Re: [PATCHv2 0/3] direct io alignment relax

From: Jens Axboe <axboe@kernel.dk>
To: Christoph Hellwig <hch@lst.de>
Cc: Keith Busch <kbusch@fb.com>,
	linux-fsdevel@vger.kernel.org, linux-block@vger.kernel.org,
	Kernel Team <Kernel-team@fb.com>,
	bvanassche@acm.org, damien.lemoal@opensource.wdc.com,
	Keith Busch <kbusch@kernel.org>
Subject: Re: [PATCHv2 0/3] direct io alignment relax
Date: Thu, 19 May 2022 06:46:08 -0600	[thread overview]
Message-ID: <92ee257f-c19f-adee-7bd2-409546b95d47@kernel.dk> (raw)
In-Reply-To: <20220519074225.GH22301@lst.de>

On 5/19/22 1:42 AM, Christoph Hellwig wrote:
> On Wed, May 18, 2022 at 04:45:10PM -0600, Jens Axboe wrote:
>> On 5/18/22 11:11 AM, Keith Busch wrote:
>>> From: Keith Busch <kbusch@kernel.org>
>>>
>>> Including the fs list this time.
>>>
>>> I am still working on a better interface to report the dio alignment to
>>> an application. The most recent suggestion of using statx is proving to
>>> be less straight forward than I thought, but I don't want to hold this
>>> series up for that.
>>
>> This looks good to me. Anyone object to queueing this one up?
> 
> Yes.  I really do like this feature, but I don't think it is ready to
> rush it in.  In addition to the ongoing discussions in this thread
> we absolutely need proper statx support for the alignments to avoid
> userspace growing all kinds of sysfs growling crap to make use of it.

OK fair enough, I do agree that we need a better story for exposing this
data, and in fact for a whole bunch of other stuff that is currently
hard to get programatically. We can defer to 5.20 and get the statx side
hashed out at the same time.

-- 
Jens Axboe