netdev - Re: Fwd: After kernel 6.3.7 or 6.3.8 b43 driver fails

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Sun, 25 Jun 2023 11:58:00 -0500
From: Sardonimous <sardonimous@...mail.com>
To: Larry Finger <Larry.Finger@...inger.net>,
 Bagas Sanjaya <bagasdotme@...il.com>, Arnd Bergmann <arnd@...db.de>,
 Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
 Linux Regressions <regressions@...ts.linux.dev>,
 Linux Wireless <linux-wireless@...r.kernel.org>,
 Netdev <netdev@...r.kernel.org>
Cc: Michael Büsch <m@...s.ch>,
 kernel test robot <lkp@...el.com>, Simon Horman <simon.horman@...igine.com>,
 Kalle Valo <kvalo@...nel.org>
Subject: Re: Fwd: After kernel 6.3.7 or 6.3.8 b43 driver fails

Hello Larry,

On 6/24/23 23:41, Larry Finger wrote:
>
> Sardonimous,
>
> The critical line is:
>> Jun 20 18:20:11 askasleikir kernel: b43_pio_tx+0x373/0x390
>
> I certainly have not used PIO for a long time. I expect that your 
> MacBook Pro should do DMA on the b43. Apple makes wierd hardware, but 
> not likely that wierd.

I have been unable to get DMA to work in the past.  So I have been 
configuring it with PIO=1 (/etc/modprobe,d/b43.conf):

     options b43 pio=1 qos=0

>
> Does dmesg offer any clues as to what is happening?
>
If I try with, say:

     options b43 pio=0 qos=0 verbose=3

Then

     rmmod b43
     rmmod ssb
     modprobe b43

I see the following:

     insmod /lib/modules/6.1.12-arch1-1/kernel/drivers/ssb/ssb.ko.zst
     insmod 
/lib/modules/6.1.12-arch1-1/kernel/drivers/net/wireless/broadcom/b43/b43.ko.zst 
pio=0 qos=0 verbose=3

...

[Jun25 11:51] ssb: Found chip with id 0x4322, rev 0x01 and package 0x0A
[  +0.096950] b43-pci-bridge 0000:02:00.0: Sonics Silicon Backplane 
found on PCI device 0000:02:00.0
[  +0.411653] b43-phy1: Broadcom 4322 WLAN found (core revision 16)
[  +0.038206] b43-phy1: Found PHY: Analog 8, Type 4 (N), Revision 4
[  +0.000032] b43-phy1: Found Radio: Manuf 0x17F, ID 0x2056, Revision 3, 
Version 0
[  +0.015660] Broadcom 43xx driver loaded [ Features: PNLS ]
[  +0.520933] ieee80211 phy1: Selected rate control algorithm 'minstrel_ht'
[  +0.726559] b43-phy1: Loading firmware version 784.2 (2012-08-15 21:35:19)
[  +0.276596] b43-phy1 debug: Chip initialized
[  +0.000562] b43-phy1 debug: 64-bit DMA initialized
[  +0.000013] b43-phy1 debug: QoS disabled
[  +0.023145] b43-phy1 debug: Wireless interface started
[  +0.056159] b43-phy1 debug: Adding Interface type 2
[  +0.086691] b43-phy1 debug: Removing Interface type 2
[  +0.000105] b43-phy1 debug: Wireless interface stopped
[  +0.099849] b43-phy1 ERROR: DMA RX reset timed out
[  +0.065067] b43 ssb0:0: Timeout waiting for bitmask 01800000 on 
register 0F90 to clear
[  +0.204905] b43-phy1: Loading firmware version 784.2 (2012-08-15 21:35:19)
[  +0.296681] b43-phy1 debug: Chip initialized
[  +0.000555] b43-phy1 debug: 64-bit DMA initialized
[  +0.000010] b43-phy1 debug: QoS disabled
[  +0.027099] b43-phy1 debug: Wireless interface started
[  +0.062503] b43-phy1 debug: Adding Interface type 2

The wlan0 device never seems to work, no IP address is obtained, etc.

> If there is nothing shown in the log, you definitely need to do a 
> proper bisection from the mainline git tree to isolate the change that 
> led to this failure.

Still working on initial kernel build.

> ADDED WITH EDIT: I looked at the code and b43 will not be built for 
> any hardware without DMA, thus it appears that adding "b43.pio=1" is 
> the only way to get PIO mode. Please check the output of dmesg for PIO 
> messages.
>
> Larry

Thanks & regards,

Sardonimous