Dell XPS 13 Kabylake Makes For A Great Linux Laptop

[Hi-Angel]

Well, I not sure of non-russia support will make any difference. Do you think it's worth it?

Well, honestly, I don't hold much hope you personally gonna get them to fix this. But I'd still do that for another reason: such actions create demand. If people often asking for something, it means there is interest, and if a company cares about its customers, the support gonna improve with time.

[zoomblab]

Correctable AER is not a problem. It's just reporting it, but it was corrected and life goes on. My Dell laptop also reports these but only one every other hour or so, usually from the NVMe drive.

It's like ECC RAM. Having it shows you the corrected errors. Not having ECC you get the errors anyway, but no log messages. So I guess you're just happier in ignorance?

Ah, sorry, so the bug is not occasional correctable AER errors but spam of many per second? That does sound like a problem.

Yes there was a spam of hundrend messages per second, though that happened in a not constant manner. There were pauses for seconds or minutes and then flood again. Anyway having reports of errors is a bad sign even if they are being corrected. Besides being a bad smell that can lead to degraded performance.

https://www.kernel.org/doc/Documentation/PCI/pcieaer-howto.txt

The solution to disable the error reporting is not a solution IMO. The actual solution was to set the kernel flag pci_aspm=off but that was also bad IMO. And also the idea of setting kernel parameter for the computer to work properly is absud to me. Because of that laptop went bye bye.

[RussianNeuroMancer]

try updating BIOS or use this param "i8042.dumbkbd" to kernel (inside "/etc/default/grub")

i8042.dumbkbd does not help, unfortunately
Next in the line will be "i8042.reset=1" and I also will try to disable TPM controller.

But I'd still do that for another reason: such actions create demand. If people often asking for something, it means there is interest, and if a company cares about its customers, the support gonna improve with time.

You are right about this, will do then. Right now I trying to convince Alan by using this case when I help a little bit with getting better support for Dell hardware in Linux. Anyway, I don't expect it to work out, because they stopped selling Dell 5855 by then, just hope there is a little chance it convince him.

[wizard69]

wizard69

The situation with Ryzen mobile and linux is rather hopeless. I wanted to buy one and done my hoework. If you search for it you will find the same problems from owners of all brands such as HP, Acer and Lenovo, random freezes. E.g.,

https://ubuntuforums.org/showthread.php?t=2394464

Linux was absolutely terrible for most of the first part of the year no doubt at all there and to be honest I tried more than a few distros and releases. At this point though iti is a far better if less than perfect experience. Would I suggest such a system to somebody with little technical background - nope - but right now I'm pretty happy. I do realize that it might take something beyond kernel 4.19 to fully iron out every issue. The difference today and the beginning of the year though is huge.

[Guest]

I have a XPS 13 9360 (i5-7200U) with Fedora Workstation 28:

- Suspend as-is is broken (it goes to "sleep", and either locks-up or shuts down); as I understand, this laptop expects "Modern Standby", and only (officially?) supports states of S0 and nothing else; using "mem_sleep_default=s2idle" works fine and seemingly without issue (can shut the lid and open it to suspend/wake the system and all devices seem to return and work fine after suspend)

- Along with the suspend note above; on Windows, the laptop only reports S0 support via powercfg; under Linux, it reports S0, S3, S4, and S5 support

- BIOS chip is a W25Q256FV but it's flat; smaller than the usual W25Q256FV chips I've seen. Was going to attempt me_cleaner's soft-disable-only option on it (since Boot Guard), but my SOIC clip won't attach

- Disassembly was relatively easy; unscrew the bottom panel (including the hidden screw), plastic pry-tool to un-wedge the metal from the plastic, rip it off, and you're in and also presented with the most important aspects right there (CPU heatsink for reseating, SSD and Wifi chips for swapping out).

- SSD (Sandisk X400 128GB in my case) gets ATA frozen on boot and doesn't unfreeze with suspend. I perform security-erase and security-erase-enhanced on SSDs prior to installing OSs and need an unfrozen drive for that; setting a ATA password from firmware settings and using that works fine though

- Since my SSD isn't NVMe, I left the SATA mode at "Raid On" without issues (drive shows up and boots under Linux Live installers; can even switch between RAID and AHCI in firmware settings and still boot existing installs no problem)

- PSR with i915 reportedly works best at 2 (I forget specifics; 1 had a bit of noticeable flicker but the results reported at some sys for 2 looked better and had no flicker)

- Speakers sound nice under Linux, but I was pretty impressed with them under Windows with the default Waves MaxxAudio EQ settings

- It feels very durable; and it's impressively lightweight, along with having pretty nice battery life; keyboard and touchpad feel great too

- Here's some notable kernel boot options I'm using: mem_sleep_default='s2idle' intel_iommu='on' pcie_aspm='force' pcie_aspm.policy='powersupersave' reboot='efi' i915.enable_fbc='1' i915.fastboot='1' i915.enable_guc='3' i915.enable_psr='2'

[Hi-Angel]

I have a XPS 13 9360 (i5-7200U) with Fedora Workstation 28:

- Suspend as-is is broken (it goes to "sleep", and either locks-up or shuts down); as I understand, this laptop expects "Modern Standby", and only (officially?) supports states of S0 and nothing else; using "mem_sleep_default=s2idle" works fine and seemingly without issue (can shut the lid and open it to suspend/wake the system and all devices seem to return and work fine after suspend)

- Along with the suspend note above; on Windows, the laptop only reports S0 support via powercfg; under Linux, it reports S0, S3, S4, and S5 support

Did you report a bug to kernel?

[RussianNeuroMancer]

Did you report a bug to kernel?

Usual firmware bug, which is why systemd default to s2idle, at least on Ubuntu. Not sure why it's different for Espionage724, as it's should be the same on Fedora.

[Hi-Angel]

Usual firmware bug, which is why systemd default to s2idle, at least on Ubuntu. Not sure why it's different for Espionage724, as it's should be the same on Fedora.

Given on Windows it reports only S0, I think ACPI handling needs to be tweaked a bit, because per mjg59 Linux ACPI implementation purposely follows the Windows one.

[Hi-Angel]

Usual firmware bug, which is why systemd default to s2idle, at least on Ubuntu. Not sure why it's different for Espionage724, as it's should be the same on Fedora.

Oops, I probably linked a bit wrong post (though still somewhat relevant), it's rather this one https://mjg59.livejournal.com/85923.html The relevant quote-snips:

As I mentioned, recent versions of Linux no longer claim to support the Linux interface. There's a couple of reasons for this. Firstly, some platforms break if they're told that they're running on Linux. It's usually an entirely untested codepath. Secondly, the whole point of the _OSI method is to provide the platform with information about the interface supported by an OS. "Linux" is too coarse grained - we make no guarantees that the interface between the OS and the platform will remain constant.

…

So what kind of interface can we claim to provide? The sad truth of the matter is that most vendors will only test their code against Windows, and that as a result Windows provides the de-facto specification for OS/platform interaction. We could skip claiming to be Windows, but (in the best case) that means that the system will assume it's working with an OS that supports no advanced features. Rather than getting support for brightness changes being handled by the OS, we get stuck with SMM traps and unexpectedly missing time. If we want hardware to work to its full potential, we need to claim to be a recent version of Windows.

…

Message to vendors: If we're not behaving in the same way as Windows does, let us know. We'll fix it.

[speculatrix]

Doesn't have a middle mouse button. Try the Latitude 7490 instead!