Message-ID: <1412338512.19413.85.camel@x220>
Subject: Re: Why do we still have 32 bit counters? Interrupt counters
 overflow within 50 days
From: Paul Bolle <pebolle@tiscali.nl>
To: Christoph Lameter <cl@linux.com>
Cc: Thomas Gleixner <tglx@linutronix.de>, linux-kernel@vger.kernel.org
Date: Fri, 03 Oct 2014 14:15:12 +0200
In-Reply-To: <alpine.DEB.2.11.1410030653040.8496@gentwo.org>
References: <alpine.DEB.2.11.1410030435260.8324@gentwo.org>
	 <alpine.DEB.2.11.1410031231420.4383@nanos>
	 <alpine.DEB.2.11.1410030653040.8496@gentwo.org>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org

On Fri, 2014-10-03 at 06:54 -0500, Christoph Lameter wrote:
> On Fri, 3 Oct 2014, Thomas Gleixner wrote:
> 
> > > Is this the way its intended or should the counters be expanded to 64 bit?
> >
> > There is no reason why we cannot or should not expand them.
> 
> Ok here is a patch to do just that:
> 
> 
> Subject: Increase irq counters to 64 bit
> 
> 
> Irq counters can overflow easily if they are just 32 bit.
> 
> For example the timer interrupt occurs 1000 times per second, so
> it is predictable that the timer interrupt will overflow in
> 
> 
> 2^ 32 / 1000 [interrupts per second] / 86400 [seconds in a day]
> 
> which results in 46 days.

dc -e "1 k 2 32 ^ 1000 / 86400 / p"
49.7

(That was the number I remembered from stories about a ancient Windows
lockup.)

> Other irq counters for devices may wrap even faster for example
> those for high speed networking devices.
> 
> This patch is needed to avoid the counter overflow by increasing
> the counters to 64 bit.
> 
> Signed-off-by: Christoph Lameter <cl@linux.com>


Paul Bolle


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/