From: Paolo Bonzini <pbonzini-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
To: Jason Baron <jbaron-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org>,
Fam Zheng <famz-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
Thomas Gleixner <tglx-hfZtesqFncYOwBW4kG4KsQ@public.gmane.org>,
Ingo Molnar <mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
"H. Peter Anvin" <hpa-YMNOUZJC4hwAvxtiuMwx3w@public.gmane.org>,
x86-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org,
Alexander Viro
<viro-RmSDqhL/yNMiFSDQTTA3OLVCufUGDwFn@public.gmane.org>,
Andrew Morton
<akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>,
Kees Cook <keescook-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>,
Andy Lutomirski <luto-kltTT9wpgjJwATOyAt5JVQ@public.gmane.org>,
David Herrmann
<dh.herrmann-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Alexei Starovoitov <ast-uqk4Ao+rVK5Wk0Htik3J/w@public.gmane.org>,
Miklos Szeredi <mszeredi-AlSwsSmVLrQ@public.gmane.org>,
David Drysdale <drysdale-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>,
Oleg Nesterov <oleg-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
"David S. Miller" <davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>,
Vivek Goyal <vgoyal-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
Mike Frysinger <vapier-aBrp7R+bbdUdnm+yROfE0A@public.gmane.org>,
Theodore Ts'o <tytso-3s7WtUTddSA@public.gmane.org>,
Heiko Carstens
<heiko.carstens-tA70FqPdS9bQT0dZR+AlfA@public.gmane.org>,
Rasmus Villemoes
<linux-qQsb+v5E8BnlAoU/VqSP6n9LOBIZ5rWg@public.gmane.org>,
Rashika Kheria
<rashika.kheria-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Hugh Dickins <hughd-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>,
Mathieu Desnoyers
<mathieu.desnoyers-vg+e7yoeK/dWk0Htik3J/w@public.gmane.org>,
Peter Zijlstra <peter>
Subject: Re: [PATCH v4 0/9] epoll: Introduce new syscalls, epoll_ctl_batch and epoll_pwait1
Date: Fri, 13 Mar 2015 15:56:33 +0100 [thread overview]
Message-ID: <5502FAA1.6000206@redhat.com> (raw)
In-Reply-To: <5502F857.6050505-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org>
On 13/03/2015 15:46, Jason Baron wrote:
> > The throttling algorithm computes a duration for the next IO, which is used to
> > arm a timer in order to delay the request a bit. As timers are always rounded
> > *UP* to the effective granularity, the timeout being 1ms in epoll_pwait is just
> > too coarse and will lead to severe inaccuracy. With epoll_pwait1, we can avoid
> > the rounding-up.
>
> right, but we could use the timerfd here to get the desired precision.
Fam, didn't you see slowdowns with few file descriptors
epoll_ctl+epoll_wait+timerfd compared to ppoll?
Do they disappear or improve with epoll_ctl_batch and epoll_pwait1?
Paolo
WARNING: multiple messages have this Message-ID (diff)
From: Paolo Bonzini <pbonzini-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
To: Jason Baron <jbaron-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org>,
Fam Zheng <famz-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
Thomas Gleixner <tglx-hfZtesqFncYOwBW4kG4KsQ@public.gmane.org>,
Ingo Molnar <mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
"H. Peter Anvin" <hpa-YMNOUZJC4hwAvxtiuMwx3w@public.gmane.org>,
x86-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org,
Alexander Viro
<viro-RmSDqhL/yNMiFSDQTTA3OLVCufUGDwFn@public.gmane.org>,
Andrew Morton
<akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>,
Kees Cook <keescook-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>,
Andy Lutomirski <luto-kltTT9wpgjJwATOyAt5JVQ@public.gmane.org>,
David Herrmann
<dh.herrmann-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Alexei Starovoitov <ast-uqk4Ao+rVK5Wk0Htik3J/w@public.gmane.org>,
Miklos Szeredi <mszeredi-AlSwsSmVLrQ@public.gmane.org>,
David Drysdale <drysdale-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>,
Oleg Nesterov <oleg-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
"David S. Miller" <davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>,
Vivek Goyal <vgoyal-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
Mike Frysinger <vapier-aBrp7R+bbdUdnm+yROfE0A@public.gmane.org>,
"Theodore Ts'o" <tytso-3s7WtUTddSA@public.gmane.org>,
Heiko Carstens
<heiko.carstens-tA70FqPdS9bQT0dZR+AlfA@public.gmane.org>,
Rasmus Villemoes
<linux-qQsb+v5E8BnlAoU/VqSP6n9LOBIZ5rWg@public.gmane.org>,
Rashika Kheria
<rashika.kheria-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Hugh Dickins <hughd-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>,
Mathieu Desnoyers
<mathieu.desnoyers-vg+e7yoeK/dWk0Htik3J/w@public.gmane.org>,
Peter Zijlstra <peter
Subject: Re: [PATCH v4 0/9] epoll: Introduce new syscalls, epoll_ctl_batch and epoll_pwait1
Date: Fri, 13 Mar 2015 15:56:33 +0100 [thread overview]
Message-ID: <5502FAA1.6000206@redhat.com> (raw)
In-Reply-To: <5502F857.6050505-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org>
On 13/03/2015 15:46, Jason Baron wrote:
> > The throttling algorithm computes a duration for the next IO, which is used to
> > arm a timer in order to delay the request a bit. As timers are always rounded
> > *UP* to the effective granularity, the timeout being 1ms in epoll_pwait is just
> > too coarse and will lead to severe inaccuracy. With epoll_pwait1, we can avoid
> > the rounding-up.
>
> right, but we could use the timerfd here to get the desired precision.
Fam, didn't you see slowdowns with few file descriptors
epoll_ctl+epoll_wait+timerfd compared to ppoll?
Do they disappear or improve with epoll_ctl_batch and epoll_pwait1?
Paolo
WARNING: multiple messages have this Message-ID (diff)
From: Paolo Bonzini <pbonzini@redhat.com>
To: Jason Baron <jbaron@akamai.com>, Fam Zheng <famz@redhat.com>
Cc: linux-kernel@vger.kernel.org,
Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@redhat.com>, "H. Peter Anvin" <hpa@zytor.com>,
x86@kernel.org, Alexander Viro <viro@zeniv.linux.org.uk>,
Andrew Morton <akpm@linux-foundation.org>,
Kees Cook <keescook@chromium.org>,
Andy Lutomirski <luto@amacapital.net>,
David Herrmann <dh.herrmann@gmail.com>,
Alexei Starovoitov <ast@plumgrid.com>,
Miklos Szeredi <mszeredi@suse.cz>,
David Drysdale <drysdale@google.com>,
Oleg Nesterov <oleg@redhat.com>,
"David S. Miller" <davem@davemloft.net>,
Vivek Goyal <vgoyal@redhat.com>,
Mike Frysinger <vapier@gentoo.org>,
"Theodore Ts'o" <tytso@mit.edu>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
Rasmus Villemoes <linux@rasmusvillemoes.dk>,
Rashika Kheria <rashika.kheria@gmail.com>,
Hugh Dickins <hughd@google.com>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
Peter Zijlstra <peterz@infradead.org>,
linux-fsdevel@vger.kernel.org, linux-api@vger.kernel.org,
Josh Triplett <josh@joshtriplett.org>,
"Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>,
Omar Sandoval <osandov@osandov.com>,
Jonathan Corbet <corbet@lwn.net>,
shane.seymour@hp.com, dan.j.rosenberg@gmail.com
Subject: Re: [PATCH v4 0/9] epoll: Introduce new syscalls, epoll_ctl_batch and epoll_pwait1
Date: Fri, 13 Mar 2015 15:56:33 +0100 [thread overview]
Message-ID: <5502FAA1.6000206@redhat.com> (raw)
In-Reply-To: <5502F857.6050505@akamai.com>
On 13/03/2015 15:46, Jason Baron wrote:
> > The throttling algorithm computes a duration for the next IO, which is used to
> > arm a timer in order to delay the request a bit. As timers are always rounded
> > *UP* to the effective granularity, the timeout being 1ms in epoll_pwait is just
> > too coarse and will lead to severe inaccuracy. With epoll_pwait1, we can avoid
> > the rounding-up.
>
> right, but we could use the timerfd here to get the desired precision.
Fam, didn't you see slowdowns with few file descriptors
epoll_ctl+epoll_wait+timerfd compared to ppoll?
Do they disappear or improve with epoll_ctl_batch and epoll_pwait1?
Paolo
next prev parent reply other threads:[~2015-03-13 14:56 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-10 1:49 [PATCH v4 0/9] epoll: Introduce new syscalls, epoll_ctl_batch and epoll_pwait1 Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 1/9] epoll: Extract epoll_wait_do and epoll_pwait_do Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 2/9] epoll: Specify clockid explicitly Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 3/9] epoll: Extract ep_ctl_do Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 4/9] epoll: Add implementation for epoll_ctl_batch Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
[not found] ` <1425952155-27603-5-git-send-email-famz-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2015-03-10 13:59 ` Dan Rosenberg
2015-03-10 13:59 ` Dan Rosenberg
2015-03-11 2:23 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 5/9] x86: Hook up epoll_ctl_batch syscall Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 6/9] epoll: Add implementation for epoll_pwait1 Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 7/9] x86: Hook up epoll_pwait1 syscall Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
[not found] ` <1425952155-27603-1-git-send-email-famz-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2015-03-10 1:49 ` [PATCH v4 8/9] epoll: Add compat version implementation of epoll_pwait1 Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` [PATCH v4 9/9] x86: Hook up 32 bit compat epoll_pwait1 syscall Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-10 1:49 ` Fam Zheng
2015-03-12 15:02 ` [PATCH v4 0/9] epoll: Introduce new syscalls, epoll_ctl_batch and epoll_pwait1 Jason Baron
2015-03-12 15:02 ` Jason Baron
[not found] ` <5501AA6B.2020209-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org>
2015-03-13 11:31 ` Fam Zheng
2015-03-13 11:31 ` Fam Zheng
2015-03-13 11:31 ` Fam Zheng
[not found] ` <20150313113122.GA7427-ZfWej9ACyHUXGNroddHbYwC/G2K4zDHf@public.gmane.org>
2015-03-13 14:46 ` Jason Baron
2015-03-13 14:46 ` Jason Baron
2015-03-13 14:46 ` Jason Baron
[not found] ` <5502F857.6050505-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org>
2015-03-13 14:56 ` Paolo Bonzini [this message]
2015-03-13 14:56 ` Paolo Bonzini
2015-03-13 14:56 ` Paolo Bonzini
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5502FAA1.6000206@redhat.com \
--to=pbonzini-h+wxahxf7alqt0dzr+alfa@public.gmane.org \
--cc=akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org \
--cc=ast-uqk4Ao+rVK5Wk0Htik3J/w@public.gmane.org \
--cc=davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org \
--cc=dh.herrmann-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
--cc=drysdale-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org \
--cc=famz-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=heiko.carstens-tA70FqPdS9bQT0dZR+AlfA@public.gmane.org \
--cc=hpa-YMNOUZJC4hwAvxtiuMwx3w@public.gmane.org \
--cc=hughd-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org \
--cc=jbaron-JqFfY2XvxFXQT0dZR+AlfA@public.gmane.org \
--cc=keescook-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org \
--cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=linux-qQsb+v5E8BnlAoU/VqSP6n9LOBIZ5rWg@public.gmane.org \
--cc=luto-kltTT9wpgjJwATOyAt5JVQ@public.gmane.org \
--cc=mathieu.desnoyers-vg+e7yoeK/dWk0Htik3J/w@public.gmane.org \
--cc=mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=mszeredi-AlSwsSmVLrQ@public.gmane.org \
--cc=oleg-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=rashika.kheria-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
--cc=tglx-hfZtesqFncYOwBW4kG4KsQ@public.gmane.org \
--cc=tytso-3s7WtUTddSA@public.gmane.org \
--cc=vapier-aBrp7R+bbdUdnm+yROfE0A@public.gmane.org \
--cc=vgoyal-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=viro-RmSDqhL/yNMiFSDQTTA3OLVCufUGDwFn@public.gmane.org \
--cc=x86-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.