Subject: Re: [PATCH/RFC] mm: don't cap request size based on read-ahead
 setting
To: Andrew Morton <akpm@linux-foundation.org>
References: <7d8739c2-09ea-8c1f-cef7-9b8b40766c6a@kernel.dk>
Cc: linux-mm@kvack.org,
        "linux-block@vger.kernel.org" <linux-block@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
From: Jens Axboe <axboe@kernel.dk>
Message-ID: <6e924b0e-a2fc-5983-fd7d-80c956308937@kernel.dk>
Date: Tue, 15 Nov 2016 14:18:12 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101
 Thunderbird/45.4.0
MIME-Version: 1.0
In-Reply-To: <7d8739c2-09ea-8c1f-cef7-9b8b40766c6a@kernel.dk>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1426
Lines: 30

On 11/10/2016 10:00 AM, Jens Axboe wrote:
> Hi,
>
> We ran into a funky issue, where someone doing 256K buffered reads saw
> 128K requests at the device level. Turns out it is read-ahead capping
> the request size, since we use 128K as the default setting. This doesn't
> make a lot of sense - if someone is issuing 256K reads, they should see
> 256K reads, regardless of the read-ahead setting.
>
> To make matters more confusing, there's an odd interaction with the
> fadvise hint setting. If we tell the kernel we're doing sequential IO on
> this file descriptor, we can get twice the read-ahead size. But if we
> tell the kernel that we are doing random IO, hence disabling read-ahead,
> we do get nice 256K requests at the lower level. An application
> developer will be, rightfully, scratching his head at this point,
> wondering wtf is going on. A good one will dive into the kernel source,
> and silently weep.
>
> This patch introduces a bdi hint, io_pages. This is the soft max IO size
> for the lower level, I've hooked it up to the bdev settings here.
> Read-ahead is modified to issue the maximum of the user request size,
> and the read-ahead max size, but capped to the max request size on the
> device side. The latter is done to avoid reading ahead too much, if the
> application asks for a huge read. With this patch, the kernel behaves
> like the application expects.

Any comments on this?

-- 
Jens Axboe