Message-ID: <474F30D5.8070404@trash.net>
Date: Thu, 29 Nov 2007 22:36:21 +0100
From: Patrick McHardy <kaber@trash.net>
User-Agent: Mozilla-Thunderbird 2.0.0.6 (X11/20071009)
MIME-Version: 1.0
To: "H. Peter Anvin" <hpa@zytor.com>
CC: =?UTF-8?B?IlBhdWwgUm9sbGFuZCAo44Od44O844Or44O744Ot44Op44OzKSI=?= 
	<rol@witbe.net>,
       Linux Kernel <linux-kernel@vger.kernel.org>
Subject: Re: constant_tsc and TSC unstable
References: <20071129171142.56c7890a@tux.DEF.witbe.net> <474F0D47.7020006@zytor.com>
In-Reply-To: <474F0D47.7020006@zytor.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1375
Lines: 38

H. Peter Anvin wrote:
> Paul Rolland (ポール・ロラン) wrote:
>> Hello,
>>
>> I've a machine with a Core2Duo CPU. /proc/cpuinfo reports the flag
>> constant_tsc, but at boot time, I have the log :
>>
>> ...
>> Total of 2 processors activated (6919.15 BogoMIPS).
>> ENABLING IO-APIC IRQs
>> ..TIMER: vector=0x31 apic1=0 pin1=2 apic2=-1 pin2=-1
>> checking TSC synchronization [CPU#0 -> CPU#1]:
>> Measured 3978592228 cycles TSC warp between CPUs, turning off TSC clock.
>> Marking TSC unstable due to: check_tsc_sync_source failed.
>> Brought up 2 CPUs
>> ...
>>
>> This machine is running 2.6.23.1-21.fc7. I know I should report to 
>> Fedora,
>> but I was wondering if this is a bug or a feature ;)
>>
>
> constant_tsc means the TSC doesn't change with CPU core frequency.
>
> The problem you're having is that the TSCs of your two cores are 
> completely different, over a second apart.  This is a bug, unrelated 
> to constant_tsc.

On my core2duo I noticed that the TSC of the second core doesn't
start counting until the CPU is activated. Since they're not
synchronized manually anymore this makes the TSC check fail.


-
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/