[Hardy]Recent kernel update to 2.6.24-11 breaks b43 (with bcm4312)

Okay, now it works again... Maybe my university network was just rebooting, just when I did to test the new kernel? ;)

Okay, in fact, this seems to be a random issue. Booting the last two times into 2.6.24-11, I could not even find the network, 2.6.24-10 works instantly.

My problem is that until kernel 2.6.24-4 the integrated wireless card Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02) worked smoothly. Since update to 2.6.24-8, then 2.6.24-10, 2.6.24-11 still the same problem which is that wireless networks are visible, but I cannot connect to any of them. KNetworkManager hangs on message "Activation stage: Configuring device.". When booting back the 2.6.24-4 kernel, the wireless works smoothly again. Must be related to kernel updates after 2.6.24-4.  Strange thing is that all the network scanning stuff works, just the attempt to establish a connection fails, does not seem to even reach the IP configuration stage.

dmesg:

[ 1077.748844] wlan0: RX AssocResp from 00:18:f3:57:f5:3c (capab=0x411 status=18 aid=1)
[ 1077.748852] wlan0: AP denied association (code=18)
[ 1077.882494] wlan0: association with AP 00:18:f3:57:f5:3c timed out
[ 1092.535767] wlan0: RX deauthentication from 00:18:f3:57:f5:3c (reason=7)
[ 1092.535774] wlan0: deauthenticated
[ 1092.538031] wlan0: Initial auth_alg=0
[ 1092.538037] wlan0: authenticate with AP 00:18:f3:57:f5:3c
[ 1092.538061] wlan0: RX deauthentication from 00:18:f3:57:f5:3c (reason=7)
[ 1092.538723] wlan0: Initial auth_alg=0
[ 1092.538727] wlan0: authenticate with AP 00:18:f3:57:f5:3c
[ 1092.539752] wlan0: RX authentication from 00:18:f3:57:f5:3c (alg=0 transaction=2 status=0)
[ 1092.539758] wlan0: authenticated
[ 1092.539762] wlan0: associate with AP 00:18:f3:57:f5:3c
[ 1092.541529] wlan0: authentication frame received from 00:18:f3:57:f5:3c, but not in authenticate state - ignored
[ 1092.971378] wlan0: RX AssocResp from 00:18:f3:57:f5:3c (capab=0x411 status=18 aid=1)
[ 1092.971389] wlan0: AP denied association (code=18)
[ 1093.136999] wlan0: association with AP 00:18:f3:57:f5:3c timed out

Connecting to unsecured network passes to IP Configuration Stage, which fails, but it could be caused by the fact that my MAC address is rejected. Perhaps my bug could be reevaluted as WPA problem related to upgrade to Ubuntu Hardy with kernel 2.6.24-8 and newer.

It seems that the bug I describe here is this one reported here: https://bugs.launchpad.net/ubuntu/+bug/183619

Hi All,

Mikeos, you appear to have a different wireless card than the original bug reporter. It's helpful to the development teams if bug reports target a specific issue against a specific set of hardware since in most cases the fix will require code changes to different drivers. I see you seem to have found another existing bug report which has the same hardware as you. I'd enourage you to continue to follow up there.

der_vegi, it seems quite a few changes for the broadcom b43 driver went in between 2.6.24-10 and 2.6.24-11. Refer to: https://launchpad.net/ubuntu/+source/linux/ . Just curious how comfortable you are building your own kernel and doing a git bisect to help narrow down the specific offending patch that may be causing you issues? The following links may help:

https://wiki.ubuntu.com/KernelTeam/GitKernelBuild - rather than using the upstream mainline kernel use the Ubuntu kernel source to build
http://www.kernel.org/doc/local/git-quick.html#bisect

Let us know if this is something you'd be willing to do as I realize it is a bit much to ask of you a bug reporter. I just know it would be extremely helpful to the kernel team. Thanks.

Mikeos, I would like to do a bisect, but I am behind a very restrictive proxy so I cannot download the git tree... Maybe I get the chance to connect to a less restrictive network these days, then I will try.

The thing that happened between -10 and -11 was that a patch to the module "ssb" was applied, to let the revision 02 of the broadcom card 4311 work. (though it does not work well, since I have continuous disconnects)

Still present in -12. Unfortunately, I am still not able to do a bisect, as the less restricted network, where I could do a git-clone, is too slow and unstable to do it in one run and I haven't found out yet how to resume an interrupted git transfer.

After the latest update from linux-image-2.6.24-12-generic 2.6.24-12.20 to 2.6.24-12.22, the b43 driver works now for me.

Can anyone confirm this?

It works for me too, but I have disabled gnome-power-manager which was
causing lots of problems to b43. I will try to re-enable it and see
what happens

Strange. I have rebooted three times to confirm that the new kernel's b43 is working, which it did. I even had the impression, that it worked with a better performance than usual.

But now, after shutting the computer down and starting a while later, it did not work any more, even after several reboots and reloading the module...

try to kill gnome-power-manager and see what happens.

Killing gnome-power-manager doesn't work for me.

Okay, I have now build my own kernel following the link above (Ubuntu wiki), using the latest 2.6.24 stable (git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.24.y.git) based on 2.6.24.3. This kernel works for me, and this time I did not just reboot... *g Booting the Ubuntu-kernel 2.6.24-12 in between resulted in a not-working b43 driver again.

I confirm that gnome-power-manager is not the problem.
can you share your kernel .deb?

Well, I don't know why, but my linux-image[...].deb is 200 MB. Is it possible (and reasonable) to upload such a big file to launchpad? By the way, do you have amd64 as well?

The steps described at https://wiki.ubuntu.com/KernelTeam/GitKernelBuild to build your own git-kernel are quite easy to follow. But using the link for git-clone to the git-repo described there, results in downloading the newest kernel, right now 2.6.25-rc_something. Just take git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.24.y.git instead and you will get your state-of-the-art 2.6.24 kernel.

I will try it myswlf. Thanks for the link

Leann, do you still need a bisect? I could do one, if you want. From the kernel.org git or the http://kernel.ubuntu.com/git ?

Hi der_vegi,

Thanks for testing upstream. It's a least a little comforting knowing that Hardy+1 (the Intrepid Ibex) should resolve this issue since it will sync with the upstream kernel. However, it would be nice to see Hardy working. If you would be comfortable doing the git bisect on the Ubuntu kernel: git://kernel.ubuntu.com/ubuntu/ubuntu-hardy.git that would be great. Or if you want to narrow down the upstream patch that fixes this we could also just pull in that patch to Hardy. I'd maybe suggest doing the latter. The following git command can be helpful to see what patches and changes have been applied: git log -p [directory/file]

For example:

git log -p drivers/net/wireless/b43

or

git log -p drivers/net/wireless/b43/main.c

Take out the -p flag if you just want to see the patch descriptions.

I hope that helps. Thanks.

Ah, okay. I thought, you would sync with upstream kernel 2.6.24 until kernel freeze...

So I will try a kind of inverse bisect on the upstream kernel 2.6.24, setting a working kernel to "bad" and a not-working one to "good" to narrow down the patch...? This will take some time, as I cannot be in front of my computer all the time these days.

When has the ubuntu-kernel last time been synced with upstream? I've done a bisect between v2.6.24 and the newest kernel, not finding a kernel, that did not work. Well, maybe the steps of the bisect were too big?

I am starting a bisect of the ubuntu-kernel now...

Well, my poor-mans-testing showed that I have to do a full power cycle between 2.6.24-10 and the buggy kernel, otherwise the bug will not happen. Sorry if I'm stating the obvious :-)

Hi der_vegi,

My apologies, for some reason I was thinking you were testing an upstream 2.6.25-rc kernel. However, it seems you were testing the upstream 2.6.24.x stable kernel (which is still very helpful). It seems the ubuntu kernel has already been synced with the upstream 2.6.24.3 stable kernel release. So it's odd that you would still be experiencing an error. Let us know the results of your bisect on the ubuntu-kernel. There have been additional patches pulled in to the ubuntu kernel that will make it differ from a vanilla 2.6.24.3 kernel from upstream. Thanks.

Well, right now after some steps I get an build error:

 CC [M] fs/xfs/support/uuid.o
  LD [M] fs/xfs/xfs.o
make[1]: Verlasse Verzeichnis '/home/mm/ubuntu-kernel/hardy'
make: *** [debian/stamp-build-kernel] Fehler 2

Diggin further into this...

Okay, I finished the bisect, here the output:
git bisect good
6421a9e1676a3b189bef1f3d539f5b9a36ad5843 is first bad commit
commit 6421a9e1676a3b189bef1f3d539f5b9a36ad5843
Author: Colin Ian King <email address hidden>
Date: Wed Feb 27 15:38:14 2008 -0500

    UBUNTU: BCM4311 Revision 2 fix
    OriginalAuthor: Larry Finger
    OriginalLocation: http://linuxwireless.org/download/b43/patch_2.6.24_for_4311_2
    Bug: #184600

    This patch includes support for the BCM4311 Rev 2.

    Signed-off-by: Colin Ian King <email address hidden>

:040000 040000 25af92b1d77d2087d4f6f8251371498c5a208f4b 3e47611e4ed8f69aac13113aea8bb0c8e2591d64 M drivers
:040000 040000 99a08002fced336defd01fd9e982956b433abcfd 4d80045d2292ac4f132eed5b1397ca718b905067 M include

So it was indeed the patch for the bcm4311 rev 2.

As I said in my first post!

Regression[0] reproducible since -11 on local hardware. Symptoms not present in 2.6.24.x or 2.6.25-rc4-git3. On local hardware, I've been running ubuntu-hardy.git with the above changeset reverted for about two weeks. My primary-use home hardware has been running 2.6.25-rc4-git3.

[0] From `lspci -vn':
02:00.0 0280: 14e4:4311 (rev 01)
 Subsystem: 103c:1363

I'm using the Lenovo 3000 N100 series and it seems that my newly, freshly installed Ubuntu 8.04 beta seems that it becomes to difficult to make the wireless connections work. As a matter of fact, no hotspots are actually detected. This newer version of Ubuntu detects a firmware for bcm4312 but it freezes my computer every time I enable the firmware. Hardware Drivers detects the firmware as in status "in use" but the check box is not checked for some reason, and when I do enable it, it freezes my laptop.

If you are familiar with compiling your own kernel (see https://wiki.ubuntu.com/KernelTeam/GitKernelBuild), you could build a vanilla kernel for you, that works for me. Or if you have the possibility, to fetch the old ubuntu kernel 2.6.24-10, that should work fine, as well.

I'm very interesting on what it will be done. Will you remove the
patch that adds support for b43 rev02 cards?

Reverting patch for original bug https://bugs.launchpad.net/ubuntu/+bug/184600 as it does more harm than good - e.g. breaks BCM4312

All right. people with BCM4312 rev02 cards will have to use
ndiswrapper so. The thing that is strange is that this patch has been
included in 2.6.25 main tree.. how did they cope with this problem?

I confirm that feroda 9 beta (which has kernel 2.6.25) is capable of
supporting without problems both b43 rv01 and rev02.
Is it possible that some developer finds the way to support both, as in fedora?

Removing the linux-source-2.6.24 task. Beginning with the Hardy development cycle the kernel source naming convention changed from linux-source-2.6.xx to just linux. Sorry for any confusion. Thanks.

I have been updating faithfully through several kernel revisions waiting for a fix now.
As of the very latest version (linux-image-2.6.24-15-generic), b43 is still not working for me.

Installing b43fwcutter (or using dpkg-reconfigure) claims that it will get the drivers, but then exits with no further output. modprobe b43 does nothing apparent. The wireless light does not come on, and NetworkManager does not show anything at all. When booting, there is a long delay (30s) at "Loading hardware drivers."

This is marked as Fix Released but I have noticed nothing in the changelogs about b43 either.

It says fix committed, not released.

I just wanted to also mention that i'm running hardy on 2.6.24-12-generic, since that is the most recent kernel (the kernel i actually installed Hardy with) that works with the b43-fwcutter and firmware, no delay while booting the -14 or -15 kernels, but no wireless, and removing b43-fwcutter, the firmware, and re-installing, re-compiling, and everything else fails. Also, b43 has disappeared from the hardware drivers menu in both -14 and -15, might be linked to that. Would be nice if this were fixed before the actual release came out, or i may end up re-installing with the newest kernel and see if it picks it up this time, we'll see if i have time. And i just want to remark: this is the first release in which my wireless card actually worked without ndiswrapper, which is nice.

oh, and i forgot to say, my card is a 4311.

Once again, (March 22, 2008). My wireless seems to be working fine, however, my wpa encrypted wireless connection is shown as a regular open wireless network. To access my wpa wireless network, I select "Connect to other Wireless Network" and fill in the data as a wpa.

Alrexrudd, you might be experiencing bug 197819 .? Jockey did not fetch the firmware properly, the fix was released after beta came out.

Lswest, which version of the bcm 4311 are you using? Rev 01 or 02? The patch, that introduced support for rev 02, broke other cards, so it was removed again.

Linux-image-2.6.24-15 works for me, but quite unstable as usual, this seems to be related to network-manager, though. Classical ifconfig, iwconfig and command-line vpnc seem to work more stable.

I probably am experiencing 19819; I'll check out the other report.

I've gone back to using bcm43xx and it works, mostly.

same problem here... this wireless device doesn't work with the latest hardy updates and b43-fwcutter (with the lates updates the wireless device is not recognized at all).

My card is a bcm4311 rev 02. Any where i can find that patch still?

Disregard the last request, found the link, hope it works ^^

2.6.24-12 b43 works for BCM94311MCG (rev. 02)
No wireless for 2.6.24-15 -16
No way to make them works.

I have exactly the same problem as ettorevi

My wireless works with 2.6.24-12 but NOT with -15 or -16

lspci:
03:00.0 Network controller: Broadcom Corporation BCM4312 802.11a/b/g (rev 02)

in messages I found the following, no idea if this is related to the problem, but it does not show up in the -12 kernel:

Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038796] WARNING: at /build/buildd/linux-2.6.24/drivers/ssb/main.c:883 ssb_tmslow_reject_bitmask()
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038802] Pid: 1604, comm: modprobe Not tainted 2.6.24-16-generic #1
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038804]
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038804] Call Trace:
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038818] [ssb:ssb_tmslow_reject_bitmask+0x4a/0x60] :ssb:ssb_tmslow_reject_bitmask+0x4a/0x60
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038825] [b43:ssb_device_is_enabled+0x16/0x50] :ssb:ssb_device_is_enabled+0x16/0x50
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038831] [ssb:ssb_pcicore_init+0x21/0x70] :ssb:ssb_pcicore_init+0x21/0x70
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038837] [ssb:ssb_attach_queued_buses+0x106/0x2d0] :ssb:ssb_attach_queued_buses+0x106/0x2d0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038845] [ssb:ssb_pci_get_invariants+0x0/0x2d0] :ssb:ssb_pci_get_invariants+0x0/0x2d0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038850] [ssb:ssb_bus_register+0x17e/0x200] :ssb:ssb_bus_register+0x17e/0x200
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038859] [ssb:ssb_bus_pcibus_register+0x32/0x60] :ssb:ssb_bus_pcibus_register+0x32/0x60
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038865] [ssb:ssb_pcihost_probe+0x7d/0xc0] :ssb:ssb_pcihost_probe+0x7d/0xc0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038872] [pci_device_probe+0xf8/0x170] pci_device_probe+0xf8/0x170
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038878] [driver_probe_device+0x9c/0x1b0] driver_probe_device+0x9c/0x1b0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038882] [__driver_attach+0xc9/0xd0] __driver_attach+0xc9/0xd0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038885] [__driver_attach+0x0/0xd0] __driver_attach+0x0/0xd0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038888] [scsi_mod:bus_for_each_dev+0x4d/0x100] bus_for_each_dev+0x4d/0x80
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038893] [bus_add_driver+0xac/0x220] bus_add_driver+0xac/0x220
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038897] [nvidia:__pci_register_driver+0x69/0x4c0] __pci_register_driver+0x69/0xb0
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038903] [ssb:init_module+0x52/0x80] :ssb:ssb_modinit+0x52/0x80
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038908] [sys_init_module+0x18e/0x1a90] sys_init_module+0x18e/0x1a90
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038916] [<ffffffff803b97c0>] bus_register+0x0/0x270
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038922] [system_call+0x7e/0x83] system_call+0x7e/0x83
Apr 15 17:35:23 bm-laptop1 kernel: [ 23.038927]

Same with bcm4312 (rev. 02). It worked fine with 2.6.24-12 but in 2.6.25-15 -16 it's broken, but that seems to be bug #212933.
Is that by chance your problem too ettorevi?

Hi people

I think I have the same sort of problem on my acer aspire 5023 notebook with a broadcom 4318 wireless chipset.
After I freshly installed Ubuntu Hardy from the live disk and installed acer_acpi and the b43 driver the wireless worked.(but the range treble!)
after some ubuntu-updates it stopped working in total. I tried installing the old bcm43xx-fwcutter and the ndiswrapper but these two also don't work. It still sees the wireless networks but when I try to connect... I see that blue thing rotating at the nm-applet but those lights stay gray... no green :(

I have attached my syslog and dmesg

I get the same, or a very very similar, problem with an amd64 install on a Gateway MX7515 with a broadcom 4318 wireless chipset and the 2.6.24-15-generic kernel. (using the b43 driver of course)

nm-applet sometimes shows that it connects to a wireless network but (almost) nothing works. I say almost because i can still ping my router and other machines on the network, but domain names do not resolve. However, other times the lights just stay gray and it never connects.

Plugging in an ethernet cable and allowing it to connect that way fixes the problem, and allows a wireless connection to work as expected once the cable is unplugged and nm-applet reconnects to the wireless network.

No changes are made to /etc/resolv.conf when doing this, so the apparent dns problem is likely a symptom of whatever really gets 'fixed' when connecting then disconnecting from a wired network.

Frankly, I have no idea what the problem is, but it happened to me in gutsy with bcm43xx part way through the development cycle and I ultimately resolved to use ndiswrapper.

My mistake, i meant to say the 2.6.24-16-generic kernel.

My bcm4312 (rev 02) worked with 2.6.24-10 -11 and -12, upgrading to -15 and -16 broke it.
Is the rev 02 problem affecting all bcm43xx cards or just the bcm4311?

Is there any chance the bcm4312 will be working in hardy?

I just found a workaround for the ones not wanting to use ndiswrapper:

get http://linuxwireless.org/download/compat-wireless-2.6/compat-wireless-2.6.tar.bz2 and compile the modules. So you don't have to compile an own kernel to get the b43 module working.

The disadvantage is you have to do a new compile on every kernel update, but with hardy getting stable this shouldn't happen very often.

I just wanted to add that I also use Hardy on a AMD64 computer (HP Laptop) could this be a problem with the b43 and AMD64 arch?

Is everybody using AMD64 or are there also problems with i386?

using AMD64 architecture too...

using i386 on HP 2510p Laptop

2008/4/18, Carsten Kollmeier <email address hidden>:
>
> using AMD64 architecture too...
>
>
> --
> [Hardy]Recent kernel update to 2.6.24-11 breaks b43 (with bcm4312)
> https://bugs.launchpad.net/bugs/197959
> You received this bug notification because you are a direct subscriber
> of the bug.
>

I have exactly the same problem, that wireless isn't recognized in kernel 2.6.24-16

 Network controller: Broadcom Corporation BCM4312 802.11a/b/g (rev 02)

And i also use amd64

Sorry, but I really did not understand why is this bug still open (and
targeted to 8.04.1): this bug has been resolved after the removal of
the ssb patch for bcm4311 rev02

first off: the bug does not seem to be resolved yet (after all, check the comment above yours), and also, removing the patch for bcm4311 rev 2 (incidentally, MY card) doesn't seem like a great way to solve a bug, so the fix is probably still committed until they patch it so all the cards work, and you don't lose support for one card while gaining support with another.

The patch that was isolated, was not complete, so it did not solve the
problem for rev 2 cards (at least in my case it did not work at all,
though it was detected).
Also I think that most of the people have the rev 2 bug, and not this
(just an opinion).

Well, this bug was filed, because the additional patch taken from 2.6.25 to make the rev 02 cards work, broke the rev 01 cards. So the developers decided to revert that patch and leave the kernel in its 2.6.24 state. As a consequence, the rev 01 cards work now and so this bug is actually fixed. To make the rev 02 cards work, there are 2 possibilities: the 2.6.24 kernel could be patched in a different way with the code taken from 2.6.25 or you compile your own 2.6.25 kernel.

Hi there,

I've got a related problem with an eMachines M6811 laptop. From one kernel version to the next, wireless completely stopped to work. The device is properly registered, but neither does it show any networks nor do manual connection attemps succeed. Oddly, the name of device is now "Unknown Unknown", whereas before it was something along of "Broadcom 4306."

It would be nice if we could a kernel update that fixed this.

There is now a test kernel in my PPA with these 4 patches at http://ppa.launchpad.net/timg-tpi/ubuntu. Please give it a try and report the results.

SRU Justification:

Impact: b43 fails to function in a variety of scenarios

Fix Description:

1) b43: Workaround invalid bluetooth settings
This adds a workaround for invalid bluetooth SPROM settings on ASUS PCI cards. This will stop the microcode from poking with the BT GPIO line. This fixes data transmission on this device, as the BT GPIO line is used for something TX related on this device (probably the power amplifier or the radio). This patch was applied to mainline kernels with commit 1855ba7812dbd294fcfc083dc7d3b14d3b1f38db and was authored by Michael Buesch <email address hidden>.

2) ssb: Fix IRQ vectors enable for early cards.
After the release of 2.6.24, an error was found in the setup for IRQ vectors in the ssb driver. This error was fixed in the mainline kernels with commit b63009b456c8d9abe684bdf8d4bd8f27eb040019, which was authored by Larry Finger. This bug prevents any data transmission by the affected cards.

3) ssb: Fix TMS low bitmask reject code.
The Ubuntu code base for Hardy Heron includes the patch from mainline that enables the usage of BCM4311 rev 2 devices. Unfortunately, a portion of a separate patch, commit c272ef4403c271799a7f09a4ab7a236c86643843 by Larry Finger, was not incorporated; however, it is needed for these devices to function properly.

4) ssb: Fix all-ones boardflags
In the SSB SPROM a field set to all ones means the value is not defined in the SPROM. In case of the boardflags, we need to set them to zero to avoid confusing drivers. For certain cards, leaving the flags all set to 1 will prevent the card from transmitting, as has been seen for the PCI version of a BCM4301. This patch was included in mainline kernels with commit 4503183aa32e6886400d82282292934fa64a81b0.

Patch:
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=4cf07487b2ecf1a6e1185d94ef3eb1eed8d23a7b
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=465c3bad182b2a44e9eb0d99a0df7b839e919a52
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=d04340b2063813a0e4aaf32725f845a424abf870
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=d8f69fd28ba852dd49e40d5959b8c587a3812926

TEST CASE: See Bug Description

http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=4cf07487b2ecf1a6e1185d94ef3eb1eed8d23a7b
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=465c3bad182b2a44e9eb0d99a0df7b839e919a52
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=d04340b2063813a0e4aaf32725f845a424abf870
http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=commit;h=d8f69fd28ba852dd49e40d5959b8c587a3812926

Thanks a lot for this nice kernel ;)

Now my Broadcom BCM94311MCG rev 02 works fine with Ubuntu Hardy Heron

are you sure it works?

Yes I'm pretty sure that it works because I'm using it since ~5 days and didn't get any error

Accepted into -proposed, please test and give feedback here

I might be asking a stuppid question but how do I test this patch.

I have updated (4th of june) my system and uname shows:

Linux bm-laptop1 2.6.24-18-generic #1 SMP Wed May 28 19:28:38 UTC 2008 x86_64 GNU/Linux

Should this include that patch or do I need to do something else?

i tried to write an howto that should help everyone to compile a patched kernel for Ubuntu 8.04 Hardy Heron, in order to get bcm4312 (rev 02) working.

You can find it here:

http://ubuntuforums.org/showpost.php?p=5126810&postcount=58

test it and please post your experience ;)

i forgot to ask you if this card works with a hardy heron patched unlocking the proposed repositories... (I cannot try at the time)

Morten,

The kernel package that includes this fix is linux 2.6.24-19.33, which is currently in the hardy-proposed archive. Updates to stable releases are normally only promoted to the -updates archive (i.e., the updates that all users see by default) after we've gotten confirmation that the changes correctly fix the bug, so we definitely need the help of users like you in testing our fix!

You can enable the hardy-proposed repository on your system by going to System->Administration->Software Sources, and enabling "Pre-released updates" under the "Updates" tab.

Slayer,

Thanks for the howto, I managed to compile everything, however, because I have an nvidia graphics chip I have problems using my own build kernel because of the binary nvidia drivers which are not available for that custom kernel (at least I do not know how). But again, the compilation did indeed work.

Steve,

Thanks for the info. I always have the proposed archive active. I did not know the process though of kernel updates through proposed to normal updates.

For your information the new 2.6.24.19 kernel does indeed work and my wireless is now working again. Thank you all very much for the effort.

I installed the pre-release 2.6.24-19 and I'm sending this message using a functioning bcm4312 rev02 and b43. Thanks!!
drew

The pre-release 2.6.24-19 woks for me too. The bcm4312 rev02 is working, finaly.

Thanks all !

Rodrigo

Copied to hardy-updates.

works great, thanks!

Hello there!

I installed Hardy Heron few months ago, I have a bcm 4311 (rev 02) ang after a lot of googling, finally appeared linux kernel 2.6.24-19 and solved all the problems around the detection of the wireless card. Well, my question is if anyone have a problem with networks with WEP security, because I detect my wireless but doesn't accept the code.

Thanks for all!