T43 2200 b/g wireless card won't initiate under ACPI when starting on battery power

[dedr79]

Thinkpad T43 w/ Intel 2200 b/g wireless (8086:4220). The wireless card works fine under APM. Using ACPI there are never error messages during boot, but if starting on battery power the card will fail to initialize even though the drivers seem to load OK (only indication is a long pause last thing before the WPS comes up) and xWLAN displays a "No WLAN card detected" error, and the card ID.

The BIOS is obviously at fault; if I wait to plug it in until after POST but before starting OS boot, wireless LAN fails. If I plug it in after power up but before POST finishes, it usually works OK. Evidently, something is different between power up on AC and power up on battery that only affects the wireless card.

[eco]

e-co:

Please show us BIOS Setup settings related to Power, WakeUp, ..

Do you use official Genmac drivers? Or self-made package for 2200?

[pasha]

Can you boot with AC on, after boot remove power, wiat 1 minute, insert power cable, then produce acpi.log?

[pasha]

Please download experimental ACPI build from Mensys site:

• Experimental build for you:

ACPI-APIC-BAT.ZIP

• Read instructions: ​http://ecomstation.ru/projects/acpitools/?action=testcase

[dedr79]

Replying to pasha:

Can you boot with AC on, after boot remove power, wiat 1 minute, insert power cable, then produce acpi.log?

I'm using the official Genmac drivers included with RC5 and 6a. I've uploaded the ACPI log file (APCI_AC-DC.log), switching to DC after boot-up, as requested, and 3 photos of the BIOS Power screen. I'll try the experimental ACPI version provided as soon as possible. Thanks.

[eco]

e-co:

what happens if you change the value of Mode for battery = .. and other modes related to battery.

Does it influence to WiFi?

[aschn]

On the first BIOS screenshot, I recognized that you have SpeedStep activated. I would disable that at first.

(The R52 and other similar models, e.g. the T43p, had a problem some years ago, which I don't remember anymore. You'll find it in this Bugtracker, it was reported by Roderick.)

Andreas

[aschn]

Here's the link to the SpeedStep ticket: #142. (I'm quite sure that this has never been fixed.)

Of course it's better to use APM instead of ACPI for models of that generation. APM-supperted suspend and resume work well, the same applies to SpeedStep support.

[eco]

e-co:

​http://ecomstation.ru/hardware.php?action=item&id=309

[dedr79]

Replying to aschn:

Here's the link to the SpeedStep ticket: #142. (I'm quite sure that this has never been fixed.)

You are right, this is the same problem. I tried a few BIOS settings, but nothing made any difference until I disabled SpeedStep. I hadn't noticed that ticket.

Of course it's better to use APM instead of ACPI for models of that generation.

Yes, I knew this all along because I used it first and saw the huge advantage to a properly functioning suspend/resume. However, there are also valid reasons to having a fully functioning ACPI, such as eventual access to the voltage table for the processor, etc. I know that also helps greatly for heat production and battery life because I've messed with those using Windows based ACPI utilities. That's why I've been hoping some of these things get looked at and fixed.

So the wireless card is now working properly with a battery start-up. Does disabling SpeedStep also disable throttling?

[dedr79]

Replying to pasha:

Please download experimental ACPI build from Mensys site:

• Experimental build for you:

ACPI-APIC-BAT.ZIP

I installed the experimental ACPI drivers mentioned above, but I'm not sure what to look for or how to follow up. They don't make any difference in the startup, except in the output on screen. The output is totally different between battery and AC startups. I also repeated the log gathering, if you want that.

I did follow others' suggestions that I disable SpeedStep, and that makes it work on a battery startup - I'm sorry I hadn't noticed the older ticket. That hardly seems like a reasonable conclusion though; we need it all working together to call it fixed, don't we?

[pasha]

acpi log from new version? About BIOS. I want 2 log. With all disable in BIOS and with all enable.

[dedr79]

Replying to pasha:

acpi log from new version? About BIOS. I want 2 log. With all disable in BIOS and with all enable.

OK, I've added the logs with the experimental ACPI drivers, three with everything power related off in BIOS and three with it all turned on. Of course, with everything off the wifi works and with it enabled it doesn't. Thanks.

[eco]

e-co:

Pasha: Yes. Log is digfference.

Steven: What is the difference and does it explain the reporter's issue?

[David Azarewicz]

This ticket is extremely old and for an unsupported version. Have you tried the current version?

[David Azarewicz]

No response from reporter.