Another BCM4322 (FreeBSD 12)

MrRichBOB

MrRichBOB

I would like to apologise in advance, if this is straight-forward, and/or answered elsewhere on the forum--I have tried for the past few hours to look hard before I dared ask any questions, and I combed through the WiFi section of the Handbook, etc.

Basically, in short, I am using an EliteBook 6930p, which has troubles with all its drivers even on Linux, and no one has even managed to get the fingerprint scanner working there. It took me about a week before I found a solution for getting the WiFi card in it working, on OpenSUSE. So, I looked around, and I found this post here: CLICK ME. So, I followed along, trying to diagnose similar issues, and ultimately compiled my first kernel--that was both fun and scary at the same time. Thankfully, after doing that, the OS can finally see my WiFi card and can even scan around (barely better than it could on Linux--not very well.)

Here are my diagnostics, after running the custom kernel:

First, I tried this command:
$ sudo wpa_supplicant -i wlan0 -c /etc/wpa_supplicant.conf

And Got this out:
Code: 
Successfully initialized wpa_supplicant
wlan0: Trying to associate with e8:fc:af:fd:cc:9e (SSID='NETGEAR78' freq=5765 MHz)
wlan0: Authentication with e8:fc:af:fd:cc:9e timed out.
wlan0: CTRL-EVENT-DISCONNECTED bssid=e8:fc:af:fd:cc:9e reason=3 locally_generated=1
- - - -
Secondly, I checked dmesg and it gave me this out:

Code: 
Error reading triso2g from NVRAM: 0
bwn0: ucode fw: ucode16_mimo
bwn0: firmware version (rev 666 patch 2 date 0xb217 time 0x9e7)
bwn0: ucode fw: ucode16_mimo
bwn0: firmware version (rev 666 patch 2 date 0xb217 time 0x9e7)

... over and over again.

So, Imma take it the the firmware is either incompatible with the driver, or it is not actually a low-power card, like I thought it was. (I drew that conclusion based off of dmesg spitting out at me that the kernel wasn't PHY, whatever, compatible before I actually made my custom kernel.

- - - -

Lastly, I have attached any relevant files I actually touched. The custom kernel I have attached as WIFI.txt, and the only things I actually messed with in it were adding lines 298-301, which simply say:

Code: 
# Options for BCM4322
option BWN_DEBUG
option BWN_GPL_PHY

I wanted to uncomment line 289, and I really feel like I should have, but I opted not to, as that seemed to be the purpose of modifying /boot/loader.conf

- - - -

THANK ANYONE IN ADVANCE!
 

Attachments

SirDice said:
If you read bwn(4) you'll see that you need to install the firmware; net/bwn-firmware-kmod. Did you install that?
Click to expand...

That is a lovely idea, I didn't think of that--I did it before making the custom kernel, but didn't think to do so afterward.

I shall do that first thing tomorrow.

EDIT:

So yeah, I cd'd to /usr/ports/net/bwn-firmware-kmod and did $ sudo make clean and $ sudo make reinstall, and still no luck, literally doing the same exact thing.
 
Please apologize my horrible English.

I have got the same issue on a Dell Latitude 4300, BCM 4322 wireless controller, FreeBSD 13.1
- compiled /usr/ports/net/bwn-firmware-kmod
- compiled the kernel with options BWN_DEBUG & BWN_GPL_PHY

ifconfig wlan create wlandev bwn0 correctly creates wlan0
ifconfig wlan0 up is ok (with a bunch and flowless "error reading triso2g from nvram:0")
wpa_supplicant -i wlan0 -c /usr/wpa_supplicant.conf works (associated) the second time
dhclient wlan0 => unknown dhcp option value 0x7d => PANIC

Can anyone help ?
Many thanks in advance 8=))

(note : same procedure on a Lenovo T61 with wpi parent interface works flowlessly)
 
context
i have a Broadcom BCM43228 which is not supported by bwn(4).
i tried to make it work by adding missing bits from linux b43 driver
after i managed to get it probed and attached it always panic-ed in ifconfig

while doing so i've seen that linux driver aligns dma buffers to 64 bits for some chipsets and the freebsd doesn't so i changed the code
in the end it worked but with very poor performance (10Mb/s) but linux performance wasn't any better
i'll post the modified if_bwn.c
 

Attachments

covacat : I picked up your file as is => worked straightforward after recompilation of bwn firmware

Thank you very much, I cannot thank you enough !

Is there something I can do in return ?
 
Hi guys.

I’ve been having almost exactly the same issue with my MacBookPro 5,5 (which has a BCM 4322) under FreeBSD 14. It worked fine under 13, but now panics when I can get it to connect at all. seems to be a page fault in the attach() function.

I’ve tried Covacat’s modified if_bwn.c and I can get it to compile (with a few changes to fix errors) and run but it has no effect.

Is there anyone with an old BCM card and enough knowledge of the drivers out there to work with me on debugging?
 
You must log in or register to reply here.
Back
Top