007RR-Qf for linux-mtd@lists.infradead.org; Mon, 14 Nov 2005 05:40:25 -0500 Message-ID: <43786975.9000204@yandex.ru> Date: Mon, 14 Nov 2005 13:39:49 +0300 From: "Artem B. Bityutskiy" MIME-Version: 1.0 To: "zhao, forrest" References: <1131961875.3417.35.camel@localhost.localdomain> In-Reply-To: <1131961875.3417.35.camel@localhost.localdomain> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Cc: linux-mtd@lists.infradead.org Subject: Re: Some thoughts about JFFS3 design doc List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , zhao, forrest wrote: > Hi, Artem > > I read through your JFFS3 design doc and had some thoughts I'm flattered you're reading my writings :-) > 1 in Section 4.1->4, there's a typo: > direntries contain {xattr name =>xattr ID}...... Oh, thanks. > 2 in section 3.6, I think it's better if you could > give a description or an example of how to replay > the journal to build the journal tree in RAM and > recover the inconsistent meta data on flash. Yes, there is supposed a distinct big chapter related to this topic. I just have not had time to write it so far. > 3 In section 4, I think it's better if you could > give a description or an example of what happens > for an update operation. Such as what happens when > a write operation update offset from 100 to 200 for > a file of length 400 bytes. I guess maybe you are > working on this now :) Nothing interesting - the new node just goes to the journal and the old one is considered obsolete. I'm planning to add a distinct chapter with description of JFFS3-in0action where this will be described. Thanks for the note. > 4 I think your design doc is well-organized and give > reader a very clear picture and illustration of the > JFFS3 design. Thanks, the only really major missing thing is Garbage Collection. Will describe it as soon as I completely design it. Feel free to send any comments, and even patches - the design sources are in MTD CVS (Documentation/mtd/). -- Best Regards, Artem B. Bityutskiy, St.-Petersburg, Russia.