Start a Conversation

Unsolved

This post is more than 5 years old

5964268

October 31st, 2015 01:00

XPS13 9350 skylake no wifi on Ubuntu 15.10

With Ubuntu 15.10 (default kernel 4.2), 9350 almost works out of box, except this DW1820A wireless card. It has been a few days, I still couldn't get the wireless card working. From what I have read on Linux Wireless, BCM4350 driver is queued to be included in kernel 4.4. That's a long wait. 

Could anyone help?

20 Posts

November 3rd, 2015 08:00

On my Ubuntu 15.10, I did:

    sudo mv brcmfmac4350-pcie.bin /lib/firmware/brcm/

    sudo reboot

but, still no wireless. Then, I open "Additional Drivers" UI, still don't see anything new. Anything else I should try?

20 Posts

November 3rd, 2015 08:00

Could you please write down how you compiled the driver from kernel 4.4 on git? Maybe someone else here can push forward based on your work.

18 Posts

November 3rd, 2015 08:00

Maybe someone wants to try this out. I was able to generate the firmware from the offiicial git repository. No warranty. Has to go in /lib/firmware/brcm/

EDIT: I can give instructions on how to compile this on your own, if someone needs it.

1 Attachment

November 3rd, 2015 18:00

Computaholic's description of how to extract the firmware works well, and I can confirm that I applied the same process to obtain the following firmware file:

$ md5sum /lib/firmware/brcm/brcmfmac4350-pcie.bin 

7ad59f4b16485a17a8e7cfcfb83906ab /lib/firmware/brcm/brcmfmac4350-pcie.bin

But to use this file, you first need to use a kernel (or kernel module) that supports the 4350 device. The patch has been incorporated into the "net-next" kernel branch, so you can follow the instructions at https://wiki.ubuntu.com/KernelTeam/GitKernelBuild 

18 Posts

November 4th, 2015 04:00

I found the according patch

http://git.kernel.org/cgit/linux/kernel/git/davem/net-next.git/commit/?id=e3c92cb2eea395c9fbec7ceb678626565cbac0fe 

I am not sure if this results in a kernel module or if the kernel needs to be recompiled. Again, kernel module is fine, recompiling should be avoided, since any kernel update will break wifi. I'll probably go with the 7265 anyway...

Can someone tell if its possible to compile the brcmfmac alone or do we need a full kernel compilation?

EDIT: Ok, seems like this can only be included by recompiling the kernel. So personally, I will go with the 30$ intel wireless....

18 Posts

November 4th, 2015 06:00

Do you have FHD or QHD? EDIT: B/C I think the FHD version is OK with the 4.3 Kernel and does not show the described effect.

17 Posts

November 4th, 2015 06:00

I found the according patch

http://git.kernel.org/cgit/linux/kernel/git/davem/net-next.git/commit/?id=e3c92cb2eea395c9fbec7ceb678626565cbac0fe 

I am not sure if this results in a kernel module or if the kernel needs to be recompiled. Again, kernel module is fine, recompiling should be avoided, since any kernel update will break wifi. I'll probably go with the 7265 anyway...

Can someone tell if its possible to compile the brcmfmac alone or do we need a full kernel compilation?

EDIT: Ok, seems like this can only be included by recompiling the kernel. So personally, I will go with the 30$ intel wireless....

Is there something I'm missing? brcmfmac can be built as a module, so you shouldn't need to recompile the whole kernel from what I can see. I was already able to build the module:

$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git
$ cd net-next/
$ cp /boot/config-`uname -r` .config
$ ls
$ make oldconfig
$ ls
$ make prepare
$ make modules_prepare
$ make SUBDIRS=scripts/mod
$ make SUBDIRS=drivers/net/wireless/brcm80211/brcmfmac modules
$ sudo cp drivers/net/wireless/brcm80211/brcmfmac/brcmfmac.ko /lib/modules/4.2.0-16-generic/kernel/drivers/net/wireless/brcm80211/brcmfmac/brcmfmac.ko
$ depmod
$ sudo depmod
$ sudo modprobe brcmfmac

But I haven't had time to test it out. Will report back when I get the chance.

November 4th, 2015 06:00

The issue with this procedure is that depending on what kernel version you are actually running, the module that you compile in a step above may or may not load.

While I have had success recompiling the kernel (wifi works), a new problem appears that also occurs with the Ubuntu 4.3 carnal as well. On boot, to the splash screen disappears and X starts, I got a blank screen. The system is operating fine in that I can login and I can even adjust the brightness of the screen, but the entire screen is black. The only work around I have found so far has been to include the nomodeset boot option as a kernel parameter, but this prevents me from getting hardware accelerated graphics, which means that everything slows to a crawl with the high resolution screen.

November 4th, 2015 06:00

@Xps13-9350-User - You must have the QHD+ display. Check out my new post at en.community.dell.com/techcenter/os-applications/f/4613/t/19659568

I am not sure how we are going to combine the screen and wifi solutions yet such that we have a fully working 9350 QHD+ machine yet.

November 4th, 2015 19:00

That's great -- more progress. And more issues. I'm currently seeing a CPU-hosing behavior that occurs intermittently after 5-30min of use; once it starts, I can only kill it by suspending/resuming.

Top shows:

    12886 root 20 0 0 0 0 R 73.2 0.0 3:49.08 kworker/3:3 

    23 root 20 0 0 0 0 R 27.0 0.0 6:35.97 ksoftirqd/3

And "echo l > /proc/sysrq-trigger" reveals the following trace in "/var/log/syslog":

Nov 4 22:39:32 morel kernel: [ 7896.149983] CPU: 3 PID: 11394 Comm: kworker/3:1 Tainted: G W OE 4.3.0-mydell #1
Nov 4 22:39:32 morel kernel: [ 7896.149984] Hardware name: Dell Inc. XPS 13 9350/0VM5NC, BIOS 1.0.0 09/10/2015
Nov 4 22:39:32 morel kernel: [ 7896.149984] Workqueue: pm pm_runtime_work
Nov 4 22:39:32 morel kernel: [ 7896.149985] task: ffff8801ba3ce900 ti: ffff8801b9758000 task.ti: ffff8801b9758000
Nov 4 22:39:32 morel kernel: [ 7896.149986] RIP: 0010:[ ] [ ] __schedule+0x109/0x980
Nov 4 22:39:32 morel kernel: [ 7896.149986] RSP: 0018:ffff8801b975b9b0 EFLAGS: 00000086
Nov 4 22:39:32 morel kernel: [ 7896.149987] RAX: ffff8804ae360f00 RBX: ffff8804c0d969c0 RCX: 0000000000082a1f
Nov 4 22:39:32 morel kernel: [ 7896.149987] RDX: 000000613ccb05ee RSI: ffff8804c0d96a58 RDI: ffff8804ae360f70
Nov 4 22:39:32 morel kernel: [ 7896.149988] RBP: ffff8801b975b9e8 R08: 0000000000000000 R09: 0000000000000000
Nov 4 22:39:32 morel kernel: [ 7896.149989] R10: ffff8801ba3ce960 R11: 0000000000000000 R12: 00000000000169c0
Nov 4 22:39:32 morel kernel: [ 7896.149989] R13: ffff8804c0d969c0 R14: ffff8801ba3cee58 R15: 0000000000000003
Nov 4 22:39:32 morel kernel: [ 7896.149990] FS: 0000000000000000(0000) GS:ffff8804c0d80000(0000) knlGS:0000000000000000
Nov 4 22:39:32 morel kernel: [ 7896.149991] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Nov 4 22:39:32 morel kernel: [ 7896.149991] CR2: 000055b56a360768 CR3: 0000000002c0c000 CR4: 00000000003406e0
Nov 4 22:39:32 morel kernel: [ 7896.149992] Stack:
Nov 4 22:39:32 morel kernel: [ 7896.149992] ffff8804c0d969c0 ffff8801ba3ce900 ffff8801b975c000 ffff8801b975ba90
Nov 4 22:39:32 morel kernel: [ 7896.149993] ffff8801b975ba88 ffff8801b975bafc 00000000000003e8 ffff8801b975ba00
Nov 4 22:39:32 morel kernel: [ 7896.149994] ffffffff817b148e 00000000000000fa ffff8801b975ba10 ffffffff817b14cc
Nov 4 22:39:32 morel kernel: [ 7896.149994] Call Trace:
Nov 4 22:39:32 morel kernel: [ 7896.149995] [ ] preempt_schedule_common+0x1e/0x40
Nov 4 22:39:32 morel kernel: [ 7896.149995] [ ] _cond_resched+0x1c/0x30
Nov 4 22:39:32 morel kernel: [ 7896.149996] [ ] wait_for_completion_timeout+0x23/0x120
Nov 4 22:39:32 morel kernel: [ 7896.149997] [ ] usb_start_wait_urb+0x7d/0x150
Nov 4 22:39:32 morel kernel: [ 7896.149997] [ ] usb_control_msg+0xc1/0x100
Nov 4 22:39:32 morel kernel: [ 7896.149998] [ ] hub_port_status+0x7a/0x100
Nov 4 22:39:32 morel kernel: [ 7896.149999] [ ] hub_activate+0xbd/0x590
Nov 4 22:39:32 morel kernel: [ 7896.149999] [ ] hub_resume+0x27/0x50
Nov 4 22:39:32 morel kernel: [ 7896.150000] [ ] usb_resume_interface.isra.6+0x9f/0x100
Nov 4 22:39:32 morel kernel: [ 7896.150000] [ ] usb_suspend_both+0xef/0x1d0
Nov 4 22:39:32 morel kernel: [ 7896.150001] [ ] usb_runtime_suspend+0x38/0x70
Nov 4 22:39:32 morel kernel: [ 7896.150002] [ ] ? usb_probe_interface+0x2d0/0x2d0
Nov 4 22:39:32 morel kernel: [ 7896.150002] [ ] __rpm_callback+0x33/0x70
Nov 4 22:39:32 morel kernel: [ 7896.150003] [ ] ? usb_probe_interface+0x2d0/0x2d0
Nov 4 22:39:32 morel kernel: [ 7896.150003] [ ] rpm_callback+0x24/0x80
Nov 4 22:39:32 morel kernel: [ 7896.150004] [ ] ? usb_probe_interface+0x2d0/0x2d0
Nov 4 22:39:32 morel kernel: [ 7896.150005] [ ] rpm_suspend+0x11c/0x5e0
Nov 4 22:39:32 morel kernel: [ 7896.150005] [ ] __pm_runtime_suspend+0x5b/0x80
Nov 4 22:39:32 morel kernel: [ 7896.150006] [ ] usb_runtime_idle+0x2a/0x40
Nov 4 22:39:32 morel kernel: [ 7896.150006] [ ] __rpm_callback+0x33/0x70
Nov 4 22:39:32 morel kernel: [ 7896.150007] [ ] rpm_idle+0x188/0x200
Nov 4 22:39:32 morel kernel: [ 7896.150008] [ ] pm_runtime_work+0x95/0xa0
Nov 4 22:39:32 morel kernel: [ 7896.150008] [ ] process_one_work+0x1aa/0x440
Nov 4 22:39:32 morel kernel: [ 7896.150009] [ ] worker_thread+0x4b/0x4c0
Nov 4 22:39:32 morel kernel: [ 7896.150009] [ ] ? process_one_work+0x440/0x440
Nov 4 22:39:32 morel kernel: [ 7896.150010] [ ] kthread+0xc9/0xe0
Nov 4 22:39:32 morel kernel: [ 7896.150011] [ ] ? kthread_create_on_node+0x180/0x180
Nov 4 22:39:32 morel kernel: [ 7896.150011] [ ] ret_from_fork+0x3f/0x70
Nov 4 22:39:32 morel kernel: [ 7896.150012] [ ] ? kthread_create_on_node+0x180/0x180
Nov 4 22:39:32 morel kernel: [ 7896.150013] Code: 45 d0 48 81 78 58 60 6b 81 81 0f 85 27 04 00 00 8b 93 84 00 00 00 39 53 04 0f 85 18 04 00 00 48 89 c6 48 89 df ff 15 b7 61 06 00 <48> 83 f8 ff 0f 84 02 04 00 00 48 85 c0 48 89 45 c8 0f 84 bb 06

 

There are various posts like https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1488426 

18 Posts

November 5th, 2015 02:00

Seems like an USB problem. Have you tried to plug in usb devices in all your usb sockets? 

Anyhow, I suggest that all posts state their exact xps 13 model and Kernel version, since the fhd and qhd are different in some ways, maybe like this (example):

XPS 13 9350 FHD i7 WIFI BCM 4.3-rc1 Ubuntu 15.10 or

XPS 13 9350 QHD i5 WIFI INTEL 4.2 Ubuntu 15.10

Also, new Issues should reside in new threads, otherwise this gets totally overloaded. But thats just my opinion;)

17 Posts

November 8th, 2015 15:00

I made a lot of progress this past weekend.

I'm on the XPS 13 QHD i7 WIFI BCM Ubuntu 15.10.

Here's what I did:

Unzipped and copied over binary that computaholic and others provided.

$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git
$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
$ cd net-next
$ git format-patch -1 e3c92cb
$ cp 0001-brcmfmac-Add-support-for-the-BCM4350-PCIE-device.patch ../linux/
$ cd ../linux
$ git fetch git://git.launchpad.net/~ubuntu-kernel-test/ubuntu/+source/linux/+git/mainline-crack cod/tip/drm-intel-nightly/2015-11-05
$ git checkout FETCH_HEAD
$ git am --reject 0001-brcmfmac-Add-support-for-the-BCM4350-PCIE-device.patch
$ cp /boot/config-4.2.0-16-generic .config
$ make oldconfig
$ yes '' | make oldconfig
$ make clean
$ make -j `getconf _NPROCESSORS_ONLN` deb-pkg LOCALVERSION=-custom
$ sudo dpkg -i linux-image-4.3.0-custom-dbg_4.3.0-custom-1_amd64.deb
$ sudo dpkg -i linux-headers-4.3.0-custom_4.3.0-custom-1_amd64.deb

I now have working WiFi and no issues with suspending/resuming. I suspect that my battery life isn't quite as good as it was. I also attempted to get my Dell D1000 dock working by installing this DisplayLink driver, and though I'm able to see the output device, my computer locks up when I plug the monitor in.

Anyway, hope this was helpful to someone.

By "this DisplayLink driver" I mean:

http://support.displaylink.com/knowledgebase/articles/684649-how-to-install-displaylink-software-on-ubuntu

17 Posts

November 8th, 2015 15:00

I made a lot of progress this past weekend.

I'm on the XPS 13 QHD i7 WIFI BCM Ubuntu 15.10.

Here's what I did:

Unzipped and copied over binary that computaholic and others provided.

$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git
$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
$ cd net-next
$ git format-patch -1 e3c92cb
$ cp 0001-brcmfmac-Add-support-for-the-BCM4350-PCIE-device.patch ../linux/
$ cd ../linux
$ git fetch git://git.launchpad.net/~ubuntu-kernel-test/ubuntu/+source/linux/+git/mainline-crack cod/tip/drm-intel-nightly/2015-11-05
$ git checkout FETCH_HEAD
$ git am --reject 0001-brcmfmac-Add-support-for-the-BCM4350-PCIE-device.patch
$ cp /boot/config-4.2.0-16-generic .config
$ make oldconfig
$ yes '' | make oldconfig
$ make clean
$ make -j `getconf _NPROCESSORS_ONLN` deb-pkg LOCALVERSION=-custom
$ sudo dpkg -i linux-image-4.3.0-custom-dbg_4.3.0-custom-1_amd64.deb
$ sudo dpkg -i linux-headers-4.3.0-custom_4.3.0-custom-1_amd64.deb

I now have working WiFi and no issues with suspending/resuming. I suspect that my battery life isn't quite as good as it was. I also attempted to get my Dell D1000 dock working by installing this DisplayLink driver, and though I'm able to see the output device, my computer locks up when I plug the monitor in.

Anyway, hope this was helpful to someone.

17 Posts

November 8th, 2015 16:00

mpalmiotto

I made a lot of progress this past weekend.

I'm on the XPS 13 QHD i7 WIFI BCM Ubuntu 15.10.

Here's what I did:

[snip]

I now have working WiFi and no issues with suspending/resuming. I suspect that my battery life isn't quite as good as it was. I also attempted to get my Dell D1000 dock working by installing this DisplayLink driver, and though I'm able to see the output device, my computer locks up when I plug the monitor in.

Anyway, hope this was helpful to someone.

By "this DisplayLink driver" I mean:

http://support.displaylink.com/knowledgebase/articles/684649-how-to-install-displaylink-software-on-ubuntu

Also, for the record, it looks like the displaylink manager is what was eating up most of my battery life, so I went ahead and uninstalled it. Looks like I'm just going to have to get a USB-C -> HDMI adapter, as the D1000 dock's HDMI isn't working anyway.

17 Posts

November 8th, 2015 16:00

This doesn't seem to work nor generate the 0001-* file, even with the suggested "--" added after the -1.

root@ubuntu:/usr/src/net-next# git format-patch -1 e3c92db
fatal: ambiguous argument 'e3c92db': unknown revision or path not in the working tree.
Use '--' to separate paths from revisions, like this:
'git [ ...] -- [ ...]'

any idea?

D'oh. Looks like I accidentally typoed. I fixed the original post -- sorry about that. The commit-ish is actually e3c92cb.

Also, you really shouldn't be logged in as root >.<
Try to avoid that, and instead use `sudo [command]` for root access (when you need to).

No Events found!

Top