Presario V3000 aka p/n RM666AV aka V3100 CTO onboard wireless support?

Frank Durda IV

I have been lumbered with a task involving this laptop (Presario V3000 aka
p/n RM666AV aka V3100 CTO), which has a Broadcom wireless chipset.
(You hear maniacal laughter in the background, possibly having to do
with the '666' in the laptops' part number...)

With much blending of the advice of several different web pages
(all of which gave part of the solution but no one procedure was
complete or worked alone), I have been finally been able to build
a kernel from 6.2 STABLE 200607 which contains the ndis patches and
some other magic, and built a ndis wrapper driver for the hardware
in question, and the driver finally loads without a panic*:


Id Refs Address Size Name
1 15 0xc0400000 71dda0 kernel
2 1 0xc0b1e000 5af54 acpi.ko
3 1 0xc0b79000 aca24 bcmwl5_sys.ko
4 3 0xc0c26000 17734 ndis.ko
5 2 0xc0c3e000 ca04 if_ndis.ko
....

Better still, on boot the system spits out some trivia suggesting the
driver is engaged (which it didn't do for a while after not panicing):

....
ndis0: <Broadcom 802.11b/g WLAN> mem 0xd4000000-0xd4003fff irq 19
at device 0.0 on pci5
ndis0: NDIS API version: 5.1
ndis0: Ethernet address: 00:14:a5:f1:71:78
....

and ifconfig shows the device, as in:

ndis0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> mtu 1500
ether 00:14:a5:f1:71:78
media: IEEE 802.11 Wireless Ethernet autoselect
status: no carrier
ssid "" channel 1
authmode OPEN privacy OFF txpowmax 100 bmiss 7 protmode CTS

Which all seems great, and far better than the panics
or modules that would load but complain about missing this or
that which earlier procedures caused.

However, before celebrating, all of the above still doesn't actually
work.


The problem appears to be with the mechanical radio on/off slide
switch on the front of the notebook case. Under Windows, when this
switch is moved to the right, the blue light next to the switch comes on
and Windows can detect wireless networks. Moving the switch to the
left and the light turns orange, and Windows goes radio blind,
saying it doesn't see any networks. (Windows doesn't seem to be aware of
why it can't see any networks.)

Shutting down Windows with the switch in the ON position and the
light blue and rebooting into FreeBSD, the light remains blue
throughout the BIOS and POST periods and into the initial FreeBSD
boot/probe phase. Then the light turns Orange about the time that the ndis
driver & bcwml5 driver load, and the switch does not appear to control
the behavior of the radio from that point on. (Unless you go back to
Windows, where the blue light comes on and the radio goes active.)
The .sys and .inf files were taken from the Windows partition of the
same laptop, and happen to be the same files as those offered on the
HP web site, so I believe those files to be correct for this notebook
model.

Now, I should also mention that under 6.2 RELEASE or 6.2 STABLE 200706,
this system whines repeatedly about ACPI things from boot up for the
first five minutes or so with messages like these:

ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBIF] (Node 0xc3620d00), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BIF] (Node 0xc3620bc0), AE_NO_HARDWARE_RESPONSE
Trying to mount root from ufs:/dev/ad4s2a
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBST] (Node 0xc3620ce0), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BST] (Node 0xc3620ba0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBIF] (Node 0xc3620d00), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BIF] (Node 0xc3620bc0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBIF] (Node 0xc3620d00), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BIF] (Node 0xc3620bc0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBIF] (Node 0xc3620d00), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BIF] (Node 0xc3620bc0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBIF] (Node 0xc3620d00), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BIF] (Node 0xc3620bc0), AE_NO_HARDWARE_RESPONSE
ACPI-0501: *** Error: Handler for [EmbeddedControl] returned AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.GBIF] (Node 0xc3620d00), AE_NO_HARDWARE_RESPONSE
ACPI-1304: *** Error: Method execution failed [\\_SB_.PCI0.LPCB.EC0_.BAT0._BIF] (Node 0xc3620bc0), AE_NO_HARDWARE_RESPONSE

and after five minutes, these messages then appear only occasionally,
say one or two every six minutes or so.


During a period of the trial and error to get this far (mainly on
RELEASE 6.2), any change in the position of the radio switch immediately
generated one of the above messages. Moving it the other direction
generated another. In 6.2 STABLE 200706, this does not seem to happen,
and although two of these messages do pop up every so often, they don't
seem occur instantly and obviously in response to touching the switch.
This is a mechanical slide switch, and not buttons.

Using 6.2 RELEASE isn't an option because the bcmwl5_sys.ko driver
panics the system immediately, and if you load the so-called ndis patch
I found in a number of places and apply that and build/install, it no
longer panics, but whines about missing functions when you load
bcmwl5_sys.ko and no device is shown in ifconfig. It may or may not
report that it has found hardware. Subsequently, I seemed forced to use
the newer 200706 STABLE platform, which seems to have more of the needed
pieces in place.


The fact that the radio light remains Orange which (indicates it
is disabled under Windows) makes me suspect that operation of the
switch is not directly part of the radio circuitry, and is not reported
back to the wireless chipset via OS or driver, so the wireless
chipset leaves itself off. That is a theory based on behavior,
nothing more.


I've wasted a few days getting to this point on the internal wireless
hardware. I have also tried to find alternate plug-in USB or PCExpress
(not PCCARD) adapters that are recognized by FreeBSD 6.2 that would
function, but so far the local stores have nothing I can buy that I have
found listed anywhere in Google searches that might work under FreeBSD.

Everything that is listed as known or believed to work isn't made anymore
or is an obscure brand that a Frys, BestBuy or Circuit City wouldn't sell.
When they do sell the same brand, the devices on the "works" list are
typically at least two card generations behind what is on the store
shelves, and that's just looking at 802.11G products they are clearing
out, never mind looking at the newer 802.11N models that will be all you
can buy in a few more months.

So, to get out of this box, I'm looking for any of these:

1. A way to get the onboard radio transmitter to actually come
on. I've got it pretty close so far, and maybe I need to just
go kluge the driver to not turn the radio off. Or maybe there is
an existing secret command that already does something like this?
OR
2. A USB 802.11G adapter that works with FreeBSD 5.5 or 6.2 RELEASE
or 6.2 STABLE. (Shoot, I'll go to 7.0 200706 if it gave me more
options, but the laptop did act a tad quirky under 7.0 in other
areas.)
OR
3. A PCExpress (the narrow card that some laptops use instead
of the PCCard/PCMCIA interface), that works with FreeBSD 5.5
or 6.2 RELEASE, or 6.2 STABLE. As above, if 7.0 offers a way
out I will consider it.)

This is for a "crash cart" system that has to be flung across the country
at short notice and perform tunneling/routing to bypass a damaged/failed
fixed network in emergencies, and the fewer plug-in and loose parts that
the flingee needs to carry with him and get past the airport guards, the
better. An ethernet-connected wireless solution isn't possible as the
ethernet port will be needed for other duties. I will likely be making
more than one of these, so I need a source of a small quantity of the
appropriate items, not just the last one on the shelf somewhere, AND
THEY HAVE TO WORK. I've burned up too much time getting to this point.


Oh, and getting all that ACPI noise to go away would be nice too,
but I suppose I can just take the module out of the kernel...


Thanks in advance for any suggestions on what has proven to be a
seriously under-documented area.


Frank Durda IV - send mail to this address and remove the "LOSE":
<uhclemLOSE.aug07%nemesis.lonestar.org>
VISTA - Virtually Impossible to Secure or Trust Anyway
Copyright 2007, ask before reprinting.