[arch64] broadcom 4319 miniPCI (a/b/g)

pobudz · 2006-04-20 16:54:02

Problem:
ndiswrapper AND bcm43xx not working
ndiswrapper compiles just fine... trying to load bcmwl5.inf, bcmwl5a.inf, netbc564.inf... with the 64bit broadcom drivers... I've got about 3 or 4 different versions of the broadcom driver... I just can't get it to load the interface.

I either get driver loaded or driver loaded, hardware present... either way no new interfaces.

So I gave up... tried bcm43xx project... installed the 80211softmac, modprobe'd it into the kernel... built bcm43xx... try loading it in the kernel and it fails out (see dmesg) and complains a LOT about 80211softmac and then exits. Now the one thing that might make sense is firmware, not entirely sure fwcutter did what it was supposed to do.

If anyone has any suggestions, I'd love to hear it. btw this is with a amd turion64 hp dv5000.

AndyRTR · 2006-04-20 18:17:00

pacman -Ss ndis
extra/ndiswrapper 1.13-1
Module for NDIS (Windows Network Drivers) drivers supplied by vendors. For
stock arch 2.6 kernel.
extra/ndiswrapper-beyond 1.13-1
Module for NDIS (Windows Network Drivers) drivers supplied by vendors. For
Beyond 2.6 kernel.
extra/ndiswrapper-utils 1.13-1
Binaries for ndiswrapper module

Have you tried our packages?

iphitus · 2006-04-20 23:38:21

pobudz wrote:

Problem:
ndiswrapper AND bcm43xx not working
ndiswrapper compiles just fine... trying to load bcmwl5.inf, bcmwl5a.inf, netbc564.inf... with the 64bit broadcom drivers... I've got about 3 or 4 different versions of the broadcom driver... I just can't get it to load the interface.
I either get driver loaded or driver loaded, hardware present... either way no new interfaces.
So I gave up... tried bcm43xx project... installed the 80211softmac, modprobe'd it into the kernel... built bcm43xx... try loading it in the kernel and it fails out (see dmesg) and complains a LOT about 80211softmac and then exits. Now the one thing that might make sense is firmware, not entirely sure fwcutter did what it was supposed to do.
If anyone has any suggestions, I'd love to hear it. btw this is with a amd turion64 hp dv5000.

Have you tried modprobing softmac first?

elahav · 2006-04-21 00:20:02

I am trying to get bcm43xx working on my laptop. It worked fine on FC5.
My current status: bcm43xx and ieee80211softmac compiled and installed. When I modprobe bcm43xx the module and its dependencies are loaded and the card is detected. No errors.
Unfortunately, I can't get eth1 up. I have the firmware installed in /lib/firmware , but I don't think it is being read (I moved it around and I don't get the expected kernel errors). I suspect that the card is not being initialised.

elahav · 2006-04-21 01:54:19

Some progress: I managed to get the card working. It turns out that the udev package for Arch64 was missing firmware_helper.
I still can't get it to recognise any networks, though. 'iwlist scan' returns 'no scan results', even though there are several wireless networks operating here.

Cam · 2006-04-21 04:51:17

I might try building tpowa's 2.6.17 kernels for Arch64, it has a new Broadcom driver merged that is supposed to support a lot of the cards ndiswrapper currently supports. That might make life easier.

pobudz · 2006-04-21 07:09:36

mmm I'm willing to try anything. I tried our packages... and all of the broadcom drivers I was able to find 64bit-wise.

I think it has something to do with the fact I have a/b/g bcm4319? I cant find actual a/b/g 64-bit drivers... infact the ones I'm using in WinXP64 only do b/g as it is.

If anyone has the a/b/g drivers for bcm4319 (64bit) perhaps ndiswrapper will enjoy that more.

I did try modprobing all of the softmac stuff... that didnt seem to help, infact most of it complained when I did try loading it. Now that I think about it... bcm43xx complained about ieee80211softmac while it was compiling but didnt error out.. the complaints show up in dmesg when I try to modprobe it.

Perhaps I'm missing something here? Well, I give up!!! I downloaded a vanilla 2.6.17-rc2 kernel... threw the kconfig26 over, added in softmac/broadcom support ... off we go! Status report in t-minus 30 minutes

Cam · 2006-04-21 08:38:36

You'll need to do the research as to which module to load but here is a build of tpowa's kernel snapshots. I haven't tested this either so 100% at your own risk but it might help

pobudz · 2006-04-21 08:52:16

okay... so i have bcm43xx loaded into the kernel as well as softmac extensions thanks to the 2.6.17-rc2 kernel which is running beautifully!

also, I rebuilt the udev pkg with extras/firmware... the only thing I really need is the exact syntax for the udev.rules file. I found something but it was from debian and I don't think its working.

When I modprobe bcm43xx it loads ieee80211, and it displays "bcm43xx driver" but thats it. I ran fwcutter and all the firmware was dumped to /lib/firmware

I am about to try ndiswrapper with the new kernel to see if theres any improvement. I found one more set of drivers as well... need to unrar them.

pobudz · 2006-04-21 09:37:27

new kernel broke ndiswrapper, whoops! heres what happens in dmesg when i modprobe the softmac stuff then broadcom... im thinking I might just build them INTO the damn kernel... enough module bs

dmesg:

ieee80211_crypt: registered algorithm 'NULL'
ieee80211: 802.11 data/management/control stack, git-1.1.7
ieee80211: Copyright (C) 2004-2005 Intel Corporation <jketreno>
ieee80211_crypt: registered algorithm 'CCMP'
ieee80211_crypt: registered algorithm 'TKIP'
ieee80211_crypt: registered algorithm 'WEP'
bcm43xx driver

lspci -v:

06:02.0 Network controller: Broadcom Corporation Unknown device 4319 (rev 02)
Subsystem: Hewlett-Packard Company Unknown device 1358
Flags: bus master, fast devsel, latency 64, IRQ 10
Memory at b0204000 (32-bit, non-prefetchable) [size=8K]

lsmod:

bcm43xx 430192 0
ieee80211_crypt_wep 5248 0
ieee80211_crypt_tkip 10880 0
ieee80211_crypt_ccmp 8064 0
ieee80211softmac 26496 1 bcm43xx
ieee80211 32712 2 bcm43xx,ieee80211softmac
ieee80211_crypt 6144 4 ieee80211_crypt_wep,ieee80211_crypt_tkip,ieee80211_crypt_ccmp,ieee80211

elahav · 2006-04-21 13:24:25

pobudz wrote:

When I modprobe bcm43xx it loads ieee80211, and it displays "bcm43xx driver" but thats it.

This driver has a debug option. Did you turn it on? If so, you should get some more information in /var/log/messages.log.
What happens when you run 'ifconfig ethX up'?

pobudz · 2006-04-21 19:16:53

I built bcm43xx and its debugging counterpart into the kernel as well as softmac... its built in, and it was compiled as a module previous.

dmesg STILL only says that one line.

iwconfig says no interfaces, ifconfig says no interfaces (No NEW interfaces anyway)... so theres nothing to bring up. its not complaining or even WHINING! Nothing at all... I'm not quite sure what to think. The card works fine in Windows XP 64bit... It is the a/b/g card which from what I understood was supported.

If anyone had an EASIER fix than what is available for the HP Bios whitelist... I have a prismgt minipci and an atheros minipci card I can pull... but stupid HP has a whitelist on the bios!!! lame.

elahav · 2006-04-21 20:02:57

I finally managed to get my card working with ndiswrapper.
In your first post, you mentioned that ndiswrapper confirms that the driver is loaded and the hardware is present. What happens when you load the module?
Also, ndiswrapper is supposed to attach itself to wlan0. Did you try 'ifconfig wlan0 up' with the module loaded?

pobudz · 2006-04-21 20:23:09

I have the DMESG output at home but basically ndiswrapper complains that it can't initiate the windows driver. so obviously when ndiswrapper says "driver loaded, hardware present" it really means config file was made, it looks like the hardware is there but cant be sure.

obviously if dmesg says "failed to initialize windows driver" theres a larger error at hand. I think I am alone here, does anyone have the a/b/g card (bcm4319) actually WORKING?

i even tried ndiswrapper 1.14... considering 1.15 rc1... the driver works in windows, it has to be something with how ndiswrapper is handling the driver in linux.

oh, and its hard to bring up an interface that doesnt exist. neither the bcm43xx builtin module nor the ndiswrapper module creates an actual interface.

pobudz · 2006-04-22 10:23:19

fixed ndiswrapper.

LILO:

append="acpi=noirq"

when I feel like breaking stuff I'll try to fix bcm43xx

Arch Linux

#1 2006-04-20 16:54:02

[arch64] broadcom 4319 miniPCI (a/b/g)

#2 2006-04-20 18:17:00

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#3 2006-04-20 23:38:21

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#4 2006-04-21 00:20:02

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#5 2006-04-21 01:54:19

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#6 2006-04-21 04:51:17

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#7 2006-04-21 07:09:36

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#8 2006-04-21 08:38:36

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#9 2006-04-21 08:52:16

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#10 2006-04-21 09:37:27

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#11 2006-04-21 13:24:25

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#12 2006-04-21 19:16:53

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#13 2006-04-21 20:02:57

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#14 2006-04-21 20:23:09

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

#15 2006-04-22 10:23:19

Re: [arch64] broadcom 4319 miniPCI (a/b/g)

Board footer