Re: [PATCH 0/2] Remove kiocb ki_complete

From: Matthew Wilcox <willy@infradead.org>
To: Pavel Begunkov <asml.silence@gmail.com>
Cc: Christoph Hellwig <hch@infradead.org>,
	Jens Axboe <axboe@kernel.dk>,
	linux-block@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-aio@kvack.org, io-uring@vger.kernel.org,
	linux-kernel@vger.kernel.org, Kanchan Joshi <joshi.k@samsung.com>,
	Javier Gonzalez <javier.gonz@samsung.com>
Subject: Re: [PATCH 0/2] Remove kiocb ki_complete
Date: Thu, 9 Jul 2020 14:43:19 +0100	[thread overview]
Message-ID: <20200709134319.GD12769@casper.infradead.org> (raw)
In-Reply-To: <ffbd272c-32f3-8c8c-6395-5eab47725929@gmail.com>

On Thu, Jul 09, 2020 at 04:37:59PM +0300, Pavel Begunkov wrote:
> On 09/07/2020 16:26, Christoph Hellwig wrote:
> > On Thu, Jul 09, 2020 at 12:10:36PM +0100, Matthew Wilcox wrote:
> >> On Thu, Jul 09, 2020 at 11:17:05AM +0100, Christoph Hellwig wrote:
> >>> I really don't like this series at all.  If saves a single pointer
> >>> but introduces a complicated machinery that just doesn't follow any
> >>> natural flow.  And there doesn't seem to be any good reason for it to
> >>> start with.
> >>
> >> Jens doesn't want the kiocb to grow beyond a single cacheline, and we
> >> want the ability to set the loff_t in userspace for an appending write,
> >> so the plan was to replace the ki_complete member in kiocb with an
> >> loff_t __user *ki_posp.
> >>
> >> I don't think it's worth worrying about growing kiocb, personally,
> >> but this seemed like the easiest way to make room for a new pointer.
> > 
> > The user offset pointer has absolutely no business in the the kiocb
> > itself - it is a io_uring concept which needs to go into the io_kiocb,
> > which has 14 bytes left in the last cache line in my build.  It would
> > fit in very well there right next to the result and user pointer.
> 
> After getting a valid offset, io_uring shouldn't do anything but
> complete the request. And as io_kiocb implicitly contains a CQE entry,
> not sure we need @append_offset in the first place.
> 
> Kanchan, could you take a look if you can hide it in req->cflags?

No, that's not what cflags are for.  And besides, there's only 32 bits
there.