From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-fsdevel-owner@vger.kernel.org>
Received: from verein.lst.de ([213.95.11.211]:41111 "EHLO newverein.lst.de"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1752488AbeAIM2w (ORCPT <rfc822;linux-fsdevel@vger.kernel.org>);
        Tue, 9 Jan 2018 07:28:52 -0500
Date: Tue, 9 Jan 2018 13:28:50 +0100
From: Christoph Hellwig <hch@lst.de>
To: viro@zeniv.linux.org.uk
Cc: Avi Kivity <avi@scylladb.com>, linux-aio@kvack.org,
        linux-fsdevel@vger.kernel.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: aio poll, io_pgetevents and a new in-kernel poll API
Message-ID: <20180109122850.GA6567@lst.de>
References: <20180104080043.14506-1-hch@lst.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180104080043.14506-1-hch@lst.de>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Thu, Jan 04, 2018 at 09:00:12AM +0100, Christoph Hellwig wrote:
> The changes were sponsored by Scylladb, and improve performance
> of the seastar framework up to 10%, while also removing the need
> for a privileged SCHED_FIFO epoll listener thread.

Due to the current events:

With KPTI enabled the aio poll code is almost 16% faster than epoll
with the special SCHED_FIFO listener thread.

The Scylladb https example is still more than 4% faster with KPTI
and aio poll vs non-KPTI with epoll.