Re: KB835732 breaks S3 standby timer

From: Andrew Aronoff (NOSPAM_WRONG.ADDRESS_at_yahoo.com)
Date: 04/26/04 

Date: Mon, 26 Apr 2004 23:32:51 +0200

Hi, Rock.

Thanks for your feedback.

>If you're running multi-boot environment on one PC, you'll have some
>trouble with your settings.

I am, but I (usually) don't. ;-) See below.

>1. Please disable APM for win2k, Winme, winXP.

I'm not sure why I'd want to do this.

I have W98SE, NT4, W2K and WXP installed on separate partitions.
Excluding NT4, of course, ACPI works very well in all of them (when
not otherwise provoked by buggy hotfixes) with APM enabled. Sure, I
*could* disable APM, but *why*?

>If you have APM enabled, Win98's default will use APM and I think
>ignore ACPI.

Not here -- W98SE uses ACPI and ignores APM.

>2. Please enable APM for Win98 and Win98SE, and disable ACPI.

In W98SE, ACPI appears to work correctly as long as I leave my Adaptec
29160N disabled in Device Manager. That's not a problem, since the
devices connected to the SCSI controller are not needed in that O/S.

regards, Andy

Rock <rock1@hotmail.com> wrote:

>If you're running multi-boot environment on one PC, you'll have some
>trouble with your settings.
>
>1. Please disable APM for win2k, Winme, winXP. APM is a BIOS
>interface where Windows drivers will look to send BIOS APM messages
>for power control. ACPI is a WDM construct. This means:
> a. WDM drivers must support ACPI power messaging to pass
>WHQL. (USB, 1394, PCI bridges, etc)
> b. SCSI, and NDIS drivers do not have to support ACPI
>messaging to pass WHQL. Lately, most do support it, because there is
>some confusion about NDIS support and NDIS power OIDs (object
>identifiers). Most net drivers hook the WDM message and convert to
>NDIS power control OIDs. This is a time-phased statement, early NDIS
>drivers (Intel Pro/Wireless 2011 Lan PC Cards , for example) don't use
>NDIS oids, but I believe they have defined their own proprietary OIDs.
>Anyway, most late SCSI and NDIS drivers hook WDM ACPI messages.
>I frankly don't know if Win2K eliminates support for APM or not, or
>how well it does support APM if it is there. If you have APM enabled,
>Win98's default will use APM and I think ignore ACPI. A 3rd party
>driver may not, and try to use ACPI, so you might get bizarre
>operation.
>
>2. Please enable APM for Win98 and Win98SE, and disable ACPI. This
>is a religious statement, and may offend many people in Redmond. They
>worked very hard to make ACPI work, but frankly weren't supported by
>the hardware community. So to avoid 3rd party hardware, problems, and
>Windows driver interaction issues, you should disable ACPI for these
>operating systems. Now, there should be some spin-down timers
>available in your powr schemes tab under power options in the control
>panel. Set that to the timeout you want. If you're using SCSI and
>that item doesn't appear, please contact your SCSI driver
>manufacturer. They probably have that fixed. (Based on your comment
>about Adaptec, maybe not.)
>
>3. Keyboard and mouse wakeup should work. We made sure it always
>worked in our notebooks, but now most NBs come from China and who
>knows how much quality testing they get.
>
>Rock
>
>
>On Sat, 24 Apr 2004 13:26:38 +0200, Andrew Aronoff
><NOSPAM_WRONG.ADDRESS@yahoo.com> wrote:
>
>>Hi, Rock.
>>
>>FOA, thanks very much for your post.
>>
>>< probably more than you ever wanted to know about Power Management>
>>
>>I'd like to know a lot more than I do, so thanks again.
>>
>>< if you're using Windows power management, disable BIOS power
>>management in your hardware setup. >
>>
>>I'm not sure what you mean by this. Here's the way my AMI BIOS is
>>configured for working S3 standby in W98SE, W2K and WXP:
>>
>>IPCA Aware O/S Yes
>>ACPI Standby State S3/STR
>>Re-Call VGA BIOS at S3 Resuming Disabled
>>Power Management/APM Enabled <-- could be DISABLED
>>Power Button Function On/Off
>>Restore on AC/Power Loss Last State
>>> Set Monitor Events ...
>>[I'll skip this screen]
>>> Set Wake Up Events ...
>>Resume On Ring Disabled
>>Resume On PME# Disabled
>>Resume On RTC Alarm Disabled
>>RTC Alarm Date <-- RTC Alarm is DISABLED
>>RTC Alarm Hour <-- RTC Alarm is DISABLED
>>RTC Alarm Minute <-- RTC Alarm is DISABLED
>>RTC Alarm Second <-- RTC Alarm is DISABLED
>>USB Device Wakeup From S3 Disabled
>>Keyboard Wakeup From S3 Specific Key <-- optional
>> Specific Key for Wakeup Password
>>Mouse Wakeup From S3 Left-button <-- optional
>>
>>Which settings, if any, would you advise to change? I'll change 'em
>>and see what happens. My hunch is that everything set above is
>>necessary with this motherboard and BIOS except the seven commented
>>lines. Of course, *one* of the lines marked "optional" must be enabled
>>for resume from S3 to be possible.
>>
>>< Power Management is absolutely the last thing that goes into a
>>driver development >
>>
>>If it makes it in at all. Just ask Adaptec. They'll tell you
>>everything they know about PM and it won't take long. ;-)
>>
>>< Check device manager for each device >
>>
>>I had indeed found PM controls for my mouse, modem and NIC. I believe
>>that any checkboxes to "Allow the computer to turn off this device to
>>save power" are a waste of energy ;-) , since these devices consume
>>very little power. However, checking "Allow this device to bring the
>>computer out of standby" on my NIC proved very helpful for the PC that
>>acts as a server. It can remain in S3 standby until a client on the
>>LAN calls for it.
>>
>>< Consider how hard it is to restore a VPN after you've saved to file
>>>
>>
>>True, but that's the nature of TCP/IP connections. What counts for the
>>user in this case isn't reestablishing the connection but a graceful
>>recovery -- no data loss and a clear error message.
>>
>>< HTH >
>>
>>Sure did.
>>
>>regards, Andy
[snipped] 

-- 
                                  **********
       Please send e-mail to: usenet (dot) post (at) aaronoff (dot) com
                                  **********