Start a Conversation

Unsolved

This post is more than 5 years old

13558

July 15th, 2016 21:00

XPS 15, Linux no longer resumes from sleep w/ Bios 1.2.10

I'm running Ubuntu 16.04 on an XPS 15 9550, dual-booting with Windows 10.  It was working reasonably well straight out of the box, and even better after using some of the tips here.   I recently upgraded bios to 1.2.10 (Jul 5th release) in the hopes that it would fix the inability to drive a mini-display-port adapter from the usb-c port (neither Windows nor Linux were able to).  Now Ubuntu is no longer able to resume from sleep.  It works fine when booted, and goes to sleep when the lid closes, but on wakeup the keyboard lights come on and no other visible activity takes place.

Are there any bios settings I should try, or just try reinstalling 1.2.0? 

350 Posts

July 28th, 2016 14:00

Your link doesn't work.

Very disappointed that something reported in a support forum doesn't get tracked.  

Sorry for the slow response. I was out the past couple of days. I fixed the link. the "-wip" needed to be removed.

26 Posts

August 8th, 2016 13:00

Jared, do you know if this is this an issue that anyone at Dell has been able to reproduce? It seems from my end (and those who have participated in this thread) to be a consistent annoyance and regression of a fairly standard laptop function. However, it also seems like several of us are running different distributions and OS configurations than the default Ubuntu installation on the 5510. For example, I am running Gentoo and have configured my laptop for legacy booting instead of UEFI.

In the event that users do not experience this bug under the 1.02.10 BIOS and the default Ubuntu installation, do you think you could provide us some information on what mechanism is being used to suspend the laptop successfully and any specifics such as pre and post scripts that are being run, modules that are being re-inserted, and whether Dell has customized anything with respect to suspend-to-RAM and the laptop's backlight?

Otherwise, as I seem to get the impression from other threads that project Sputnik laptops are "dogfooded" by your team members, could you confirm whether these are issues that have been identified and whether they are in the process of being addressed? Thanks for your help.

Also, in the event that Canonical is indeed involved with tackling this issue, is there anywhere that we might be able to track that progress?

10 Posts

August 23rd, 2016 15:00

A new BIOS is out today: 1.2.13

Has anybody tested it? Does it fix the black screen after resume issue?

21 Posts

August 23rd, 2016 17:00

Just tested it.

My experience is it failed to resume after suspend twice out of 3 trials. The only time it actually resumed, brightness was max and could not be decreased.

Don't know whether it is an improvement over the previous update which i skipped due to negative feedback in this forum.

36 Posts

August 23rd, 2016 18:00

Yep, it fails.  This BIOS is also bad.  Dell, don't you test this stuff?

Here's what I found with the latest:

- Resume fails to enable display.  This happens frequently.  Exactly the same as 1.2.10.

As with all prior BIOS versions:

- Thunderbolt port does not work after resume.  The hotplug event is never delivered.

- Anything plugged into the thunderbolt port after resume will inhibit the package from reaching any C-State greater than C3.  This causes great power drain, and great heat.  Doubt it's really safe, but unlikely to cause a problem since, as noted, the port doesn't actually *work* after a resume. So you are unlikely to plug anything in, I guess?  

- A suspend resume cycle corrupts the BIOS memory.  This can be readily seen by running dmidecode before and after a resume.

This is really a mess.  Not only has the thunderbolt port never worked, Dell is not making *any* forward progress and seems to be taking strides backwards.

26 Posts

August 24th, 2016 10:00

Upon installation of the A12/1.2.13 BIOS I'm seeing the

acpi PNP0C0B:00: Failed to set initial power state
acpi PNP0C0B:01: Failed to set initial power state
acpi PNP0C0B:01: Failed to set initial power state
acpi PNP0C0B:03: Failed to set initial power state
acpi PNP0C0B:04: Failed to set initial power state

errors again at boot.  I saw them in A07 but they went away in A10/1.2.10 but have returned now.

Otherwise, I can echo the same issues with an inability to resume from suspend and also the question of testing.  I'm guessing that the most formal testing probably occurs with Windows and not Ubuntu, but I'd imagine that we'd at least receive confirmation and/or feedback from Sputnik team members and others at Dell who are using this machines under Ubuntu (or another distribution).

2 Posts

September 1st, 2016 07:00

Just tried the 1.2.13 bios yesterday and can confirm the resume issue. On resume the PC wakes up with the screen turned off. The screen came on after turning the backlight to full brightness, but after that you can't adjust brightness. It's either off or full brightness until the next reboot.

So I have reverted back to 1.2.0. This has issues with wifi not initializing after resume though. Come on Dell can't you spare just a minimal number of resources to test this fine piece of hardware with Linux???

10 Posts

September 7th, 2016 13:00

Dell seems to have rolled back the latest BIOS for the XPS 15.

In their website they list the latest bios as 1.2.10 from Jul 5th. There's no mention of 1.2.13.

Progres??

4 Posts

September 13th, 2016 17:00

BIOS 1.2.14 is now listed w/ a Sept 9 release date.  Has anyone tried this release?

2 Posts

September 14th, 2016 01:00

Nope, still the same.

Also noticing that with latest kernel (4.8) it doesn't go to sleep.

2 Posts

September 14th, 2016 05:00

Yes, I tried it. It still has the issue.

2 Posts

September 14th, 2016 09:00

Having the same problem on Dell Precision 5510 with 1.2.14

5 Posts

September 18th, 2016 03:00

10 Posts

September 18th, 2016 18:00

Cool, that works. Thanks cregganna!


The funny thing is that the register was also set to 0 on suspend/resume with an older BIOS. I was running 1.2.0, which did not suffer from the black display problem on resume, and I saw the register was also set to zero without causing any trouble. It makes me wonder now if this is a linux driver bug and not a BIOS bug.

7 Posts

September 19th, 2016 15:00

I've used the hsamed script (systemd) and put it in /lib/systemd/system-sleep/fixbacklight and can suspend/resume without the monitor turned off (BIOS 1.2.14). Thanks a lot for everyone involved!

On caveat I still have is that on lowest brightness setting the monitor becames so-dark that I need a flashlight (or I suspect direct sunlight at midday) to see what's in the monitor. Is that an intended functionality from Dell?

No Events found!

Top