From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chao Yu Subject: Re: [PATCH 2/2] mkfs.f2fs: make the default extensions list much more sensical Date: Wed, 29 May 2019 09:57:18 +0800 Message-ID: References: <20190416064355.29712-1-qkrwngud825@gmail.com> <20190416064355.29712-2-qkrwngud825@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from [172.30.20.202] (helo=mx.sourceforge.net) by sfs-ml-4.v29.lw.sourceforge.com with esmtps (TLSv1.2:ECDHE-RSA-AES256-GCM-SHA384:256) (Exim 4.90_1) (envelope-from ) id 1hVnqH-0002ge-QI for linux-f2fs-devel@lists.sourceforge.net; Wed, 29 May 2019 01:57:25 +0000 Received: from szxga06-in.huawei.com ([45.249.212.32] helo=huawei.com) by sfi-mx-3.v28.lw.sourceforge.com with esmtps (TLSv1.2:ECDHE-RSA-AES256-GCM-SHA384:256) (Exim 4.90_1) id 1hVnq4-002o3l-4e for linux-f2fs-devel@lists.sourceforge.net; Wed, 29 May 2019 01:57:25 +0000 Received: from DGGEMS405-HUB.china.huawei.com (unknown [172.30.72.58]) by Forcepoint Email with ESMTP id 496F474DC45634C33274 for ; Wed, 29 May 2019 09:57:04 +0800 (CST) In-Reply-To: Content-Language: en-US List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: linux-f2fs-devel-bounces@lists.sourceforge.net To: Ju Hyung Park Cc: linux-f2fs-devel@lists.sourceforge.net Hi Ju Hyung, On 2019/5/28 18:28, Ju Hyung Park wrote: > Hi Chao, > > On Tue, May 28, 2019 at 7:19 PM Chao Yu wrote: >> How about adding below extensions: >> >> "zip", >> "bin", >> "dat", >> "txt", > > zip is capable of random updates. I didn't add bz2 for the same reason. > But I do agree that most users won't be constantly updating zip files. Yup, I think the most possible case is using zip/bz2 to pack log files or for some downloaded resources, which should be write-once file. > > I personally use my Android device with zip treated as cold, but I'm > not sure if it makes good sense to make it as the default that's > supposed to run under various scenarios. > > How much different is the random write performance from cold to hot? We use different write policies: trigger OPU for hot data mostly, and IPU for cold data, so performance is quite relating to the test scenario. > > But I'm against the idea of adding the rest 3 extensions. > "bin" and "dat" is way too generic. You wouldn't know if a program > happens to heavily update files named .bin/.dat. > > For txt, it won't be uncommon for a user to update it frequently. > Moreover, most txt size is pretty small anyways. I think your concern is right, those extensions are too common and it needs to consider various platforms instead of just android. So, let's leave those three extensions. Thanks, > > And finally, circling back to your original concern, we should be more > careful adding extensions as there's a limit.> > Thanks. > . >