[P4M900][HP 2133] wl module causes hard lock

[Justin Chudgar]

Binary package hint: bcmwl-kernel-source

Using the latest, as of now, version of dkms and bcmwl-kernel-source, the wl module builds seemingly find. Doing a depmod -a and update-initramfs cycle, the module loads on boot; the laptop (HP Mininote 2133) WiFi light transitions to blue indicating that the driver loaded.

Upon kdm starting up, however, the entire system hard locks. The three finger salute does nothing, nor does <ALT-PRTSCR
N+(REISUB)>. The only way to recover is to force a power cycle.

Making things even more wierd, this is very repeatable; however, if one keeps the cursor moving continuously using the integrated Synaptics touchpad while kdm/X11 are loading, the system boots without locking.

In this latter case, the eth1 wireless interface shows in the logs, but any attempt to interact with it, e.g. iwconfig, causes an immediate hard lock.

This particular BCM4312 device worked well with ndiswrapper through hardy and with the wl module in LRM in intrepid. It never worked with the Jaunty kernels; which motivated my upgrade to karmic. Unfortunately, it still does not work. Being a laptop, this is a serious regression.

Note that with the Jaunty and Karmic kernels, neither wl, nor ndiswrapper, nor b43 work. With Intrepid and earlier, I used the VIA modules for the Chrome9HC video. For the Jaunty and Karmic kernels, this is not available from VIAarena, so I am using the openchrome module that ships with Ubuntu.

Let me know how I can help as this is driving me (more) bonkers (than normal :)).

[Bartosz Kosiorek (gang65)]

Please attach the output of `lspci -vvnn`, and attach your /var/log/Xorg.0.log (and maybe Xorg.0.log.old) file from after reproducing this issue. If you've made any customizations to your /etc/X11/xorg.conf please attach that as well.

[Justin Chudgar (justinzane)]

As you requested, lspci -vvnn and the Xorg logs. Sorry for the delay. End of semester work to do.

[Justin Chudgar (justinzane)]

Well, I remembered that I have disabled the WLAN (BCM4312) via BIOS so that I can use my computer. So, to get you a version that shows the bug, I re-enabled WLAN and rebooted. The system hard-locked as expected. I powered off manually and, after disbabling the WLAN controller, I am attaching Xorg.1.log which shows where the lock occurs. I hope this helps.

[Justin Chudgar (justinzane)]

FWIW, Andrew Aylett's patched openchrome module did not resolve the problem for me. You can find his patch at https://launchpad.net/~andrew-aylett/+archive/ppa/+sourcepub/608672/+listing-archive-extra .

In addition to disabling the WLAN device in BIOS, I can also boot into single user mode, "modprobe -r wl; modprobe -r ssb" and then continue booting and avoid the hard lock.

This problem looks superficially similar to https://bugs.launchpad.net/bugs/331952 however, that bug references the Jaunty kernel not the Karmic one.

It seems like the wl module and the openchrome module are stomping on each-others toes; but, I am not sophisticated enough to debug their operations. For those of us on the Mininote 2133 platform, this is a killer bug, since netbooks without WiFi lose much of their appeal.

[Justin Chudgar (justinzane)]

Here is the lspci -vvnn from a boot with the WLAN module enabled and the wl and ssb modules manually removed:

02:00.0 Network controller [0280]: Broadcom Corporation BCM4312 802.11a/b/g [14e4:4312] (rev 02)
        Subsystem: Hewlett-Packard Company Device [103c:1370]
        Control: I/O- Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Interrupt: pin A routed to IRQ 24
        Region 0: Memory at fdffc000 (64-bit, non-prefetchable) [size=16K]
        Capabilities: [40] Power Management version 3
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=2 PME-
        Capabilities: [58] Vendor Specific Information <?>
        Capabilities: [e8] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
                Address: 0000000000000000 Data: 0000
        Capabilities: [d0] Express (v1) Endpoint, MSI 00
                DevCap: MaxPayload 128 bytes, PhantFunc 0, Latency L0s <4us, L1 unlimited
                        ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset-
                DevCtl: Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
                        RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop-
                        MaxPayload 128 bytes, MaxReadReq 128 bytes
                DevSta: CorrErr+ UncorrErr- FatalErr- UnsuppReq+ AuxPwr- TransPend-
                LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Latency L0 <4us, L1 <64us
                        ClockPM+ Suprise+ LLActRep- BwNot-
                LnkCtl: ASPM L0s Enabled; RCB 64 bytes Disabled- Retrain- CommClk-
                        ExtSynch- ClockPM+ AutWidDis- BWInt- AutBWInt-
                LnkSta: Speed 2.5GT/s, Width x1, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
        Capabilities: [100] Advanced Error Reporting <?>
        Capabilities: [13c] Virtual Channel <?>
        Capabilities: [160] Device Serial Number 21-00-35-ff-ff-00-96-09
        Capabilities: [16c] Power Budgeting <?>
        Kernel modules: wl, ssb ...

[Justin Chudgar (justinzane)]

A bugzilla.kernel.org bug about the same BCM4312 wireless module and the Chrome9HC video module on the Mininote 2133 platform.

Don't know how related they are...

[Justin Chudgar (justinzane)]

I do not believe that this is a duplicate. I have not had the same screen resolution problems observed in 355918. Instead my system is inoperable, completely locked if both openchrome and bcmwl are loaded together.

[Bartosz Kosiorek (gang65)]

Please build/install the latest openchrome driver and enable debug mode.
The detail instruction is available at:
https://help.ubuntu.com/community/OpenChrome

After that attach your Xorg.0.log file, with enabled and disabled WLAN.

[Justin Chudgar (justinzane)]

I would love to; however, due to another karmic bug; I cannot. There is an issue with the current version of libgcrypt11 that causes "stack smashing" errors in anything that uses it. This includes svn. So I get:

justin@justin-mininote:~/src$ svn checkout http://svn.openchrome.org/svn/trunk openchrome
*** stack smashing detected ***: svn terminated
======= Backtrace: =========
/lib/tls/i686/cmov/libc.so.6(__fortify_fail+0x48)[0x2bb038]
/lib/tls/i686/cmov/libc.so.6(__fortify_fail+0x0)[0x2baff0]
/lib/libgcrypt.so.11[0x82b4a4]
/lib/libgcrypt.so.11[0x822d14]
[0x0]
======= Memory map: ========
00110000-00131000 r-xp 00000000 08:01 1548521 /usr/lib/libaprutil-1.so.0.3.7
00131000-00132000 r--p 00020000 08:01 1548521 /usr/lib/libaprutil-1.so.0.3.7
00132000-00133000 rw-p 00021000 08:01 1548521 /usr/lib/libaprutil-1.so.0.3.7
00133000-00145000 r-xp 00000000 08:01 2959951 /lib/tls/i686/cmov/libresolv-2.9.so
00145000-00146000 r--p 00011000 08:01 2959951 /lib/tls/i686/cmov/libresolv-2.9.so
00146000-00147000 rw-p 00012000 08:01 2959951 /lib/tls/i686/cmov/libresolv-2.9.so
00147000-00149000 rw-p 00000000 00:00 0
00149000-00159000 r-xp 00000000 08:01 1548843 /usr/lib/libtasn1.so.3.1.5
00159000-0015a000 r--p 0000f000 08:01 1548843 /usr/lib/libtasn1.so.3.1.5
0015a000-0015b000 rw-p 00010000 08:01 1548843 /usr/lib/libtasn1.so.3.1.5
0015b000-0015d000 r-xp 00000000 08:01 2941184 /lib/libcom_err.so.2.1
0015d000-0015e000 r--p 00001000 08:01 2941184 /lib/libcom_err.so.2.1
0015e000-0015f000 rw-p 00002000 08:01 2941184 /lib/libcom_err.so.2.1
0015f000-00166000 r-xp 00000000 08:01 1550767 /usr/lib/libkrb5support.so.0.1
00166000-00167000 r--p 00006000 08:01 1550767 /usr/lib/libkrb5support.so.0.1
00167000-00168000 rw-p 00007000 08:01 1550767 /usr/lib/libkrb5support.so.0.1
00168000-0016a000 r-xp 00000000 08:01 2941272 /lib/libkeyutils-1.2.so
0016a000-0016b000 r--p 00001000 08:01 2941272 /lib/libkeyutils-1.2.so
0016b000-0016c000 rw-p 00002000 08:01 2941272 /lib/libkeyutils-1.2.so
0016f000-001bb000 r-xp 00000000 08:01 1554698 /usr/lib/libsvn_subr-1.so.1.0.0
001bb000-001bc000 r--p 0004c000 08:01 1554698 /usr/lib/libsvn_subr-1.so.1.0.0
001bc000-001bd000 rw-p 0004d000 08:01 1554698 /usr/lib/libsvn_subr-1.so.1.0.0
001bd000-00319000 r-xp 00000000 08:01 2959936 /lib/tls/i686/cmov/libc-2.9.so
00319000-0031a000 ---p 0015c000 08:01 2959936 /lib/tls/i686/cmov/libc-2.9.so
0031a000-0031c000...

[Justin Chudgar (justinzane)]

Ok, fixed the libgcrypt thing... Built and installed openchrome r751 from subversion. Still am not able to use the broadcom wireless with openchrome. I have attached the Xorg.0.log with WiFi enabled.

[Justin Chudgar (justinzane)]

If I reboot into single user mode and manually (modprobe -r wl) remove the bcmwl driver, then continue booting, I get the wrong resolution as mentioned in bug 355918. I've attached the log from this boot.

[Justin Chudgar (justinzane)]

Again, rebooting, this time with the BCM4312 wireless disabled in BIOS. I still get the wrong resolution unless I include:
ActiveDevice LCD
PanelSize 1280x768

I've attached the log for this one as well.

[Bartosz Kosiorek (gang65)]

Please apply this patch and check if it helps.

[Justin Chudgar (justinzane)]

Woo! Hoo! Hunk 1 of the patch failed, but hunk 2 applied. With the patched openchrome (r751 + hp2133 patch), I'm surfing on WiFi with accelerated graphics for the first time since Intrepid.

Thank you very much. One of the things I love most about Ubuntu and FOSS is the wealth of really smart, helpful people like yourself.

Will this likely be included in the standard Ubuntu packages or will HP 2133 users need to patch and build their own going forward?

[Bartosz Kosiorek (gang65)]

Please apply this patch.
Does it help?

[Bartosz Kosiorek (gang65)]

Of course remove openchrome_741_hp_2133.patch patch before apply the openchrome-switch-on-lvds-pads-only-for-active-channels.patch.

[Bartosz Kosiorek (gang65)]

Hi.
Please check the latest openchrome revision.
The problem can be resolved at openchrome 758 revision.

The detail build instruction is available at:
https://help.ubuntu.com/community/OpenChrome

[Bartosz Kosiorek (gang65)]

Please build the latest openchrome revision.

[Justin Chudgar (justinzane)]

As of kernels 2.6.31-4 and 2.6.31-5, the bug reappears. I hope that the patch can be integrated into the relevant binary .debs so that upgrades will not cause unpleasant regressions.

I am unsure what patch I should apply, please advise. I am using the most current karmic kernel with
ii xserver-xorg-video-openchrome 1:0.2.903+svn758-0ubuntu1 X.Org X server -- VIA display driver

If there is any debug info I can provide, let me know what would be helpful.

Thanks, Bartosz

[Bartosz Kosiorek (gang65)]

This bug was fixed in Karmic Koala (openchrome 1:0.2.903+svn758-0ubuntu1)

Thanks