Dell XPS 13 9350 Late 2015

belette · 2016-12-13 20:34:30

thanks frank!
Yep I realized sometimes wd15 bugs or make my XPS crazy/reboot...but it is not the main issues (even if it occurs about every 10 times I connect/disconnect).
I made some tests and here are my results :
-XPS off (completely off) / WD15 connected pre-boot, cold boot OK
-XPS on / WD15 connected for the first time OK
-XPS off (sleep) / connecting WD15, OK (it wakes up XPS correctly and screen+USB+ETH OK)
-XPS on / WD15 connected: putting into sleep/waking up OK
-XPS on / disconnecting / connecting WD15 : USB/ETH OK, external screen stays black : I need to Ctrl+Alt+F1(or any other tty) and Ctrl+Alt+F7 to get my screen appears on my external screen.
The good news is that it seems to be an issue with Xorg or a configuration problem, the bad news is that I am unable to figure out the problem… (same issue on XFCE with SLIM or with straight i3)
If you have any idea

frank604 wrote:

When wd15 bugs out, I have to remove power for 30seconds and then power back on and replug. Just a heads up. Normal use doesn't trigger this bug but constant configuration changes with xrandr makes this issue occur. Haven't nailed down steps to reproduce.

damige · 2016-12-13 22:39:41

I have done some new linux-nvme testing on 4.9 kernel, i seem to have some extra power benefits compared to previously!

Some quick comparisons, these are not scientifc and your milage may vary, do you own testing too!

XPS 13 9350 | i7-6560U | 8GB | SM951 256G NVME | Intel 8260 | FHD
Plasma 5.8.4 | X11 | Keyboard backlight off | Wifi Connected (2bars) | BT Connected (Headset) | 25% brightness | Idle 
Kernel entries: i915.enable_psr=2

Linux-Mainline 4.9.0:

  Time    User  Nice   Sys  Idle    IO  Run Ctxt/s  IRQ/s Fork Exec Exit  Watts
23:28:04   0.2   0.0   0.1  99.8   0.0    1    192     37    2    0    0   5.60 
23:28:14   0.0   0.0   0.1  99.8   0.0    1    107     22    0    0    0   6.14 
23:28:24   0.1   0.0   0.1  99.7   0.0    1    221     79    0    0    1   4.97 
23:28:34   0.1   0.0   0.1  99.8   0.0    1    115     26    0    0    0   5.00 
23:28:44   0.1   0.0   0.1  99.8   0.0    1    108     23    0    0    0   5.23 
23:28:54   0.1   0.0   0.1  99.8   0.0    1    128     18    0    0    0   5.95 
23:29:04   0.2   0.0   0.2  99.7   0.0    1    194     40    0    0    0   5.04 
-------- ----- ----- ----- ----- ----- ---- ------ ------ ---- ---- ---- ------ 
 Average   0.1   0.0   0.1  99.8   0.0  1.0  152.2   35.1  0.3  0.0  0.1   5.42 
  StdDev   0.0   0.0   0.0   0.1   0.0  0.0   44.9   19.4  0.7  0.0  0.3   0.45 
-------- ----- ----- ----- ----- ----- ---- ------ ------ ---- ---- ---- ------ 
 Minimum   0.0   0.0   0.1  99.7   0.0  1.0  106.6   18.5  0.0  0.0  0.0   4.97 
 Maximum   0.2   0.0   0.2  99.8   0.0  1.0  221.4   79.3  2.0  0.0  1.0   6.14 
-------- ----- ----- ----- ----- ----- ---- ------ ------ ---- ---- ---- ------ 
Summary:
System:   5.42 Watts on average with standard deviation 0.45 

Linux-NVME 4.9.0:

  Time    User  Nice   Sys  Idle    IO  Run Ctxt/s  IRQ/s Fork Exec Exit  Watts
23:18:00   0.1   0.0   0.1  99.7   0.1    1     79     40    0    0    3   3.70 
23:18:10   0.1   0.0   0.1  99.5   0.3    1    121     69    3    1   10   3.80 
23:18:20   0.2   0.0   0.1  99.7   0.0    1    159     44    1    0    2   3.72 
23:18:30   0.1   0.0   0.1  99.9   0.0    1     56     20    0    0    1   3.74 
23:18:40   0.1   0.0   0.2  99.7   0.1    1    205    117    1    0    0   3.69 
23:18:50   0.1   0.0   0.1  99.9   0.0    1     97     21    0    0    0   4.64 
23:19:00   0.2   0.0   0.2  99.7   0.0    1    193     43    1    0    0   3.75 
^C-------- ----- ----- ----- ----- ----- ---- ------ ------ ---- ---- ---- ------ 
 Average   0.1   0.0   0.1  99.7   0.1  1.0  130.0   50.6  0.9  0.1  2.3   3.86 
  StdDev   0.1   0.0   0.0   0.1   0.1  0.0   53.1   31.3  1.0  0.3  3.3   0.32 
-------- ----- ----- ----- ----- ----- ---- ------ ------ ---- ---- ---- ------ 
 Minimum   0.1   0.0   0.1  99.5   0.0  1.0   55.9   20.2  0.0  0.0  0.0   3.69 
 Maximum   0.2   0.0   0.2  99.9   0.3  1.0  204.9  117.4  3.0  1.0 10.0   4.64 
-------- ----- ----- ----- ----- ----- ---- ------ ------ ---- ---- ---- ------ 
Summary:
System:   3.86 Watts on average with standard deviation 0.32 

a 1.56W gain when running idle, great!

source: https://github.com/damige/linux-nvme/bl … BENCHMARKS

Edit: almost 30% power reduction when idle!

On further note it would be great if we can figure out what bugs psr=1, if anyone has an idea let us know!

Last edited by damige (2016-12-13 22:44:59)

ror191505 · 2016-12-14 08:38:21

damige wrote:

I have done some new linux-nvme testing on 4.9 kernel, i seem to have some extra power benefits compared to previously!
...
Edit: almost 30% power reduction when idle!
On further note it would be great if we can figure out what bugs psr=1, if anyone has an idea let us know!

May you add some info\script how do you collect all that stats? I want to do that on my own laptop.

Plawucht · 2016-12-14 10:36:45

ror191505 wrote:

May you add some info\script how do you collect all that stats? I want to do that on my own laptop.

I would also like to add to that, is there any way to get the grade of deterioration of the battery?

Lastly, quick question: My mkinitcpio.conf now has:

MODULES="i915 i8042 nvme"
[...]
BINARIES=""
[...]
FILES="/etc/modprobe.d/modprobe.conf"
[...]
HOOKS="base udev autodetect modconf block keyboard keymap filesystems fsck"

Everything else is commented out.
Further, /etc/modprobe.d/modprobe.conf:

blacklist psmouse

Anything that might need adding? Can I leave the "i915" out in the MODULES?

EDIT:
As belette stated, he left it out. I will try it, too.

belette wrote:

For tearing, installing compton (pacman -S compton) is giving me very good results and I got now the same performance/tearing free with modesetting DMX than my old intel drivers!

[strike]Do you use gdm or lightdm?[/strike] Yeah, no, forget about it. What are you using it with?

Last edited by Plawucht (2016-12-14 10:53:48)

OdinEidolon · 2016-12-14 11:00:38

Plawucht wrote:

ror191505 wrote:
May you add some info\script how do you collect all that stats? I want to do that on my own laptop.
I would also like to add to that, is there any way to get the grade of deterioration of the battery?

upower -d

kang · 2016-12-15 17:39:15

belette wrote:

Reply to myself for future reference and to save some time for the others.
I have deleted i915 from my mkinitcpio and it is ok.
For tearing, installing compton (pacman -S compton) is giving me very good results and I got now the same performance/tearing free with modesetting DMX than my old intel drivers!
The last and very annoying point is external screen detection with the WD15 dock.
The first time I plug the WD15 through usb type c everything is working great, I can put my laptop to sleep&wake up 10 times with no problem.
When I disconnect the usb type c and re-plug it : most of the time my external screen stay black.
Doing a xrandr shows me sometimes only my XPS pannel and nothing else, sometimes I can still see my external screen but no signal is sent even if I do a xrandr --auto.

Sorry for the late reply. I do have i915 in my mkinitcpio.
It sounds like to me that you're running the intel ddx due to performance and black screen issues. Did you check what is loaded in the Xorg X11 logs?

belette · 2016-12-16 08:52:01

I entererly deleted xf86-video-intel and put on my Xordg conf (which is not necessary normaly):

Section "Device"
        Identifier "Intel Graphics"
        Driver "modesetting"
EndSection

logs confirms I am using modesetting ddx:

[    11.915] (II) LoadModule: "modesetting"
[    11.916] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[    11.917] (II) modesetting: Driver for Modesetting Kernel Drivers: kms

kang wrote:

belette wrote:
Reply to myself for future reference and to save some time for the others.
I have deleted i915 from my mkinitcpio and it is ok.
For tearing, installing compton (pacman -S compton) is giving me very good results and I got now the same performance/tearing free with modesetting DMX than my old intel drivers!
The last and very annoying point is external screen detection with the WD15 dock.
The first time I plug the WD15 through usb type c everything is working great, I can put my laptop to sleep&wake up 10 times with no problem.
When I disconnect the usb type c and re-plug it : most of the time my external screen stay black.
Doing a xrandr shows me sometimes only my XPS pannel and nothing else, sometimes I can still see my external screen but no signal is sent even if I do a xrandr --auto.
Sorry for the late reply. I do have i915 in my mkinitcpio.
It sounds like to me that you're running the intel ddx due to performance and black screen issues. Did you check what is loaded in the Xorg X11 logs?

davidgf · 2016-12-17 20:29:08

Hey damige, I applied your patches (4.9 from github) into my fedora kernel (see https://copr.fedorainfracloud.org/coprs … d/489228/) and I don't see as much improvement as you see. While it might be my laptop (is slightly different), I am not seeing this var (/sys/class/nvme/nvme0/power/pm_qos_latency_tolerance_us) exported.

Any ideas? Maybe I didn't quite apply the patches properly. I didn't add the patch about log verbosity though.

Thanks for your help mate!

damige · 2016-12-18 23:20:56

ror191505 wrote:

May you add some info\script how do you collect all that stats? I want to do that on my own laptop.

install and run powerstat, it will show you the output for battery usage over time. (-d0 for no delay when starting)

davidgf wrote:

Hey damige, I applied your patches (4.9 from github) into my fedora kernel (see https://copr.fedorainfracloud.org/coprs … d/489228/) and I don't see as much improvement as you see. While it might be my laptop (is slightly different), I am not seeing this var (/sys/class/nvme/nvme0/power/pm_qos_latency_tolerance_us) exported.
Any ideas? Maybe I didn't quite apply the patches properly. I didn't add the patch about log verbosity though.
Thanks for your help mate!

Hi! i did a quick check on your link to see if you applied all patches, which it seems you have. Just a random guess:

Maybe your ssd controller does not support APST? Or are you using the SM951 like me?
If you have a different SSD maybe let someone try your kernel with an nvme SSD that has been confirmed to work with APST.

Did it work for you before on the 4.8 kernel?

I do not know fedora enough to know if there might be something in their build process / kernel that could conflict with these patches.

kubrick · 2016-12-21 12:12:51

Since the last bios upgrade, I'm having a few issues:

If I disable the SATA adapter in the BIOS, I get these errors at boot:

Dec 18 09:49:48 XPS16 kernel: ACPI Error: [\_SB_.PCI0.SAT1] Namespace lookup failure, AE_NOT_FOUND (20160831/dswload-210)
Dec 18 09:49:48 XPS16 kernel: ACPI Exception: AE_NOT_FOUND, During name lookup/catalog (20160831/psobject-227)
Dec 18 09:49:48 XPS16 kernel: ACPI Exception: AE_NOT_FOUND, (SSDT:IdeTable) while loading table (20160831/tbxfload-228)
Dec 18 09:49:48 XPS16 kernel: ACPI Error: 1 table load failures, 8 successful (20160831/tbxfload-246)

Also, when I switch to battery (with or without SATA enabled), I can't set min_perf_pct without waiting for a minute or so.
If I do

echo 12 | sudo tee /sys/devices/system/cpu/intel_pstate/min_perf_pct

and read the value immediately afterwards, it's set to 12, but one second later, it reverts to 100.
I have to use this command every time I switch to battery now:

while [ $(cat /sys/devices/system/cpu/intel_pstate/min_perf_pct) -ne "12" ]; do sudo tlp bat; sleep 3; done

and it usually loops a dozen times before succeeding.

Any suggestion?

frank604 · 2016-12-22 18:53:56

@kubrick are you using some power management tools like tlp, udev rules, lmt ?

slackcub · 2016-12-22 22:55:27

I'm not sure if this is the right place to put this, but I figured it's as good a place as any. I just acquired an XPS 13 9350, and I'm having issues even installing Arch (which I've done many times throughout the years). It seems to related to secureboot. According to the wiki entry for the 9350, "The installation process for Arch on the XPS 13 does not differ from any other PC.". When I boot archiso with secureboot installed, I get an error saying that my USB stick isn't signed (according to the wiki it hasn't been since archlinux-2016.06.01-dual.iso) and can't be booted. If I disable secureboot, I get it up and running, but for some reason the hard drive isn't detected. When I do an lsblk, the only entries are for the install media. I checked out the install guide and UEFI wiki pages for help, but I couldn't find anything in there.

Last edited by slackcub (2016-12-22 22:55:53)

slackcub · 2016-12-23 02:20:31

I figured it out. Was reading around some more and it turns out it's because the SATA mode in the BIOS was set to RAID not AHCI. I followed the instructions here and was able to get it switched over without needing to re-install Windows. Archiso now sees the hard drive.

Procrat · 2016-12-23 06:33:16

I read through the whole thread hoping, in vain, that someone would have found a solution for that buggy DA200 adapter.

I noticed however that no one responded to @bellete's question whether someone tried to the Thunderbolt firmware update to fix the issue. So I'm gonna try that know and if it fails, I guess I'll just send it back to Dell?

belette wrote:

USB type C Dell DA200...
Anyone who has been able to make it work correctly ?
I saw on different past posts that a lot are complaining, on my side in some previous kernel I was able to get Ethernet connectivity, then it was unstable, now on 4.7.3 no connection at all.
Speaking about HDMI, maximum resolution is 832x624 !!
Only VGA is working fine (full HD ok).
So at the end this multi adapter is only working on one slot for me it is too bad
I remember I saw on Dell support website that the firmware can be upgraded (only from Windows), someone tried and make it better after the upgrade ?
Many thanks

ror191505 · 2016-12-23 22:45:44

damige wrote:

ror191505 wrote:
May you add some info\script how do you collect all that stats? I want to do that on my own laptop.
install and run powerstat, it will show you the output for battery usage over time. (-d0 for no delay when starting)

Thank you for such cool tool!
I've performed bunch of tests for my XPS 15 (9550) using it. You can take a look into it in XPS 15 thread.

kubrick · 2016-12-24 15:32:53

Procrat wrote:

I read through the whole thread hoping, in vain, that someone would have found a solution for that buggy DA200 adapter.
I noticed however that no one responded to @bellete's question whether someone tried to the Thunderbolt firmware update to fix the issue. So I'm gonna try that know and if it fails, I guess I'll just send it back to Dell?

The Thunderbolt update didn't change anything for me. Have you tried booting with i915.enable_guc_loading=1 i915.enable_guc_submission=1 kernel parameters? Some people have reported it allowed them to achieve 1680x1050@60Hz on kernel 4.4 with these parameters.

The problem with the DA200 is quite well documented, and comes from buggy/incomplete i915 driver, the adapter in itself is fine. See:
https://bugs.freedesktop.org/show_bug.cgi?id=93578
https://bugs.freedesktop.org/show_bug.cgi?id=94567

damige · 2016-12-24 17:12:07

ror191505 wrote:

damige wrote:
ror191505 wrote:
May you add some info\script how do you collect all that stats? I want to do that on my own laptop.
install and run powerstat, it will show you the output for battery usage over time. (-d0 for no delay when starting)
Thank you for such cool tool!
I've performed bunch of tests for my XPS 15 (9550) using it. You can take a look into it in XPS 15 thread.

Some nice testing there! I did not expect the 15inch FHD screen to be so compareble to the 13 inch FHD power wise.

Just something i found out by fiddling today:
setting, 'i915.disable_power_well=0' allows 'i915.enable_psr=1' to work.
Atleast on my system.

Im not sure if there is less battery usage with this setting currently. I would need some more time for testing. Maybe you would like to test it also? :-)

Just for reference my current kernel options: "i915.modeset=1 i915.enable_psr=1 i915.enable_rc6=7 i915.enable_fbc=1 i915.semaphores=1 i915.disable_power_well=0 i915.lvds_downclock=1"

Asriel · 2016-12-24 19:08:56

New bios is released

http://www.dell.com/support/home/us/en/ … egoryId=BI

Changelog claims fixes for DA200 - maybe it will help?

bkuri · 2016-12-25 18:36:17

Just wanted to mention that the following USB-C to HDMI adapter works great with the 9350:
https://www.amazon.com/VicTsing-Female- … 00&sr=1-17

I'm using kernel 4.8.15-2-ck-skylake. HTH

Asriel · 2016-12-25 19:22:50

http://www.dell.com/en-us/shop/accessor … u=470-abmz
This adapter is working perfectly with Linux as well - but it is HDMI only, not combo dock.

belette · 2016-12-26 13:17:19

Thanks for the news, I got the hope it would solve my WD15 issues, it doesn't ...
I don't have my DA200 anymore as it was a pain so I change it you can find in old past the link of a perfectly working solution

Asriel wrote:

New bios is released
http://www.dell.com/support/home/us/en/ … egoryId=BI
Changelog claims fixes for DA200 - maybe it will help?

silles · 2016-12-28 19:10:54

When I connect my XPS-13 to a HDTV via usb-c I'm getting very strange colours:
https://goo.gl/photos/9x1M9bzp49faovae8

It always works in 800x600 and sometimes works in 1920x1080 when reverting from 800x600
Tried 2 different adaptors with no luck
I installed Windows 10 just to try if it's a hardware issue, but it worked all the time

anyone seen anything like that?

thx

kgizdov · 2016-12-29 10:21:46

@silles, that looks more like bad cable, loose connection. Are you sure your USB-C cable is good? What adaptors are you using? I had an issue like that and then I bought a good cable and it's all fine now. I got the "Plugable USB-C to DisplayPort Adapter Cable".

silles · 2016-12-29 10:55:54

@kgizdov

I'm using
1. CableCreation Gold USB 3.1 Type C (USB-C) to HDMI Adapter from Amazon
2. Hama USB 3.1 Type C (USB-C) to HDMI Adapter from a local shop

as I said before same hardware works fine on Windows

kgizdov · 2016-12-29 11:30:51

@silles, sorry, I didn't quite get what you meant originally. I can't be sure what the issue is exactly, but there are a number of things you might want to check. First, make sure the following packages are installed correctly - linux, linux-firmware, intel-ucode (this needs config as per Wiki), xf86-video-intel (if not using the kernel driver), mesa (usually you cannot go without having this installed). Then make sure your BIOS is up to date - 1.4.12. After that, make sure your GUI - Gnome, Xfce, KDE, etc is properly configured and recognises the monitor correctly. Meaning double-check that "xrandr --current" returns the same resolutions, modes and frequencies as in your chosen GUI. Check before and after plugging the cable and/or monitor. To give you an idea, this is what I get before external monitor:

Screen 0: minimum 320 x 200, current 3200 x 1800, maximum 8192 x 8192
XWAYLAND0 connected 3200x1800+0+0 290mm x 170mm
   3200x1800     59.96*+

And this is what I get with VGA external monitor over a CableMatters adaptor:

Screen 0: minimum 320 x 200, current 4880 x 1800, maximum 8192 x 8192
XWAYLAND0 connected 3200x1800+0+0 290mm x 170mm
   3200x1800     59.96*+
XWAYLAND1 connected 1680x1050+3200+0 470mm x 300mm
   1680x1050     59.85*+

Obviously, this is in Wayland, so yours might be a bit different, but you get what I mean. Check that all the listed stats agree with the ones in your monitor menu as well. For example, check the resolution (obviously), the refresh rate and the listed connection type. If they don't match, then your monitor is not correctly recognised and that will explain your issues. Any deviation from the exact settings will produce false colours, due to the nature of monitor technology and stuff like HDCP.

In the case that those settings don't match between devices, you have to investigate what is causing it. If your packages are configured correctly, then it is for sure a hardware issue. Depending on your location you can get away with trying out stuff from Amazon until it works. In the UK (and maybe around EU, not sure), you can return anything you bought online during the first 14 days for full refund, no questions asked. So I did this until I got hardware that worked. Bare in mind that having tested two adaptors in Linux and Windows, does not mean it's still not a hardware fault. It makes it very unlikely, but still can happen that bought adaptors are not working well with LInux, or with Linux and the certain modes the monitor requires, or a combination of those. The USB-C and Thunderbolt 3 standards and their execution by manufactures and Linux devs is quite shaky still.

Sorry, I cannot be of much more help, but these things seem to be very individual to the particular situation. I wish you the best of luck and hopefully it's an easy fix.

Last edited by kgizdov (2016-12-29 11:31:11)

Arch Linux

#1176 2016-12-13 20:34:30

Re: Dell XPS 13 9350 Late 2015

#1177 2016-12-13 22:39:41

Re: Dell XPS 13 9350 Late 2015

#1178 2016-12-14 08:38:21

Re: Dell XPS 13 9350 Late 2015

#1179 2016-12-14 10:36:45

Re: Dell XPS 13 9350 Late 2015

#1180 2016-12-14 11:00:38

Re: Dell XPS 13 9350 Late 2015

#1181 2016-12-15 17:39:15

Re: Dell XPS 13 9350 Late 2015

#1182 2016-12-16 08:52:01

Re: Dell XPS 13 9350 Late 2015

#1183 2016-12-17 20:29:08

Re: Dell XPS 13 9350 Late 2015

#1184 2016-12-18 23:20:56

Re: Dell XPS 13 9350 Late 2015

#1185 2016-12-21 12:12:51

Re: Dell XPS 13 9350 Late 2015

#1186 2016-12-22 18:53:56

Re: Dell XPS 13 9350 Late 2015

#1187 2016-12-22 22:55:27

Re: Dell XPS 13 9350 Late 2015

#1188 2016-12-23 02:20:31

Re: Dell XPS 13 9350 Late 2015

#1189 2016-12-23 06:33:16

Re: Dell XPS 13 9350 Late 2015

#1190 2016-12-23 22:45:44

Re: Dell XPS 13 9350 Late 2015

#1191 2016-12-24 15:32:53

Re: Dell XPS 13 9350 Late 2015

#1192 2016-12-24 17:12:07

Re: Dell XPS 13 9350 Late 2015

#1193 2016-12-24 19:08:56

Re: Dell XPS 13 9350 Late 2015

#1194 2016-12-25 18:36:17

Re: Dell XPS 13 9350 Late 2015

#1195 2016-12-25 19:22:50

Re: Dell XPS 13 9350 Late 2015

#1196 2016-12-26 13:17:19

Re: Dell XPS 13 9350 Late 2015

#1197 2016-12-28 19:10:54

Re: Dell XPS 13 9350 Late 2015

#1198 2016-12-29 10:21:46

Re: Dell XPS 13 9350 Late 2015

#1199 2016-12-29 10:55:54

Re: Dell XPS 13 9350 Late 2015

#1200 2016-12-29 11:30:51

Re: Dell XPS 13 9350 Late 2015

Board footer