Bug 10398 - b43: Critical DMA error
Summary: b43: Critical DMA error
Status: RESOLVED OBSOLETE
Alias: None
Product: Drivers
Classification: Unclassified
Component: network-wireless (show other bugs)
Hardware: All Linux
: P1 high
Assignee: Michael Buesch
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-04-05 10:28 UTC by Patrick Matthäi
Modified: 2013-12-19 15:54 UTC (History)
4 users (show)

See Also:
Kernel Version: 2.6.38
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
dmesg (29.94 KB, application/octet-stream)
2008-04-05 10:28 UTC, Patrick Matthäi
Details
loaded modules (2.59 KB, application/octet-stream)
2008-04-05 10:29 UTC, Patrick Matthäi
Details
lspci -vv (22.24 KB, application/octet-stream)
2008-04-05 10:30 UTC, Patrick Matthäi
Details
My Kernel config (43.78 KB, application/octet-stream)
2008-04-05 10:30 UTC, Patrick Matthäi
Details

Description Patrick Matthäi 2008-04-05 10:28:25 UTC
Distribution: Debian Sid
Hardware Environment: 30:00.0 Network controller: Broadcom Corporation BCM4312 802.11a/b/g (rev 02)
Software Environment: i386
Problem Description:

Hello,

I'm using now the b43 driver instead of ndiswrapper and today my network failed.
My wireless led on my HP Compaq 6715s just blinks up and the link was down, then I've saw following in dmesg:


b43-phy0 ERROR: PHY transmission error
b43-phy0 ERROR: Fatal DMA error: 0x00000000, 0x00000800, 0x00000000, 0x00000000, 0x00000000, 0x00000000
b43-phy0: Controller RESET (DMA error) ...
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
b43-phy0: Controller RESET (DMA error) ...
b43-phy0: Controller restarted
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
b43-phy0: Controller RESET (DMA error) ...
b43-phy0: Controller restarted
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
b43-phy0: Controller RESET (DMA error) ...
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
printk: 1 messages suppressed.
b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
wlan0: No ProbeResp from current AP 00:04:0e:45:80:9b - assume out of range
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
printk: 8 messages suppressed.
b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
printk: 8 messages suppressed.
b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
printk: 8 messages suppressed.
b43-phy0: Controller restarted
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)


My full dmesg and kernel configuration is attached.

These transmission errors were there before this blocker occurs.
Comment 1 Patrick Matthäi 2008-04-05 10:28:58 UTC
Created attachment 15623 [details]
dmesg
Comment 2 Patrick Matthäi 2008-04-05 10:29:58 UTC
Created attachment 15624 [details]
loaded modules
Comment 3 Patrick Matthäi 2008-04-05 10:30:18 UTC
Created attachment 15625 [details]
lspci -vv
Comment 4 Patrick Matthäi 2008-04-05 10:30:53 UTC
Created attachment 15626 [details]
My Kernel config
Comment 5 Patrick Matthäi 2008-04-10 11:25:23 UTC
It happened again with 2.6.25-rc8-git8 and my notebook wasn't responsible, I had to hard reset it:

Apr 10 20:18:18 gnu kernel: printk: 8 messages suppressed.
Apr 10 20:18:18 gnu kernel: b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
Apr 10 20:18:18 gnu kernel: b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
Apr 10 20:18:22 gnu last message repeated 2 times
Apr 10 20:18:23 gnu kernel: printk: 8 messages suppressed.
Apr 10 20:18:23 gnu kernel: b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
Apr 10 20:18:23 gnu kernel: b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
Apr 10 20:18:26 gnu last message repeated 2 times
Apr 10 20:18:28 gnu kernel: printk: 8 messages suppressed.
Apr 10 20:18:28 gnu kernel: b43-phy0 ERROR: Fatal DMA error: 0x00000400, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
Apr 10 20:18:28 gnu kernel: b43-phy0: Loading firmware version 410.2160 (2007-05-26 15:32:10)
Comment 6 Patrick Matthäi 2008-05-02 00:52:05 UTC
Also in 2.6.25 but there's an new 2.6.25.1 kernel which seems as it fixes such a problem.
I will test it and respond again if I've got some news.
Comment 7 John W. Linville 2008-07-02 07:46:01 UTC
Is this now resolved?
Comment 8 Patrick Matthäi 2008-07-02 09:30:06 UTC
It is hard to say, I get still random crashes if I'm connected via wlan and downloading about torrent for example.
The problem now is, that it completly hard locks and nothing will be logged to dmesg, maybe because of another firmware.

Currently I'm using version 351.126 (2006-07-29 05:54:02) on 2.6.26-rc7-git2.

But two times I noticed that my wlan LED was blinking, maybe because of this bug (several reloads).
Comment 9 Giuliani Vito Ivan 2008-11-04 07:56:24 UTC
I can confirm this bug under 2.6.26
Comment 10 Patrick Matthäi 2010-02-14 12:42:43 UTC
Hm since a longer time I do not get anymore such random crashes.

But yesterday I had it for two times, where I used ktorrent with many opened connections.
Looks like the chip has to be under fire to reproduce it.
Comment 11 cool_sch.alex 2010-02-15 02:16:40 UTC
Same problem on kernel 2.6.32-13.
On every start I get this Problem. Had to remove b43-fwcutter and blacklist b43/ssb, else the notebook hangs every ~10 seconds.
Greetz

PS: rmmod b43 didn't work it said that the device were busy. But after that the device was "unloaded". --Don't know much about linux, trying to make WLAN working with b43 driver.
Comment 12 Patrick Matthäi 2011-02-03 20:06:33 UTC
I am now on 2.6.37 and the problem still does not occured again, what is with you, Loader?
Comment 13 cool_sch.alex 2011-03-28 19:52:37 UTC
I've tried it today, installed a 2.6.37 Kernel version for Ubuntu from their kernel-ppa.
No luck, the DMA error occurs.
I have also made a cool start (shutdown, wait a few seconds and started again), it didn't help.

If I don't make a cool start and the had linux started before, the device seems to be crashed. If I just reboot from Linux into Windows, WLAN doesn't work, until I make a cool start.
Greetz
Comment 14 Patrick Matthäi 2011-03-28 19:56:10 UTC
Hmm, speaking for my case, I'm thinking that it is more a hardware error.

I reproduced it a second time with 2.6.37 and it occured where I put my notebook maybe a bit fast on my desk, maybe it is also a HW error with your notebook?
Comment 15 cool_sch.alex 2011-03-28 20:04:47 UTC
That's a good thought.
I've a Dell Studio 1555.
It is known that it has problems with sound + wifi, if used both together.
But on Windows it is still useable, with the b43 driver on linux it is only in PIO mode usable. (Very slow)
Since my knowledge with linux is very small, I can only guess.
But I am able to give the logs, if needed. Maybe test something.
Currently Ubuntu isn't installed for productive use, I can test much.
Greetz
Comment 16 Patrick Matthäi 2011-03-29 18:47:32 UTC
(In reply to comment #14)
> Hmm, speaking for my case, I'm thinking that it is more a hardware error.
> 
> I reproduced it a second time with 2.6.37 and it occured where I put my
> notebook maybe a bit fast on my desk, maybe it is also a HW error with your
> notebook?

Args, sorry, I replied to the wrong bug! This should go to https://bugzilla.kernel.org/show_bug.cgi?id=26802 instead of 10398 :(
Comment 17 cool_sch.alex 2011-03-29 18:54:49 UTC
It's good you do that, because I saw this "MAC suspend failed" in my logs too, but didn't thought of another bugreport.
I'll look if I can post something usefull there.
Greetz
Comment 18 Jason Vas Dias 2011-06-18 15:43:07 UTC
I've begun seeing these errors after starting to run 2.6.38.8 on my HP 6715b 
laptop ( CPU: AMD x86_64 TL-64 X2 ) with its on-board wireless card:
  BCM4312 802.11a/b/g (rev 02)
I'm pretty sure it did not happen with 2.6.38.2 and earlier under similar
conditions .

What happens is that when I leave the machine inactive for several hours
(no keystrokes or mouse activity, but the wireless interface is associated
with AP, bound to a DHCP IP,  and is serving a website I want to keep online)
eventually I get these error messages in the logs:

[ 2622.241065] b43-phy0 ERROR: Fatal DMA error: 0x00000800, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000
[ 2622.241073] b43-phy0 ERROR: This device does not support DMA on your system. It will now be switched to PIO.
[ 2622.241079] b43-phy0: Controller RESET (DMA error) ...
[ 2622.241711] b43-phy0 debug: Wireless interface stopped
[ 2622.241871] device: 'hw_random': device_unregister
[ 2622.241875] PM: Removing info for No Bus:hw_random
[ 2622.243164] device: 'hw_random': device_create_release
[ 2622.243179] b43-phy0 debug: DMA-64 rx_ring: Used slots 12/64, Failed frames 0/0 = 0.0%, Average tries 0.00
[ 2622.243394] b43-phy0 debug: DMA-64 tx_ring_AC_BK: Used slots 0/256, Failed frames 0/0 = 0.0%, Average tries 0.00
[ 2622.246095] b43-phy0 debug: DMA-64 tx_ring_AC_BE: Used slots 68/256, Failed frames 2053/74366 = 2.7%, Average tries 1.36
[ 2622.249066] b43-phy0 debug: DMA-64 tx_ring_AC_VI: Used slots 0/256, Failed frames 0/0 = 0.0%, Average tries 0.00
[ 2622.251064] b43-phy0 debug: DMA-64 tx_ring_AC_VO: Used slots 2/256, Failed frames 0/22 = 0.0%, Average tries 1.45
[ 2622.253068] b43-phy0 debug: DMA-64 tx_ring_mcast: Used slots 0/256, Failed frames 0/0 = 0.0%, Average tries 0.00
[ 2622.404082] b43-phy0: Loading firmware version 478.104 (2008-07-01 00:50:23)
[ 2622.438130] b43-phy0 debug: Chip initialized
[ 2622.438201] b43-phy0 debug: PIO initialized
[ 2622.438280] b43-phy0 debug: QoS enabled
[ 2622.446648] device: 'hw_random': device_add
[ 2622.446964] PM: Adding info for No Bus:hw_random
[ 2622.447681] b43-phy0 debug: Wireless interface started
[ 2622.447689] b43-phy0: Controller restarted
[ 2623.704970] wlan0: detected beacon loss from AP - sending probe request
[ 2624.206653] ieee80211 phy0: wlan0: No probe response from AP 50:67:f0:94:f3:55 after 500ms, try 1
[ 2624.708129] ieee80211 phy0: wlan0: No probe response from AP 50:67:f0:94:f3:55 after 500ms, try 2
[ 2625.209158] ieee80211 phy0: wlan0: No probe response from AP 50:67:f0:94:f3:55 after 500ms, try 3
[ 2625.710180] ieee80211 phy0: wlan0: No probe response from AP 50:67:f0:94:f3:55 after 500ms, try 4
[ 2626.211128] ieee80211 phy0: wlan0: No probe response from AP 50:67:f0:94:f3:55 after 500ms, disconnecting.
[ 2626.213863] ieee80211 phy0: Removed STA 50:67:f0:94:f3:55
[ 2626.215385] ieee80211 phy0: Destroyed STA 50:67:f0:94:f3:55
[ 2626.216414] ieee80211 phy0: device now idle


And after such 'Fatal DMA Error' messages appear in the log, it never 
responds to an IP packet again until I do a full power-down and reboot . 

Oddly, after the DMA error, I can reset the iwconfig parameters and
a  DHCP / BOOTP request succeeds, (this involves only ethernet level packets) ,  and the interface reports as 'UP', but attempting to send any IP packet on
the interface (eg. a 'ping' to the ADSL router that sent it the DHCP  response),
 results in an ICMP 'Network Unreachable' error, even though the routes are correct.

I've tried leaving a "while sleep 10; do ping -c 1 router; done" going ,
but the problem still occurs, so it does not seem related to some interface
inactivity timer ; yet the problem only occurs when I leave the machine
untouched (no input) for over two hours .

Any ideas / suggestions gratefully received.
Comment 19 cool_sch.alex 2011-06-18 15:53:37 UTC
"And after such 'Fatal DMA Error' messages appear in the log, it never 
responds to an IP packet again until I do a full power-down and reboot"


That's right.
If you "soft"-reboot to windows, the adapter doesn't work too...
It seems to be crashed.
Only after a full power-down the (Wlan-)card seems to be resetted correctly and I am able to use it.
(I noticed same thing with my SoundCard, but other PC; the SoundCard "crashes" in linux and is until a full power-down not usable. -- Soundcard is an AuzenTech X-Fi 5.1 Forte, PCIe)

Also I have to notice, my WLAN-Card does not support 802.11a, only b/g.
Greetings
Comment 20 Alan 2013-12-19 13:51:35 UTC
If this is still seen with a modern kernel please update

Note You need to log in before you can comment on or make changes to this bug.