Re: [PATCH] ia64: disable preemption in udelay()

From: Lee Revell <rlrevell@joe-job.com>
To: Zwane Mwaikambo <zwane@arm.linux.org.uk>
Cc: "Luck, Tony" <tony.luck@intel.com>,
	hawkes@sgi.com, Tony Luck <tony.luck@gmail.com>,
	Andrew Morton <akpm@osdl.org>,
	linux-ia64@vger.kernel.org, linux-kernel@vger.kernel.org,
	Jack Steiner <steiner@sgi.com>, Keith Owens <kaos@sgi.com>,
	Dimitri Sivanich <sivanich@sgi.com>
Subject: Re: [PATCH] ia64: disable preemption in udelay()
Date: Thu, 15 Dec 2005 21:03:55 -0500	[thread overview]
Message-ID: <1134698636.12086.222.camel@mindpipe> (raw)
In-Reply-To: <Pine.LNX.4.64.0512151750500.1678@montezuma.fsmlabs.com>

On Thu, 2005-12-15 at 17:52 -0800, Zwane Mwaikambo wrote:
> On Thu, 15 Dec 2005, Luck, Tony wrote:
> 
> > On Wed, Dec 14, 2005 at 03:25:26PM -0800, hawkes@sgi.com wrote:
> > > Sending this to a wider audience:
> > > 
> > > The udelay() inline for ia64 uses the ITC.  If CONFIG_PREEMPT is enabled
> > > and the platform has unsynchronized ITCs and the calling task migrates
> > > to another CPU while doing the udelay loop, then the effective delay may
> > > be too short or very, very long.
> > > 
> > > The most simple fix is to disable preemption around the udelay looping.
> > > The downside is that this inhibits realtime preemption for cases of long
> > > udelays.  One datapoint:  an SGI realtime engineer reports that if
> > > CONFIG_PREEMPT is turned off, that no significant holdoffs are
> > > are attributed to udelay().
> > > 
> > > I am reluctant to propose a much more complicated patch (that disables
> > > preemption only for "short" delays, and uses the global RTC as the time
> > > base for longer, preemptible delays) unless this patch introduces
> > > significant and unacceptable preemption delays.
> > 
> > Stuck between a rock and the proverbial hard place.
> > 
> > I think that the more complex patch is needed though.  If some crazy
> > driver has a pre-emptible udelay(10000), then you really don't want
> > to spin for that long without allowing preemption.
> 
> If it's a preemptible sleep period it should just use msleep.

There are 10 drivers that udelay(10000) or more and a TON that
udelay(1000).  Turning those all into 1ms+ non preemptible sections will
be very bad.

Lee