Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 21 Jun 2023 21:08:03 -0700
From:   Christoph Hellwig <hch@infradead.org>
To:     Matthew Wilcox <willy@infradead.org>
Cc:     Dave Chinner <david@fromorbit.com>,
        Jeremy Bongio <bongiojp@gmail.com>, Ted Tso <tytso@mit.edu>,
        "Darrick J . Wong" <djwong@kernel.org>,
        Allison Henderson <allison.henderson@oracle.com>,
        linux-ext4@vger.kernel.org, linux-fsdevel@vger.kernel.org,
        linux-xfs@vger.kernel.org
Subject: Re: [PATCH 0/1] iomap regression for aio dio 4k writes
Message-ID: <ZJPJI1t/wdj/L1W2@infradead.org>
References: <20230621174114.1320834-1-bongiojp@gmail.com>
 <ZJOO4SobNFaQ+C5g@dread.disaster.area>
 <ZJOqC7Cfjr5AoW7S@dread.disaster.area>
 <ZJO4OAYhJlXOBXMf@casper.infradead.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ZJO4OAYhJlXOBXMf@casper.infradead.org>
Precedence: bulk

On Thu, Jun 22, 2023 at 03:55:52AM +0100, Matthew Wilcox wrote:
> Latency is important for reads, but why is it important for writes?
> There's such a thing as a dependent read, but writes are usually buffered
> and we can wait as long as we like for a write to complete.

That was exactly my reasoning on why I did always defer the write
completions in the initial iomap direct I/O code, and until now no
one has complained.

I could see why people care either about synchronous writes, or polled
io_uring writes, for which we might be able to do the work in the
completion thread context, but for aio writes this feels a bit odd.