From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Weinberger Subject: Re: [PATCH 2/2] ubifs: Allow O_DIRECT Date: Thu, 20 Aug 2015 08:42:06 +0200 Message-ID: <55D576BE.5040207@nod.at> References: <1440016553-26481-1-git-send-email-richard@nod.at> <1440016553-26481-2-git-send-email-richard@nod.at> <55D542C5.6040500@cn.fujitsu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, dedekind1@gmail.com To: Dongsheng Yang , linux-mtd@lists.infradead.org Return-path: In-Reply-To: <55D542C5.6040500@cn.fujitsu.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Yang, (Sorry if I've used your last name lately) Am 20.08.2015 um 05:00 schrieb Dongsheng Yang: > On 08/20/2015 04:35 AM, Richard Weinberger wrote: >> Currently UBIFS does not support direct IO, but some applications >> blindly use the O_DIRECT flag. >> Instead of failing upon open() we can do better and fall back >> to buffered IO. > > Hmmmm, to be honest, I am not sure we have to do it as Dave > suggested. I think that's just a work-around for current fstests. > > IMHO, perform a buffered IO when user request direct IO without > any warning sounds not a good idea. Maybe adding a warning would > make it better. Well, how would you inform the user? A printk() to dmesg is useless are the vast majority of open() callers do not check dmesg... :) Major filesystems implement ->direct_IO these days and having a "return 0"-stub seems to be legit. For example exofs does too. So, I really don't consider it a work around. > I think we need more discussion about AIO&DIO in ubifs, and actually > I have a plan for it. But I have not listed the all cons and pros of > it so far. Sure, having a real ->direct_IO would be be best solution. My patch won't block this. Thanks, //richard