system remains frozen with the chip bcm4320

dissident_goodchild · 11-29-2007, 09:09 AM

I need a help, the system remains frozen. I am hopeless ;-(

I have a USB Belkin G+ F5D7051 wireless adapter with Broadcom bcm4320 chip. I load win driver with ndiswrapper 1.47 in openSUSE 10.3.

The driver was installed correctly and hardware is present.

If run 'ndiswrapper -ma', the system booting returns 'Kernel panic - not syncing...', in other case, if run 'modprobe ndiswrapper' to turn the device, the system remains frozen when device is detect.

I have read that the kernel of Linux manage the usb energy, some devices need more energy and the kernel blocks it. I have added this line in the file /etc/udev/rules.d/99-custom.rules or .../05-udev-early.rules, but the problem is not resolved.

Code:

BUS=="usb", SYSFS{idProduct}=="7051", SYSFS{idVendor}=="050d", RUN+="/bin/sh -c 'echo 1 > /sys/$devpath/device/bConfigurationValue'"

I try also without result:

.../sys/bus/usb/devices/usb%k/bConf...
.../sys/bus/usb/devices/%k/bConf...
.../sys/%p/device/bConf...

How can I solve this problem?

dissident_goodchild · 11-30-2007, 07:05 AM

Please, someone knows of this?, I am patient but have to solve the problem of a third person to which I have to return him the computer as soon as possible.

I have observed that from time to time opens the nautilus window of the usb-storages, as if resumed the usb electrical sockets, perhaps be this the problem?, where the configuration hardware of the usb's can be modified in openSUSE?

Thanks in advance.

PTrenholme · 11-30-2007, 09:50 AM

I don't know if this is a problem on SUSE, but

1) Some (older) versions of ndiswrapper are incompatible with kernel versions after 2.6.21
2) Fedora automatically installs the FOSS b43 driver for Broadcom chipsets. If SUSE also does this, or installs the older bcm43xx driver, and you then install ndiswarpper, you could be trying to run two different drivers for the same hardware. Check ps -e | grep b43 to see if you've got any of the b43 drivers running.

Edit: And, since the b43 drivers are available, you could junk ndiswrapper and go with the FOSS solution.

dissident_goodchild · 11-30-2007, 10:38 AM

Many thanks, I am going to revise all, although already I have verified that I do not have running bc43.

I have got that the system not be frozen with starting the kernel with apic = off, although some time remains hanging, I have to test it well and seek the drivers b43.

Umm I see that b43 is only for PCI... my chipset is a bcm4320 for usb.

b43 devices list supported

dissident_goodchild · 11-30-2007, 03:57 PM

Delays more in being frozen but the problem persists.

Also I have a problem with VirtualBox that returns me 'error inserting vboxdrv' and I have seen a solution charging ‘nmi_watchdog=0' in the boot kernel.

What I have understood of NMI is a controller of interruptions. He can have something to do with the problem of the wireless adapter?

PTrenholme · 11-30-2007, 04:23 PM

Well, yes, it's possible that NMI could be your problem. What (if anything) do your log files, especially /var/log/messages, tell you?

Have you run a memchk to verify that you have no memory problems? One of the more common causes of "random" system freeze is bad memory.

Another common problem is a failing disk drive. Is your problem system's drive S.M.A.R.T., and are you running smartd to catch those messages?

Another common cause, and, I think, more likely in your case, is an apparent system freeze caused by a kernel "wait" for an interrupt from a network device. Those "waits," however, usually time-out after thirty to sixty seconds, although I've seen some systems where the time-out was defaulted to 300 seconds. Have you tried waiting for 5 to 10 minutes when the system seems "frozen" to see if it "thaws" and, hopefully, produces an error message that can point you in the correct direction?

dissident_goodchild · 11-30-2007, 06:10 PM

I have expected to an hour but the system did not leave the comma.

I tested with the mentioned NMI but did not solve anything.

I am going to verify everything that you explain me. The truth is that the disk does some very rare sounds... but yes it is clear that the failure produces it the adaptor Belkin, if I do not plug in the adaptor the system does not freeze.

Memchk is not available. Smartd is running, How catching I those messages?

Your answer is very satisfactory because you have centered it in the possible causes, thanks.

PTrenholme · 12-01-2007, 10:05 AM

smartd will post any error messages to the log file, usually /var/log/messages.

Have you look at that log file for any helpful messages?

Have you tried booting from a "live CD" with the card plugged in to see if it can be handled? The Ubuntu 6.10 CD or similar might be helpful.

-------------------------------------------------------

What follows may not be possible for you to do, since the system is freezing when the card is plugged in. But it might be helpful, so . . .

Can you look at the output of the pstree command? It should be something like this:

Code:

$ pstree
init─┬─/usr/bin/sealer
     ├─NetworkManager───{NetworkManager}
     ├─acpid
     ├─anacron───run-parts─┬─000-delay.cron───sleep
     │                     └─awk
     ├─atd
     ├─auditd─┬─audispd───{audispd}
     │        └─{auditd}
     ├─avahi-daemon───avahi-daemon
     ├─console-kit-dae───61*[{console-kit-dae}]
     ├─crond
     ├─cupsd
     ├─2*[dbus-daemon───{dbus-daemon}]
     ├─dbus-launch
     ├─dcopserver
     ├─dhclient
     ├─exe───2*[{exe}]
     ├─gconfd-2
     ├─gnome-keyring-d
     ├─gpg-agent
     ├─hald───hald-runner─┬─hald-addon-acpi
     │                    ├─hald-addon-inpu
     │                    ├─2*[hald-addon-stor]
     │                    └─hald-addon-usb-
     ├─hcid───bluetoothd-serv
     ├─irqbalance
     ├─kaccess
     ├─kded
     ├─kdeinit─┬─artsd
     │         ├─consolehelper-g───userhelper───system-config-s
     │         ├─firefox───run-mozilla.sh───firefox-bin───8*[{firefox-bin}]
     │         ├─gkrellm
     │         ├─kdf
     │         ├─kio_file
     │         ├─klauncher
     │         ├─konsole───bash
     │         ├─kwin
     │         ├─nm-applet
     │         ├─puplet
     │         └─python
     ├─kdesktop───konsole───bash───pstree
     ├─kdm─┬─X
     │     └─kdm───startkde─┬─kwrapper
     │                      └─ssh-agent
     ├─kicker
     ├─kio_uiserver
     ├─klipper
     ├─knotify
     ├─konqueror
     ├─kpowersave
     ├─ksmserver
     ├─6*[mingetty]
     ├─2*[mount.ntfs-3g]
     ├─mysqld_safe───mysqld───9*[{mysqld}]
     ├─nasd
     ├─nmbd
     ├─notification-da
     ├─ntpd
     ├─restorecond
     ├─rklogd
     ├─rpc.idmapd
     ├─rpc.statd
     ├─rpcbind
     ├─rsyslogd───{rsyslogd}
     ├─2*[sendmail]
     ├─setroubleshootd───2*[{setroubleshootd}]
     ├─smbd───smbd
     ├─start_kdeinit
     ├─udevd
     ├─winbindd───winbindd
     ├─wpa_supplicant
     └─xinetd

I added the red highlights to some of the network processes that should be running. (Note that I'm at my desktop system, which uses a wired network, so no "wireless" processes are shown.)

Also look at the dmesg output for duplicated drivers. Again, for my desktop system with two Ethernet ports (one disconnected), I get this:

Code:

$ dmesg | grep -i '\(net\)\|\(eth\)'
NET: Registered protocol family 16
NetLabel: Initializing
NetLabel:  domain hash size = 128
NetLabel:  protocols = UNLABELED CIPSOv4
NetLabel:  unlabeled traffic allowed by default
NET: Registered protocol family 2
audit: initializing netlink socket (disabled)
SELinux:  Registering netfilter hooks
Initializing XFRM netlink socket
NET: Registered protocol family 1
NET: Registered protocol family 17
8139too Fast Ethernet driver 0.9.28
8139cp: 10/100 PCI Ethernet driver v1.3 (Mar 22, 2004)
eth0: RealTek RTL8139 at 0xf897cc00, 00:11:2f:7d:e6:3c, IRQ 22
eth0:  Identified 8139 chip type 'RTL-8101'
eth1: RealTek RTL8139 at 0xf899c800, 00:30:bd:1e:95:3e, IRQ 22
eth1:  Identified 8139 chip type 'RTL-8100B/8139D'
NET: Registered protocol family 10
IA-32 Microcode Update Driver: v1.14a <tigran@aivazian.fsnet.co.uk>
ip6_tables: (C) 2000-2006 Netfilter Core Team
Netfilter messages via NETLINK v0.30.
ip_tables: (C) 2000-2006 Netfilter Core Team
NET: Registered protocol family 31
eth1: link down
ADDRCONF(NETDEV_UP): eth1: link is not ready
eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
eth0: no IPv6 routers present

dissident_goodchild · 12-01-2007, 08:18 PM

Results of pstree:

Code:

init─┬─acpid
     ├─auditd─┬─audispd
     │        └─{auditd}
     ├─avahi-daemon
     ├─avahi-dnsconfd
     ├─beagled───10*[{beagled}]
     ├─beagled-helper───5*[{beagled-helper}]
     ├─bonobo-activati───{bonobo-activati}
     ├─console-kit-dae───61*[{console-kit-dae}]
     ├─cron
     ├─cupsd
     ├─2*[dbus-daemon]
     ├─dbus-launch
     ├─dhcpcd
     ├─evolution-data-───2*[{evolution-data-}]
     ├─firefox───firefox-bin───7*[{firefox-bin}]
     ├─gconfd-2
     ├─gdm───gdm─┬─X───Xorg
     │           └─gnome-session─┬─compiz───gtk-window-deco
     │                           ├─esd
     │                           ├─gnome-cups-icon
     │                           ├─gnome-panel
     │                           ├─nautilus
     │                           ├─nm-applet
     │                           ├─opensuse-update
     │                           ├─ssh-agent
     │                           └─{gnome-session}
     ├─gnome-keyring-d
     ├─gnome-power-man
     ├─gnome-screensav
     ├─gnome-settings-───{gnome-settings-}
     ├─gnome-terminal─┬─bash───su───bash───pstree
     │                ├─gnome-pty-helpe
     │                └─{gnome-terminal}
     ├─gnome-vfs-daemo
     ├─gnome-volume-ma
     ├─gweather-applet
     ├─hald───hald-runner─┬─hald-addon-acpi
     │                    ├─hald-addon-inpu
     │                    └─2*[hald-addon-stor]
     ├─ifplugd
     ├─irqbalance
     ├─klogd
     ├─main-menu
     ├─mapping-daemon
     ├─master─┬─pickup
     │        └─qmgr
     ├─6*[mingetty]
     ├─mixer_applet2
     ├─nscd───8*[{nscd}]
     ├─ntpd
     ├─portmap
     ├─powersaved
     ├─resmgrd
     ├─rpc.rquotad
     ├─smartd
     ├─smpppd
     ├─sshd
     ├─syslog-ng
     ├─tomboy───2*[{tomboy}]
     ├─udevd
     └─vino-server

Resutls of dmesg | grep...:

Code:

Parsing all Control Methods:
Table [DSDT](id 0001) - 412 Objects with 47 Devices 90 Methods 11 Regions
NET: Registered protocol family 16
Initializing Device/Processor/Thermal objects by executing _INI methods:..
Executed 2 _INI methods requiring 3 _STA executions (examined 51 objects)
NET: Registered protocol family 2
audit: initializing netlink socket (disabled)
NET: Registered protocol family 1
eth0: ADMtek Comet rev 17 at Port 0x1800, 00:30:05:5A:3B:25, IRQ 11.
udev: renamed network interface ath0 to wlan2
IA-32 Microcode Update Driver: v1.14a <tigran@aivazian.fsnet.co.uk>
NET: Registered protocol family 10
ip6_tables: (C) 2000-2006 Netfilter Core Team
ip_tables: (C) 2000-2006 Netfilter Core Team
Netfilter messages via NETLINK v0.30.
NET: Registered protocol family 17
eth0: no IPv6 routers present

dissident_goodchild · 12-01-2007, 08:33 PM

I have bought a PCI OVISLINK with chip Atheros, I have abandoned by impossible the Belkin adapter.

Results of hwinfo --pci:

Code:

23: PCI 10.0: 0282 WLAN controller
  [Created at pci.301]
  UDI: /org/freedesktop/Hal/devices/pci_168c_13
  Unique ID: sWMw.+xpX4rQtix5
  SysFS ID: /devices/pci0000:00/0000:00:10.0
  SysFS BusID: 0000:00:10.0
  Hardware Class: network
  Model: "Atheros EZ Connect g 802.11g 108Mbps Wireless PCI Adapter"
  Vendor: pci 0x168c "Atheros Communications, Inc."
  Device: pci 0x0013 "AR5212/AR5213 Multiprotocol MAC/baseband processor"
  SubVendor: pci 0x168c "Atheros Communications, Inc."
  SubDevice: pci 0x1051 "EZ Connect g 802.11g 108Mbps Wireless PCI Adapter"
  Revision: 0x01
  Driver: "ath_pci"
  Driver Modules: "ath_pci"
  Device File: wlan2
  Device Files: wlan2, wifi0
  Features: WLAN
  Memory Range: 0xc4000000-0xc400ffff (rw,non-prefetchable)
  IRQ: 9 (43322 events)
  HW Address: 00:14:78:74:bd:9c
  Link detected: yes
  Module Alias: "pci:v0000168Cd00000013sv0000168Csd00001051bc02sc00i00"
  Driver Info #0:
    Driver Status: ath_pci is active
    Driver Activation Cmd: "modprobe ath_pci"
  Config Status: cfg=no, avail=yes, need=no, active=unknown

PTrenholme · 12-02-2007, 08:12 AM

Well, that seems like a solution that worked.

I'm puzzled about one thing: Why, in your first dmsg output above, did udev rename the device as wlan2? Is it possible that the computer you're using has a "built-in" Atheros WiFi device? Did you ever try the computer without a WiFi card in the slot?

dissident_goodchild · 12-02-2007, 04:12 PM

I saw wlan2 name. It is a computer of a third person, I had to utilize my usb D-Link to connect to internet and to be able to seek helps (wlan1). Also I installed the Belkin (wlan2) and perhaps therefore appears like wlan2, although I opt more by the process "force persistent names" of the system.

Belkin has left me a bad taste in one's mouth, because I am not sure if is a problem of the adapter or I know not to run.

dissident_goodchild · 12-02-2007, 07:22 PM

I have had a problem with VirtuaBox and in another post have commented that the virtualbox-kmp release should be equal to kernel release.

I believe that this did not happen with the ndiswrapper-kmp. For our information, Could be the problem?

Fatal disaster!, already had bought a new pci adapter.