Package Details: broadcom-wl-dkms 6.30.223.271-9

Git Clone URL: https://aur.archlinux.org/broadcom-wl-dkms.git (read-only)
Package Base: broadcom-wl-dkms
Description: Broadcom 802.11 Linux STA wireless driver
Upstream URL: https://www.broadcom.com/support/802.11
Keywords: network wireless
Licenses: custom
Conflicts: broadcom-wl
Submitter: felixonmars
Maintainer: andreyv
Last Packager: andreyv
Votes: 169
Popularity: 5.592550
First Submitted: 2013-07-08 14:59
Last Updated: 2016-08-11 12:21

Latest Comments

ranger commented on 2016-09-29 20:10

Yes the latest broadcom-wl-dkms 6.30.223.271-9

I tried to build after I rebooted into linux-lts and it builds ok now

andreyv commented on 2016-09-29 20:04

@ranger: What is your version of this package, is it the latest one? Can you pastebin a build log?

ranger commented on 2016-09-29 19:43

same problem here. doesn't compile with 4.7.5-1, compiles with no problem with linux-lts

make KERNELRELEASE=4.7.5-1-ARCH -C /usr/lib/modules/4.7.5-1-ARCH/build M=/var/lib/
Error! Bad return status for module build on kernel: 4.7.5-1-ARCH (x86_64)

andreyv commented on 2016-09-26 18:27

@tpolgrabia: It builds fine for me on kernel 4.7.5. Try to update your package version.

tpolgrabia commented on 2016-09-26 18:20

Doesn't compile within the newest version of linux (build failed) - 4.7.4-1

Jujstme commented on 2016-08-28 19:18

Too bad I cannot register in the forum right now since it requires a terminal command and I'm not in linux right now. I'll update the wiki though. By the way this bug will be likely resolved in the next update to nm.

andreyv commented on 2016-08-28 19:12

@Jujstme, @tck42: Nice. You might want to add this information to https://bbs.archlinux.org/viewtopic.php?id=216366 too.

Jujstme commented on 2016-08-28 19:12

@tck42:
The upstream bug: https://bugzilla.gnome.org/show_bug.cgi?id=770456
And the patch that has been added: https://cgit.freedesktop.org/NetworkManager/NetworkManager/commit/?h=th/platform-set-address-enfile-bgo770456&id=1902c0b88f52302a1b5e253a7352577e7e5ff682

tck42 commented on 2016-08-28 19:02

@Jujstme - do you have the bug #? I'm not finding it.
The commit that broke this is 767abfa69040c62ae1dc0989a77df8c5e0e489c1
Sorry to keep spamming the wrong packages comments.

Jujstme commented on 2016-08-28 18:34

Same problem here. Had to downgrade networkmanager to 1.2.5

EDIT: this is a bug in nm that has been fixed upsream a few hours ago.
In the meantime, an alternative solution is to add the following to /etc/NetworkManager/NetworkManager.conf

[device]
wifi.scan-rand-mac-address=no

RiadVargas commented on 2016-08-28 15:23

Same here, downgraded to Network Manager 1.2.5dev+5+g99e34d7 and Network Manger Applet 1.2.4-1.

fincan commented on 2016-08-28 08:53

Same here I am back to the nm 1.2.5

zerophase commented on 2016-08-28 04:56

@hobarrera I'm getting that issue too now that I've upgraded Network Manager. I just rolled back to networkmanager 1.2.5dev+5+g99e34d7-1, and Network Manger Applet 1.2.4-1, and wireless works again.

zerophase commented on 2016-08-27 05:45

@hobarrera hmm, it seems to work fine for me with the ck patch.

hobarrera commented on 2016-08-27 05:45

I downgraded networkmanager to the previous version and my network works again, so:

* The issue is some interaction between networkmanager and wl. NM works fine with other drivers.
* Even if you skip NM and use the cli, it seems that it somehow gets in the way and breaks things.

I'm currently back to networkmanger-1.2.5dev+5+g99e34d7.

tck42 commented on 2016-08-27 05:22

In my case failure is trying to set the MAC, which is working with b43.

hobarrera commented on 2016-08-27 04:57

I tried 4.7.1-1-ARCH. Same result.

tck42 commented on 2016-08-27 04:36

Fwiw NM gives the same error here. Journal shows:
<error> [1472271742.3671] platform-linux: do-change-link[3]: failure changing link: failure 23 (Too many open files in system)

Haven't had a chance to look into it much further yet. 4.7.2-1-ARCH

hobarrera commented on 2016-08-27 03:06

@zerophase:

I have now. NM shows this message when trying to connect to a new network: "Failed to add/activate connection. The access point /org/freedesktop/NetworkManager/AccessPoint/1 was not in the scan list".

When connecting to a saved connection, I get "Active connection removed before it was initialized"

Note however, that I tried using wpa_supplicant[1] directy, but it still fails, so I don't think it's anything NM specific (I also think this since NM works with another, USB adapter).

[1] https://wiki.archlinux.org/index.php/WPA_supplicant#Connecting_with_wpa_passphrase

zerophase commented on 2016-08-27 01:03

Have you tried deleting the connection and adding it back. Works for me sometimes if I can't connect.

hobarrera commented on 2016-08-27 00:35

I tried to connect via CLI and kept getting wrong passphrase, even though it was definitely right.

I plugged in a USB 802.11 adapter, and it worked fine (with that same command/passphrase!).

Looks like there's something failing with the last driver, on linux-4.7.2-1-ARCH, with a BCM4360.

hobarrera commented on 2016-08-26 23:08

Builds fine but does not connect since my last system update (running linux 4.7.2-2): "Active connection removed before it was initialized".

andreyv commented on 2016-08-21 16:18

@ikke: Thanks, good to know.

ikke commented on 2016-08-21 15:08

@andreyv I've downloaded a fresh copy of the package files, and now the package seems to build properly.

andreyv commented on 2016-08-20 21:15

@ikke: Before making changes, I would like to understand the cause of the problem. dkms.conf is generated from dkms.conf.in in prepare() and should not be a symlink from the source dir. Can you provide precise steps for how to reproduce this with makepkg in a clean checkout?

ikke commented on 2016-08-20 19:45

/usr/src/broadcom-wl-6.30.223.271/dkms.conf is a symlink pointing to the dkms.conf in the package dir (which doesn't exist on the machine I'm installing the package on).

src/dkms.conf in the package dir is a symlink and when you cp that symlink, it just stays a symlink. Adding -L to the copy command makes sure the actual file is copied.

$ ls -l pkg/broadcom-wl-dkms/usr/src/broadcom-wl-6.30.223.271/dkms.conf
lrwxrwxrwx 1 build users 41 Jul 23 08:15 pkg/broadcom-wl-dkms/usr/src/broadcom-wl-6.30.223.271/dkms.conf -> /data/packages/broadcom-wl-dkms/dkms.conf

andreyv commented on 2016-08-20 18:13

@ikke: I don't observe this when I run "makepkg -sf" in the package directory. The generated package doesn't contain symlinks.

How do you build the package? Which symlinks are created and to where?

ikke commented on 2016-08-20 14:56

PKGBUILD creating broken symlinks for me. This patch fixes that:

--- PKGBUILD 2016-08-20 14:51:45.651959470 +0000
+++ PKGBUILD 2016-08-20 14:53:40.650991334 +0000
@@ -41,7 +41,7 @@
package() {
local dest="${pkgdir}/usr/src/${pkgname/-dkms/}-${pkgver}"
mkdir -p "${dest}"
- cp -a src Makefile dkms.conf "${dest}"
+ cp -a -L src Makefile dkms.conf "${dest}"
install -D -m 0644 -t "${dest}/patches" *.patch

install -D -m 0644 lib/wlc_hybrid.o_shipped "${pkgdir}/usr/lib/${pkgname}/wlc_hybrid.o_shipped"

Fixed commented on 2016-08-18 09:34

Cпасибог, Андрей! Помогло.

andreyv commented on 2016-08-17 15:57

@Fixed: Update to the latest version of the package.

Fixed commented on 2016-08-17 13:06

After

makepkg -s

curl fails with

==> ERROR: Failure while downloading https://www.broadcom.com/docs/linux_sta/hybrid-v35_64-nodebug-pcoem-6_30_223_271.tar.gz


What to do?

marius-arc commented on 2016-08-15 17:52

@CelticSpice & @andreyv: Confirmed, I installed the latest package and it's working again.

CelticSpice commented on 2016-08-15 17:41

@andreyv: I didn't notice that the package had been updated. I can confirm that the latest version works fine on my end.

CelticSpice commented on 2016-08-15 17:40

@marius-arc: I got the same error. I'm not certain why, but cloning the package and reinstalling the driver seemed to work.

andreyv commented on 2016-08-15 17:35

@marius-arc: Do you use the latest version of this package (6.30.223.271-9)?

marius-arc commented on 2016-08-15 17:29

@andreyv: Here is the contents of /var/lib/dkms/broadcom-wl/6.30.223.271/build/make.log:
https://ptpb.pw/GcOE

andreyv commented on 2016-08-15 17:07

@marius-arc: Yes, you should paste the full log on some pastebin (you can try http://www.pastie.org/).

marius-arc commented on 2016-08-15 16:47

@andreyv: it is very long, I am guessing you don't want me to post it here. Here are the last few lines of it at least:
make[1]: *** [scripts/Makefile.build:290: /var/lib/dkms/broadcom-wl/6.30.223.271/build/src/wl/sys/wl_cfg80211_hybrid.o] Error 1
make: *** [Makefile:1457: _module_/var/lib/dkms/broadcom-wl/6.30.223.271/build] Error 2
make: Leaving directory '/usr/lib/modules/4.7.0-1-ARCH/build'
----
edit: also a whole lot of "error: array subscript is not an integer" and "error: initializer element is not constant"

andreyv commented on 2016-08-15 16:40

@marius-arc: As the error message says, what is in /var/lib/dkms/broadcom-wl/6.30.223.271/build/make.log?

marius-arc commented on 2016-08-15 16:23

I'm getting this just now when running pacman -Syu:
(1/2) Install DKMS modules
==> dkms install -m broadcom-wl -v 6.30.223.271 -k 4.7.0-1-ARCH
Error! Bad return status for module build on kernel: 4.7.0-1-ARCH (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.30.223.271/build/make.log for more information.

andreyv commented on 2016-08-11 11:45

I usually test the new kernel release when it enters [testing]. If I get time, I might look at 4.8 support sooner.

jlanzobr commented on 2016-08-10 00:20

Does not work on Linux 4.8rc1 from linux-mainline AUR package.

zerophase commented on 2016-07-29 11:44

I'm trying to check if the Asus pce-ac88 wireless card is supported on Linux yet. Wikidevi has been down for at least a week, so I don't know which broadcom chip is in there. Does anyone know if this card works with Linux yet?

USA-RedDragon commented on 2016-06-20 05:15

@nTia89 I'm looking into it now using the ck kernel. In the meantime a quick search shows that this isn't an uncommon or new issue, with threads dating back to 2014. https://www.google.com/#q=cfg80211_inform_bss_frame+error

nTia89 commented on 2016-06-19 08:55

$ uname -a
$ Linux arch-xps 4.5.7-1-ck #1 SMP PREEMPT Thu Jun 9 10:55:48 CEST 2016 x86_64 GNU/Linux

USA-RedDragon commented on 2016-06-19 08:52

@nTia89 what kernel version are you on. If you'd please share the output of "uname -a"

nTia89 commented on 2016-06-16 15:17

@andreyv I've never seen it before!

andreyv commented on 2016-06-16 15:11

@nTia89: Can you confirm that this specifically fails with package version 6.30.223.271-7 and works with version 6.30.223.271-6?

nTia89 commented on 2016-06-16 15:05

with latest update (seventh release) I get this new error:
`ERROR @wl_inform_single_bss : cfg80211_inform_bss_frame error`

beside omnipresent "module crash" traces like this:
http://pastebin.com/NyUJpgTf

andreyv commented on 2016-06-15 17:04

@USA-RedDragon: Thanks, the changes look good to me. I merged the patch to the repository.

severoraz commented on 2016-06-15 12:44

Has anybody else noticed significant improvement in connection stability with increased signal strength? With a strong signal my connection is virtually stable.

jmx commented on 2016-06-15 08:35

I am also facing severe connection stability issues since the kernel update to 4.6.

USA-RedDragon commented on 2016-06-13 17:21

Hey there, I forked your package and updated it to work on future Linux kernel versions while maintaining compatibility. If you would please look at it and possibly merge the commit I will link below.

Thank you,
Jacob (USA-RedDragon)

https://github.com/USA-RedDragon/broadcom-wl-dkms/commit/52ef285fd9c688bb8b161b64d5dd1c4a8f2a1c4f

fincan commented on 2016-06-07 21:40

I have connection stability problem here is the log.

00:37:42 kernel: ---[ end trace c03f9055661bd6ff ]---
00:37:42 kernel: [<ffffffff81097120>] ? kthread_worker_fn+0x170/0x170
00:37:42 kernel: [<ffffffff815b59ff>] ret_from_fork+0x3f/0x70
00:37:42 kernel: [<ffffffff81097120>] ? kthread_worker_fn+0x170/0x170
00:37:42 kernel: [<ffffffff810971f8>] kthread+0xd8/0xf0
00:37:42 kernel: [<ffffffffa092beb0>] ? wl_deinit_priv_mem+0xa0/0xa0 [wl]
00:37:42 kernel: [<ffffffffa092bf2c>] wl_event_handler+0x7c/0x1e0 [wl]
00:37:42 kernel: [<ffffffffa092f505>] wl_notify_roaming_status+0xc5/0x140 [wl]
00:37:42 kernel: [<ffffffffa057dee6>] cfg80211_roamed+0x86/0xa0 [cfg80211]
00:37:42 kernel: [<ffffffff8107918a>] warn_slowpath_null+0x1a/0x20
00:37:42 kernel: [<ffffffff81079042>] warn_slowpath_common+0x82/0xc0
00:37:42 kernel: [<ffffffff812dd982>] dump_stack+0x63/0x81
00:37:42 kernel: Call Trace:
00:37:42 kernel: ffff880253af9000 ffff88021af50600 000000000000007c ffff88017801d240
00:37:42 kernel: 0000000000000000 ffffffffa05a230b ffff8802544c3df0 ffffffff81079042
00:37:42 kernel: 0000000000000286 000000003ffe4cdb ffff8802544c3db8 ffffffff812dd982
00:37:42 kernel: Hardware name: Hewlett-Packard HP Pavilion 15 Notebook PC /2281, BIOS F.36 02/02/2015
00:37:42 kernel: CPU: 1 PID: 373 Comm: wl_event_handle Tainted: P U W O 4.5.4-1-ARCH #1
00:37:42 kernel: drm
00:37:42 kernel: glue_helper ablk_helper cryptd rfkill input_leds mii serio_raw led_class thermal snd_timer wmi snd battery xt_limit soundcore mei_me mei evdev ac mac_hid lpc_ich xt_tcpudp lis3lv02d input_polldev hp_wireless tpm_tis tpm fjes i2c_i801 shpchp processor xt_addrtype nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack ip6table_filter ip6_tables nf_conntrack_netbios_ns nf_conntrack_broadcast nf_nat_ftp nf_nat nf_conntrack_ftp nf_conntrack iptable_filter sch_fq_codel coretemp msr ip_tables x_tables ext4 crc16 mbcache jbd2 hid_logitech_hidpp hid_logitech_dj hid_generic usbhid hid sr_mod cdrom sd_mod atkbd libps2 xhci_pci xhci_hcd ahci libahci libata ehci_pci ehci_hcd scsi_mod usbcore usb_common i8042 serio i915 video button intel_gtt i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops
00:37:42 kernel: Modules linked in: cmac ecb nvidia_modeset(PO) nvidia(PO) fuse rfcomm bbswitch(O) bnep nls_iso8859_1 nls_cp437 vfat fat ip6t_REJECT uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 videobuf2_core btusb videodev btrtl btbcm btintel bluetooth media nf_reject_ipv6 nf_log_ipv6 xt_hl nvram ip6t_rt nf_conntrack_ipv6 nf_defrag_ipv6 ipt_REJECT nf_reject_ipv4 snd_hda_codec_realtek mousedev snd_hda_codec_hdmi snd_hda_codec_generic joydev hp_wmi sparse_keymap nf_log_ipv4 nf_log_common wl(PO) snd_hda_intel iTCO_wdt snd_hda_codec iTCO_vendor_support xt_LOG snd_hda_core intel_rapl psmouse snd_hwdep hp_accel r8169 pcspkr x86_pkg_temp_thermal intel_powerclamp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel aes_x86_64 lrw snd_pcm gf128mul cfg80211
00:37:42 kernel: WARNING: CPU: 1 PID: 373 at net/wireless/sme.c:850 cfg80211_roamed+0x86/0xa0 [cfg80211]()
00:37:42 kernel: ------------[ cut here ]------------

bimmlerd commented on 2016-05-02 08:34

@severoraz I was under the same impression, and was getting some kernel panics in the journal as well, but rebuilding the package helped. Maybe try that?

EDIT: Actually, rebuilding did in fact not help. I'm experiencing frequent 'slow_path' kernel panics (?) in the journal.

severoraz commented on 2016-05-01 19:45

Does anybody else feel connection stability with this driver has regressed?

Universebenzene commented on 2016-03-15 16:02

Hello, I'm running archlinux on my MacBook Air, with the wireless device BCM4360. I want to create a wifi hotspot by using create_ap, but it said "Your adapter does not support AP (master) mode". Also when I use iw list, the AP doesn't appear. However, when I switch to OSX, the wifi hotspot can be created and it works well. So the problem is not from the wireless device itself. Then is that because this driver for does not support AP? Or do I need to do something else to make it support AP in linux?

Sorry for my poor English expression.

andreyv commented on 2016-02-29 15:49

@analca3: It may be a limitation of the Broadcom driver. When I last tested it (with BCM4312), I could not see channel 13 with this driver, but it worked with the b43 driver.

Part of this driver is provided in the source form — you can try to read and modify it.

analca3 commented on 2016-02-29 15:42

Hi all! Is there a reason for my BCM4360 can't reach networks inside channel 13? Regdomain is correctly set. There is more info in this post.

https://bbs.archlinux.org/viewtopic.php?pid=1608376#p1608376

I think without broadcom driver source it hasn't a fast solution/workaround, but any help is appreciated. There isn't a open source driver for my device.

Thank you!

0x414A commented on 2016-02-22 18:42

@andreyv huh, I had to change the source_* attrs to point to the locally-downloaded tarballs for some reason.

Glad to know it's working again, though.

nnnn20430 commented on 2016-02-22 09:29

huh all of the sudden i get curl: (60) SSL certificate problem: unable to get local issuer certificate

https://www.broadcom.com/ seems to have an untrusted cert now ?

edit: ok they fixed it

andreyv commented on 2016-02-22 06:59

Thanks for the heads-up.

You don't need to modify the PKGBUILD. Just download the .tar.gz for your respective platform manually, place it in the same directory and run makepkg.

If using yaourt: when yaourt asks whether you want to modify the PKGBUILD, place the downloaded file in /tmp/yaourt-tmp-<user>/aur-broadcom-wl-dkms, and _then_ press y/n to continue.

0x414A commented on 2016-02-22 00:22

Follow loqs' instructions to manually download the *.tar.gz from broadcom (https://bbs.archlinux.org/viewtopic.php?pid=1605719#p1605719), if you are having trouble at the curl step.

You probably want https://www.broadcom.com/docs/linux_sta/hybrid-v35_64-nodebug-pcoem-6_30_223_271.tar.gz if you're running a 64-bit install.

Modified PKGBUILD https://gist.github.com/0x414A/9463a5e8be54ba0463a4#file-pkgbuild

Punky commented on 2016-02-18 16:17

I can not build this package...First script can not download package from link, than after I download tarball and download package my self, than can not be build...Using kernel 4.4.1-2-ARCH-686

cemsbr commented on 2016-02-18 14:44

I'm also having the same error as last reported by @mamantoha (-22). I'm running kernel 4.4.1-2-ARCH x86_64

andreyv commented on 2016-02-17 19:19

Hi @waldman,

This is not specific to this driver — it is a dependency of the Linux kernel itself. You can see it in the "linux" package:

crda (optional) - to set the correct wireless channels of your country


@trusktr: Your first problem is likely specific to your environment, I have not seen it before. In your second problem, you need linux-headers for your kernel as suggested by this package's optdepends.

waldman commented on 2016-02-17 13:51

Hi,

Can you please add the package crda as a dependency?

It's installation will fix the message bellow on dmesg:
[11711.824792] cfg80211: Calling CRDA to update world regulatory domain
[11714.976073] cfg80211: Exceeded CRDA call max attempts. Not calling CRDA

Thanks! :)

trusktr commented on 2016-02-17 03:57

I'm not sure if this is a problem with the package, but I get this error when trying `makepkg`:



```
❯ makepkg -f [19:51:16]
==> Making package: broadcom-wl-dkms 6.30.223.271-4 (Tue Feb 16 19:51:17 PST 2016)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
-> Found broadcom-wl-dkms.conf
-> Found dkms.conf.in
-> Found 001-null-pointer-fix.patch
-> Found 002-rdtscl.patch
-> Found hybrid-v35_64-nodebug-pcoem-6_30_223_271.tar.gz
==> Validating source files with sha256sums...
broadcom-wl-dkms.conf ... Passed
dkms.conf.in ... Passed
001-null-pointer-fix.patch ... Passed
002-rdtscl.patch ... Passed
==> Validating source_x86_64 files with sha256sums...
hybrid-v35_64-nodebug-pcoem-6_30_223_271.tar.gz ... Passed
==> Extracting sources...
-> Extracting hybrid-v35_64-nodebug-pcoem-6_30_223_271.tar.gz with bsdtar
==> Starting prepare()...

File src/wl/sys/wl_linux.c is read-only; trying to patch anyway
patch: **** can't open file src/wl/sys/wl_linux.c : Permission denied
==> ERROR: A failure occurred in prepare().
Aborting...
```



In order to fix that problem, I had to modify PKGBUILD so the `prepare()` function looks like the following; note the two `chmod` lines:



```sh
prepare() {
sed -i -e "/BRCM_WLAN_IFNAME/s:eth:wlan:" src/wl/sys/wl_linux.c

chmod a+r src/wl/sys/wl_linux.c
chmod u+w src/wl/sys/wl_linux.c

patch -p1 < 001-null-pointer-fix.patch
patch -p1 < 002-rdtscl.patch

sed -e "s/@PACKAGE_VERSION@/${pkgver}/" dkms.conf.in > dkms.conf
}
```



I was then able to install the resulting package, which appears to have installed, but it then gives this error:



```sh
'(1/1) reinstalling broadcom-wl-dkms [----------------------------] 100%

Creating symlink /var/lib/dkms/broadcom-wl/6.30.223.271/source ->
/usr/src/broadcom-wl-6.30.223.271

DKMS: add completed.
Error! echo
Your kernel headers for kernel 4.3.3-3-ARCH cannot be found at
/usr/lib/modules/4.3.3-3-ARCH/build or /usr/lib/modules/4.3.3-3-ARCH/source.
error: command failed to execute correctly
```

andreyv commented on 2016-02-04 16:03

According to https://source.android.com/security/bulletin/2016-02-01.html, Broadcom Wi-Fi driver on Android had two vulnerabilities (CVE-2016-0801, CVE-2016-0802) which have been patched. I have no information whether the driver in this package is affected too.

EgidioCaprino commented on 2015-12-20 19:01

I installed this for Arch running on a Macbook and I can't see the wireless device anymore.

mamantoha commented on 2015-12-14 14:21

I have the same error 22 with latest broadcom-wl-dkms and BCM43228:
kernel: ERROR @wl_notify_scan_status : wlp4s0 Scan_results error (-22)

After this error i need to manually disconnect and connect back in NetworkManage.

swaraj commented on 2015-12-08 21:54

@andreyv: Sorry, my previous comment is useless as having reinstalled the latest version of broadcom-wl-dkms and rebooting my system (4.2.5-1-ARCH, which is the latest), everything seems to be working now. I'll keep you updated in case anything breaks!

andreyv commented on 2015-12-07 15:55

@swaraj: Which kernel version and arch are you running? What did you update? How does error 22 look like?

swaraj commented on 2015-12-07 13:55

broadcom-wl-dkms has been weird for me as well for around the last week (maybe it's only since then because I did an update...). Ever since then it sorta just stops working after a while, and then I go through the hassle of reinstalling and/or rebooting. Recently though, it hasn't been working at all (I get error 22). How can I (like severoraz) help debug/solve this problem?

severoraz commented on 2015-11-29 15:02

I still have unstability :-/ how can I help debug? I see no significant messages in dmesg usually.

graysky commented on 2015-11-21 15:33

@andreyv - Works now, thank you. I will use your solution in broadcom-wl-ck. Note - You can test it in a clean chroot yourself. See clean-chroot-manager in the AUR.

andreyv commented on 2015-11-21 14:53

@graysky: Try again now.

graysky commented on 2015-11-21 14:04

@andreyv - Doesn't build for me.
Building module:
cleaning build area...
make KERNELRELEASE=4.3.0-1-ARCH -C /usr/lib/modules/4.3.0-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.30.223.271/build.....(bad exit status: 2)
Error! Bad return status for module build on kernel: 4.3.0-1-ARCH (i686)
Consult /var/lib/dkms/broadcom-wl/6.30.223.271/build/make.log for more information.
error: command failed to execute correctly

And the log:
DKMS make.log for broadcom-wl-6.30.223.271 for kernel 4.3.0-1-ARCH (i686)
Sat Nov 21 09:03:16 EST 2015
make: Entering directory '/usr/lib/modules/4.3.0-1-ARCH/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.271/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.o
/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.c: In function ‘osl_getcycles’:
/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.c:936:16: error: implicit declaration of function ‘native_read_tsc’ [-Werror=implicit-function-declaration]
cycles = (u32)native_read_tsc();
^
cc1: some warnings being treated as errors
scripts/Makefile.build:258: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.o] Error 1
Makefile:1378: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.271/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.271/build] Error 2
make: Leaving directory '/usr/lib/modules/4.3.0-1-ARCH/build'

andreyv commented on 2015-11-21 13:33

I added a patch for kernel 4.3 on i686. Someone should test it as I don't have an i686 system.

andreyv commented on 2015-11-19 22:10

@graysky: I only tested it on 4.3-1 x86_64.

I found this in Google: https://lkml.org/lkml/2015/6/12/725. So the workaround for now would be to just replace the rdtscl() call with the macro contents. Can you test if that works for you?

graysky commented on 2015-11-15 20:29

@andreyv - Are you able to build this against linux-4.3 on i686 currently in [testing]? I cannot.

% cat /var/lib/dkms/broadcom-wl/6.30.223.271/build/make.log
DKMS make.log for broadcom-wl-6.30.223.271 for kernel 4.3.0-1-ARCH (i686)
Sun Nov 15 15:27:32 EST 2015
make: Entering directory '/usr/lib/modules/4.3.0-1-ARCH/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.271/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.o
/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.c: In function ‘osl_getcycles’:
/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.c:935:2: error: implicit declaration of function ‘rdtscl’ [-Werror=implicit-function-declaration]
rdtscl(cycles);
^
/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.c:935:2: warning: ‘cycles’ is used uninitialized in this function [-Wuninitialized]
cc1: some warnings being treated as errors
scripts/Makefile.build:258: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.271/build/src/shared/linux_osl.o] Error 1
Makefile:1382: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.271/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.271/build] Error 2
make: Leaving directory '/usr/lib/modules/4.3.0-1-ARCH/build'

xdom commented on 2015-11-11 07:06

When AP signal strength gets under ~50%, I get connected for ~10sec then immediatelly disconnected with following errors going on:
Getting ERROR @wl_cfg80211_get_station : Wrong Mac address, mac = a4:6c:2a:a7:e8:51 profile =a4:6c:2a:c0:3f:61

From time to time I see this error http://pastie.org/private/nas0zkgorixnajkha1z6ka

Using BCM43142, on 4.2.5-1-ARCH

CjK commented on 2015-11-07 18:27

@severoraz I'm getting the same error-message as @alonhar at bootup. However, my wireless works if I unplug my wired cable before boot. If I leave it plugged in, the above error occurs and can't connect using a wireless connection.

I'm using a BCM4352 as well.

andreyv commented on 2015-11-07 10:59

@severoraz: Just read the comments, there are several people with this problem. So far no solution has been mentioned.

alonhar commented on 2015-11-06 15:35

I am using BCM4352 and I am getting this error
ERROR @wl_cfg80211_scan : WLC_SCAN error (-22)..

severoraz commented on 2015-11-05 22:42

So is nobody else still getting an unstable connection? I am using BCM4352 14e4:43b1

andreyv commented on 2015-11-01 16:54

@hobarrera: Currently it already conflicts with broadcom-wl. I'm not sure if I should add a "provides" too, because neither of both packages is "main", they just implement two different approaches. It's also unlikely someone will want to depend on broadcom-wl.

petruschka commented on 2015-11-01 00:23

@gourdcaptain --- thanks, it seems that was that!

gourdcaptain commented on 2015-10-31 23:06

@petruschka - you need to enable the dkms.service to compile it at boot, probably.
@hobarrera - Actually, I've got a use case where I use broadcom-wl for the stock kernel and broadcom-wl-dkms for some others without a specialized package, so that wouldn't really work the best.

hobarrera commented on 2015-10-31 23:04

Package should state "provides=broadcom-wl" to make this more obvious.

petruschka commented on 2015-10-31 22:58

Greetings all,

@joehillen, this never seems to work for me: my wireless is never recognised when linux is updated --- what am I missing? This is sort of driving nuts...

Thanks in advanced!
P

joehillen commented on 2015-10-31 22:42

@trusktr Yes, use this package instead of broadcom-wl if you don't want to have to reinstall broadcom-wl after kernel updates.

trusktr commented on 2015-10-31 20:30

Is this package meant to allow broadcom-wl to work after linux updates? For example, sometimes I have to manually install broadcom-wl after updates to linux as Wifi stops working.

attractedtofire commented on 2015-10-29 21:48

Hello andrevy,

I am running a BCM4352 (rev 3) in a Dell XPS13. I switched to 271-3 after having the same kernel panic issues as everyone else. Generally speaking things are working fine now but the .271 version seemed to have introduced one more issue I had already noticed with 271-1: iwconfig does no longer show the signal quality.

journalctl -xe shows
ERROR @wl_dev_intvar_get : error (-1)
ERROR @wl_cfg80211_get_tx_power : error (-1)

as well as a regular (every 6 seconds) but maybe unrelated:
net_ratelimit: 2 callbacks suppressed

Can anybody confirm this?

All the very best,
a

Haxxa commented on 2015-10-22 01:08

Works well on my Dell XPS 13 as of Version: broadcom-wl-dkms 6.30.223.271-3

js4113434 commented on 2015-10-17 11:22

@andreyv no it doesn't.

andreyv commented on 2015-10-15 06:10

@js4113434: It's not a device but an interface. Does "prism0" appear in "ip link show"?

joehillen commented on 2015-10-14 20:23

@js4113434 I don't know for sure, but generally monitor mode isn't supported on proprietary drivers.

js4113434 commented on 2015-10-14 19:57

Hi, I'm using a BCM4352 and am unable to get it into monitor mode. I echoed 1 to /proc/brcm_monitor0, and it shows it as such when i open with nano, but there is no prism0 device or any other that is created. Is monitor mode unsupported for this chipset?

wendell commented on 2015-10-12 15:26

I had roaming issues for as long as I've used this driver/computer (MacBook Pro, BCM4360 [14e4:43a0]), and the previous build (6.30.223.271-2) had kernel panics. The kernel panics seem fixed now, but they were sporadic so I can't be sure.

I also applied the roaming patch, and uploaded the build here: https://github.com/wackywendell/arch_broadcom_wl_noroaming

For those wanting a quick fix for roaming:

$ git clone https://github.com/wackywendell/arch_broadcom_wl_noroaming.git noroaming
$ cd noroaming
$ makepkg -fi

Pull requests / issues / etc. welcome.

matteoguglielmi commented on 2015-10-11 21:42

kernel panic solved for me as well. Now testing stability.

severoraz commented on 2015-10-11 13:02

6.30.223.271-3 solved the kernel panic issue for me (Lenovo 80K9 a.k.a. Edge 15, Linux kernel 4.2.2-1-ARCH, BCM4312) but I still have unstable connection. Am I the only one?

hobarrera commented on 2015-10-11 09:16

6.30.223.271-2 works fine on my 2013 MBA as well, though there's lots of log polluting due to some errors (network works fine though).

The same applies for 6.30.223.271-3.

Iwmrby commented on 2015-10-11 09:06

Can confirm, the last version works fine on my 11,2 MacBook Pro with BCM4360.

andreyv commented on 2015-10-11 08:14

Added the patch in 6.30.223.271-3.

Yvanoff commented on 2015-10-10 19:19

@andreyv the pastie fixes it for me (BCM43142, with latests updates (firefox 41.0.1, on kernel 4.2.2-1))
at least, firefox, pacman and yaourt work fine

neosilky commented on 2015-10-10 12:44

@andreyv, works for me. I have a BCM43142. Noticed this change whilst diffing between -248 and -271.

andreyv commented on 2015-10-09 17:01

Alright people,

Today I went through the patches that were meant for the old version, and all of them seemed to be applied upstream (or fixed otherwise). But there was one exception.

Those who experience crashes with the new version, can you please check if http://pastie.org/10470745 fixes it?

CjK commented on 2015-10-09 16:36

Here is a success report: 6.30.223.271-2 works well with the current LTS-Kernel and my Broadcom 4352-chipset.
Never got it to work with Kernel 4.2.x though.

matteoguglielmi commented on 2015-10-09 16:02

Same here (kernel panic, 4.2.2-1, 6.30.223.271-2) on macbook pro (15-inch, early 2013) and macbook air (11-inch, mid 2012).

troydwill commented on 2015-10-08 22:08

Kernel panic with latest kernel (linux 4.2.2-1) on a Dell XPS 13. (broadcom-wl-dkms 6.30.223.271-2)

andreyv commented on 2015-10-07 19:11

@trustin: Thanks for verifying this.

I want to add your patch to the package. For now, I will make it optional because it disables a feature which might be functional for other people.

Can you please adapt your patch so that it compiles on both 4.1 and 4.2 kernels?

aravind commented on 2015-10-06 18:54

The kernel (4.2.2-1) seems to be fine. I think it has more to do with the new rev of the broadcom-wl module. In one of the earlier comments, someone pointed out that https://aur.archlinux.org/cgit/aur.git/snapshot/aur-2b7a13592e2d22f46d1f1a8bb957985506ca2a8d.tar.gz is working fine.

The new version 6.30.223.271-{1,2} breaks for me, but the one from the link above 6.30.223.248-7 works fine with the new/previous kernels.

Doctor_Mooch commented on 2015-10-06 14:12

@lonelywalker: Same issue here, I have a late 2014 MBP. Ever since the latest kernel update, I've been receiving panics from this interface.

lonelywalker commented on 2015-10-06 14:03

I have a late 2013 MBP with BCM4360 (14e4:43a0). I get kernel panics when I try to receive data from this interface. My kernel version is 4.2.2-1, is anyone else experiencing the problem? I'm thinking of downgrading to kernel 3.16 or 3.18 if that helps.

trustin commented on 2015-10-06 08:24

I tried the driver in a multi-AP environment and it still crashes with the same error. Applying the patch seems to fix the crash. I'm not sure about the connection stability though. (I get disconnected quite often, but this seems unrelated with the patch)

trustin commented on 2015-10-06 07:58

The patch was not adopted by Ubuntu or any other community as far as I know. I basically didn't get a response (yet).

Anyways, let me update to the new release and let you know if the new version doesn't require the patch.

andreyv commented on 2015-10-05 19:17

@arcanis: This is probably the same problem as for the previous commenters. Hopefully someone shares a kernel panic stacktrace to give a hint where exactly does the driver crash.

andreyv commented on 2015-10-05 19:12

Thanks, @joehillen.

No one has reported back about roaming with the new version yet, so we just need to wait and see if the patch is still necessary.

arcanis commented on 2015-10-05 19:07

not sure is it my system-specific issue or not... But upgrade to 6.30.223.271-{1,2} causes kernel panic on 4.2.2-1-ARCH (and possible lts one as well). WiFi card is:

03:00.0 Network controller: Broadcom Corporation BCM4313 802.11bgn Wireless Network Adapter (rev 01)

Rollback to the previous version solves the problem.

joehillen commented on 2015-10-05 18:40

Reporting back with my experience with the no-roaming patch. It worked like a charm, and I didn't have any issues with wifi.

I don't know what this means for the new driver version. Does the driver still crash when roaming on the new version?

andreyv commented on 2015-10-03 17:47

New version: 6.30.223.271-2
- Various PKGBUILD improvements

DO NOT update to this version if 6.30.223.271-1 does not work for you.

CDF-H commented on 2015-10-03 15:27

Seeing the same as lwmrby, same card but a macbook air 6,2

andreyv commented on 2015-10-03 08:21

FWIW, here is a diff between the old patched driver and the new driver: http://hastebin.com/anoqejiriv.md

l33tlinuxh4x0r commented on 2015-10-03 02:48

AC wireless doesn't work with my BCM4352 card. with version 6.30.223.271 it worked with previous versions... Also not a new issue but a chronic one. It takes about 30 seconds to connect with this driver. I have an intel card too and it connects in no time at all.

Iwmrby commented on 2015-10-02 20:21

New version is severely broken: as soon as you try to use the wireless( it connects fine, sending or receiving data seems to be a problem) a kernel panic is caused and the system becomes unresponsive. This is on rMBP 11,2 using BCM4360.

Unfortunately I can't provide debug logs for now.

Iwmrby commented on 2015-10-02 20:21

New version is severely broken: as soon as you try to use the wireless( it connect fine, sending or receiving data is the problem) a kernel panic is caused and the system becomes unresponsive. This is on rMBP 11,2 using BCM4360.

Unfortunately I can't provide debug logs for now.

delic commented on 2015-10-02 20:19

I am seeing the exact same behavior as Yvanoff. Running a 4352 on a Dell XPS 13. Tried with both 4.1.8-1-lts and 4.2.2-1

Yvanoff commented on 2015-10-02 20:17

I don't recommend the new version. I can connect to network fine with it, but whenever I try to launch firefox or chromium, my laptop freezes.
Switching back to the older version (I had saved it on my PC) for the time being, until broadcom (eventually) fixes it
Running a BCM43142, with latests updates (firefox 41.0.1, tested with chromium 45.0.2454.101-1, on kernel 4.2.2-1)

Iwmrby commented on 2015-10-02 18:56

New version is severely broken: as soon as you try to use the wireless a kernel panic is caused and the system becomes unresponsive. This is on rMBP 11,2 using BCM4360.

Unfortunately I can't provide debug logs for now.

delic commented on 2015-10-02 18:21

I am seeing the exact same behavior as Yvanoff. Running a 4352 on a Dell XPS 13.

delic commented on 2015-10-02 18:00

I am seeing exact same behavior as Yvanoff. Anything that starts to send data kills the system.

Yvanoff commented on 2015-10-02 17:47

I don't recommend the new version. I can connect to network fine with it, but whenever I try to launch firefox or chromium, my laptop freezes.
Switching back to the older version (I had saved it on my PC) for the time being, until braodcom (eventually) fixes it

andreyv commented on 2015-10-02 16:37

New upstream version: 6.30.223.271

Upstream release notes:
- Upgraded to support 3.19 kernel APIs.

What to do if the new version does not work:
- Post the details here; see if someone has a similar problem
- You can try the linux-lts kernel
- For the time being, you can also use the previous driver version. Get the old PKGBUILD here: https://aur.archlinux.org/cgit/aur.git/snapshot/aur-2b7a13592e2d22f46d1f1a8bb957985506ca2a8d.tar.gz.
- For general information and alternative drivers, see https://wiki.archlinux.org/index.php/Broadcom_wireless.

andreyv commented on 2015-10-02 15:07

Alright, thanks for the information!

And thank you for maintaining this package for all the time until now :)

twelveeighty commented on 2015-10-02 15:05

@andreyv: I discovered that when you first add a co-maintainer and then disown it, it will automatically be transfered to the co-maintainer. Therefore, you should now be the official maintainer!

linux-recent.patch I believe contains only compat changes, no functionality. Most of the patches come from Canonical and they used to be adopted from the AUR broadcom-wl package, but since the maintainer changed there, there haven't been too many updates there to adopt from.

andreyv commented on 2015-10-02 15:01

@twelveeighty: Do I assume correctly that linux-recent.patch only contains compatibility changes and no functional changes?

andreyv commented on 2015-10-02 14:30

Also, of course, you can leave it as is if you want.

andreyv commented on 2015-10-02 14:29

The typical procedure is that a current maintainer disowns the package, and a new maintainer adopts it. I don't know how it works with the new git system, though :)

twelveeighty commented on 2015-10-02 14:27

@andreyv: I added you as a co-maintainer, so you should be able to commit now. I will probably have to "disown" it to fully transfer ownership to you.

twelveeighty commented on 2015-10-02 14:24

@andreyv - perfect! Now if I can only find out how to transfer the maintainer over to you....

Seems like you are volunteering at the worst possible time, with an upstream version bump, unless you like the challenge, of course.

@bryn: thanks for the heads-up. If the previous upstream version bump is any indication, it may be a while before we can run that new version on kernel 4.2.x or better. It's typically someone at Canonical that finds a way to patch it, so keep an eye on their bug tracker and let's help them out where we can.

andreyv commented on 2015-10-02 06:42

Hi @twelveeighty,

I just noticed your message about maintainership. I can maintain the package — I have a Broadcom card, usually run testing kernels and have some C knowledge.

bryn commented on 2015-10-01 22:16

Coincidentally there's just been an upstream release (so I flagged out-of-date :o), compiles but gives me kernel panic with BCM4352 on 4.2.2-1 :(

patrl commented on 2015-10-01 00:26

@twelveeighty god...I feel like a bit of the idiot. After paying a bit more attention to the version numbers, I realized the maintainer of the repository I'm using hadn't updated this package to the latest version. Everything's working great now. I'm so sorry!

twelveeighty commented on 2015-09-30 22:06

@patrl, that exact error was the reason for the update to rel 7 for this driver. Your error code actually states the old version of the line that was patched in the driver, so something went wrong when you updated the driver. In doubt, I always recommend to use makepkg and not an AUR helper tool to update.

patrl commented on 2015-09-30 15:54

@twelveeighty I haven't marked the driver as out of date, and yes, I updated the driver(!).

twelveeighty commented on 2015-09-30 14:15

This driver compiles without problems on kernel 4.2.1-1. For those that are reporting problems, did you actually update the driver since it went to version -7?

twelveeighty commented on 2015-09-30 14:08

Who marked this out of date? DO NOT mark a package out of date unless upstream has changed. Broadcom hasn't released an update to 6.30.223.248.

patrl commented on 2015-09-30 12:45

@CjK

I'm also using the 4352 Chipset, and after the upgrade to 4.2.1-1 my wifi stopped working too. Attempting to re-install broadcom-wl-dkms resulted in the error I posted previously. I can recommend switching to the lts kernel as a temporary fix.

Raimar commented on 2015-09-29 19:06

May be the following patch will help:
https://github.com/magnific0/broadcom-wl/blob/master/linux42.patch

CjK commented on 2015-09-29 16:16

I'm using the 4352 Chipset and both drivers (w/wo dkms) load fine. But recently my wireless connections stopped working all together, perhaps this is due to Kernel 4.2.x ?!

zerophase commented on 2015-09-29 15:56

Has anyone else had any issues detecting the wireless card BCM4360 with this driver? Broadcom-wl works fine with BCM4360, but not the dkms.

patrl commented on 2015-09-28 21:51

Error on kernel 4.2.1-1.

===========================================================================
DKMS make.log for broadcom-wl-6.30.223.248 for kernel 4.2.1-1-ARCH (x86_64)
Mon 28 Sep 22:47:38 BST 2015
make: Entering directory '/usr/lib/modules/4.2.1-1-ARCH/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.248/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c: In function ‘wl_notify_connect_status’:
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:2075:4: error: too few arguments to function ‘cfg80211_disconnected’
cfg80211_disconnected(ndev, 0, NULL, 0, GFP_KERNEL);
^
In file included from /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:33:0:
include/net/cfg80211.h:4584:6: note: declared here
void cfg80211_disconnected(struct net_device *dev, u16 reason,
^
scripts/Makefile.build:258: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o] Error 1
Makefile:1386: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.248/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.248/build] Error 2
make: Leaving directory '/usr/lib/modules/4.2.1-1-ARCH/build'
===========================================================================

trulex commented on 2015-09-27 20:36

My BCM43142 does not have AP mode with this driver. Is that normal?

joehillen commented on 2015-09-26 04:55

I'm going on a week long business trip so I will have lots of opportunity to try out roaming this week. I usually have lots of wireless crashes when I travel. I have a BCM4352. I'll report back in 7 days at the latest. Ping me if I forget.

gourdcaptain commented on 2015-09-20 19:59

Installed the patched version on my netbook, seems to be working but I really don't have an environment to test roaming. I'll try it tomorrow when I head into university.

Gonzih commented on 2015-09-19 08:50

I'm also getting "ERROR @wl_notify_scan_status : wlp3s0 Scan_results error (-22)" pretty often. After this error i need to manually disconnect and connect back to AP. Running arch on macbook pro 8.2 (BCM4331). Any hints where to look for a possible cause? Thanks!

jtyr commented on 2015-09-17 11:20

For those who want to test the patch and don't know how:

$ sudo pacman -S --needed base-devel
$ mkdir ~/pacman
$ curl -o - https://aur.archlinux.org/cgit/aur.git/snapshot/broadcom-wl-dkms.tar.gz | tar xzf - -C ~/pacman
$ curl -o ~/pacman/broadcom-wl-dkms/noroaming.patch https://launchpadlibrarian.net/217382168/noroaming.patch
$ echo "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" | base64 -d | patch -d ~/pacman/broadcom-wl-dkms -p2
$ cd ~/pacman/broadcom-wl-dkms
$ makepkg
$ sudo pacman -U ./broadcom-wl-dkms-6.30.223.248-8-x86_64.pkg.tar.xz
$ sudo netctl stop wireless
$ sudo rmmod wl
$ sudo modprobe wl
$ sudo netctl start wireless

mailme45 commented on 2015-09-15 15:38

I'd be willing to test that patch, if someone could point to a how-to.

jpike commented on 2015-09-13 20:41

Have a bcm43142 and this driver works but get terrible latency and lots of packet loss. Card works great in (sigh) windows 10.

twelveeighty commented on 2015-09-12 18:14

@trustin - thanks for the patch. Since it looks like there's a fair bit of code there, before I incorporate it: did Unbuntu adopt the patch into their packaging? I'm looking for some level of testing by others before we roll it in. Maybe there's some others here that want to test it? I no longer have this card, so I can't test it myself.

trustin commented on 2015-09-12 08:02

Would you be interested in including this patch:
- https://gist.github.com/trustin/aee6dedb8f96654eaa08

It fixes the module crash whenever it tries to roam from one AP to the other. Related information:
- https://bugs.launchpad.net/ubuntu/+source/bcmwl/+bug/1276651/

ashwin.vbs commented on 2015-09-01 01:18

On restart, Makefile in the linux-headers source gets replaced by a null file causing dkms build to fail on subsequent boots. Any idea whats causing this? or how to fix it?
This happens when building against linux-headers version 4.1.6-1.

Sracinas commented on 2015-08-31 16:51

@gokcen how did you figure it out?

Sracinas commented on 2015-08-31 16:49

@gokcen how did you figere it out?

gokcen commented on 2015-08-30 20:57

Nevermind, figured it out.

gokcen commented on 2015-08-30 20:47

@twelveeighty it seems it is coming in 4.3:

https://git.kernel.org/cgit/linux/kernel/git/tj/wq.git/commit/?h=for-4.3&id=1dadafa86a779884f14a6e7a3ddde1a57b0a0a65

Besides that, I am getting the following error on dmesg, even if I change bcmwl's license to GPL temporarily on 4.2-rc8:

wl: Unknown symbol flush_workqueue (err 0)

Does anyone know why?

man1gold commented on 2015-08-24 12:38

Nice... Running in kernel 4.1.6-1.

twelveeighty commented on 2015-08-17 03:11

@gokcen - see https://github.com/longsleep/bcmwl-ubuntu/issues/6, specifically the comment starting with "They were supposed to fix it in 4.2rc6 :".

Either the kernel devs commit the promised "Make flush_workqueue() available again to non GPL modules", or Broadcom wakes up and contributes GPL code.

Otherwise, you can break the rules and change the license type to "fix" this problem.

gokcen commented on 2015-08-16 21:44

Now there is another error:

DKMS make.log for broadcom-wl-6.30.223.248 for kernel 4.2.0-rc6-mainline (x86_64)
Sun Aug 16 23:38:09 CEST 2015
make: Entering directory '/usr/lib/modules/4.2.0-rc6-mainline/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.248/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o
LD [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/wl.o
Building modules, stage 2.
CFG80211 API is prefered for this kernel version
Using CFG80211 API
MODPOST 1 modules
FATAL: modpost: GPL-incompatible module wl.ko uses GPL-only symbol 'flush_workqueue'
scripts/Makefile.modpost:90: recipe for target '__modpost' failed
make[1]: *** [__modpost] Error 1
Makefile:1389: recipe for target 'modules' failed
make: *** [modules] Error 2
make: Leaving directory '/usr/lib/modules/4.2.0-rc6-mainline/build'

twelveeighty commented on 2015-08-16 20:48

NEW VERSION: 6.30.223.248-7

Added patch for 4.2 support, adopted from Ubuntu. Thanks for @gokcen for providing the links!

NOTE: I can no longer test anything myself, since I no longer have Broadcom hardware. Please report back with the results. Also: anyone who is interested in taking over maintenance for this package, please let me know? I'll maintain it until we find a replacement.

gokcen commented on 2015-08-16 13:55

Ubuntu added a patch for 4.2 support:

https://launchpad.net/ubuntu/+source/bcmwl/6.30.223.248+bdcom-0ubuntu4

which is debian/patches/0018-cfg80211_disconnected.patch.

gokcen commented on 2015-08-16 13:40

Error on kernel 4.2.0-rc6-mainline:

==================================
DKMS make.log for broadcom-wl-6.30.223.248 for kernel 4.2.0-rc6-mainline (x86_64)
Sun Aug 16 15:32:27 CEST 2015
make: Entering directory '/usr/lib/modules/4.2.0-rc6-mainline/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.248/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c: In function ‘wl_notify_connect_status’:
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:2075:4: error: too few arguments to function ‘cfg80211_disconnected’
cfg80211_disconnected(ndev, 0, NULL, 0, GFP_KERNEL);
^
In file included from /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:33:0:
include/net/cfg80211.h:4584:6: note: declared here
void cfg80211_disconnected(struct net_device *dev, u16 reason,
^
scripts/Makefile.build:258: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o] Error 1
Makefile:1386: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.248/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.248/build] Error 2
make: Leaving directory '/usr/lib/modules/4.2.0-rc6-mainline/build'
==================================

Looks similar to this:

https://lkml.org/lkml/2015/6/3/76

twelveeighty commented on 2015-08-10 00:22

I dunno - I started a Forum thread: https://bbs.archlinux.org/viewtopic.php?id=200871

hobarrera commented on 2015-08-10 00:01

broadcom-wl got deleted today, does anyone know why? No news about this anywhere.

twelveeighty commented on 2015-08-06 17:19

Folks, after a couple of years of maintaining broadcom-wl-dkms and enduring the Broadcom driver pains, I have finally switched to an Intel card. I don't mind staying as maintainer of this package, but since I can no longer test the driver myself, if there is a volunteer to take over, I'm okay with that as well. Please respond if you have an interest.

benob commented on 2015-08-06 08:06

After updating to linux-4.1.4-1, my BCM43b1 won't associate to a wpa2 network. It works fine with 4.1.3.

Atsuri commented on 2015-07-23 17:42

I am using the most recent available kernel and I have the exact error as Kuci. Only the wl module shows progress. kernel-contained b43 is associated with the device and firmware is loaded, but my bcm4322 14e4:432b wireless adapter is not recognized by Network Manager.

What is the state of drivers for broadcom wifi pci adapters? Is anyone even working on them still? :(

Kuci commented on 2015-06-04 13:24

Getting "ERROR @wl_notify_scan_status : wlp2s0 Scan_results error (-22)" and "ERROR @wl_cfg80211_scan : WLC_SCAN error (-22)" multiple times a day and my connection does not work afterwards. I have to turn off and turn on the Wi-Fi on my notebook every time it happens in order to get it to work again. I use BCM43228 [14e4:4359], any ideas ?

stiols commented on 2015-05-30 16:42

To get Bluetooth working i did the following:

git clone git://github.com/jessesung/hex2hcd.git
cd hex2hcd
make

wget http://www.iogear.com/support/driver/GBU521_Windows_XP_Driver_v5.6.0.7700.zip
unzip GBU521_Windows_XP_Driver_v5.6.0.7700.zip
cd GBU521_Windows_XP_Driver_v5.6.0.7700
cp GBU521_Windows_XP_Driver_v5.6.0.7700/Win64/drivers/BCM43142A0_001.001.011.0049.0056.hex .
./hex2hcd BCM43142A0_001.001.011.0049.0056.hex BCM43142A0_001.001.011.0049.0056.hcd
sudo cp BCM43142A0-0a5c-21d7.hcd /lib/firmware/brcm/
sudo modprobe -r btusb
sudo modprobe btusb

(credit http://askubuntu.com/users/300665/jeremy31)

If you have no controllers yet, try to run:
sudo systemctl restart bluetooth

leafac commented on 2015-05-20 19:50

@h3st, I can build this on 4.0.4, but I'm getting an error while running the driver, which I reported to the broadcom-wl package at https://aur.archlinux.org/packages/broadcom-wl.

h3st commented on 2015-05-20 19:10

Broken on 4.0.4 with another error

```
DKMS make.log for broadcom-wl-6.30.223.248 for kernel 4.0.4-1-ARCH (x86_64)
Wed 20 May 21:06:24 CEST 2015
make: Entering directory '/usr/lib/modules/4.0.4-1-ARCH/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.248/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c: In function ‘wl_cfg80211_get_station’:
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:1455:20: error: ‘STATION_INFO_TX_BITRATE’ undeclared (first use in this function)
sinfo->filled |= STATION_INFO_TX_BITRATE;
^
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:1455:20: note: each undeclared identifier is reported only once for each function it appears in
/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.c:1468:20: error: ‘STATION_INFO_SIGNAL’ undeclared (first use in this function)
sinfo->filled |= STATION_INFO_SIGNAL;
^
scripts/Makefile.build:258: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o] Error 1
Makefile:1390: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.248/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.248/build] Error 2
make: Leaving directory '/usr/lib/modules/4.0.4-1-ARCH/build'
```

twelveeighty commented on 2015-05-02 13:05

OK- thanks for confirming. I adopted a patch from @Leryan a while ago for our -6 release but broadcom-wl did not, I think, so that maybe why this dkms version works on 4.0.1-1.

hobarrera commented on 2015-05-02 10:19

Yes, it works fine on 4.0.1. The report on broadcom-wl seems to have been mistaken.

andreyv commented on 2015-05-02 07:36

I can also confirm that the module builds on 4.0.1-1.

pauldcox commented on 2015-05-02 00:30

I'm on kernel 4.0.1-1 and this builds and runs just fine so I don't know what that one guy who commented on broadcom-wl is talking about.

hobarrera commented on 2015-05-01 21:45

Comments on broadcom-wl state that the source does not build. This package merely adds dkms to that one (ie: it's the same source), so I'd guess it doesn't either.

twelveeighty commented on 2015-05-01 16:44

Has anyone had a chance to test the driver on the current kernel 4.0.1-1? I haven't seen any posts here, so I wonder if @Leryan patch still holds up?

hobarrera commented on 2015-04-29 19:53

@gokcen
This really needs to be reported to broadcome.

As a workaround, you can alter the licence setting in the downloaded source (sorry, I don't know *how* to do this), however, be warned that you cannot legally redistribute that source, nor the resulting module!

gokcen commented on 2015-04-27 14:13

broadcom-wl dkms gives the following error on linux-mainline 4.1RC:

$ sudo dkms install broadcom-wl/6.30.223.248
Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=4.1.0-1-mainline -C /usr/lib/modules/4.1.0-1-mainline/build M=/var/lib/dkms/broadcom-wl/6.30.223.248/build....(bad exit status: 2)
Error! Bad return status for module build on kernel: 4.1.0-1-mainline (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.30.223.248/build/make.log for more information.

$ cat /var/lib/dkms/broadcom-wl/6.30.223.248/build/make.log
DKMS make.log for broadcom-wl-6.30.223.248 for kernel 4.1.0-1-mainline (x86_64)
Mon Apr 27 15:57:23 CEST 2015
make: Entering directory '/usr/lib/modules/4.1.0-1-mainline/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.248/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/src/wl/sys/wl_cfg80211_hybrid.o
LD [M] /var/lib/dkms/broadcom-wl/6.30.223.248/build/wl.o
Building modules, stage 2.
CFG80211 API is prefered for this kernel version
Using CFG80211 API
MODPOST 1 modules
FATAL: modpost: GPL-incompatible module wl.ko uses GPL-only symbol 'cpu_tss'
scripts/Makefile.modpost:90: recipe for target '__modpost' failed
make[1]: *** [__modpost] Error 1
Makefile:1386: recipe for target 'modules' failed
make: *** [modules] Error 2
make: Leaving directory '/usr/lib/modules/4.1.0-1-mainline/build'

twelveeighty commented on 2015-03-21 21:23

NEW VERSION: 6.30.223.248-6

Merged changes from @Leryan to support kernel 4.0.0.

twelveeighty commented on 2015-03-03 01:14

@nextsux: thanks for the patch. I'll look at incorporating it. By the way, feel free to fork and submit merges at GitHub: https://github.com/twelveeighty/broadcom-wl-dkms, makes it a little easier.

nextsux commented on 2015-03-02 20:07

Yes, I can confirm it works with 3.18.6-1-ARCH too. I'm using exactly this patch: http://pastebin.com/Sy97gj04

twelveeighty commented on 2015-02-27 21:39

@nextsux: that patch doesn't encloses the fixes in #if LINUX_VERSION_CODE blocks. Have you tried building the driver with that fix on kernel 3.18? I'd hate to break the build for the current kernel.

tsukikari commented on 2015-02-26 23:52

Recently my wifi interface keeps turning on and off and won't connect to anything when this package is installed (although it was working before). Is there a way to fix this?

hobarrera commented on 2015-02-25 22:45

@nextsux: It's nice to know that somebody's testing ahead of time, and warns us that before 4.0 comes out and it breaks all of our setups. Thanks.

nextsux commented on 2015-02-25 22:22

Does not work with linux-mainline (4.0). You can find patch here: https://bugs.launchpad.net/ubuntu/+source/bcmwl/+bug/1424676

Sracinas commented on 2015-02-10 16:28

you can make a file modprobe.conf in /etc/modules-load.d/ and write lib80211_crypt_tkip inside it. Now it should automatically modprobe it on boot.

chrisdunder commented on 2015-02-10 07:01

WPA2 networks fail for me unless I modprobe lib80211_crypt_tkip. Is there something we can do about this?

Sracinas commented on 2015-02-09 14:56

Works with kernel 3.19 no need to patch it

alejandro.perez commented on 2015-02-06 15:01

Just use version .141. This is the only know version to work fairly well with BCM43142. At least for a couple of us here. And yes, I've made Bluethz to work.

debanshuk commented on 2015-02-03 11:22

I am still having frequent disconnections and slow speed issue with BCM43142 on kernel 3.18.4. I have started a new thread for this here https://bbs.archlinux.org/viewtopic.php?pid=1499313. Please reply to the thread if you are facing the similar issue(s) and/or have a solution.

trulex commented on 2015-02-01 17:47

Did anyone manage to get Bluetooth to work on BCM43142 with this driver maybe? Please help.

amir.autorun commented on 2015-01-23 11:03

@twelveeighty hey thanks man! looks like it's working pretty well for me.

hobarrera commented on 2015-01-21 15:39

@twelveeighty Proper credit goes to the guys at gentoo that actually posted the patch in the first place, but you're welcome.

twelveeighty commented on 2015-01-21 15:31

NEW VERSION

I've adapted the suggested patch - thanks to @hobarrera!

Please post back and confirm the kernel panic is fixed.

huiyiqun commented on 2015-01-21 01:22

I have tried the PKGBUILD below and everything goes well.

hobarrera commented on 2015-01-20 20:07

Here's an updated PKGBUILD and it's patch to fix this. Works fine on 3.18, no panics after over an hour. Many other have reported this patch working fine on the relevante forum thread:

https://gist.github.com/hobarrera/ac0e6225210ac5bb13f6

Kuci commented on 2015-01-20 18:26

What about trying this patch ? https://bugs.gentoo.org/show_bug.cgi?id=523326#c24

Kuci commented on 2015-01-20 18:26

What about trying this patch ? [url]https://bugs.gentoo.org/show_bug.cgi?id=523326#c24[/url]

twelveeighty commented on 2015-01-16 05:12

All - I started a thread on the Forums to discuss the kernel 3.18 problems. Rather than posting here, please visit:

https://bbs.archlinux.org/viewtopic.php?id=192287

Please post a response there with your specific Broadcom card as reported by "lspci -nn" (only the line for the Wifi card) and if your card is 1) experiencing the kernel panic running the broadcom-wl or broadcom-wl-dkms driver on 3.18; 2) working fine with that driver and kernel combination; or 3) if you switched to the b43 driver and that works. This will be a great help troubleshooting and providing workarounds for all affected.

andreyv commented on 2015-01-15 14:25

A small correction: it works here with "Network controller: Broadcom Corporation BCM4312 802.11b/g LP-PHY (rev 01)" and core/linux 3.18.2-2. However, I did experience a kernel panic once when I switched networks.

hobarrera commented on 2015-01-15 13:38

As I suggested om broadcom-wl: can we set depends=("linux<3.18") to avoid unwanted upgrades (after all 3.18 does not work anyway).

twelveeighty commented on 2014-12-31 15:43

NEW VERSION 6.30.223.248-4

Adopted same patches and changes as with AUR broadcom-wl.

twelveeighty commented on 2014-12-25 14:42

@step-2: please see https://bbs.archlinux.org/viewtopic.php?id=57855 for some useful tips on how to get started with troubleshooting problems like that.

step-2 commented on 2014-12-24 18:29

i can't get this driver to work with the core repo kernel for 3.17.6 nor earlier kernels , i install it & nothing happen , PLEASE HELP

am running gnome on 2013 macbook air .

twelveeighty commented on 2014-12-18 21:32

Folks - please see the comments in https://aur.archlinux.org/packages/broadcom-wl/, we are essentially downstream from that driver. If someone has a patch that fixes the problems with 3.18, I'd be happy to include it, but until then, this driver doesn't appear to work with kernel 3.18. Again, as I've said many many times before: complain to Broadcom and/or do not buy their hardware.

xuhdev commented on 2014-12-18 20:43

The whole system would be stuck once connecting to a wireless network after upgrading to 3.18.1. Here is the dmesg info: http://pastebin.com/7p4ww0e6

twelveeighty commented on 2014-12-18 03:39

NEW VERSION 6.30.223.248-3

Adopted same patch and changes as with AUR broadcom-wl.

Sracinas commented on 2014-12-17 19:14

3.18.1 is already in testing repo

Sracinas commented on 2014-12-13 18:51

The thing is, that I tried applying the patch and it didn't work, there is already a patch made for 3.18 bcmwl in launchpad, but I can't seem to apply it correctly.

twelveeighty commented on 2014-12-13 13:59

@Scracinas: https://github.com/twelveeighty/broadcom-wl-dkms.

After you've tested your patch, post a pull request.

Sracinas commented on 2014-12-11 03:30

Hello, could you add a patch for kernel 3.18 or at least write how to do it myself? Thank You

ranger commented on 2014-11-26 09:16

I was to lazy to make a patch so I just edit directly the sources and did a makepkg -e

yochai commented on 2014-11-26 05:57

Ranger, thank you for the effort, my BCM43142 is useless with the current driver.
Are you willing to upload the PKGBUILD file ?

ranger commented on 2014-11-25 12:56

With the help of alejandro.perez I managed to build the previous version (6.30.223.141) now fully working with latest kernel. No more disconnections or slow speed with my BCM43142.
If anyone is interested, I uploaded the package (x86_64) to archlinuxgr repo
https://wiki.archlinux.org/index.php/unofficial_user_repositories#archlinuxgr

xuhdev commented on 2014-11-25 10:07

Just let everyone know that this driver is quite unstable (speed down to 0 often) for BCM4331. When it works normally, the speed is as fast as the OS X driver.

twelveeighty commented on 2014-11-05 22:50

@swoosh: This driver is *exactly* the same as broadcom-wl. Same patches, except this one adds configuration for dkms.

swoosh commented on 2014-11-05 20:39

quick update:
broadcom-wl works like charm on 3.17.2

twelveeighty commented on 2014-11-05 17:23

<sigh>

Please ONLY mark packages out of date if upstream is at a higher version. This rule applies throughout ArchLinux, not just for this package.

Please check the comment thread in AUR broadcom-wl: Ubuntu is using an older version of this package that seems to work for some cards. However, this 6.30.223.248 version works fine on my laptop here.

If someone wants to maintain a different package matched to an older version using Ubuntu's patches, go ahead, but I don't have time for it. Maybe call it broadcom-wl-dkms-legacy or something.

In the mean time, keep the pressure on Broadcom to either open-source or fix their crappy drivers, or, as I've said many times before, do NOT buy their hardware.

swoosh commented on 2014-11-05 12:03

@ranger
just saw that our chip is kinda well supported on ubuntu.
i just love arch but also need wlan to work with my new system.
did you try if ndiswrapper would maybe work?
any1 who is able to provide a patch? :/

ranger commented on 2014-11-05 11:56

@swoosh, twelveeighty maintains the latest version. If we want the previous version patched we will have to build it ourselves, or make a new package. But I don't know if there is a patch.

swoosh commented on 2014-11-05 11:49

@ranger
thanks for your reply. thinking about headin to ubuntu after my new notebook with bcm43142 arrvies.
@twelveeighty
is a patch ranger talked about going to happen soon?
thanks in advance

ranger commented on 2014-11-05 10:27

With BCM43142 wifi doesn't work.
It tries to connect but fails. And when it connects (1 out of 10) I have very low speeds and disconnections.

We need 141 driver patched for kernel 3.17

hobarrera commented on 2014-11-05 08:17

@swoosh: Latest kernel works for me with a BCM4360 (MacBook Air 2013), but other have reported issues. Not sure about a BCM43142.

swoosh commented on 2014-11-04 22:57

could anyone tell me if this works with an up2date kernel?
card: BCM43142

hobarrera commented on 2014-10-31 16:01

I've had no wifi issues with 3.17.1-1-ARCH. Haven't rebooted since upgrading to 3.17.2-1 though.
Lemme know if there's some details that might help.

KubAT commented on 2014-10-30 16:15

So I confirm, on 3.17.1-1 wifi fails over and over.

twelveeighty commented on 2014-10-28 22:54

@eduardoeae: would you mind adding your changes to https://github.com/twelveeighty/broadcom-wl-dkms and submitting a pull request? Makes it so much easier to manage.

eduardoeae commented on 2014-10-25 12:12

Please take this into consideration:
http://pastebin.com/idVHFBtF

It is a patch I sumbitted to you on 2014-08-05, so the module works on -lts kernels as well.

ranger commented on 2014-10-24 15:44

No luck even with this kernel. Most of the times wifi fails to connect and if it gets connected it is very unstable.Slow speeds and disconnecting all the time. With 141 I had no problems, but unless someone patches 141 for kernel 3.17 we are stuck.

ranger commented on 2014-10-24 14:00

Thanks. Builds fine now. Let's hope for no problems with BCM43142 because now I can't downgrade

twelveeighty commented on 2014-10-24 13:44

New version with the latest kernel patch has been uploaded: 6.30.223.248-2.

Kuci commented on 2014-10-23 16:08

You can use linux-recent.patch from https://aur.archlinux.org/packages/broadcom-wl/.

andreyv commented on 2014-10-23 14:27

Looks like a patch is available here: https://bugs.launchpad.net/ubuntu/+source/bcmwl/+bug/1358966

stathisq commented on 2014-10-23 09:35

same here with linux 3.17.1-1

ranger commented on 2014-10-23 08:46

Module doesn't build with linux 3.17.1-1

Building module:
cleaning build area....
make KERNELRELEASE=3.17.1-1-ARCH -C /usr/lib/modules/3.17.1-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.30.223.248/build....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.17.1-1-ARCH (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.30.223.248/build/make.log for more information.

twelveeighty commented on 2014-10-04 13:28

@tiagofalcao - can you please check out https://github.com/twelveeighty/broadcom-wl-dkms. I got a pull request there for this driver a while ago that looks very similar to yours, but not identical. Can you check out the master branch and compare/test? Once 3.17 enters Testing, I'll cut a new release here with that 3.17 patch.

tiagofalcao commented on 2014-10-03 22:21

Please, add this patch: https://launchpadlibrarian.net/183190658/0013-add-support-for-Linux-3.17.patch

Koins commented on 2014-09-28 01:07

Ok so I'm not particularly versed in this field, so you may have to bear with me a little bit, but here it goes. I have a post mid 2013 Macbook Air, which has a BCM4360 as its wireless card. I often boot of a persistent Live Kali USB, and want to be able to use tools like airmon, airodump, etc. My problem so far is that I can't get the card into monitor mode with the 6.30.223.248 drivers. Will the 6.30.223.141 drivers work for me? Is there anything I can do to get monitor mode support for the Broadcom card?

Thanks!

MissingNo commented on 2014-09-19 08:22

I would say someone with a broken wifi driver flagged it who thought it meant something different.

I really hope Broadcom fixes this sh*t soon

twelveeighty commented on 2014-09-12 16:30

Who flagged this out-of-date? Upstream is still at 6.30.223.248

ranger commented on 2014-08-20 13:37

With my BCM43142 I have same problems as parachuter and alejandro.perez. I had to downgrade

azrdev commented on 2014-08-13 20:05

thanks @twelveeighty!
still broken with linux 3.16.0-2 kernel

*BUT* the brcmsmac driver in-kernel is currently working for me, as opposed to earlier kernel versions… so no need for broadcom-wl anymore. \o/

azrdev commented on 2014-08-13 18:57

thanks @twelveeighty!
still broken with linux 3.16.0-2 kernel

twelveeighty commented on 2014-08-13 15:43

@memeplex: https://github.com/twelveeighty/broadcom-wl-dkms. I don't know if Broadcom keeps their old version around, so the older PKGBUILD may not work.

memeplex commented on 2014-08-11 03:47

@parachuter or @twelveeighty, could you share the pkgbuild / tarball for downgrading to 141? Thank you in advance.

eduardoeae commented on 2014-08-05 21:02

Here it is the linux-recent.patch fixed for compiling with both -ARCH and -lts kernels
http://pastebin.com/idVHFBtF

azrdev commented on 2014-07-27 11:33

upgraded broadcom-wl-dkms (6.30.223.141-7 -> 6.30.223.248-1)
upgraded linux (3.15.1-1 -> 3.15.5-2)

System hangs hard (cannot even switch to tty) some time after booting, especially when starting instant messengers.

alejandro.perez commented on 2014-07-27 07:29

I retract of my previous comment. This new version makes my connection much worse than before, as it happens to @parachuter.

alejandro.perez commented on 2014-07-26 09:43

@parachuter, I had the same problem because I just removed "wl" and the loaded it again. I didn't notice the error messages on dmesg. After reboot, this new drive works much better then 141. At least in my case.

eduardoeae commented on 2014-07-21 13:56

You are correct @twelveeighty, now it compiles and works fine on lts kernel.

parachuter commented on 2014-07-21 10:05

I'm getting performance issues with the latest version on a BCM43142. The previous version worked fine, with this one however my internet connection slows down to a crawl. dmesg shows no errors, but there's a large number of dropped packets in wireshark. I downgraded to .141 for the time being.

twelveeighty commented on 2014-07-20 17:31

@eduardoeae, I have a feeling that is because the first patch in linux-recent.patch may be for 3.15 only (cfg80211_ibss_joined). Can you please try editing linux-recent.patch, remove the first part that changes wl_cfg80211_hybrid.c and report back? You'll have to update the sha256sum for that file, of course.

eduardoeae commented on 2014-07-20 17:06

Build fails with lts kernel. Sorry can't share the log right now.

DeltaKilo commented on 2014-07-19 18:56

System silently hungs on echoing 1 to /proc/brcm_monitor0 on 3.15.5-2 kernel using BCM43227, so if you use MONITOR functionality DO NOT UPDATE until patches exist.

DeltaKilo commented on 2014-07-19 18:55

System silently hungs on echoing 1 to /proc/brcmwl0 on 3.15.5-2 kernel using BCM43227, so if you use MONITOR functionality DO NOT UPDATE until patches exist.

DeltaKilo commented on 2014-07-19 18:54

System silently hungs on echoing 1 to /proc/brcmwl0 on 3.15.5-2 kernel, so if you use MONITOR functionality DO NOT UPDATE until patches exist.

twelveeighty commented on 2014-07-19 14:37

NEW VERSION NOTIFICATION: 6.30.223.248-1

Please test and report back packaging issues if found.

izmntuk commented on 2014-07-18 08:58

New version from upstream available, please check upstream page for details.

acssilva commented on 2014-07-05 18:05

@twelveeighty: yes i did, but is still not working. All i get is "no wireless network found"

twelveeighty commented on 2014-06-29 01:11

@acssilva, @baronmog: you probably didn't update to the -7 version then?

acssilva commented on 2014-06-26 12:51

@baronmog Same problem here.

baronmog commented on 2014-06-25 06:18

Doesn't build against the 3.15.1-1-ARCH kernel:

CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.o
/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.c: In function ‘wl_notify_connect_status’:
/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.c:1844:4: warning: passing argument 3 of ‘cfg80211_ibss_joined’ makes pointer from integer without a cast
cfg80211_ibss_joined(ndev, (u8 *)&wl->bssid, GFP_KERNEL);
^
In file included from /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.c:33:0:
include/net/cfg80211.h:3933:6: note: expected ‘struct ieee80211_channel *’ but argument is of type ‘unsigned int’
void cfg80211_ibss_joined(struct net_device *dev, const u8 *bssid,
^
/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.c:1844:4: error: too few arguments to function ‘cfg80211_ibss_joined’
cfg80211_ibss_joined(ndev, (u8 *)&wl->bssid, GFP_KERNEL);
^
In file included from /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.c:33:0:
include/net/cfg80211.h:3933:6: note: declared here
void cfg80211_ibss_joined(struct net_device *dev, const u8 *bssid,
^
scripts/Makefile.build:318: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.o] Error 1
Makefile:1310: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.141/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.141/build] Error 2

step-2 commented on 2014-06-14 16:41

Hi , may i ask what's the difference between this driver & broadcom-wl ?

also , i have a macbook air 2013 , should i be using this or broadcom-wl ?

hippieben commented on 2014-06-10 18:47

I'm not sure why, but I'm not getting any 5GHz support with this driver. I used to, but not anymore. Any ideas?

twelveeighty commented on 2014-06-08 20:41

I didn't realize until today that 3.15 has been released. As such, I've updated this package to include the patch for that kernel.

twelveeighty commented on 2014-06-08 15:09

@jgke: based on @WonderWoofy comment in broadcom-wl back in April, I have his patch ready to go. Since we had a bunch of other changes happening in April, I didn't want to introduce too many patches. In lieu of the new PKGBUILD, here's the patch I have lined up:

diff -Naur a/src/wl/sys/wl_cfg80211_hybrid.c b/src/wl/sys/wl_cfg80211_hybrid.c
--- a/src/wl/sys/wl_cfg80211_hybrid.c 2013-08-01 00:52:22.000000000 -0600
+++ b/src/wl/sys/wl_cfg80211_hybrid.c 2014-05-03 08:36:50.708972367 -0600
@@ -1841,7 +1841,11 @@
wl_get_assoc_ies(wl);
memcpy(&wl->bssid, &e->addr, ETHER_ADDR_LEN);
wl_update_bss_info(wl);
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 15, 0)
+ cfg80211_ibss_joined(ndev, (u8 *)&wl->bssid, &wl->conf->channel, GFP_KERNEL);
+#else
cfg80211_ibss_joined(ndev, (u8 *)&wl->bssid, GFP_KERNEL);
+#endif
set_bit(WL_STATUS_CONNECTED, &wl->status);
wl->profile->active = true;
}

jgke commented on 2014-06-07 19:34

This package will fail to build in the future (3.15 onwards) due to a changed function. This patch should prevent that.

http://pastebin.com/hkVFEx23

twelveeighty commented on 2014-05-04 15:51

Edit: I think this settles the linux-headers debate pretty clearly:

https://wiki.archlinux.org/index.php/Dkms#Dependencies

twelveeighty commented on 2014-05-04 15:34

@hobarrera and @zrl: as I mentioned a couple of days ago, please take the linux-headers discussion to AUR mailing list - this is not the right platform for such a discussion. I get as many people asking it to be removed as a dep as I get to have it as optdep (as it is now).

The real problem is that there is no generic "provides" for linux-headers, and other kernels such as linux-lts and linux-ck have their own linux-lts-headers and linux-ck-headers.

Technically speaking, this driver should actually not even list linux-headers as an (optional) dependency, just take a look at the dkms deps: it lists linux-headers as an optional dep. Users should have examined its installation log and decided which linux-headers to install at that point prior to installing/using any dkms drivers.

When it's all said and done, I could not care less bringing in linux-headers as a full dependency, but since this is Arch, we should do things the "right" way, under the KISS principles, so drivers should be managed consistently. The moment I bring in linux-headers as a dep, I am forced to type another long note like this because folks are complaining and we're back to square 1 - so let's please hash it out on the AUR mailing list instead?

hobarrera commented on 2014-05-04 03:29

@zrl: It should. I already reported this a month ago.

zrl commented on 2014-05-03 20:11

Shouldn't this be dependent on linux-headers? I was unable to install without them installed.

zrl commented on 2014-05-03 20:11

Shouldn't this be dependent on linux-headers? I was unable to install without them installed.

twelveeighty commented on 2014-05-03 13:52

I agree with @Krejzi over at broadcom-wl: @Zombifier's patch seems more elegant by using an EXTRA_CFLAGS in Makefile vs. patching wl_linux.c itself. This update to -6 includes that fix instead.

ranger commented on 2014-05-03 10:42

thanks twelveeighty.
everything back to normal

twelveeighty commented on 2014-05-03 04:46

New version has been updated, driver now works for kernel 3.14.2.

Thanks to @bricewge and @hckr for the patches!

I also adopted some requested changes in the .install portion. I haven't changed linux-headers from optdepends to depends - it seems there's no consensus on this - maybe we should take that discussion to the AUR mailing list so all dkms drivers are consistent.

I only did some basic testing - as in installing it on my own laptop and it installed without errors, so please test it and post back if there's issues.

bricewge commented on 2014-05-02 14:53

A working patch can be found there (http://hastebin.com/wokesufoki.pl) thanks to Krejzi.

And a working source tarball here (https://drive.google.com/file/d/0BygqpGaN4JGARURtdWtBcHJ2REE/edit?usp=sharing) until this package is updated.

mid-kid commented on 2014-05-02 13:09

The maintainer of broadcom-wl has posted a patch. It'd be great if that got added here.

produnis commented on 2014-05-02 10:03

same here (like bricewge described)

bricewge commented on 2014-05-02 09:38

Like yochai I can't build this package anymore with linux-3.14.2-1 even with 4.9.0-1 which was thought to fix this bug (https://bugs.archlinux.org/task/40171).

My log is not the same:
$ cat /var/lib/dkms/broadcom-wl/6.30.223.141/build/make.log
DKMS make.log for broadcom-wl-6.30.223.141 for kernel 3.14.2-1-ARCH (x86_64)
ven. mai 2 11:28:38 CEST 2014
make: Entering directory '/usr/lib/modules/3.14.2-1-ARCH/build'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.141/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.o
/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.c: In function ‘wl_dump_ver’:
/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.c:2042:3: erreur: macro "__DATE__" might prevent reproducible builds [-Werror=date-time]
__DATE__, __TIME__, EPI_VERSION_STR);
^
/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.c:2042:13: erreur: macro "__TIME__" might prevent reproducible builds [-Werror=date-time]
__DATE__, __TIME__, EPI_VERSION_STR);
^
cc1: some warnings being treated as errors
scripts/Makefile.build:308: recipe for target '/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.o' failed
make[1]: *** [/var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.o] Error 1
Makefile:1274: recipe for target '_module_/var/lib/dkms/broadcom-wl/6.30.223.141/build' failed
make: *** [_module_/var/lib/dkms/broadcom-wl/6.30.223.141/build] Error 2
make: Leaving directory '/usr/lib/modules/3.14.2-1-ARCH/build'

altercation commented on 2014-05-02 03:41

As noted, 3.14 breaks this. Vote for and watch the relevant bug: https://bugs.archlinux.org/index.php?do=details&action=details.addvote&task_id=40171

rabidcoder commented on 2014-05-01 19:28

yochal, this link has more information: https://bbs.archlinux.org/viewtopic.php?id=180924

yochai commented on 2014-05-01 18:34

Hey,

I've ran pacman -Suy today and the driver doesn't build any more:

cat /var/lib/dkms/psmouse-alps/1.3/build/make.log
DKMS make.log for psmouse-alps-1.3 for kernel 3.14.2-1-ARCH (x86_64)
Thu May 1 20:47:00 IDT 2014
make: Entering directory '/usr/lib/modules/3.14.2-1-ARCH/build'
Makefile:615: Cannot use CONFIG_CC_STACKPROTECTOR_STRONG: -fstack-protector-strong not supported by compiler
CC [M] /var/lib/dkms/psmouse/alps-1.3/build/src/psmouse-base.o
gcc: error: unrecognized command line option ‘-fstack-protector-strong’
scripts/Makefile.build:308: recipe for target '/var/lib/dkms/psmouse/alps-1.3/build/src/psmouse-base.o' failed
make[1]: *** [/var/lib/dkms/psmouse/alps-1.3/build/src/psmouse-base.o] Error 1
Makefile:1440: recipe for target 'psmouse.ko' failed
make: *** [psmouse.ko] Error 2
make: Leaving directory '/usr/lib/modules/3.14.2-1-ARCH/build'


Any idea on how to fix that ?

twelveeighty commented on 2014-04-30 19:33

I couldn't care less about adding the dep back, but about a year or so ago (scroll back for the comments) this was a big deal to have it taken out as a dep.

andreyv commented on 2014-04-29 15:04

However, now that I think about it, another approach would be to do just that and let users edit the PKGBUILD if they want.

andreyv commented on 2014-04-29 14:55

@hobarrera: Should linux-lts-headers also be in depends then? And then linux-git-headers etc. ?

The real problem here is that the linux-*-headers packages don't 'provide' a common name.

andreyv commented on 2014-04-29 14:54

@hobarerra: Should linux-lts-headers also be in depends then? And then linux-git-headers etc. ?

The real problem here is that the linux-*-headers packages don't 'provide' a common name.

hobarrera commented on 2014-04-29 14:51

linux-headers should be in depends, not optdepends.

This package fails without it! (I cleaned all orphaned dependencies a few days ago, and this stopped working since then, since those headers are necesary).

matersci commented on 2014-04-23 15:26

Same here on my vaio Fit

kumen commented on 2014-04-23 13:59

I have 100% same problem as "acssilva" with bluetooth on my Thinkpad E531

acssilva commented on 2014-04-16 19:33

Thanks a lot for the package. First i was unable to install it, but after reading some tips i realized what was the problem. I needed to install "linux-headers" before perform the installation. After install it, Wi-Fi is working fine.

Now i am in trouble with the bluetooth. My card is a BCM43142 and i think the bluetooth adapter is bundled with it (http://www.broadcom.com/products/Wireless-LAN/802.11-Wireless-LAN-Solutions/BCM43142).

I read bluetooth page (https://wiki.archlinux.org/index.php/Bluetooth) a few times and tried some tips they give. Especially to install "bluez" and "bluez-utils" packages and then in the Troubleshooting to switch the Bluetooth mode from HID to HCI if the device is not found. After to try these ones here there are some outputs:

# lspci | grep Network
02:00.0 Network controller: Broadcom Corporation BCM43142 802.11b/g/n (rev 01)

# lsusb
Bus 002 Device 003: ID 058f:6366 Alcor Micro Corp. Multi Flash Reader
Bus 002 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 001 Device 005: ID 5986:0295 Acer, Inc
Bus 001 Device 004: ID 105b:e065
Bus 001 Device 003: ID 147e:1002 Upek Biometric Touchchip/Touchstrip Fingerprint Sensor
Bus 001 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

# hcitool dev
Devices:

# rfkill list
0: tpacpi_bluetooth_sw: Bluetooth
Soft blocked: no
Hard blocked: no
1: phy0: Wireless LAN
Soft blocked: no
Hard blocked: no

# hciconfig -a

# dmesg | grep Blue
[ 3329.124593] Bluetooth: Core ver 2.18
[ 3329.124625] Bluetooth: HCI device and connection manager initialized
[ 3329.124634] Bluetooth: HCI socket layer initialized
[ 3329.124639] Bluetooth: L2CAP socket layer initialized
[ 3329.124654] Bluetooth: SCO socket layer initialized
[ 4275.252345] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[ 4275.252352] Bluetooth: BNEP filters: protocol multicast
[ 4275.252364] Bluetooth: BNEP socket layer initialized

Any help is welcome.
Thanks a lot in advance.

mfuentesu commented on 2014-04-07 01:40

Ok, thanks for your patience, and fast response. I will know find a way that my system don't put down the network interfaces when it suspends :)...
Thanks again

bricewge commented on 2014-04-06 19:36

@mfuentesu You're welcome! I experienced the same thing, a freeze when disabling the monitor mode. But as far I remember this doesn't happen with the first version of this package which included this feature...

mfuentesu commented on 2014-04-06 19:13

@bricewge Thnxs!!, simple little things :P
by the way..do you know why when i try
echo 0 > /proc/brcm_monitor0
my system frezze??.
Just to know, i really dont need to put it down :D

bricewge commented on 2014-04-06 18:09

@mfuentesu to enable the monitor mode you have to this following command as documented here (http://www.broadcom.com/docs/linux_sta/README.txt).
# echo 1 > /proc/brcm_monitor0
This will create a 'prism0' network interface.

mfuentesu commented on 2014-04-06 00:11

I know someone in the previous comments says that the monitor mode is now working...But i just can't make it work in a BCM43142 at least with:
> iwconfig "interface" mode monitor

it says :
Error for wireless request "Set Mode" (8B06) :
SET failed on device wlp4s0 ; Operation not supported.

any help would be appreciated.

twelveeighty commented on 2014-04-03 23:03

@hobarrera, please read the previous comments on the discussion for linux-headers. They'll explain why it's optdepends.

hobarrera commented on 2014-04-02 13:43

linux-headers should be in depends, not optdepends.

twelveeighty commented on 2014-02-19 02:10

@andreyv - I'll take out the rmmod / modprobe and depmod commands from the install script. Next update will take those changes into account, but I'm not releasing a new pkgrel just for those changes.

andreyv commented on 2014-02-03 16:03

Is depmod needed in dkms_install()? "dkms install" already executes it.

In my opinion installing the package should not try to load the module as well, as this may disconnect the user without warning from the network. Perhaps it would be better to just print an advice to execute the rmmod ... modprobe commands, or reboot.

twelveeighty commented on 2014-02-02 22:22

Folks,
I updated the .install script hooks, which now should work properly when updating, installing and removing the package. Calling pacman -U should now properly dkms remove / install without the need to manually do that in case of updates.
Please test and provide feedback where appropriate.

twelveeighty commented on 2014-01-17 23:01

I think I can probably improve the pre_remove() hook which I stole from another dkms package a long time ago and never revisited. That should improve the "upgrade" functionality. I usually uninstall the old version and then install the new one, so it doesn't bother me that much, but I'll fix it in the next release.

mpalvig commented on 2014-01-09 18:21

Same here, had to do:
dkms add -m broadcom-wl -v 6.30.223.141
systemctl restart dkms.service

bricewge commented on 2014-01-05 18:08

Same issue as jyaworski here, I had to remove the package and reinstall it to be build correctly.

jyaworski commented on 2014-01-05 15:20

@andreyv: This is after reboot. The -1 version of this package was working fine. Upgrading to -2 and rebooting, there's no module.

However, I took your advice and reran the install script. It looks like the module wasn't added, built or installed. It works fine after doing those steps. Thanks.

andreyv commented on 2014-01-05 15:11

@jyaworski: You probably need to either reboot or manually run the dkms build/dkms install commands.

jyaworski commented on 2014-01-05 15:09

wl refuses to build with dkms after installing this. Actually, it doesn't even appear in the dkms status output. Any ideas?

bricewge commented on 2014-01-05 11:42

@twelveeighty: Thanks for adding this patch, the monitor mode is now working.

twelveeighty commented on 2014-01-04 23:16

I've adopted the latest linux-recent.patch from broadcom-wl as requested.

@mickael9: can you please test this package and report back the results?

mickael9 commented on 2014-01-04 17:58

Hi, can you please take the latest linux-recent.patch from broadcom-wl that fixes monitor mode ?

Thanks.

felixonmars commented on 2013-12-02 00:55

@MazeChaZer
No, as users may be using linux-lts or another kernel, so the best place for "linux-headers" should be optdepends, like what the vbox dkms package in [community] did.

MazeChaZer commented on 2013-12-01 21:31

Works flawlessly with the wifi-chip of my Mac Book Air 2013, thanks for the package. However I had to install "linux-headers" like jpala already said to be able to build the package, I'm not sure if this should be a dependency as I am quite new to Arch Linux.

hippieben commented on 2013-11-25 02:50

Seems to work pretty well (better than b43 overall, especially on my school's network), however I get these peppered though out my log. Sometimes my connection will stop working, but that could be the router (came with my apartment) or my ISP (Comcast, which I would never have if I had any choice in the matter).

ERROR @wl_notify_scan_status : wlp3s0 Scan_results error (-22)

twelveeighty commented on 2013-11-08 20:04

Works fine with WPA connections for me. Not likely to be a driver issue, but network management configuration. Check your netcfg, NetworkManager, etc. settings. Also see: https://wiki.archlinux.org/index.php/Wireless_Setup.

neonmagnets commented on 2013-11-06 12:36

I added this module to my laptop (broadcom BCM43142) and I can now find networks. Which is great and all, but I can't set up WPA or WPA2 connections, only WEP.

Can this be right, or is something wrong?

the other package in aur broadcom-wl-no-dkms specifically mentions it, but it fails to build.

twelveeighty commented on 2013-10-22 16:56

Why did someone mark this out-of-date? Upstream is at 6.30.223.141.

aruniiird commented on 2013-10-13 14:15

Hey,
Upgrade to the latest linux seems to have solved the problem...

uname -a
Linux manjaro 3.10.15-1-MANJARO #1 SMP Sat Oct 5 17:55:55 UTC 2013 x86_64 GNU/Linux

Thanks,
aruniiird

aruniiird commented on 2013-10-11 04:19

Hi twelveeighty,
I am getting the following error while installing the driver...

_______________________________

>>> DKMS: Module add, build, and install

Creating symlink /var/lib/dkms/broadcom-wl/6.30.223.141/source ->
/usr/src/broadcom-wl-6.30.223.141

DKMS: add completed.

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.10.15-1-MANJARO -C /usr/lib/modules/3.10.15-1-MANJARO/build M=/var/lib/dkms/broadcom-wl/6.30.223.141/build....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.10.15-1-MANJARO (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.30.223.141/build/make.log for more information.

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.10.15-1-MANJARO -C /usr/lib/modules/3.10.15-1-MANJARO/build M=/var/lib/dkms/broadcom-wl/6.30.223.141/build....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.10.15-1-MANJARO (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.30.223.141/build/make.log for more information.
modprobe: FATAL: Module wl not found.
==> For more information on this driver, see:
==> http://www.broadcom.com/support/802.11/linux_sta.php
___________________________________

The log has following message...
_____________________________________

DKMS make.log for broadcom-wl-6.30.223.141 for kernel 3.10.15-1-MANJARO (x86_64)
Fri Oct 11 09:44:34 IST 2013
make: Entering directory `/usr/src/linux-3.10.15-1-MANJARO'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.30.223.141/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/src/wl/sys/wl_cfg80211_hybrid.o
LD [M] /var/lib/dkms/broadcom-wl/6.30.223.141/build/wl.o
Building modules, stage 2.
CFG80211 API is prefered for this kernel version
Using CFG80211 API
MODPOST 1 modules
FATAL: vmlinux has no symtab?
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2
make: Leaving directory `/usr/src/linux-3.10.15-1-MANJARO'
_____________________________________

andreyv commented on 2013-10-07 09:41

Has anyone tried this driver with Wi-Fi channels 12 or 13? Here it doesn't seem to find networks on these channels, even after setting the correct regulatory domain.

jyaworski commented on 2013-09-27 03:43

I'm getting this repeated error on the 3.11 kernel:

@wl_cfg80211_scan : WLC_SCAN error (-22)

Is this related to the patch?

twelveeighty commented on 2013-09-24 03:21

@jpala: see the comments below around the July 26 timeframe. It *was* a dependency and it was requested to be removed. I have no problem with it either being a dep or not, but I wish someone had the final answer to that.

jpala commented on 2013-09-22 06:59

Hi, I'm not sure but is there a missing dependency for "linux-headers" in this package.

I have a HP 2140 laptop with BCM4132 chipset. I followed the instructions from the wiki (including installing dkms) but whenever I tried to load the wl module I got a "wl module not found" error.

I then installed linux-headers manually and rebuilt/reinstalled this package. Everything works fine now.

Thanks for the package and I hope this is helpful.

twelveeighty commented on 2013-09-20 23:55

@abruptadieu: it should rebuild at the next boot after a kernel update, assuming that dkms is enabled, which it is according to your systemctl status.

yochai commented on 2013-09-20 05:20

abruptadieu, you need to add it using dkms add.

abruptadieu commented on 2013-09-19 04:44

I'm still having to manually invoke
dkms install /usr/src/broadcom-wl-6.30.223.141/
every time my kernel upgrades. What am I doing wrong? I have dkms enabled at boot.

$ sudo systemctl status dkms.service

dkms.service - Dynamic Kernel Modules System
Loaded: loaded (/usr/lib/systemd/system/dkms.service; enabled)
Active: active (exited) since Wed 2013-09-18 23:28:31 CDT; 12min ago
Process: 523 ExecStart=/usr/lib/systemd/scripts/dkms start (code=exited, status=0/SUCCESS)

andreyv commented on 2013-09-17 16:04

Thanks for the update. The new driver seems to work fine with BCM4312.

I think it is useful to have a DKMS-flavored version of the driver, as it doesn't need maintainer intervention for new kernels. There are also some DKMS-based packages in the main repos (https://www.archlinux.org/packages/?q=dkms).

silvik commented on 2013-09-17 14:51

@twelveeighty: thank you for maintaining this package, some of us need a working dkms version (nonstandard/custom kernels)

SMOG commented on 2013-09-17 14:33

I've moved to using broadcom-wl... it works pretty well...
sorry guys, it has been nice to be here with you, but now we are no more orphans, now we have finally a father! :P
Thank you twelveeighty for everything you have done to let us use our wifi while broadcom didn't care.

twelveeighty commented on 2013-09-17 14:17

Folks - package has been updated to use the official Broadcom release files. This PKGBUILD should now build the exact same driver as broadcom-wl, but for dkms.
I was only able to test the 64 bit version installation. If someone can report back here when testing the 32 bit, that would be great.

twelveeighty commented on 2013-09-17 00:46

Thanks, WonderWoofy. I'll take a look at the official released code by Broadcom. We can go a couple of different ways here: we can swap the sources to point to Broadcom and I can keep maintaining it, or if someone else is really eager to maintain this other codebase using this package name, have at it! I have no emotional attachment to this package whatsoever. I just need my wireless to work.

Now, knowing Broadcom and their lack of dedication to Linux, they may have fsck'd up "their" v6 code so it won't work, in which case we'll need this v6 code until such time they've fixed it again.

WonderWoofy commented on 2013-09-16 06:06

I just want to let you know that the v6 wl is now the officially released module on the broadcom site. So you can use that and it will not require all this insane patching to make it work. The regular broadcom-wl package is already using this, as are a few of the other wl packages. The version from the broadcom site is also a newer version then the one you are shipping here.

Though, with that being the case, I am now not entirely sure if this package is necessary… though I guess this would be the most suitable name for the officially released broadcom wl module in dkms packaging form.

twelveeighty commented on 2013-09-01 16:36

I could be just paranoid and it's just a timing issue, but I have seen it where the driver wasn't loaded after the reboot following a kernel update. But either way, I don't think a reinstall is necessary.

SMOG commented on 2013-09-01 16:19

If a second reboot isn't necessary, enabling dkms.service is surely the best choice.

andreyv commented on 2013-09-01 16:06

@twelveeighty: I think a second reboot isn't even necessary, because DKMS seems to modprobe the driver after compiling it. All I do when a new kernel arrives is reboot once, and everything magically works.

SMOG commented on 2013-09-01 15:58

So, if I get it right, I do a kernel update and then dkms (even if running as service) is not going to recompile it until I first reboot once and then I have to reboot a second time for the driver to be enabled?

Two reboots seems to me worse then what I'm already doing, that is to keep a local copy of this package and after a kernel update give:

yaourt -R broadcom-wl-dkms; yaourt -U ~/broadcom-wl-dkms-*-x86_64.pkg.tar.xz

It takes a few minutes either, but at least I did not have to reboot again! :)

twelveeighty commented on 2013-09-01 15:50

Not sure why you would have to do that. In my case, all I have to do after a kernel update is:
* Reboot (obviously)
* As long as you have enabled dkms.service, dkms will automatically recompile and reinstall the driver on the new kernel. This can take a couple of seconds / minutes, depending on your hardware and how busy your startup is
* Reboot again for the driver to be enabled

terlar commented on 2013-09-01 08:55

Yes, you have to run the command mrueegg wrote.

If you had it installed before you can run this to figure out which driver version:
dkms status -m broadcom-wl

And this to see which kernelversion you have:
pacman -Q linux

And then install it with:
dkms install -m broadcom-wl -v <driverversion> -k <kernelversion>

SMOG commented on 2013-08-31 22:05

Just to be sure... do you also have to reinstall this driver after every kernel update or am I missing something?

Thank you

mrueegg commented on 2013-08-25 21:27

The driver module was not properly installed by this package. After a manual installation with

dkms install -m broadcom-wl -v <driverversion> -k <kernelversion>

I'm now able to use my Broadcom wifi.

twelveeighty commented on 2013-08-25 17:05

@mrueegg: does 'dkms status -m broadcom-wl' report that the driver is installed on the current kernel?

mrueegg commented on 2013-08-23 13:55

Hi. I have problems getting my wifi (broadcom 14e4:43a0 / MacBook Air) to work with this package. After the installation, I was not able to see the interface with "ip link". "modprobe wl" yields "FATAL: Module wl not found". I also tried to enable the systemd service DKMS as was suggested in one of the comments, but this did not help either.

Any suggestions would be highly appreciated. Thannks.

andreyv commented on 2013-08-17 07:18

@frigaut: This is just how DKMS is designed to work: the idea is that you install a package once and it works for all kernels without reinstallation. The "unowned" files, which are generated for every new kernel, are a small price to pay for this. But note that this package's "install" script removes all installed versions when the package is uninstalled. You can also clean them manually with the dkms command from time to time.

frigaut commented on 2013-08-17 04:05

@andreyv: thanks. Turns out after fighting with it a bit more, I got it to work. Yes, I ended up understanding that dkms builds it, and that the build is not done before the package installation phase, as the bunch of broadcom files in /var/lib/dkms/broadcom-wl does not belong to the package:

513:~ $ pacman -Qo /var/lib/dkms/broadcom-wl/
error: No package owns /var/lib/dkms/broadcom-wl

which is somehow a problem, no? I mean when you get rid of the package, it doesn't get rid of these files and neither does it get rid of /usr/lib/modules/3.10.5-1-ARCH/kernel/drivers/net/wireless/wl.ko, which is the module built and installed by broadcom-wl-dkms (indirectly thru dkms). Isn't that a problem? Or may be I'm missiing something.

Also, I didn't know about enabling dkms.service, and it was not enabled on my system. May be the problem outlined above is actually a consequence of this. Anyway, it works now, and I know a bit more today that yesterday (although obviously I'm far from understanding everything dkms :-)

frigaut commented on 2013-08-16 23:04

@andreyv: thanks. Turns out after fighting with it a bit more, I got it to work. Yes, I ended up understanding that dkms builds it, and that the build is not done before the package installation phase, as the bunch of broadcom files in /var/lib/dkms/broadcom-wl does not belong to the package:

513:~ $ pacman -Qo /var/lib/dkms/broadcom-wl/
error: No package owns /var/lib/dkms/broadcom-wl

which is somehow a problem, no? I mean when you get rid of the package, it doesn't get rid of these files and neither does it get rid of /usr/lib/modules/3.10.5-1-ARCH/kernel/drivers/net/wireless/wl.ko, which is the module built and installed by broadcom-wl-dkms (indirectly thru dkms). Isn't that a problem? Or may be I'm missiing something.

Also, I didn't know about enabling dkms.service, and it was not enabled on my system. May be the problem outlined above is actually a consequence of this. Anyway, it works now, and I know a bit more today that yesterday (although obviously I'm far from understanding everything dkms :-)

andreyv commented on 2013-08-16 12:59

@frigaut: The package only installs the module sources, but DKMS builds them. Just do "sudo systemctl enable dkms.service" and reboot, then it should take care of itself. You can also compile the module manually, see "man dkms".

andreyv commented on 2013-08-16 12:58

@frigaut: The package only install the module sources, but DKMS builds them. Just do "sudo systemctl enable dkms.service" and reboot, then it should take care of itself. You can also compile the module manually, see "man dkms".

frigaut commented on 2013-08-16 12:55

ok, I give up and have to ask you guys.
the package build fine, and install fine, but then I have no wl module:

$ sudo modprobe wl
modprobe: FATAL: Module wl not found.

This happened to me already repeatedly a couple of weeks ago, but somehow it ended working (meaning I ended up having a wl.ko on my file system). now, nada.
It is not even listed in the owned file from this package:

533:~ $ pacman -Ql broadcom-wl-dkms | grep ko
534:~ $

I'm totally at loss here. Not using any exotic config. well, I downgraded the kernel, but -as I said- this happened to me a short while ago in a non-downgraded config, so I tend to believe the downgrade doesn't have anything to do with it.
Any hints?

Just in case:
538:~ $ pacman -Qs linux | grep local | grep linux
local/archlinux-keyring 20130525-2
local/archlinux-themes-slim 1.2-3
local/cpupower 3.10-2 (linux-tools)
local/lib32-util-linux 2.23.2-1
local/linux 3.10.5-1 (base)
local/linux-api-headers 3.8.4-1
local/linux-docs 3.10.5-1
local/linux-firmware 20130725-1
local/linux-headers 3.10.5-1
local/util-linux 2.23.2-1 (base base-devel)
539:~ $

alejandro.perez commented on 2013-08-14 19:36

Is anyone else having reception/power problems? I have 3 more laptops at home, and this one (the only one with this card) is getting 5Mbit/s in places where others have 54Mbit or even 72Mbit.

alejandro.perez commented on 2013-08-14 19:35

Anyone else is having reception problems. I have 3 more laptops at home, and this one (the only one with this card) is getting 5Mbit/s in places where others have 54Mbit or even 72Mbit.

Anonymous comment on 2013-08-06 12:55

Works like a charm! Thank you twelveeighty!

twelveeighty commented on 2013-07-26 21:29

If we're all convinced that this package will install and build without linux-headers, I'm fine with removing it. I thought this was a requirement as well, as @andreyv says too.

felixonmars commented on 2013-07-26 14:55

@andreyv: However, other dkms packages (nvidia-dkms and virtualbox-host-dkms that I use) and the dkms package itself only optdepends on linux-headers, so I think using optdepends for this could be more appropriate.

andreyv commented on 2013-07-26 14:42

@sudaraka: But this package *is* dependent on Linux headers, isn't it? If your custom kernel package contains the headers, then it should "provide" the linux-headers package.

sudaraka commented on 2013-07-26 13:37

I see this has been brought up before, removing the dependancy on "linux-headers". Just wondering is it will be done? bcs, I'm using a custom kernel and there's no point in installing that package for me.

twelveeighty commented on 2013-07-22 21:05

@WonderWoofy: thanks!

WonderWoofy commented on 2013-07-22 02:08

Just want to let you know that this so far builds fine against 3.11 (rc2).

twelveeighty commented on 2013-07-09 20:40

Good point, @taifunbrowser, I'll add that as a conflict next time I edit the PKGBUILD.

taifunbrowser commented on 2013-07-09 06:52

Great, worked for me with a BCM 43228 wireless card, when broadcom-wl stopped working after an upgrade.

I had to pacman -R broadcom-wl before I could successfully install this. Can we mark that package as a conflict?

andreyv commented on 2013-07-08 16:47

Great, now this package has a decent name and the latest driver version :)

twelveeighty commented on 2013-07-08 15:41

We should be all moved over to this new package (broadcom-wl-dkms) now. There was a bit of a mix-up with the merge, which was fixed today, but everything should be in place now with this package.

triplc commented on 2013-07-06 01:56

Could you please remove the "linux-headers" dependency (as it was in the old "dkms-broadcom-wl" package)? With "linux-pae-headers" installed, it compiled okay with "linux-headers" dependency removed.

twelveeighty commented on 2013-07-05 21:29

@WonderWoofy: had not noticed power management issue before. I have "Power Management:off" listed under iwconfig. Maybe you just need to turn it on with iwconfig <card> power <settings> (see man ifconfig)?

yochai commented on 2013-07-05 20:18

After using th driver (the old AUR package) for some time with no issue, I'm now facing the following problem.

I connect to the wireless network (which work for other computers) but I have no access out. I can't ping anything on the netowrk or outside of it.
I found the following error:
Jul 5 22:29:41 tartop kernel: [ 442.341085] WARNING: at net/wireless/sme.c:645 cfg80211_roamed+0x8c/0x90 [cfg80211]()
Jul 5 22:29:41 tartop kernel: [ 442.341088] Hardware name: Vostro 3460
Jul 5 22:29:41 tartop kernel: [ 442.341090] Modules linked in: lib80211_crypt_tkip wl(O) lib80211 cfg80211 fuse michael_mic arc4 ecb bbswitch(O) btusb snd_hda_codec_hdmi uvcvideo videobuf2_vmalloc bluetooth snd_hda_codec_conexant intel_powerclamp videobuf2_memops videobuf2_core coretemp videodev kvm_intel rts5139(C) media kvm crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel aes_x86_64 iTCO_wdt dell_wmi sparse_keymap iTCO_vendor_support xts snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_page_alloc snd_timer snd dell_laptop mei alx(O) i915 rfkill intel_agp intel_gtt lrw gf128mul ablk_helper drm_kms_helper cryptd drm dcdbas i2c_algo_bit compat(O) i2c_i801 soundcore mdio lpc_ich pcspkr i2c_core microcode wmi thermal video battery acpi_cpufreq mperf processor button ac psmouse(O) serio_raw evdev ext4 crc16 mbcache jbd2 sr_mod sd_mod cdrom ahci libahci ehci_pci ehci_hcd xhci_hcd libata scsi_mod usbcore usb_common [last unloaded: lib80211]
Jul 5 22:29:41 tartop kernel: [ 442.341180] Pid: 2107, comm: wl_event_handle Tainted: G WC O 3.9.9-1-ARCH #1
Jul 5 22:29:41 tartop kernel: [ 442.341183] Call Trace:
Jul 5 22:29:41 tartop kernel: [ 442.341193] [<ffffffff81057f60>] warn_slowpath_common+0x70/0xa0
Jul 5 22:29:41 tartop kernel: [ 442.341197] [<ffffffff8105804a>] warn_slowpath_null+0x1a/0x20
Jul 5 22:29:41 tartop kernel: [ 442.341206] [<ffffffffa0db190c>] cfg80211_roamed+0x8c/0x90 [cfg80211]
Jul 5 22:29:41 tartop kernel: [ 442.341227] [<ffffffffa0f98d9c>] wl_notify_roaming_status+0xac/0x130 [wl]
Jul 5 22:29:41 tartop kernel: [ 442.341242] [<ffffffffa0f95c05>] wl_event_handler+0x55/0x1f0 [wl]
Jul 5 22:29:41 tartop kernel: [ 442.341256] [<ffffffffa0f95bb0>] ? wl_deinit_priv_mem+0xa0/0xa0 [wl]
Jul 5 22:29:41 tartop kernel: [ 442.341261] [<ffffffff8107b580>] kthread+0xc0/0xd0
Jul 5 22:29:41 tartop kernel: [ 442.341266] [<ffffffff8107b4c0>] ? kthread_create_on_node+0x120/0x120
Jul 5 22:29:41 tartop kernel: [ 442.341272] [<ffffffff814da7ec>] ret_from_fork+0x7c/0xb0
Jul 5 22:29:41 tartop kernel: [ 442.341276] [<ffffffff8107b4c0>] ? kthread_create_on_node+0x120/0x120
Jul 5 22:29:41 tartop kernel: [ 442.341279] ---[ end trace 7f42c3e699175611 ]---
Jul 5 22:29:42 tartop NetworkManager[274]: <warn> Connection disconnected (reason -3)

Anyone knows how to fix it ?

WonderWoofy commented on 2013-07-05 18:22

This works just as before for my BCM43228. Though I feel like I should mention that although it seems to perform just fine, everything that is typically below the "Power Management" line of iwconfig is missing. When I use the 5.100.* version, that is all there as usual.

twelveeighty commented on 2013-07-05 17:13

Switched from broadcom-wl6-dkms, this package will replace wireless-bcm43142-dkms

twelveeighty commented on 2013-07-05 16:52

@timothy.lee: be sure this v6 driver works for your hardware; there are known cases where the v6 driver does not work and you must use the v5 driver. Otherwise we'll still need the old v5 as a separate package.

If everyone is okay with it, I'll do the following:

1. grab the broadcom-wl-dkms for ownership and 'replace' it with the current broadcom-wl6-dkms PKGBUILD

2. ask the mailing list to merge wireless-bcm43142-dkms with the new broadcom-wl-dkms

twelveeighty commented on 2013-07-05 16:52

@timothy.lee: be sure this v6 driver works for your hardware; there are known cases where the v6 driver does not work and you must use the v5 driver. Otherwise we'll still need the old v5 as a separate package.

If everyone is okay with it, I'll do the following:

1. grab the broadcom-wl-dkms for ownership and 'replace' it with the current broadcom-wl6-dkms PKGBUILD

2. ask the mailing list to merge wireless-bcm43142-dkms with the new broadcom-wl-dkms

timothy.lee commented on 2013-07-05 01:52

I've disowned the broadcom-wl-dkms package. Feel free to own it and merge with that package.

timothy.lee commented on 2013-07-04 15:52

Disowning package. Switched to broadcom-wl6-dkms.

timothy.lee commented on 2013-07-04 15:51

Disowning package. Switching to broadcom-wl6-dkms.

yochai commented on 2013-06-30 05:34

AFAIK there is really no need to reboot, just install the new package. As the packages conflicts the old one will be removed by pacman. That what I did here and it seems to cause no issue.

twelveeighty commented on 2013-06-29 17:23

Folks, I have created the new package broadcom-wl6-dkms: https://aur.archlinux.org/packages/broadcom-wl6-dkms/ . This package also includes the latest upstream version for 3.10 compatibility. I have not tested this against a 3.10 kernel, but the driver runs fine on the current kernel.

Please test it and I'll ask for a merge sometime next week.

To use the new version, I recommend first uninstalling the old wireless-bcm43142-dkms driver with pacman -R wireless-bcm43142-dkms. Then reboot. Then install the new broadcom-wl6-dkms driver. Reboot again.

twelveeighty commented on 2013-06-29 17:06

This is the renamed package for https://aur.archlinux.org/packages/wireless-bcm43142-dkms/. A merge will be performed soon. This package also has the latest upstream version and some suggested changes to use $pkgver instead of hard coded numbers.

timothy.lee commented on 2013-06-28 06:49

Renamed from dkms-broadcom-wl

timothy.lee commented on 2013-06-28 06:35

The package has now been renamed broadcom-wl-dkms in accordance with module naming convention. Please proceed to https://aur.archlinux.org/packages/broadcom-wl-dkms

ryley commented on 2013-06-19 11:50

New release available at Launchpad: https://launchpad.net/ubuntu/+source/bcmwl . I haven't tried it yet, the main change seems to be support for kernel 3.10.

ryley commented on 2013-06-16 16:51

In terms of determing which version supports which wireless cards, version 5 of the driver has a table of PCI ID's in wl_linux.c (http://pastebin.com/raw.php?i=YgvQyY8h) . These ID's correspond to the following cards (from the B43 page):

BCM4311
BCM4312
BCM4318
BCM4321
BCM4322
BCM43224
BCM43241?
BCM43225
BCM4313
BCM43227
BCM43228

I though from this I could see which cards version 6 dropped support for. Unfortuanately, the equivalent PCI ID table in the version 6 driver is http://pastebin.com/raw.php?i=2pPJmjUa . (It's basically an empty table). I can't see a table with the equivalent info from v5 anywhere in the v6 source. So I guess the only way to know what cards this supports is gonna be trial and error.

BTW: This driver has heaps of references to 80MHZ channels and 802.11ac, so this might be the best driver for the new Haswell Macbook airs!

Krejzi commented on 2013-06-16 11:10

As the driver more correct name is Broadcom STA, I maintain broadcom-sta-dkms which differs from this one in interface naming (ethx vs wlanx in -sta). And guess what, I've created my package just because I couldn't find any dkms one since this one was weird named.

andreyv commented on 2013-06-15 20:51

I don't think there is a strict need to rename other packages: just name this one as, say, broadcom-wl-v6-dkms. This will be enough for people to notice a difference and compare versions of both packages.

twelveeighty commented on 2013-06-15 18:03

Folks, I maintain the v6 branch of this driver on the AUR https://aur.archlinux.org/packages/wireless-bcm43142-dkms/). I would like to rename that package to be more in line with the naming standards used in the repos, which would be broadcom-wl-dkms. That would be very confusing with this package (dkms-broadcom-wl). Would you be okay to rename your package to broadcom-wl5-dkms and we'll take broadcom-wl-dkms? We need both packages, because some hardware only works on the v5 version and some hardware only works on the v6 version.

twelveeighty commented on 2013-06-15 17:39

Good points, @tchebb. I'll post a comment on the dkms-broadcom-wl if they are interested in renaming to broadcom-wl5-dkms. If not, we'll end up with dkms-broadcom-wl for the v5 driver and broadcom-wl-dkms for v6 and I'm sure nobody likes that confusion.

tchebb commented on 2013-06-15 16:12

It seems to me that the best naming convention would be one which mirrors that used in the official repos: broadcom-wl for this package and broadcom-wl5 for v5. Appropriate prefixes and suffixes (such as dkms) can then be added to those base package names as needed. I dislike the idea of naming either version after a single chipset it supports, since that leads to confusion for those who have other chipsets (for example, I'm using this package for my BCM4331).

According to the guidelines at https://wiki.archlinux.org/index.php/Dynamic_Kernel_Module_Support#Guidelines and dkms packages in the official repos, dkms should always be added as a suffix, so, ideally, this package should be renamed broadcom-wl-dkms and dkms-broadcom-wl should be renamed broadcom-wl5-dkms.

That's my $0.02. Feel free to use or ignore it as you please.

twelveeighty commented on 2013-06-15 15:25

All - here's an update on the package renaming: read the upstream bug reports: https://bugs.launchpad.net/ubuntu/quantal/+source/bcmwl/+bug/923809/comments/36 and further down. As I suspected, and this is reported here on Arch AUR as well, the v5 and v6 versions of this driver should really be seen as separate drivers, since there are many reports of older cards that only work with v5 and we know for certain that there are newer cards that only work with v6. Therefore a merge is out of the question.

I feel I need to keep a clearer division between the already existing AUR dkms-broadcom-wl (v5) and the "broadcom-wl-dkms" name we're suggesting for this v6 driver. The real, big, issue is that nobody has been able to define exactly which cards are compatible with v5 and which are v6. The only thing I know for certain is that the bcm43142 works only with v6 and not with v5, so in light of that, I think the best possible name would be this monster: broadcom-wl-bcm43142-dkms, because it identifies at least one card we know for certain that needs this driver over the others but at least it will be listed alongside the other broadcom drivers when sorted alphabetical.

Thoughts?

cassfalg commented on 2013-06-12 08:44

I've had to make and make install the driver from /usr/src manually - is that intended? Shouldn't that be done automatically?

tchebb commented on 2013-06-11 06:18

I'm definitely in favor of a rename. As long as there are no major regressions from 5.100.82.112, I don't see any reason why this version of the driver shouldn't entirely replace the older one in the existing broadcom-wl-* packages.

There is already a package named dkms-broadcom-wl, so perhaps that one should be updated and this one should be removed altogether. To be consistent with the official repositories, it ought to be named "broadcom-wl-dkms", but I don't think a rename purely for consistency is worth it.

twelveeighty commented on 2013-06-10 14:30

OK - while I'm changing this, let's also standardize the package description, etc. Any suggestions? I don't think there is a list somewhere which cards this driver supports specifically. Does anyone know?

andreyv commented on 2013-06-09 17:06

I think changing the name to a more general one is good, because this driver seems to work with other wireless cards too. It seems that Ubuntu uses this driver for all Broadcom cards. But if this driver is newer and always works, what should happen to the broadcom-wl and broadcom-wl-dkms packages which provide an older version?

twelveeighty commented on 2013-06-09 16:34

@OboeNerd: I'm okay with a rename if that's the consensus.

As for the licensing question: that's something you need to decide for yourself; right now, the driver will NOT build AT ALL if the license is changed to MIXED. There is supposed to be an upstream kernel change coming that allows us to use the non-GPL calls, but so far nothing yet. Legal or not, we need to run this driver.

I can change the filepath the next time we need to update it.

OboeNerd commented on 2013-06-09 00:20

Hi all, I have a number of concerns that I believe should be addressed collectively:

First of all, I am personally using this PKGBUILD because my BCM4331 card is only supported by the newer driver packages provided by Canonical. However, I only found it by accident, as this was the only bcmwl PKGBUILD of its kind on the AUR. But why is it named after a specific card? Should this package be renamed to something like broadcom-wl-dkms to more accurately describe its contents? Also, should this updated Canonical version 6.30.223.30 replace broadcom-wl on the AUR entirely?

Also, the licensing: I understand that in order to get the driver to build cleanly we needed to change the license in the source from MIXED/Proprietary to GPL. Where does this stand legally? Is there a need to add a disclaimer, something like "You may not redistribute this build"?

Lastly, a very minor and mostly cosmetic suggestion: should the PKGBUILD use file paths such as ./bcmwl-${pkgver}+bdcom instead of ./bcmwl-6.30.223.30+bdcom ? Would this be easier for individuals to maintain?

I'd like to hear what you think.

silvik commented on 2013-06-02 21:24

like ottoshmidt, after each of the latest updates the module failed to build and after rebooting I had to uninstall and reinstall the package, even if dkms service was OK. Maybe the module fails to build when in use..?

silvik commented on 2013-06-02 21:24

like ottoshmidt, after each of the latest updates the module failed to build and after reinstalling I had to uninstall and reinstall the package, even if dkms service was OK. Maybe the module fails to build when in use..?

ottoshmidt commented on 2013-06-02 16:44

I have this installed, but after each linux update I need to reinstall this anyway ((

timothy.lee commented on 2013-06-02 15:34

Added uname-r.patch from lknix to support compilation against non-current kernel (rev. 13)

timothy.lee commented on 2013-06-02 15:33

Added uname-r.patch from lknix to support compilation against non-current kernel

lknix commented on 2013-05-30 12:45

Forcing a build for a specific kernel version is not possible at the moment (look at alex-eri's comment).

This fails if you use any other kernel version than `uname -r`: dkms install -m broadcom-wl -v 5.100.82.112 -k <KERNEL_VER>

It fails due to hardcoded `uname -r` in:
/usr/src/broadcom-wl-5.100.82.112/Makefile

And a patch to fix this:

--- Makefile 2011-10-22 18:56:55.000000000 +0200
+++ Makefile2 2013-05-29 12:16:22.750874568 +0200
@@ -108,6 +108,10 @@
$(info Using Wireless Extension API)
endif

+ifeq ($(KERNELRELEASE),)
+ KERNELRELEASE := `uname -r`
+endif
+
obj-m += wl.o

wl-objs :=
@@ -123,10 +127,10 @@
EXTRA_LDFLAGS := $(src)/lib/wlc_hybrid.o_shipped

all:
- KBUILD_NOPEDANTIC=1 make -C /lib/modules/`uname -r`/build M=`pwd`
+ KBUILD_NOPEDANTIC=1 make -C /lib/modules/$(KERNELRELEASE)/build M=`pwd`

clean:
- KBUILD_NOPEDANTIC=1 make -C /lib/modules/`uname -r`/build M=`pwd` clean
+ KBUILD_NOPEDANTIC=1 make -C /lib/modules/$(KERNELRELEASE)/build M=`pwd` clean

install:
- install -D -m 755 wl.ko /lib/modules/`uname -r`/kernel/drivers/net/wireless/wl.ko
+ install -D -m 755 wl.ko /lib/modules/$(KERNELRELEASE)/kernel/drivers/net/wireless/wl.ko

Zombifier commented on 2013-05-26 00:58

@twelveeighty: Okay, I've just switched back to the unpatched version and strangely the error message is gone. I have no idea what solved the issue.

@ryley: Maybe you can try the ioctl patch I've created and see if it works for you.

hoerbert commented on 2013-05-21 19:59

Hi twelveeighty,

just my experiences:

[Bus 004 Device 003: ID 0a5c:21d7 Broadcom Corp. BCM43142 Bluetooth 4.0]

No debug-messages for me, everything seems to work just fine.

andreyv commented on 2013-05-21 07:06

Hi,

I don't get such messages here on BCM4312.

ryley commented on 2013-05-21 05:50

I've been getting the same message in dmesg as Zombifier. However, my WiFi card is a Broadcom 43224 (I use this package as I get higher sustained transfer speeds than I do with broadcom-wl).

twelveeighty commented on 2013-05-21 01:39

@Zombifier: I have a strong suspicion that something else is going on with your setup: I have the exact same hardware [Network controller: Broadcom Corporation BCM43142 802.11b/g/n (rev 01)]; [Bus 004 Device 003: ID 0a5c:21d7 Broadcom Corp. BCM43142 Bluetooth 4.0].

However, I have no [@wl_cfg80211_scan : WLC_SCAN error (-22)] whatsoever...

Anyone else see this particular problem that Zombifier has?

Zombifier commented on 2013-05-20 19:17

@twelveeighty: In dmesg, TTYs and even SSH PTS. I've just upgraded to 3.9 series kernel and the messages are still there.

Here is part of the dmesg output and details about the card:
http://pastebin.com/J4QGgaVq

twelveeighty commented on 2013-05-20 04:20

@Zombifier - what log are you seeing this in? /var/log/messages.log, or dmesg? I see no debug or error messages there with my version of 6.30.223.30-1...?

Zombifier commented on 2013-05-19 20:13

@twelveeighty: Do you have such debug messages on your system?

Zombifier commented on 2013-05-19 20:12

@twelveeighty: Yes, those debug messages are still showing on my machine. I'm not sure if it is card specific, as I only have one PC with bcm43142 (it would be out of mind to buy another Broadcom card :)).
Since the patch does not change the behaviour other than debug message, no harm would be done anyway. I'll run the patched driver for several days and will let you know if any issue comes out.

twelveeighty commented on 2013-05-19 15:24

@Zombifier: I have no real issue against using your patch, but be aware that this is a patched function that wasn't patched previously. In the old version, we had wl_cfg80211_get_station() patched and this has been changed to use WL_DBG now. Your patch is changing wl_cfg80211_scan(); are you sure this is really something we need to patch out, or is there something specific to your card/network happening?

twelveeighty commented on 2013-05-19 15:10

The reason I didn't re-patch ioctl is that the new code now uses WL_DBG instead of WL_ERR. Are those debug messages still showing?

Zombifier commented on 2013-05-19 01:08

@twelveeighty: I've created a patch which works on my machine.
http://pastebin.com/gKhK3mAm

Zombifier commented on 2013-05-19 00:55

@twelveeighty: Seems you dropped ioctl patch and all the scan errors are filling the syslog again, at least on 3.8 series kernel. Can you have a look on this issue? Thanks

Error message:
ERROR @wl_cfg80211_scan : WLC_SCAN error (-22)

twelveeighty commented on 2013-05-18 17:08

OK - I've packaged the new upstream version: 6.30.223.30-1; I've been running it briefly on my own laptop - no major issues so far.

twelveeighty commented on 2013-05-17 21:46

Interestingly enough, the 6.30 version upstream has the exact same issue with rcu_read_unlock() and GPL licensing. I really wish someone had a straight answer on this, because now that call is in the upstream code of the driver (line 2049 in wl/sys/wl_cfg80211_hybrid.c). And up to the current kernel 3.9.2, this errors out as "GPL-incompatible module wl.ko uses GPL-only symbol" during modpost, same as before.

twelveeighty commented on 2013-05-17 18:53

@ryley - I'll have a look at the new version. It appears Canonical patched it all the way up to 3.9, so it may work here too.

aruniiird commented on 2013-05-17 11:28

Updated to 3.9.2 kernel... Wifi is working perfectly fine... @twelveeighty and others, thanks a lot guys...

ryley commented on 2013-05-17 03:39

There's a new release (6.30.223.30) available at https://launchpad.net/ubuntu/+source/bcmwl . Seems to be working fine here, can't see any massive performance differences or instability.

andreyv commented on 2013-05-15 08:25

It works now as well.

timothy.lee commented on 2013-05-15 06:33

Updated user-ioctl.patch from broadcom-wl (rev. 12) to fix compilation against linux-3.9

twelveeighty commented on 2013-05-15 01:13

Sorry folks, I realized that my previous fix would break kernels < 3.9, so I had to change that for rel 8 now.

Can someone re-test this rel 8 against kernel 3.9?

andreyv commented on 2013-05-14 07:54

Yes, it seems to work now. Thank you.

twelveeighty commented on 2013-05-14 01:59

With the tip from @tjbp, I created a patch for the change in the cfg80211_put_bss() function for kernel 3.9.

I have NOT tested this build, but it does compile without errors on my 3.9 VM. Can someone please test this build 7 and post back with the results?

twelveeighty commented on 2013-05-13 23:45

Looking into the 3.9 kernel stuff now. Stay tuned.

tjbp commented on 2013-05-13 13:30

cfg80211_put_bss() takes two arguments in 3.9, so the current release of this code will fail to compile.

aruniiird commented on 2013-05-13 06:17

Kernel- 3.9.2-1 has come as the arch update.
@twelveeighty: Is this kernel supported for our driver (6.20.155.1-6) version?

Anonymous comment on 2013-05-08 05:35

@RozenTensai

Linux-headers is optional in DKMS's dependencies, so you should add it in this package.

For instance, this package didn't work in my computer until I installed linux-headers manually.

Anonymous comment on 2013-04-26 19:55

Disowning, can't update often and now using b43 firmware.

Ah, linux-headers isn't needed because DKMS already has it in its dependencies.

twelveeighty commented on 2013-04-06 03:34

@n4p4: I really have no idea on how to build packages for Debian. You may be better off going to the upstream package: https://launchpad.net/ubuntu/+source/bcmwl and go from there.

Anonymous comment on 2013-04-04 16:44

Hi !

I try to run the shell cmd from your PKGBUILD and nothing is working I got no *wl* working in the lsmod.

I can't execute the PKGBUILD because I'm running on Debian with the 3.7-trunk-amd64 kernel. I correctly add the dkms and linux-headers.

So, I firstly run the cmd from the build function, then the package function and finaly I run the make. No errors returns but I can't ./configure (No such file) and the make install cmd tell that make 'install' is up to date.

Maybe I am not running the right cmd to build the package

# mkdir -p /usr/src/broadcom-wl-6.20.155.1
# cp -RL * /usr/src/broadcom-wl-6.20.155.1
# install -D -m 644 lib/LICENSE.txt "/usr/share/licenses/broadcom-wl/LICENSE"
# install -D -m 644 wireless-bcm43142-dkms.conf "/etc/modprobe.d/wireless-bcm43142-dkms.conf"

Thanks a lot for this package :-)

twelveeighty commented on 2013-04-03 02:52

Excellent! The last thing I need after the ordeal of going from kernel 3.7 to 3.8 is that minor kernel releases start to break this darn driver. Let's hope we're good for a while until 3.9.

andre.vmatos commented on 2013-04-03 01:58

Oh, sorry! A lost file (maybe from other package) was in extramodules, that seems to have priority over kernel/drivers/net/wireless. I removed this package, removed the orphan file, reinstalled this package, and everything is working fine. Sorry the mistake.

andre.vmatos commented on 2013-04-03 01:54

# Compilation output, only build, stripped makepkg common output:
==> Starting build()...
patching file debian/bcmwl-kernel-source.postrm
patching file debian/bcmwl-kernel-source.install.in
patching file debian/compat
patching file debian/control
patching file debian/bcmwl-kernel-source.postinst
patching file debian/bcmwl-kernel-source.prerm
patching file debian/copyright
patching file debian/bcmwl-kernel-source.modaliases
patching file debian/dkms.conf.in
patching file debian/changelog
patching file debian/rules
patching file debian/patches/0007-nl80211-move-scan-API-to-wdev.patch
patching file debian/patches/0006-add-support-for-linux-3.8.0.patch
patching file debian/patches/0001-MODULE_LICENSE.patch
patching file debian/patches/0002-Makefile.patch
patching file debian/patches/0005-add-support-for-linux-3.4.0.patch
patching file debian/patches/0003-Make-up-for-missing-init_MUTEX.patch
patching file debian/patches/0004-Add-support-for-Linux-3.2.patch
patching file Makefile
patching file src/wl/sys/wl_linux.c
Hunk #1 succeeded at 95 with fuzz 2 (offset -4 lines).
patching file src/wl/sys/wl_linux.c
Hunk #1 succeeded at 338 with fuzz 1 (offset -47 lines).
patching file src/wl/sys/wl_linux.c
Hunk #1 succeeded at 47 (offset 7 lines).
patching file src/include/bcmutils.h
patching file src/wl/sys/wl_cfg80211.c
patching file src/wl/sys/wl_cfg80211.c
patching file bcmwl-6.20.155.1+bdcom/Makefile
patching file src/wl/sys/wl_linux.c
patching file src/wl/sys/wl_cfg80211.c
patching file src/wl/sys/wl_linux.c
patching file src/wl/sys/wl_cfg80211.c
patching file src/wl/sys/wl_linux.c

# Installation output with pacman -U:
(1/1) instalando wireless-bcm43142-dkms [############################################################] 100%
>>> DKMS: Module add, build, and install

Creating symlink /var/lib/dkms/broadcom-wl/6.20.155.1/source ->
/usr/src/broadcom-wl-6.20.155.1

DKMS: add completed.

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.8.5-1-ARCH -C /usr/lib/modules/3.8.5-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.20.155.1/build.......
cleaning build area....

DKMS: build completed.

wl:
Running module version sanity check.
- Original module
- No original module exists within this kernel
- Installation
- Installing to /usr/lib/modules/3.8.5-1-ARCH/kernel/drivers/net/wireless/

depmod......

DKMS: install completed.
modprobe: ERROR: could not insert 'wl': Invalid argument
depmod: ERROR: failed to load symbols from /lib/modules/3.8.5-1-ARCH/extramodules/wl.ko.gz: Invalid argument
==> For more information on this driver, see:
==> https://bbs.archlinux.org/viewtopic.php?id=145884

twelveeighty commented on 2013-04-03 01:47

@andre.vmatos: I just tried installing this version on a VM with 3.8.5-1 from Testing repo and there's no errors or warnings regarding depmod or modprobe. Now, this VM obviously doesn't have the wireless card installed, but I'm curious to see your logs.

twelveeighty commented on 2013-04-03 01:17

@andre.vmatos: can you post the full output from compilation and installation?

andre.vmatos commented on 2013-04-01 19:43

I'm getting Invalid symbol in depmod and modprobe with kernel 3.8.5, and broadcom-wl is panicking. Any hints?

Casper commented on 2013-03-27 17:41

@twelveeighty, it works great and I don't see errors in logs now. Good work! Thanks again!

$ uname -r
3.8.4-1-ARCH

twelveeighty commented on 2013-03-27 04:51

@aruniiird: with this driver, be sure to select to be notified for updates, comments, etc. on the AUR. You are about 4 builds behind, we're on 6.20.155.1-6, your log shows 6.20.155.1-2. Uninstall your current version, reboot and then install this new version. This driver is very unstable with regards to kernel updates, etc. so always check the notes and comments here before you update your kernel or the driver.

aruniiird commented on 2013-03-27 04:04

@twelveeighty, thanks it work great now... Thanks a lot...

aphirst commented on 2013-03-26 20:41

I think you need to add linux-headers to the dependencies.

Casper commented on 2013-03-26 19:22

@twelveeighty, it works great and I don't see errors in logs now. Good work! Thanks again!

$ uname -r
3.8.4-1-ARCH

aruniiird commented on 2013-03-26 08:24

With latest kernel 3.8.4-1, wireless-bcm43142-dkms driver fails, with the following error...
__________________________

Name Old Version New Version Net Change

wireless-bcm43142-dkms 6.20.155.1-2 7.18 MiB

Total Installed Size: 7.18 MiB

Proceed with installation? [Y/n]
(1/1) checking package integrity [###############################################] 100%
(1/1) loading package files [###############################################] 100%
(1/1) checking for file conflicts [###############################################] 100%
(1/1) checking available disk space [###############################################] 100%
(1/1) installing wireless-bcm43142-dkms [###############################################] 100%
>>> DKMS: Module add, build, and install

Creating symlink /var/lib/dkms/broadcom-wl/6.20.155.1/source ->
/usr/src/broadcom-wl-6.20.155.1

DKMS: add completed.

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.8.4-1-ARCH -C /usr/lib/modules/3.8.4-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.20.155.1/build....(bad exit status:
2)
Error! Bad return status for module build on kernel: 3.8.4-1-ARCH (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.20.155.1/build/make.log for more information.

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.8.4-1-ARCH -C /usr/lib/modules/3.8.4-1-ARCH/build M=/var/lib/dkms/broadcom-wl/6.20.155.1/build....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.8.4-1-ARCH (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.20.155.1/build/make.log for more information.
modprobe: FATAL: Module wl not found.
==> For more information on this driver, see:
==> https://bbs.archlinux.org/viewtopic.php?id=145884
__________________________

/var/lib/dkms/broadcom-wl/6.20.155.1/build/make.log showed the following error

__________________________

DKMS make.log for broadcom-wl-6.20.155.1 for kernel 3.8.4-1-ARCH (x86_64)
Tue Mar 26 13:02:47 IST 2013
make: Entering directory `/usr/src/linux-3.8.4-1-ARCH'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
Building for x86_64
LD /var/lib/dkms/broadcom-wl/6.20.155.1/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.155.1/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_cfg80211.o
/var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_cfg80211.c:1578:2: warning: initialization from incompatible pointer type [enabled by default]
/var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_cfg80211.c:1578:2: warning: (near initialization for ‘wl_cfg80211_ops.set_tx_power’) [enabled by default]
/var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_cfg80211.c:1579:2: warning: initialization from incompatible pointer type [enabled by default]
/var/lib/dkms/broadcom-wl/6.20.155.1/build/src/wl/sys/wl_cfg80211.c:1579:2: warning: (near initialization for ‘wl_cfg80211_ops.get_tx_power’) [enabled by default]
LD [M] /var/lib/dkms/broadcom-wl/6.20.155.1/build/wl.o
Building modules, stage 2.
CFG80211 API is prefered for this kernel version
Using CFG80211 API
Building for x86_64
MODPOST 1 modules
FATAL: modpost: GPL-incompatible module wl.ko uses GPL-only symbol '__rcu_read_unlock'
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2
make: Leaving directory `/usr/src/linux-3.8.4-1-ARCH'
__________________________

I am running archlinux on Dell-Vostro laptop, please let me know if there is anything I missed...

Thanks in advance,
Arun

andreyv commented on 2013-03-24 18:32

It's only a warning, though… it shouldn't cause kernel panics. BTW, @magus, though I've also got "Broadcom Corporation BCM4312 802.11b/g LP-PHY (rev 01)", it doesn't show up in lsusb, meaning it's a PCI device... Perhaps this is the difference with your system.

twelveeighty commented on 2013-03-24 18:29

@magus - this is probably the same bug as you're experiencing: https://bugs.launchpad.net/ubuntu/+source/bcmwl/+bug/1159306. I recommend following the issue there if that matches your problem.

twelveeighty commented on 2013-03-24 18:20

To @Zombifier, @Casper and all others that tested this driver - Thanks! Be aware that I will have to update this driver as soon as the issues with read_unlock_special and the GPL have been sorted out but this requires a patch to the kernel to get all the way from Canonical upstream and then back into Arch.

magus commented on 2013-03-24 17:21

Despite a report in the comments on 'broadcom-wl' that this package might work with my BCM4312 [14e4:4315] it doesn't.

I get an IP for a little while, then it goes away. I've not been able to see if the network is functional for that short period of time though. Turning off wireless (in NM) and then back on again causes a kernel panic.

Furthermore, reconnecting cases the following to appear in dmesg: http://pastebin.com/a528dZAV

Zombifier commented on 2013-03-24 14:21

It works great. Thank you!

twelveeighty commented on 2013-03-23 20:29

@Casper, @Zombifier, with regards to that system log error issue: I took the user-ioctl patch from broadcom-wl (assuming that was the one) and worked it into this driver as well. Can you please let me know if this solves that problem?

Casper commented on 2013-03-20 10:24

@SMOG
yes, need to uninstall previous version first

Anonymous comment on 2013-03-19 23:47

Updated to match build fixes done in package broadcom-wl (rev. 11)

SMOG commented on 2013-03-19 18:30

Just updating doesn't work here, after that wifi isn't working anymore.
I need to uninstall previous version and then install the package (updated) again.
Is anyone else noticing this behavior?

SMOG commented on 2013-03-19 18:11

On my system after installing kernel 3.8.3 and updating this package, wireless isn't working anymore...
"journalctl -f" doesn't give any additional info.

Casper commented on 2013-03-18 18:57

it works with 3.8.3-2-ARCH kernel, but in system logs every second:

[casper@casper-pc ~]$ journalctl -f
-- Logs begin at Сб 2013-03-16 04:03:16 MSK. --
мар 18 22:54:32 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rate (-1)
мар 18 22:54:32 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rssi (-1)
мар 18 22:54:32 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rate (-1)
мар 18 22:54:32 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rssi (-1)
мар 18 22:54:34 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rate (-1)
мар 18 22:54:34 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rssi (-1)
мар 18 22:54:36 casper-pc kernel: ERROR @wl_cfg80211_get_station : Could not get rate (-1)

and at boot time regarding bluetooth:
Bluetooth: hci0 command 0x1003 tx timeout

after this error bluetooth cannot find any devices.

Anyway many thanks for your job.

twelveeighty commented on 2013-03-18 17:36

@andreyv: see https://bbs.archlinux.org/viewtopic.php?id=145884; this is the reason we're all in this pickle: Broadcom has two different code bases, with this one (6.20) treated as a "special" thing for Dell, maintained by Canonical (as far as I can tell). It's a mess, but for us unfortunate ones with this 43142 card, this is the only driver that works. I've asked many times to be put in touch with the Broadcom people that "gave" Canonical the code, but no luck. I do have access to the developer at Canonical, though, I believe he's trying to help as much as he can.

andreyv commented on 2013-03-18 10:09

I just looked at this package and saw that it has version 6.20.155 (as Ubuntu does, too). I'm curious, where does this version come from? The only official Broadcom page I can find still lists 5.100.82.112: http://www.broadcom.com/support/802.11/linux_sta.php

twelveeighty commented on 2013-03-11 14:20

Another update: I may have to take down this version; Canonical doesn't seem to agree with my assessment that the posted Broadcom copyright/license notice publishes the src/ folder as GPL. If I take out the MODULE_LICENSE("GPL"), we're back to square 1 on kernel 3.8. I'll keep it up in until Canonical posts an alternative upstream link.

@Zombifier: I'll look at the syslog issue

Zombifier commented on 2013-03-11 10:00

Yes, works for me as well. Really appreciate your effort.
Can you have a look at the syslog problem as well? This broadcom-wl thing is flooding my syslog with:
[ 270.582032] ERROR @wl_cfg80211_get_station : Could not get rate (-1)
[ 270.582040] ERROR @wl_cfg80211_get_station : Could not get rssi (-1)
every several seconds.
There is a patch in broadcom-wl(AUR) which solves this issue.

ryley commented on 2013-03-11 09:49

Thanks for the update twelveeighty. No issues so far with the updated PKGBUILD, I no longer get rcu symbol errors and the panic whenever I would run "iwconfig" is gone. This is with a broadcom 43224 on linux-ck 3.8.2 .

twelveeighty commented on 2013-03-11 02:00

FOR 3.8 TESTING ONLY - communication with the folks from Canonical haven't resolved anything yet. Instead, to make sure we're only dealing with a remote, potential GPL issue, I've created this version for testing against the 3.8 kernel.

@Zombifier: does this one work on 3.8? It should compile/install fine, but I cannot test against the real hardware on 3.8, so please post back with actual network testing on the card.

Zombifier commented on 2013-03-08 21:47

No, same warning message.

twelveeighty commented on 2013-03-08 19:12

I've received an updated patch from Canonical - can you retest this version?

Zombifier commented on 2013-03-08 16:58

It does compile and system loads the driver, however it still does not work.
As seen in the dmesg output, there is a missing symbol in the code.

dmesg: http://pastebin.com/st6pyFf0
build log: http://pastebin.com/VRfpJRSt

twelveeighty commented on 2013-03-08 16:10

I have received a potential patch from Canonical for the kernel 3.8 / GPL licensing problem. I do not have a 3.8 kernel myself, so can someone test this and let me know if it works?

@Zombifier: can you test it, perhaps?

twelveeighty commented on 2013-03-08 04:21

Thanks for testing and posting. Kernel 3.8 is going to present a huge problem for this driver. Upstream (Canonical) had already patched for 3.8, but in doing so is using a GPL-only call. I've contacted the upstream maintainer, but I'm not sure how we can fix this.

Zombifier commented on 2013-03-04 00:17

dkms failed to build in the latest testing kernel(3.8.1-1-ARCH).
Make log: http://pastebin.com/dSNU3Xkn

twelveeighty commented on 2013-02-18 17:14

This version no longer uses dpkg to detect 32 or 64 bit; also has a "conflicts" clause with the old package.

twelveeighty commented on 2013-02-17 20:16

UPDATE: I have now created a newer version of this driver with a different package name in AUR called wireless-bcm43142-dkms (https://aur.archlinux.org/packages/wireless-bcm43142-dkms/). I recommend people switching to this newer version (test it first, of course). That version now has the following improvements:

0. No more use of the "oneiric" name in the package name, which was unfortunate to begin with, since that references a specific Ubuntu version

1. Upstream links to Ubuntu; which is where this dreaded driver's source is "maintained" with Dell; still no proper open-source version from Broadcom, but it's much more reliable

2. This version should be compatible with 2.x kernels up to at least 3.8

3. Both 32-bit and 64 bit versions should work now, but please be aware I have only tested the 64 bit version

4. This is now the latest available version (6.20.155.1)

YOU MUST UNINSTALL the old package first with pacman -R since this is a different package name and I'm sure the two will conflict.

As always, if you have a choice: DO NOT BUY BROADCOM HARDWARE, it's just losers like you and I that have to deal with these proprietary, closed drivers.

I will submit a request to the AUR group to have the package history merged

twelveeighty commented on 2013-02-17 20:09

PLEASE NOTE: this is the newer version of the original package in AUR called wireless-bcm43142-oneiric-dkms. This version now has the following improvements:

0. No more use of the "oneiric" name in the package name, which was unfortunate to begin with, since that references a specific Ubuntu version

1. Upstream links to Ubuntu; which is where this dreaded driver's source is "maintained" with Dell; still no proper open-source version from Broadcom, but it's much more reliable

2. This version should be compatible with 2.x kernels up to at least 3.8

3. Both 32-bit and 64 bit versions should work now, but please be aware I have only tested the 64 bit version

4. This is now the latest available version (6.20.155.1)

YOU MUST UNINSTALL the old package first with pacman -R since this is a different package name and I'm sure the two will conflict.

As always, if you have a choice: DO NOT BUY BROADCOM HARDWARE, it's just losers like you and I that have to deal with these proprietary, closed drivers.

alex-eri commented on 2013-02-16 12:22

if i not reboted in new kernel, i not able to install wl via dkms.
how to reconfigure dkms to build for not loaded kernel?

$ sudo dkms autoinstall -k 3.7.8-1-ARCH
[sudo] password for eri:

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.7.8-1-ARCH....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.7.8-1-ARCH (i686)
Consult /var/lib/dkms/broadcom-wl/5.100.82.112/build/make.log for more information.

$ cat /var/lib/dkms/broadcom-wl/5.100.82.112/build/make.log
DKMS make.log for broadcom-wl-5.100.82.112 for kernel 3.7.8-1-ARCH (i686)
Сб фев 16 16:16:26 MSK 2013
CFG80211 API is prefered for this kernel version
Makefile:68: Neither CFG80211 nor Wireless Extension is enabled in kernel
KBUILD_NOPEDANTIC=1 make -C /lib/modules/`uname -r`/build M=`pwd`
make: *** /lib/modules/3.7.6-1-ARCH/build: Нет такого файла или каталога. Останов.
make: *** [all] Error 2

$ ls /lib/modules
3.7.6-1-ARCH 3.7.8-1-ARCH extramodules-3.7-ARCH

SMOG commented on 2013-02-15 15:25

Using this package and NetworkManager the wifi worked like a charm, until yesterday: I've updated kernel from 3.7.6 to 3.7.7 and wifi isn't working anymore...
anyone's having the same problem?

SMOG commented on 2013-02-15 15:24

Using this package and NetworkManager the wifi worked like a charm, since yesterday: I've updated kernel from 3.7.6 to 3.7.7 and wifi isn't working anymore...
anyone's having the same problem?

Anonymous comment on 2013-02-03 22:08

I'll look into it, have been busy and there were no major updates to broadcom-wl so I almost forgot about this. Will update soon.
Thanks, @Maxwel!

Maxwel commented on 2013-02-03 21:59

This package doesn't work with my kernel (I dont know exactly why, but I think that it's because make variables API=WEXT KBUILD_NOPEDANTIC=1 is missing). And (most important) module cannot be compiled for other kernel than currently loaded. Here is my version of dkms.conf, which can build module for all installed kernels and it works on all kernels, which I tested.
http://pastebin.com/tTFnZ3Ed
Please update it. Pleeeeeeeease)))

twelveeighty commented on 2013-01-11 03:03

Nobody seemed to throw stones at me after my last remark, so I've removed the out-of-date flag for now.

ryley commented on 2013-01-09 13:54

Sorry twelveeighty, can't believe I didn't notice that... The only reason I can think of updating this would be that it can support both 64 and 32 bit linux. The only difference between the 5.100.82.112 aur version and 6.20.155 I've seen is a lot more kernel panics. I think I'll keep using broadcom-wl for the power management, 40MHz channels and hardware encryption, but I'll use the older version for now. Luckily my 43224 wifi card is supported by B43, brcmsmac and broadcom-wl (both old and new).

twelveeighty commented on 2013-01-09 03:58

Okay all, we have a problem. I downloaded the source from the link provided by @ryley (https://launchpad.net/ubuntu/raring/+source/bcmwl), and, yes, the version number of that driver is higher (6.20.155 vs the current AUR 6.20.55) BUT the source code is IDENTICAL to what we have now.
The only diffs between the source posted on launchpad and this AUR build are the ones that we need to patch out again (wl_linux.c: ndo_set_multicast_list replaced by ndo_set_rx_mode as of kernel 3.2).

So, unless you all enjoy updating this terrible driver for no reason, I would much rather not touch it. Otherwise, please post back here and I'll go through the effort of updating it if we get some votes. If not, I'll remove the "out-of-date" flag in a couple of days, deal?

I hate Broadcom so much.

twelveeighty commented on 2013-01-08 04:32

Sorry guys, I wasn't monitoring this feed - best to update the thread on the forum (https://bbs.archlinux.org/viewtopic.php?pid=1215189) to get my attention. I'll look into this.

Anonymous comment on 2013-01-04 05:47

Thanks very much dude this works for me on my dell 5220 awesome

ryley commented on 2012-12-16 09:50

An updated version of Broadcom's closed driver is abailable here: https://launchpad.net/ubuntu/raring/+source/bcmwl . Debs are available for both i686 and x86_64. Would you be able to update the PKGBUILD twelveeighty?

samuaz commented on 2012-10-24 19:33

thanks for the update

i have problem with the last update

==> Validando el archivo fuente con sha1sums...
hybrid-portsrc_x86_64-v5_100_82_112.tar.gz ... FALLÓ
modprobe.d ... Aprobado
license.patch ... Aprobado
user-ioctl.patch ... Aprobado
linux-recent.patch ... Aprobado
dkms.conf ... Aprobado
==> ERROR: ¡Uno o más archivos no pasaron la verificación de integridad!
==> ERROR: Makepkg no ha podido compilar dkms-broadcom-wl.

the hybrid-portsrc_x86_64-v5_100_82_112.tar.gz fail and the integrity check fails

maybe its a external fail?

thanks

Anonymous comment on 2012-10-17 16:10

Yeah, man! The it's works! I am writing connected with WiFi. Thank you!

Anonymous comment on 2012-10-17 14:03

Package updated.
Includes new patch for kernel 3.6.2+ as in broadcom-wl.

twelveeighty commented on 2012-10-17 13:48

Last night, I realized I missed changing the parameters to the wl_cfg80211_scan function. This latest rel should fix this. I don't have a 3.6 kernel yet to test, so please post back with results.

nplatis commented on 2012-10-17 12:55

Same here: problems with this package and 3.6.2 kernel. But in the comments of the broadcom-wl package I see that a patch for this new kernel is available.

Anonymous comment on 2012-10-17 06:19

[code]
$ export LANG=en_US && sudo dkms install broadcom-wl/6.20.55.19

Kernel preparation unnecessary for this kernel. Skipping...

Building module:
cleaning build area....
make KERNELRELEASE=3.6.2-1-ARCH....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.6.2-1-ARCH (x86_64)
Consult /var/lib/dkms/broadcom-wl/6.20.55.19/build/make.log for more information.

$ cat /var/lib/dkms/broadcom-wl/6.20.55.19/build/make.log

DKMS make.log for broadcom-wl-6.20.55.19 for kernel 3.6.2-1-ARCH (x86_64)
Wed Oct 17 09:17:55 EEST 2012
CFG80211 API is prefered for this kernel version
Makefile:73: Neither CFG80211 nor Wireless Extension is enabled in kernel
KBUILD_NOPEDANTIC=1 make -C /lib/modules/`uname -r`/build M=`pwd`
make[1]: Entering directory `/usr/src/linux-3.6.2-1-ARCH'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/6.20.55.19/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.55.19/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_linux.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_iw.o
CC [M] /var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.o
/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.c: In function 'wl_cfg80211_scan':
/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.c:489:21: error: 'ndev' redeclared as different kind of symbol
/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.c:485:58: note: previous definition of 'ndev' was here
/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.c: At top level:
/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.c:1561:2: warning: initialization from incompatible pointer type [enabled by default]
/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.c:1561:2: warning: (near initialization for 'wl_cfg80211_ops.scan') [enabled by default]
make[2]: *** [/var/lib/dkms/broadcom-wl/6.20.55.19/build/src/wl/sys/wl_cfg80211.o] Error 1
make[1]: *** [_module_/var/lib/dkms/broadcom-wl/6.20.55.19/build] Error 2
make[1]: Leaving directory `/usr/src/linux-3.6.2-1-ARCH'
make: *** [all] Error 2
[/code]

can't build module

twelveeighty commented on 2012-10-17 04:27

Although I have not been able to verify this, kernel 3.6 breaks this driver. This latest version has NOT been tested, but please let me know if it works. I borrowed the same type of patch applied to other Broadcom drivers to wl_cfg80211.c.
Again, let me repeat: I have not been able to test this new version. Please let me know if this version works on kernel 3.6. It SHOULD NOT work on 3.5 and below anymore. Yeah - kernel drivers suck!

samuaz commented on 2012-10-16 23:22

whit the new kernel update 3.6.2 this package broke the system

ottoshmidt commented on 2012-10-06 16:00

I first thought that dkms-broadcom-wl is a feature of this package and that it should be installable together. :)

ottoshmidt commented on 2012-09-29 19:48

:: dkms-broadcom-wl and broadcom-wl are in conflict. Remove broadcom-wl? [y/N]

twelveeighty commented on 2012-09-22 20:30

Uploaded Rel = 2, now using the link from Dropbox to reduce package size on AUR.

twelveeighty commented on 2012-09-22 20:17

Actually, I did find another upstream tarball (.deb) on Dropbox (the other one was a download off a hunting website from one of the contributors). I can change the PKGBUILD to refer to that to reduce the size of this AUR package, but what if that file gets removed from Dropbox? Any suggestions?

twelveeighty commented on 2012-09-22 17:29

Please read the associated thread - https://bbs.archlinux.org/viewtopic.php?id=145884. The problem is that there is no reliable upstream tar ball. If someone can fix that, I'd be happy to change the PKGBUILD.

falconindy commented on 2012-09-22 16:32

Please do not host source tarballs on the AUR. This package is absurdly large.

Anonymous comment on 2012-07-24 04:46

Ok, fixed, I neglected the old INSTALL script, overwriting it with the broadcom-wl one, my fault.
Thanks for seeing that, Boris!

Updated depends, and conflicts with broadcom-wl as in previous versions.

Anonymous comment on 2012-07-23 23:36

@boppbo Hmmm, it uses dkms so it should be done automatically at boot if you have dkms daemon in the correspondant array in rc.conf.
I tested it many times and there where no problems.
Let me check that about pkgname ASAP.

boppbo commented on 2012-07-23 23:31

Hi, I had some problems installing this package. First i think ${_pkgname} needs to be defined, maybe to broadcom-wl. In addition I had to build/install the module manually. AFAIR this was done automatically in previous versions. Thanks for your work - Boris

Anonymous comment on 2012-07-22 23:16

Updated to revision 5 as on broadcom-wl AUR package.

srl commented on 2012-06-15 17:13

Disowning. This package may stop working with linux-3.4 - you have been warned.

fdservices commented on 2012-05-02 16:48

newbie time - on a kernel update the wl module is recompiled by dkms but not installed in /lib/modules. Can anyone explain why? - Andrew

srl commented on 2012-04-08 23:05

Whoops, my bad. Fixed, I think. Thanks RozenTensai.

Anonymous comment on 2012-04-08 22:56

Fails complaining the integrity check for the 'hybrid-portsrc_x86_32-v5.100.82.112.tar.gz' file
The actual sha1 for that file should be '01aa32f9e85621253a3f15cf4361bb80d41da3e8' not '5bd78c20324e6a4aa9f3fafdc6f0155e884d5131'

After that modification it installed and worked fine.
Tested on 3.3.1-1

srl commented on 2012-02-26 23:23

Updated. Tested on 3.2.5-1, at least - should work without trouble on 3.2.7-1 (testing now).

Anonymous comment on 2012-02-26 22:30

Building module:
cleaning build area....
make KERNELRELEASE=3.2.7-1-ARCH....(bad exit status: 2)
Error! Bad return status for module build on kernel: 3.2.7-1-ARCH (x86_64)
Consult /var/lib/dkms/broadcom-wl/5.100.82.112/build/make.log for more information.
==> For more information on this driver, see:
==> http://wiki.archlinux.org/index.php/Broadcom_BCM4312


cat /var/lib/dkms/broadcom-wl/5.100.82.112/build/make.log
DKMS make.log for broadcom-wl-5.100.82.112 for kernel 3.2.7-1-ARCH (x86_64)
dom 26 feb 2012, 22.59.48, CET
/bin/sh: line 0: [: : integer expression expected
/bin/sh: line 0: [: : integer expression expected
Wireless Extension is the only possible API for this kernel version
Using Wireless Extension API
KBUILD_NOPEDANTIC=1 make -C /lib/modules/`uname -r`/build M=`pwd`
make[1]: Entering directory `/usr/src/linux-3.2.7-1-ARCH'
CFG80211 API is prefered for this kernel version
Using CFG80211 API
LD /var/lib/dkms/broadcom-wl/5.100.82.112/build/built-in.o
CC [M] /var/lib/dkms/broadcom-wl/5.100.82.112/build/src/shared/linux_osl.o
CC [M] /var/lib/dkms/broadcom-wl/5.100.82.112/build/src/wl/sys/wl_linux.o
/var/lib/dkms/broadcom-wl/5.100.82.112/build/src/wl/sys/wl_linux.c:389:2: error: unknown field ‘ndo_set_multicast_list’ specified in initializer
/var/lib/dkms/broadcom-wl/5.100.82.112/build/src/wl/sys/wl_linux.c:389:2: warning: initialization from incompatible pointer type [enabled by default]
/var/lib/dkms/broadcom-wl/5.100.82.112/build/src/wl/sys/wl_linux.c:389:2: warning: (near initialization for ‘wl_netdev_ops.ndo_validate_addr’) [enabled by default]
make[2]: *** [/var/lib/dkms/broadcom-wl/5.100.82.112/build/src/wl/sys/wl_linux.o] Errore 1
make[1]: *** [_module_/var/lib/dkms/broadcom-wl/5.100.82.112/build] Errore 2
make[1]: Leaving directory `/usr/src/linux-3.2.7-1-ARCH'
make: *** [all] Errore 2

srl commented on 2012-02-26 14:28

I am still unable to build this module as-is on 3.2.5-1 - the patch I mentioned below is needed. Am I alone in this, and if so, any ideas why? (And if not, please update the package.)

ushimitsudoki commented on 2011-11-14 08:25

- fixed up depmod error in install file
- moved closer to main broadcom-wl package
- tested build and install on x86_64 (don't currently have a machine to test operation)

srl commented on 2011-11-11 17:17

The following patch is needed to fix the build.

--- old/src/wl/sys/wl_linux.c 2011-11-05 23:19:53.148240182 -0400
+++ new/src/wl/sys/wl_linux.c 2011-11-05 23:20:11.847035005 -0400
@@ -386,7 +386,7 @@
#endif
.ndo_get_stats = wl_get_stats,
.ndo_set_mac_address = wl_set_mac_address,
- .ndo_set_multicast_list = wl_set_multicast_list,
+ .ndo_set_rx_mode = wl_set_multicast_list,
.ndo_do_ioctl = wl_ioctl
};

mobad commented on 2011-11-10 04:57

Doesn't work with 3.1.
The non-dkms one has been updated so you can use their code.

ushimitsudoki commented on 2011-11-02 07:50

- Version bump
- Added patches and moved PKGBUILD closer to broadcom-wl package

ushimitsudoki commented on 2011-10-23 01:21

- Version bump

ushimitsudoki commented on 2011-07-21 22:58

- cleaned up the symlink stuff / proper file copying
- tested x86_64 build and install (but not operation, don't have correct machine for that right now)

gfrito,

I tried copying less, but ran into build errors, and don't have the time or machine right now to test it in depth. I welcome input if someone wants to trim the files copied down more.

Anonymous comment on 2011-07-21 22:00

This is faulty:

% ls -ls /usr/src/broadcom-wl-5.100.82.38/dkms.conf
0 lrwxrwxrwx 1 root root 51 Jul 21 23:50 /usr/src/broadcom-wl-5.100.82.38/dkms.conf -> /tmp/yaourt-tmp-xxx/aur-dkms-broadcom-wl/dkms.conf

Anonymous comment on 2011-07-17 23:18

It seems that the PKGBUILD copies symlinks instead of the actual target. I think it should only be copying the dkms.conf file (as opposed to the patch files and the .tar.gz), with the addition of the -L flag to cp.

ushimitsudoki commented on 2011-03-21 09:17

-added conflicts/replaces
-tested on i686
-removed spurious bits in install file