Update on using b43 with PCI ID 14e4:4315

[-Riffer-]

Originally Posted by lwfinger

Without some diagnostics, this will be hard to fix. The symptom has
not been reported by anyone else.

If you tell me how to diagnose (any debugging flags?) I will do my very best.

Yesterday I told the bios of the notebook to shut of the wlan and as expected the system boots without any problem. Naturaly there is no wlan available then and from dmesg I see it did not tried to load the b43.

[lwfinger]

-Riffer- wrote:
> lwfinger;2040120 Wrote:
>> Without some diagnostics, this will be hard to fix. The symptom has
>> not been reported by anyone else.
>
> If you tell me how to diagnose (any debugging flags?) I will do my very
> best.
>
> Yesterday I told the bios of the notebook to shut of the wlan and as
> expected the system boots without any problem. Naturaly there is no wlan
> available then and from dmesg I see it did not tried to load the b43.

Any console output would be helpful. There are no specific debug
settings for this issue.

The fix for the compilation error you mentioned earlier has been
submitted to the compat-wireless maintainers, but I don't know if it
has been incorporated yet.

[-Riffer-]

I am still on investigations on this freezing and all I can tell is that the ssb-driver is loaded for b43-pci-bridge and while the wireless card is enabled in the bios the driver freezes system on start.

[-Riffer-]

I am sorry for making an new post but this forum does not have an edit function.

Currently the situation is the following:
I have blacklistet ssb and and the entrys

modprobe ssb
modprobe b43

in rc.local

and now it works without freezing the system. But I have no idea why it is like this.
Naturally depmod has been invoked and I see that the ssb will be loaded indirectly by b43 - but it crashes my system if I let it.

[blowf1sh]

try loading the modules manualy and check if a kernel oops appears in the system logs.
What kernel are you using?

[-Riffer-]

Modprobing the module manually does not result in kernels oops and there is no freezings. Maybe it is a very special hardware issue with Compaq 615 notebook?

Currently I am using kernel 2.6.30-6. Compiled for AMD64.

I can live with it, because the driver does load stable with the situation I described (blacklist + modprobe in rc.local). Maybe the situation will be cleared by other fixes in the future.


But now I faced a new bug:
From date 17 september on the driver does no longer work for me (had to compile every version from now on backwards to test).
It does load and tries to connect, but after (/var/log/message) "ADDRCONF(NETDEV_UP): wlan0: link is not ready" there comes nothing and the ubuntu net-manager cycles after a while in trying to connect. From older versions I know there should come "ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready" a few seconds later.

I had a look into the sources between versions from 16 to 17 of september (with fldiff) and I suspect the critical difference is in sme.c (line 777):

/* we're good if we have a matching bss struct */
		if (bss) {
			wdev->conn->state = CFG80211_CONN_AUTHENTICATE_NEXT;
			err = cfg80211_conn_do_work(wdev);
			cfg80211_put_bss(bss);
		} else {

As far as I understand from the sources there was a migration from boolean to structure in function
"*cfg80211_get_conn_bss()".

I am sorry to say that my programming skills in C is not good enough to help with solving that.
But maybe my post helps to fix something :-)

[letocharm]

Hi, I've been following this thread, having the same device. I'm just writing to confirm the issue Riffer posted about the perpetual "ADDRCONF(NETDEV_UP): wlan0: link is not ready."

I've tried multiple firmwares - the one recommended by iwfinger and the one on the linux-wireless driver site. Both yield the same results.

I'm going to try the supposed "good" driver dated Sept 16 and will report back.

[letocharm]

So it works - except on resume from suspend. Then the device is never ready and i have to manually reload the driver. rmmod b43 produces

------------[ cut here ]------------
WARNING: at /root/compat-wireless-2009-09-16/net/mac80211/scan.c:267 ieee80211_scan_completed+0x2a6/0x2c0 [mac80211]()
Hardware name: Inspiron 1525
Modules linked in: usbhid hid b43(-) nls_utf8 isofs nls_iso8859_1 nls_cp437 usb_storage ipv6 snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss vfat fat cpufreq_ondemand acpi_cpufreq freq_table lp ppdev parport_pc parport fuse ide_cd_mod cdrom snd_hda_codec_idt snd_hda_codec_intelhdmi uvcvideo videodev v4l1_compat v4l2_compat_ioctl32 ide_pci_generic mac80211 joydev snd_hda_intel i915 cfg80211 snd_hda_codec snd_hwdep drm ssb snd_pcm snd_timer sdhci_pci sdhci i2c_algo_bit snd video mmc_core thermal processor soundcore psmouse firewire_ohci uhci_hcd rtc_cmos thermal_sys intel_agp piix sg ehci_hcd firewire_core pcmcia ac snd_page_alloc rtc_core dell_laptop battery button hwmon dcdbas i2c_i801 sky2 pcmcia_core rtc_lib ide_core serio_raw led_class i2c_core evdev wmi output crc_itu_t ahci libata ext4 jbd2 crc16 [last unloaded: b43]
Pid: 524, comm: rmmod Tainted: G W 2.6.31-64 #1
Call Trace:
[<ffffffffa03490a6>] ? ieee80211_scan_completed+0x2a6/0x2c0 [mac80211]
[<ffffffffa03490a6>] ? ieee80211_scan_completed+0x2a6/0x2c0 [mac80211]
[<ffffffff8103e389>] ? warn_slowpath_common+0x79/0xd0
[<ffffffffa03490a6>] ? ieee80211_scan_completed+0x2a6/0x2c0 [mac80211]
[<ffffffffa0351ce1>] ? ieee80211_stop+0x5c1/0x5f0 [mac80211]
[<ffffffff8126bf00>] ? dev_deactivate+0x1b0/0x1d0
[<ffffffff8125a1f3>] ? dev_close+0x73/0xc0
[<ffffffff8125a286>] ? rollback_registered+0x46/0x2f0
[<ffffffff8125a545>] ? unregister_netdevice+0x15/0x60
[<ffffffffa03511df>] ? ieee80211_remove_interfaces+0x7f/0xc0 [mac80211]
[<ffffffffa034511a>] ? ieee80211_unregister_hw+0x3a/0xe0 [mac80211]
[<ffffffffa0375a95>] ? b43_remove+0xa5/0xb0 [b43]
[<ffffffffa026c983>] ? ssb_device_remove+0x23/0x40 [ssb]
[<ffffffff811f0ca2>] ? __device_release_driver+0x72/0xf0
[<ffffffff811f0dd0>] ? driver_detach+0xb0/0xc0
[<ffffffff811efd80>] ? bus_remove_driver+0x90/0xd0
[<ffffffffa03909dc>] ? b43_exit+0x10/0x19 [b43]
[<ffffffff8106980e>] ? sys_delete_module+0x1be/0x290
[<ffffffff81028767>] ? do_page_fault+0x167/0x310
[<ffffffff811630c1>] ? __up_read+0x21/0xc0
[<ffffffff8100b46b>] ? system_call_fastpath+0x16/0x1b
---[ end trace 31484438c4922444 ]---

[-Riffer-]

I mosten use suspend to disk where it works for me.

Thus you can confirm my notice about the drivers later then 16th (up to 20th which was my last test) of september are not usable?

(Apart from these problems I have random drops of the connection. Maybe you have them too? The drops apear mostly in cases where the system is on high load situation (playing a flash movie in FF for example) but from time to time random.
I'm not sure wether the problem is caused by my wireless router (LinkSys WRT 54g with DD-WRT Firmware) or the wireless chip. The drops exist with any driver (b43 or Broadom STA) under linux. Testing with windows is still an open task.)

[-Riffer-]

Hmm... no progress in this thread.

Is anyone reading this who could work on the problems we have?

Else I think it would be a good idea to inform the official maintainer of the bleeding edge drivers.