Hacker News new | comments | ask | show | jobs | submit login

A few semi-relevant notes: * We've bought multiple Dell XPS 13 laptops with Windows off Amazon because the same spec there was roughly $400-500 cheaper than buying the XPS Developer Edition from the Dell website. * We tried running Linux on a Dell non-XPS Skylake Core i7 laptop and it was having many kernel panics. A quick Googling revealed people with a Skylake chipset having a similar issue. * I noticed just a couple days ago that Dell has updated their XPS Developer Edition laptop to Skylake. One difference I see on the site is it ships with Ubuntu 14.04 SP1. I haven't read much about SP1, other than it looks like it's about a year old.

So yeah, I'm still buying XPS laptops with 5th gen chipsets because we've had issues with Skylake.

Keep in that the XPS units often have Broadcom wifi chipsets while the developer edition units have Intel. Much better Linux driver support for the Intel cards. The Precision 15" now comes in the same chassis as the XPS and has the Intel chipset (and other goodies)

This x1000. Always get the Precision 15" over the 13"...the hardware is just better suited for Linux. The Sputnik team has to take chassis from other Dell products and support them as manufactured...this is why the XPS 13 DE has Broadcom. I have a XPS 13 DE from last year and I wish every day I would have just ponied up and gotten the 15.

But keep in mind that Dell is releasing their drivers into upstream kernels, so if you are on 14.04, you should be using the LTS Enablement stack in order to avail yourself of the latest firmware with stability improvements (this is how I fixed the Broadcom Null pointer bug).


You don't have to use the LTS Enablement Stack if you installed with newer service pack .isos.

The 15" also has Thunderbolt which is a must if you want to have a reasonable docking solution for a Linux laptop.

So thats what LTSE is. Saw it in the compatibility table for the TB15 thunderbolt dock.

Which, I really can't tell if it'll work or not. Really need to plug a 4k monitor into it.

Yeah, LTSE is how you can keep upgrading your kernel while staying on the LTS releases of Ubuntu. Granted you are limited to the kernel versions of the previously released Ubuntu version and there is a slight delay between when a new version is relased and when it becomes available to LTSE, but overall I have found the stability to be good.

From what I found out (if I remember correctly), I have great confidence that Thunderbolt docks should work well with Linux kernel 3.19+. But take this with a grain of salt since I ended up with the 13 DE and therefore only have a mDP and couldn't try a Thunderbolt dock.

Can confirm. Have the Precision 5510. With the 4.4 kernel, it more or less works fine. Last I checked, I was getting ~5hrs on battery. Which is pretty shit-hot for a linux laptop.

One gripe so far: Since I updated the BIOS, the entire thing performs like a Netbook-Hype-era Atom CPU when there's no AC power.

A couple months ago I bought a Dell XPS 13 (signature edition) off of Amazon because it was a few hundred dollars cheaper than buying it from Dell, and installed Linux over it.

It's a Skylake machine, and it's been working great. I'm using Debian Testing with the kernel pulled from Sid (necessary to get the wifi working).

Powertop output is:

    Package             |             Core    |            CPU 0       CPU 2
                        |                     | C0 active   6.7%        3.9%
                        |                     | POLL        0.1%    0.2 ms  0.0%    0.0 ms
                        |                     | C1E-SKL     5.5%    0.2 ms  1.3%    0.2 ms
    C2 (pc2)    9.7%    |                     |
    C3 (pc3)    2.2%    | C3 (cc3)    2.5%    | C3-SKL      3.2%    0.3 ms  0.4%    0.2 ms
    C6 (pc6)    4.1%    | C6 (cc6)    7.0%    | C6-SKL      8.0%    0.3 ms  8.4%    0.7 ms
    C7 (pc7)    0.0%    | C7 (cc7)   28.7%    | C7s-SKL     0.0%    0.1 ms  0.0%    0.7 ms
    C8 (pc8)   12.1%    |                     | C8-SKL     29.4%    1.9 ms 60.5%    2.3 ms
    C9 (pc9)    0.0%    |                     | C9-SKL      0.0%    0.0 ms  0.0%    1.3 ms
    C10 (pc10)  0.0%    |                     | C10-SKL     4.3%    3.6 ms 14.6%    4.1 ms

                        |             Core    |            CPU 1       CPU 3
                        |                     | C0 active   5.2%        4.4%
                        |                     | POLL        0.0%    0.0 ms  0.0%    0.2 ms
                        |                     | C1E-SKL     1.5%    0.2 ms  1.5%    0.3 ms
                        |                     |
                        | C3 (cc3)    0.6%    | C3-SKL      0.7%    0.2 ms  0.7%    0.2 ms
                        | C6 (cc6)   14.7%    | C6-SKL     10.8%    0.7 ms 13.8%    0.8 ms
                        | C7 (cc7)   52.0%    | C7s-SKL     0.1%    1.1 ms  0.0%    0.4 ms
                        |                     | C8-SKL     59.6%    2.1 ms 59.8%    1.8 ms
                        |                     | C9-SKL      0.0%    0.0 ms  0.0%    0.0 ms
                        |                     | C10-SKL     9.7%    4.4 ms  8.4%    3.5 ms
                        |             GPU     |
                        |                     |
                        | Powered On 13.2%    |
                        | RC6        86.8%    |
                        | RC6p        0.0%    |
                        | RC6pp       0.0%    |
                        |                     |
                        |                     |
                        |                     |
                        |                     |
                        |                     |

As far as I can see, this seems to imply C8 is working fine?

Could you paste the output of sudo lspci -vvv and cat /sys/devices/virtual/dmi/id/bios_version somewhere?

Not OP, but I have dell xps 15 skylake. (i3-6100H)


I am currently running Ubuntu 16.04 4.4.0-8 kernel which works fine. The ubuntu backport of i915 from 4.6 to 4.4 which is present in 4.4.0-9 causes weird video issues. (upstream bug: https://bugs.freedesktop.org/show_bug.cgi?id=94593)

Only the ultra-low-power chips (U and Y series) are susceptible to damage from poor power management.

any source?


lspci -vvv: http://pastebin.com/raw/FUsMqPZz

cat /sys/devices/virtual/dmi/id/bios_version: 1.2.3

Happy to provide any other info as well, I love this machine to bits and would be happy if I could help other people enjoy it too.

Could you re-run the lspci as root? Thanks!

This is probably a slightly different configuration, but I also see support for low Cstates on my Skylake processor (running 4.6.0 rc2)

sudo lspci -vvv : http://pastebin.com/79BfxQC4

Sorry about that! here you go:


also, what kernel version is (uname -a)

Linux eenie 4.4.0-1-amd64 #1 SMP Debian 4.4.6-1 (2016-03-17) x86_64 GNU/Linux

That seems to make sense to me. Running a 2 year old OS (even with 1 year old service pack) on bleeding edge laptop and you don't expect to have problems? Doesn't compute. This is not an LTS situation, live a little.

Why not try Ubuntu 15.10 or 16.04 beta? At the very least run something released in the last 6 months if you have any reasonable expecatation of things working.

Seems like all the Arch folks have this working also.

There's no such thing as SP1 from Canonical, sounds like a Dell invention.

They mean they're shipping the 14.04.01 LTS, the point release that comes out three months after the main release. Dell actually ship a custom enabled Ubuntu version - it's one of the LTS releases that's stable with any custom work that was done for them on enablement. Which is why buying from Amazon isn't going to provide the same experience.

I've been running Arch on a 2016 model XPS 13 Signature Edition. Everything worked out of the box crash-free. This includes Broadcom WiFi, Skylake CPU, media keys, multitouch tracked, and touchscreen.

Offtopic-ish, but the 2015 model recently had its sound completely demolished by a kernel release in arch. The only fixes are to either downgrade the kernel or recompile it with a flag. Both fixes are annoying as all hell and there doesn't seem to be enough support from arch nor dell to fix this problem!

Camel, meet straw.

Well, it makes sense because Arch's policy is to follow upstream with minimal patches, so they would wait for the new kernel to fix the bug rather than applying a patch themselves. If you don't like this policy, you might want to pick a different distro or use your own kernel.

Nah, I totally get it. It's still incredibly frustrating, especially considering it only broke after spending three hours fixing an issue that pacman itself caused. I'm definitely in the works of finding a new distro. :)

OpenSUSE Tumbleweed has much better testing, and is actually faster than Arch in some cases. I've switched a lot of my machines recently. And you can use OBS which actually builds "AUR" packages for you (you can branch them on OBS if you want to modify them, or you can even download and modify the package locally).

If I didn't run Fedora and CentOS on my servers at work I would probably switch to Tumbleweed on my XPS 13. In reality I shouldn't really need to keep my development environment so close to what I deploy on, but it's nice to be able to quickly test a RPM I'm packaging without needing access to my server at home/work or dealing with a VM or docker.

NixOS or Void linux are rational alternatives to Arch

What kernel version/distro were you running with Linux when you experienced the panics?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact