From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S932938AbaAaTyO (ORCPT <rfc822;w@1wt.eu>);
	Fri, 31 Jan 2014 14:54:14 -0500
Received: from merlin.infradead.org ([205.233.59.134]:54415 "EHLO
	merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S932811AbaAaTyM (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Fri, 31 Jan 2014 14:54:12 -0500
Date: Fri, 31 Jan 2014 20:54:07 +0100
From: Peter Zijlstra <peterz@infradead.org>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        paulmck@linux.vnet.ibm.com, linux-rt-users@vger.kernel.org,
        linux-kernel@vger.kernel.org, tglx@linutronix.de,
        Clark Williams <williams@redhat.com>
Subject: Re: [PATCH 2/2] timer: really raise softirq if there is irq_work to
 do
Message-ID: <20140131195407.GQ5002@laptop.programming.kicks-ass.net>
References: <1391178845-15837-1-git-send-email-bigeasy@linutronix.de>
 <1391178845-15837-2-git-send-email-bigeasy@linutronix.de>
 <20140131120757.594e24d6@gandalf.local.home>
 <20140131174227.GN9012@linux.vnet.ibm.com>
 <20140131125719.73340f6e@gandalf.local.home>
 <52EBF8FE.3080608@linutronix.de>
 <20140131143441.478f79ee@gandalf.local.home>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20140131143441.478f79ee@gandalf.local.home>
User-Agent: Mutt/1.5.21 (2012-12-30)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Fri, Jan 31, 2014 at 02:34:41PM -0500, Steven Rostedt wrote:
> On Fri, 31 Jan 2014 20:26:54 +0100
> Sebastian Andrzej Siewior <bigeasy@linutronix.de> wrote:
> 
> > On 01/31/2014 06:57 PM, Steven Rostedt wrote:
> > 
> > > In vanilla Linux, irq_work_run() is called from update_process_times()
> > > when it is called from the timer interrupt. In -rt, there's reasons we
> > 
> > and in vanilla Linux some architectures (like x86 or sparc to name just
> > a few) overwrite arch_irq_work_raise() which means they provide
> > their "own" interrupt like callback. That means on those architectures
> > irq_work_run() gets invoked twice: once via update_process_times() and
> > via and once the custom interface.
> > So my question to the original inventor of this code: Peter, do we
> > really need that arch specific callback? Wouldn't one be enough? Is it
> > that critical that it can't wait to the next timer tick?
> 
> There's flags that determine when the next call should be invoked. The
> irq_work_run() should return immediately if it was already done by the
> arch specific call. The work wont be called twice.
> 
> As I have worked on code that uses irq_work() I can say that we want
> the arch specific interrupts. For those architectures that don't have
> it will experience larger latencies for the work required. It's
> basically, a "too bad" for them.
> 
> But to answer your question, no we want the immediate response.

Yah, what Steve said. That fallback is really a you suck to have to use
this.