From: Jeremy Fitzhardinge <jeremy@goop.org>
Subject: Re: [PATCH RFC V2 3/5] jump_label: if a key has already been initialized,
 don't nop it out
Date: Thu, 06 Oct 2011 15:10:39 -0700
Message-ID: <4E8E275F.6010801@goop.org>
References: <cover.1317506051.git.jeremy.fitzhardinge@citrix.com>    <477dead9647029012f93c651f2892ed0e86b89e7.1317506051.git.jeremy.fitzhardinge@citrix.com>    <20111003150205.GB2462@redhat.com> <4E89E28C.7010700@goop.org>    <20111004141011.GA2520@redhat.com> <4E8B3489.60902@zytor.com>    <4E8CF348.4080405@goop.org> <4E8CF385.2080804@zytor.com>    <4E8DEB19.1050509@goop.org> <20111006181055.GA2505@redhat.com>   <1317925615.4729.14.camel@gandalf.stny.rr.com>   <4E8DF870.6010000@redhat.com>  <1317929321.4729.17.camel@gandalf.stny.rr.com> <4E8E20CD.5030207@goop.org> <1317938775.4729.29.camel@gandalf.stny.rr.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: 7bit
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <1317938775.4729.29.camel@gandalf.stny.rr.com>
Sender: linux-kernel-owner@vger.kernel.org
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Richard Henderson <rth@redhat.com>, Jason Baron <jbaron@redhat.com>, "H. Peter Anvin" <hpa@zytor.com>, "David S. Miller" <davem@davemloft.net>, David Daney <david.daney@cavium.com>, Michael Ellerman <michael@ellerman.id.au>, Jan Glauber <jang@linux.vnet.ibm.com>, the arch/x86 maintainers <x86@kernel.org>, Xen Devel <xen-devel@lists.xensource.com>, Linux Kernel Mailing List <linux-kernel@vger.kernel.org>, Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>, peterz@infradead.org
List-Id: xen-devel@lists.xenproject.org

On 10/06/2011 03:06 PM, Steven Rostedt wrote:
> But it only speeds up the tracing case. The non-tracing case is a nop
> and 5bytes is 5bytes regardless.
>
> Did you see a 5% speed up while tracing was happening? How did you do
> your test. I find a 5 byte compared to a 2 byte jump being negligible
> with the rest of the overhead of tracing, but I could be wrong.

You're right, this was a completely artificial microbenchmark.  In
practice the improvement would be a much smaller effect.

But bear in mind, I'm not using jump-label for tracing.  While its
important for the "disabled" state to be quick, performance of the
"enabled" state is also important.

Thanks,
    J