bcmwl-kernel-source 6.30.223.30+bdcom-0ubuntu1 fails to see 5GHz networks and is unstable on 2.5GHz

Bug #1189611 reported by Michael Zanetti
62
This bug affects 11 people
Affects Status Importance Assigned to Milestone
bcmwl (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

Hardware: Broadcom Corporation BCM4331 802.11a/b/g/n (rev 02)

This chip is working perfectly fine on 2.5GHz and 5GHz using version 6.20.155.1+bdcom-0ubuntu6 of the driver.

After upgrading to saucy and installing version 6.30.223.30+bdcom-0ubuntu1 it fails to find any 5GHz networks and connections on 2.5GHz are unstable.

Downgrading the saucy machine to kernel 3.8.x and bcmwl-kernel-source 6.20.155.1+bdcom-0ubuntu6 (both from raring) makes the chip work again.

Downgrading just the kernel to 3.8 and keeping bcmwl-kernel-source at version 6.30.223.30+bdcom-0ubuntu1 does NOT fix the issue.

Looking at the changelog its most likely caused by the update to the new upstream release.

summary: - bcmwl-kernel-source 6.30.223.30+bdcom-0ubuntu1 fails to connect to see
- 5GHz networks and is unstable on 2.5GHz
+ bcmwl-kernel-source 6.30.223.30+bdcom-0ubuntu1 fails to see 5GHz
+ networks and is unstable on 2.5GHz
Revision history for this message
Adam Porter (alphapapa) wrote :

According to the package description, it doesn't even support the BCM4331 card:

"This package contains Broadcom 802.11 Linux STA wireless driver for use with Broadcom's BCM4311-, BCM4312-, BCM4313-, BCM4321-, BCM4322-, BCM43224-, and BCM43225-, BCM43227- and BCM43228-based hardware."

Assuming that the description is incomplete, do the other drivers work, like the b43 packages, the broadcom-sta packages, or the in-tree driver (which requires uninstalling all the packages)?

Changed in bcmwl (Ubuntu):
status: New → Incomplete
Revision history for this message
Michael Zanetti (mzanetti) wrote :

The b43 does not support 5GHz on this chip and 3.4GHz connections are so unreliable and slow that I lost quite some hair because of that one. Basically the connection drops either after some time (10-20 mins) or as soon as there is more than 5 Megabytes of traffic to handle.

The standard wl driver does recognize the card (only b/g), but does not connect at all.

I'm now using this driver for around half a year (version 6.20.155.1+bdcom-0ubuntu6 ) without a single connection drop and data rates of around 13 MB/s which is something I never experienced in Linux ever before.

Here's some more people happily using this driver with the BCM4331:
http://www.hackermusings.com/2013/02/better-wireless-drivers-for-the-bcm4331/

Please, please, keep it working for the 4331 even if it might not be explicitly listed as a supported device in its README. I've been using broken WiFi drivers in Linux since 2004 and just got it working properly half a year ago. Don't take that away from me again :D.

Btw, Right now I'm using Saucy + linux-generic-3.8.0-030800 + 6.20.155.1+bdcom-0ubuntu6 since the minute I reported this bug without any connection drop on 5GHz.

Changed in bcmwl (Ubuntu):
status: Incomplete → New
Revision history for this message
Bernardo Reino (reinob) wrote :

"According to the package description, it doesn't even support the BCM4331 card"

Now read first post:

"Hardware: Broadcom Corporation BCM4331 802.11a/b/g/n (rev 02)
This chip is working perfectly fine on 2.5GHz and 5GHz using version 6.20.155.1+bdcom-0ubuntu6 of the driver."

So it does support the chip.

Now why is this bug incomplete?
It is clearly reporting a regression from 6.20.155.1+bdcom-0ubuntu6 to 6.30.223.30+bdcom-0ubuntu1.

Revision history for this message
Michael Zanetti (mzanetti) wrote :

Btw, digging a bit deeper, seems support for the BCM4331 has been added with version 6.20.155.1 release while the package description still stems from version 5.100.82.112 which did not yet support this newer chip yet.

Revision history for this message
Adam Porter (alphapapa) wrote :

Michael,

"The standard wl driver does recognize the card (only b/g), but does not connect at all."

My understanding is that bcmwl is the wl.ko driver, and you later say that bcmwl 6.20.x works. What do you mean by, "the standard wl driver"?

Bernardo,

Now read my post. I said, "Assuming that the description is incomplete..." which Michael confirmed in his last message.

I marked the bug Incomplete because it was waiting on Michael to reply to my questions. He did, and reset it to New.

It seems that it is another regression. This whole bcmwl situation keeps getting more confusing: some cards have regressions from 5.100->6.20 which may be fixed in 6.30, and other cards have regressions from 6.20->6.30.

That's why I asked Michael to test the other drivers: if brcmsmac (the in-tree, Free, non-binary driver) can be used with his card, he can dump the restricted driver and its regressions.

Michael, can you please test brcmsmac with your card?

Revision history for this message
Michael Zanetti (mzanetti) wrote :

Adam,

according to http://wireless.kernel.org/en/users/Drivers/brcm80211#Broadcom_brcmsmac_driver the brcmsmac driver does not support the B4331 chip (not to be confused with the B4313). I tried to unload wl and loaded the brcmsmac. The card does not get detected. So using the brcmsmac driver is not an option.

Scratch that "standard wl driver". Before using Ubuntu I was using ArchLinux and they shipped the wl driver too. Apparently that was the same as the current one but only an earlier version, which did not support this chip yet. Because of that I got confused and thought there would be a in-tree wl driver and additionally, this patched version in form of the bcmwl-kerner-source package.

So bottomline, bcmwl-kernel-source version 6.20 is the only driver that makes my card usable. b43 is really slow and unstable, brcmsmac does not support the card at all.

Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in bcmwl (Ubuntu):
status: New → Confirmed
Revision history for this message
Gerry Boland (gerboland) wrote :

I can concur with Michael's assessment.

I've a Macbook Pro 8,1, with broadcom 4331 wifi chip. With raring, wifi was quite stable (I was using b43 then). Doing a clean-install of saucy, my wifi was highly unstable (using wl). I had same instability using b43.

I followed Michael's advice and installed a 3.8 kernel with bcmwl-kernel-source version 6.20, and I have reliable wifi once more.

Revision history for this message
Michael Zanetti (mzanetti) wrote :

Yeah... the b43 driver somewhat worked with raring but has always been very unstable for me. Additionally it does not support the 5GHz band on the BCM4331 chip which would eliminate it as an alternative to the bcmwl, even if it would become more stable.

Revision history for this message
Nickolay Ihalainen (ihanick) wrote :

Hello,
I'm able to compile old driver
bcmwl-6.20.155.1+bdcom
with
patches for 3.10 kernel (bcmwl-6.30.223.30+bdcom)

Required changes are:
0008-add-support-for-linux-3.9.0.patch : replace wl_cfg80211_hybrid.c with wl_cfg80211.c
0009-add-support-for-linux-3.10.0.patch
Add in dkms.conf:
PATCH[7]="0008-add-support-for-linux-3.9.0.patch"
PATCH[8]="0009-add-support-for-linux-3.10.0.patch"
And rebuild module with dkms:
https://help.ubuntu.com/community/DKMS

Could you test if the problem with BCM4331 connection gone with old bcmwl driver and 3.10 kernel?

Revision history for this message
Michael Zanetti (mzanetti) wrote :

Would it be possible for you to upload a package somewhere?

Revision history for this message
Nickolay Ihalainen (ihanick) wrote :
Revision history for this message
Michael Zanetti (mzanetti) wrote :

Thanks a lot Nickolay,

Yes, your package seems to work fine with 3.10.0-6-generic on amd64.

Revision history for this message
Gao Shichao (xgdgsc) wrote :

Is it fixed? Will it be released to raring?

Revision history for this message
Gao Shichao (xgdgsc) wrote :

No, I meant saucy.

Revision history for this message
Michael Zanetti (mzanetti) wrote :

No, seems it isn't fixed. I still need to manuall install Nickolay's package posted above to have working and reliable WiFi with the BCM4331 every time I upgrade my kernel.

Revision history for this message
Marcos Nils (marcosnils) wrote :

Hi,

I have the same problem using 3.12.9-031209-generic 64bit mode. Nickolay, is it too much to ask if you could provide a deb package for this specifications?

Thanks,

Marcos.

Revision history for this message
Gao Shichao (xgdgsc) wrote :

@Michael

I followed instructions here: https://zoni.nl/posts/2013/11/09/fixing-the-broadcom-bcm4331-wireless-drivers-on-ubuntu-1310/ to build a dkms package.

Is it an option for ubuntu developers to provide this old version in the repo, or at least on some supported ppa? That would make it easier for many people with such wireless card.

Revision history for this message
Šarūnas Valaškevičius (rakatan) wrote :

I've tried 6.30.223.141+bdcom-0ubuntu1 and while I could see 5Ghz networks, both 2.5 and 5Ghz networks were unstable - connections dropping every 30s or so + ping slow..

Downgraded according to the mentioned instructions to 6.20.155.1+bdcom-0ubuntu6 and it all works fine again.

                description: Wireless interface
                product: BCM4331 802.11a/b/g/n

Revision history for this message
Michael Zanetti (mzanetti) wrote :

Just an update: Nickolay's patched 6.20.155.1 package doesn't compile against kernel 3.15 any more. So atm you either want to stay on a lower kernel version or you'll be left with broken/bad WiFi on that chip.

Anyone got the time to try updating 6.20 for the latest kernel sources?

Revision history for this message
Šarūnas Valaškevičius (rakatan) wrote :

does anyone know if there is any progress on this issue? so far I'm holding my kernel related updates with a fear to see this again :D

Revision history for this message
penalvch (penalvch) wrote :

Michael Zanetti, Saucy reached EOL on July 17, 2014.
See this document for currently supported Ubuntu releases: https://wiki.ubuntu.com/Releases

If you are having a problem in a supported release, please file a new report via a terminal:
ubuntu-bug bcmwl

Do feel free to report any other bugs you may find.

Helpful bug reporting tips:
https://wiki.ubuntu.com/ReportingBugs

Changed in bcmwl (Ubuntu):
status: Confirmed → Invalid
Revision history for this message
Gao Shichao (xgdgsc) wrote :

This one still affects trusty. Why file a new one when this has the relevant info?

Revision history for this message
penalvch (penalvch) wrote :

Gao Shichao, thank you for your comment. So your hardware and problem may be tracked, could you please file a new report with Ubuntu by executing the following in a terminal while booted into the default Ubuntu kernel (not a mainline one) via:
ubuntu-bug bcmwl

For more on this, please read the official Ubuntu documentation:
Ubuntu Bug Control and Ubuntu Bug Squad: https://wiki.ubuntu.com/Bugs/BestPractices#X.2BAC8-Reporting.Focus_on_One_Issue
Ubuntu Community: https://help.ubuntu.com/community/ReportingBugs#Bug_reporting_etiquette

When opening up the new report, please feel free to subscribe me to it.

As well, please do not announce in this report you created a new bug report.

Thank you for your understanding.

Helpful bug reporting tips:
https://wiki.ubuntu.com/ReportingBugs

Revision history for this message
Michael Zanetti (mzanetti) wrote :

I still have the same issue on utopic. Am stuck on kernel 3.13.0-23, which is the last one I could compile bcmwl-kernel-source_6.20.155.1+bdcom against.

Here's the new bug report: https://bugs.launchpad.net/ubuntu/+source/bcmwl/+bug/1387221

Changed in bcmwl (Ubuntu):
status: Invalid → Confirmed
Revision history for this message
penalvch (penalvch) wrote :

Michael Zanetti, please do not toggle the Status of this report, mark it a duplicate of another, or vice versa.

Changed in bcmwl (Ubuntu):
status: Confirmed → Invalid
Revision history for this message
Michael Zanetti (mzanetti) wrote :

@Christoper: This is a valid bug. The fact that it hasn't been fixed in 1.5 years doesn't make it invalid. By closing this and invalid and making me create a new one you just drop useful information, delete "hottness" and exclude subscribers from progress on the new bug. Not the way to go imo

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.