From mboxrd@z Thu Jan  1 00:00:00 1970
From: Avi Kivity <avi@qumranet.com>
Subject: Re: [kvm-devel] [PATCH 2/5] SCI fixes (v2)
Date: Wed, 28 May 2008 09:27:57 +0300
Message-ID: <483CFB6D.3070302@qumranet.com>
References: <1202137865-20232-1-git-send-email-aliguori@us.ibm.com> <1202137865-20232-3-git-send-email-aliguori@us.ibm.com> <1211857519.14708.30.camel@bling> <483C52D2.2080603@us.ibm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: Alex Williamson <alex.williamson@hp.com>,
	kvm-devel <kvm@vger.kernel.org>
To: Anthony Liguori <aliguori@us.ibm.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from il.qumranet.com ([212.179.150.194]:36663 "EHLO il.qumranet.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751622AbYE1G2A (ORCPT <rfc822;kvm@vger.kernel.org>);
	Wed, 28 May 2008 02:28:00 -0400
In-Reply-To: <483C52D2.2080603@us.ibm.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

Anthony Liguori wrote:
> Alex Williamson wrote:
>> On Mon, 2008-02-04 at 09:11 -0600, Anthony Liguori wrote:
>>  
>>> @@ -97,11 +100,12 @@ static void pm_update_sci(PIIX4PMState *s)
>>>      pmsts = get_pmsts(s);
>>>      sci_level = (((pmsts & s->pmen) &
>>>                    (RTC_EN | PWRBTN_EN | GBL_EN | TMROF_EN)) != 0);
>>> -    qemu_set_irq(s->dev.irq[0], sci_level);
>>> +    qemu_set_irq(s->irq, sci_level);
>>>      /* schedule a timer interruption if needed */
>>>      if ((s->pmen & TMROF_EN) && !(pmsts & TMROF_EN)) {
>>>          expire_time = muldiv64(s->tmr_overflow_time, ticks_per_sec, 
>>> PM_FREQ);
>>>          qemu_mod_timer(s->tmr_timer, expire_time);
>>> +        s->tmr_overflow_time += 0x800000;
>>>     
>>
>> Hi Anthony,
>>
>> I'm digging up a patch from a while back.  What's the purpose of bumping
>> tmr_overflow_time here?  It seems like this is intended to prevent the
>> timer overflow status from ever triggering an SCI, but that seems
>> contrary to the ACPI spec.  I've tracked down a case where Vista is
>> enabling the timer overflow and appears to get stuck if it never gets
>> the overflow interrupt.  Commenting out the last line here fixes it.
>> Suggestions?  Thanks,
>>   
>
> Avi is the original author of the patch I believe.  Perhaps he can 
> shed some light on this?
>

This is
> commit ce35c9534137b71327466fa9abc243cbe2d7e8dc
> Author: Avi Kivity <avi@qumranet.com>
> Date:   Wed Jan 2 12:52:28 2008 +0200
>
>     kvm: qemu: fix power management timer overflow handling
>     
>     The PMSTS overflow bit needs to be set each time bit 23 of the pm 
> timer
>     is toggled.  This means we need to adjust the overflow time every time
>     we have an overflow.
>     
>     Taken from qemu patch by TeLeMan in
>     
>     http://www.mail-archive.com/qemu-devel@nongnu.org/msg14680.html

And, like the explanation says, we have to advance the overflow time in 
order to get an interrupt.  Is there something horribly broken?

-- 
Do not meddle in the internals of kernels, for they are subtle and quick to panic.