Arch on New XPS 15(Late 2015)

OdinEidolon · 2017-01-02 14:22:18

sPHERE wrote:

I know.
I tried repetedly some times and about 70% of the times out of say...20, it worked.
But no thermal controls tho. Do you find any other sensors except for the "coretemp" module in "sudo sensors-detect"?
Here is my last:
[petterk@petterk-nix ~]$ sudo smbios-keyboard-ctl -g -v
Helper function to print current status of keyboard illumination
Current status of KeyBoard Illumination setting on your system:
-------------------------------------------------------------------
Get Keyboard Illumination Configuration: res[smi.cbRES1]=0x-2,res[smi.cbRES2]=0x0,res[smi.cbRES3]=0x0
Info: Unable to fetch Keyboard Illumination Configuration information: Feature Not supported on this system
Get Keyboard illumination Current State: res[smi.cbRES1]=0x0,res[smi.cbRES2]=0xA070040,res[smi.cbRES3]=0x0
Configured mode state:
Auto: Input-activity-based On (illumination level 50%); input-activity based Off
Your Keyboard will illumination on:
Any Keystroke
Touchpad activity
Pointing stick
Keyboard illumination timeout has bee set at: 10 Seconds
Note - Timeout value of 0 means its always ON
Current setting of ALS value that turns the light on or off: 0
Current ALS Reading : 0
Current keyboard light level : 0
I just also compared dmi and acpi-tables from all kinds of bootparameters, and currently the DSDT is looking for these acpi_osi=name and acpi_os_name=name .
This is a list of the OS's and what difference they make in terms of code.
Simply follow this guide (I used acpidump && acpixtract for this with iasl):

And also, the most prevalent error I usually get is this:
ACPI Error: [\_SB_.PCI0.LPCB.H_EC.ECAV] Namespace lookup failure, AE_NOT_FOUND (20160422/psargs-359)
and the ones where FNCL and FN01 etc namespaces does not add up. This is because of errors written in the actual files.
Because this is my actual DSDT table:
https://paste.pound-python.org/show/SM5 … QnDd0nIla/
and after extraction compiles into 15 SSDT tables and 1 DSDT table. The corresponding output defined based on _OSI and OSNAME (as in acpi_os_name) is actually quite crucial for it to function, and obviously, there has been major imporvements according to my DMESG in regards to errors atleast. This is from the DSDT table.
But I DO know for a fact that the thermals are quite advanced as you will see yourselves in the tables with accpidump, extract and iasl.

Sorry, which guide? Maybe you forgot a link?
Sensors-detect can detect only coretemp and 1 acpix sensor, but ksysguard can find more:
http://i.imgur.com/3HHiMIV.png

EDIT: these are in /sys/class/thermal

Last edited by OdinEidolon (2017-01-02 14:25:28)

OdinEidolon · 2017-01-15 11:34:01

Is anyone having systemd-backlight@leds:dell::kbd_backlight.service failing once every two or three reboots? When this happens, I seem to be unable to control keyboard brightness from KDE's keyboard brightness control, I am forced to use F10 (not a big deal). Also, I find sometimes the OS seems to reset (t boot) the keyboard brightness to 0 despite being set to the middle level as default in the BIOS, which is more annoying.

OdinEidolon · 2017-01-16 08:36:44

Has anyone done any tests on how different scaling drivers and governors work on our XPS? Pstate/cpufreq ondemand/performance/schedutil?

analbeard · 2017-01-17 10:02:08

Does anyone have bluetooth working successfully? I've recently got a late 2016 9550 and although wifi works, I can't get bluetooth to start:

[NEW] Controller 9C:B6:D0:D3:9C:70 laptop.domain.com [default]
[bluetooth]# scan on
Failed to start discovery: org.bluez.Error.NotReady

Jan 16 22:02:39 laptop.domain.com bluetoothd[2658]: Bluetooth management interface 1.13 initialized
Jan 16 22:02:39 laptop.domain.com bluetoothd[2658]: Failed to obtain handles for "Service Changed" characteristic
Jan 16 22:02:39 laptop.domain.com bluetoothd[2658]: Sap driver initialization failed.
Jan 16 22:02:39 laptop.domain.com bluetoothd[2658]: sap-server: Operation not permitted (1)
Jan 16 22:03:14 laptop.domain.com bluetoothd[2658]: Endpoint registered: sender=:1.42 path=/MediaEndpoint/A2DPSource
Jan 16 22:03:14 laptop.domain.com bluetoothd[2658]: Endpoint registered: sender=:1.42 path=/MediaEndpoint/A2DPSink
Jan 16 22:03:15 laptop.domain.com bluetoothd[2658]: Failed to set mode: Failed (0x03)
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: Endpoint unregistered: sender=:1.42 path=/MediaEndpoint/A2DPSource
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: Endpoint unregistered: sender=:1.42 path=/MediaEndpoint/A2DPSink
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: Failed to obtain handles for "Service Changed" characteristic
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: Sap driver initialization failed.
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: sap-server: Operation not permitted (1)
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: Endpoint registered: sender=:1.42 path=/MediaEndpoint/A2DPSource
Jan 17 08:33:07 laptop.domain.com bluetoothd[2658]: Endpoint registered: sender=:1.42 path=/MediaEndpoint/A2DPSink

02:00.0 Network controller: Qualcomm Atheros QCA6174 802.11ac Wireless Network Adapter (rev 32)
	Subsystem: Bigfoot Networks, Inc. Device 1535
	Flags: bus master, fast devsel, latency 0, IRQ 135
	Memory at dd200000 (64-bit, non-prefetchable) [size=2M]
	Capabilities: <access denied>
	Kernel driver in use: ath10k_pci
	Kernel modules: ath10k_pci

[    4.311423] ath10k_pci 0000:02:00.0: Direct firmware load for ath10k/pre-cal-pci-0000:02:00.0.bin failed with error -2
[    4.311437] ath10k_pci 0000:02:00.0: Direct firmware load for ath10k/cal-pci-0000:02:00.0.bin failed with error -2
[    4.311675] ath10k_pci 0000:02:00.0: Direct firmware load for ath10k/QCA6174/hw3.0/firmware-5.bin failed with error -2
[    4.311678] ath10k_pci 0000:02:00.0: could not fetch firmware file 'ath10k/QCA6174/hw3.0/firmware-5.bin': -2

OdinEidolon · 2017-01-17 10:13:29

analbeard wrote:

Does anyone have bluetooth working successfully? I've recently got a late 2016 9550 and although wifi works, I can't get bluetooth to start:

I did try for a bit following the XPS13 wiki page when I got the laptop, but without success. Since I do not use BT devices I gave up pretty soon.

analbeard · 2017-01-17 10:34:21

OdinEidolon wrote:

analbeard wrote:
Does anyone have bluetooth working successfully? I've recently got a late 2016 9550 and although wifi works, I can't get bluetooth to start:
I did try for a bit following the XPS13 wiki page when I got the laptop, but without success. Since I do not use BT devices I gave up pretty soon.

Damnit I use it for my headphones all day at work so if anyone has some suggestions I'll gladly try them out and report back!

OdinEidolon · 2017-01-17 10:35:56

analbeard wrote:

OdinEidolon wrote:
analbeard wrote:
Does anyone have bluetooth working successfully? I've recently got a late 2016 9550 and although wifi works, I can't get bluetooth to start:
I did try for a bit following the XPS13 wiki page when I got the laptop, but without success. Since I do not use BT devices I gave up pretty soon.
Damnit :( I use it for my headphones all day at work so if anyone has some suggestions I'll gladly try them out and report back!

You have the Dell WiFi card, right? Did you try the steps in the XPS13 wiki?

analbeard · 2017-01-17 10:48:42

OdinEidolon wrote:

You have the Dell WiFi card, right? Did you try the steps in the XPS13 wiki?

Sorry, can you point me towards the right page? I can see 4 pages relating to XPS13 variants and I can't see anything relevant to me (I may be missing something obvious!)

OdinEidolon · 2017-01-17 10:52:59

analbeard wrote:

OdinEidolon wrote:
You have the Dell WiFi card, right? Did you try the steps in the XPS13 wiki?
Sorry, can you point me towards the right page? I can see 4 pages relating to XPS13 variants and I can't see anything relevant to me (I may be missing something obvious!)

Sure:
https://wiki.archlinux.org/index.php/De … #Bluetooth

If you have the Dell card (which is Broadcom) then there are a few steps to take. For me, they did not work. The Dell cards in the XPS13 and 15 should be slightly different tho, so you may have to download a different file. I do not remember what I tried tbh.

Last edited by OdinEidolon (2017-01-17 10:53:24)

analbeard · 2017-01-17 11:02:41

OdinEidolon wrote:

analbeard wrote:
OdinEidolon wrote:
You have the Dell WiFi card, right? Did you try the steps in the XPS13 wiki?
Sorry, can you point me towards the right page? I can see 4 pages relating to XPS13 variants and I can't see anything relevant to me (I may be missing something obvious!)
Sure:
https://wiki.archlinux.org/index.php/De … #Bluetooth
If you have the Dell card (which is Broadcom) then there are a few steps to take. For me, they did not work. The Dell cards in the XPS13 and 15 should be slightly different tho, so you may have to download a different file. I do not remember what I tried tbh.

Right sorry, I'm with you - I~~'ve got a Qualcomm bluetooth card:~~ actually I'm an idiot - I ran that on the wrong machine - too many terminals open.

Last edited by analbeard (2017-01-17 11:07:22)

OdinEidolon · 2017-01-17 11:06:40

analbeard wrote:

OdinEidolon wrote:
analbeard wrote:
Sorry, can you point me towards the right page? I can see 4 pages relating to XPS13 variants and I can't see anything relevant to me (I may be missing something obvious!)
Sure:
https://wiki.archlinux.org/index.php/De … #Bluetooth
If you have the Dell card (which is Broadcom) then there are a few steps to take. For me, they did not work. The Dell cards in the XPS13 and 15 should be slightly different tho, so you may have to download a different file. I do not remember what I tried tbh.
Right sorry, I'm with you - I've got a Qualcomm bluetooth card:
Bus 004 Device 003: ID 0cf3:3004 Qualcomm Atheros Communications AR3012 Bluetooth 4.0

Weird, that's not the default card right? Or are you on the newer Killer wifi model?

analbeard · 2017-01-17 11:08:40

OdinEidolon wrote:

analbeard wrote:
OdinEidolon wrote:
Sure:
https://wiki.archlinux.org/index.php/De … #Bluetooth
If you have the Dell card (which is Broadcom) then there are a few steps to take. For me, they did not work. The Dell cards in the XPS13 and 15 should be slightly different tho, so you may have to download a different file. I do not remember what I tried tbh.
Right sorry, I'm with you - I've got a Qualcomm bluetooth card:
Bus 004 Device 003: ID 0cf3:3004 Qualcomm Atheros Communications AR3012 Bluetooth 4.0
Weird, that's not the default card right? Or are you on the newer Killer wifi model?

See my edit - i'm an idiot and ran it on the wrong machine. Just gotta rebuild the kernel on the new laptop as I changed mkinicpio.conf and broke something - I'll get the right output after I've done that. I believe I have Killer in the new one though.

analbeard · 2017-01-17 11:36:46

analbeard wrote:

OdinEidolon wrote:
Weird, that's not the default card right? Or are you on the newer Killer wifi model?
See my edit - i'm an idiot and ran it on the wrong machine. Just gotta rebuild the kernel on the new laptop as I changed mkinicpio.conf and broke something - I'll get the right output after I've done that. I believe I have Killer in the new one though.

Right, let's try that again from the right machine:

Bus 001 Device 003: ID 0cf3:e300 Qualcomm Atheros Communications

OdinEidolon · 2017-01-17 12:17:03

analbeard wrote:

analbeard wrote:
OdinEidolon wrote:
Weird, that's not the default card right? Or are you on the newer Killer wifi model?
See my edit - i'm an idiot and ran it on the wrong machine. Just gotta rebuild the kernel on the new laptop as I changed mkinicpio.conf and broke something - I'll get the right output after I've done that. I believe I have Killer in the new one though.
Right, let's try that again from the right machine:
Bus 001 Device 003: ID 0cf3:e300 Qualcomm Atheros Communications

I fear I can't help you with that...

analbeard · 2017-01-17 12:27:44

OdinEidolon wrote:

analbeard wrote:
analbeard wrote:
See my edit - i'm an idiot and ran it on the wrong machine. Just gotta rebuild the kernel on the new laptop as I changed mkinicpio.conf and broke something - I'll get the right output after I've done that. I believe I have Killer in the new one though.
Right, let's try that again from the right machine:
Bus 001 Device 003: ID 0cf3:e300 Qualcomm Atheros Communications
I fear I can't help you with that...

Not to worry, any input is appreciated!

analbeard · 2017-01-19 08:48:29

Just in case anyone else needs this, I solved my own problem - I missed installing bluez-firmware so of course it wasn't going to work.

OdinEidolon · 2017-01-19 08:53:51

analbeard wrote:

Just in case anyone else needs this, I solved my own problem - I missed installing bluez-firmware so of course it wasn't going to work.

Weird that bluez did not bring that in as a dependency. Is this a bug?

analbeard · 2017-01-19 09:38:45

OdinEidolon wrote:

analbeard wrote:
Just in case anyone else needs this, I solved my own problem - I missed installing bluez-firmware so of course it wasn't going to work.
Weird that bluez did not bring that in as a dependency. Is this a bug?

Seemingly it's only necessary for certain controllers:

Description     : Firmwares for Broadcom BCM203x and STLC2300 Bluetooth chips

It wasn't in my package list when I dumped it on my old laptop as it uses a different controller.

OdinEidolon · 2017-01-30 13:14:22

I'm trying to understand why my laptop seems to use more power than needed at idle.
Can we do a survey, running the following:

sudo powerstat -d 0 -f -g -c

?

My setup is:

i915 powersaving options: enable_psr=2 enable_rc6=7 enable_fbc=1 semaphores=1 lvds_downclock=1 enable_guc_loading=1 enable_guc_submission=1
snd_hda_intel powersaving options: power_save=1 power_save_controller=1
iwlwifi powersaving options: power_save=1 d0i3_disable=0 uapsd_disable=0
iwldvm powersaving options: force_cam=0
wireless ON, associated to a nearby AP
bluetooth OFF
pcie_aspm=force
sysctl options: kernel.sysrq = 1 kernel.nmi_watchdog = 0 vm.swappiness = 0
SATA ALPM min_power and PCI-E ASPM powersave
CPU: no turbo and max_perf_pctl 80
all auto: /sys/bus/*/devices/*/power/control; /sys/bus/*/devices/*/ata*/power/control; /sys/block/*/device/power/control
USB autosuspend
powertop: all "good"
min brightness
wifi powersave ON
SDDM login screen active, but running from TTY

And my average power is 5.66W at idle, lowest brightness, wifi ON. It looks like a lot to me!

Powerstat results:
http://pastebin.com/Nia4Jtcq

Last edited by OdinEidolon (2017-01-30 13:26:47)

OdinEidolon · 2017-02-19 11:47:14

I have 32GB or RAM installed and power consumption while suspended is HUGE. Last time I checked the battery (56Wh) went down by ~20% in 15 hours of suspend: that's almost an entire Watt when suspended.

Can anybody confirm this?

Last edited by OdinEidolon (2017-02-19 11:47:29)

ewaller · 2017-02-19 16:08:10

Seems a bit high. Let's explore some worse case numbers.

I would have expected a one, may two hundred mW. You do have a fair amount of RAM, so the time the processor will need to be awake to refresh that RAM will be a bit longer than laptops that are more memory challenged. I would expect the refresh on a typical machine to take on the order of 250nS, and that the machine need to do that every 64mS (roughly 16 times per second). Your RAM might take twice that long. We still should be talking on the order of 300mW.

I honestly have no set-of-the-pants feeling for PCI-e bus power use. I have a hunch that there is some overhead there in order to maintain state. Especially if there is a GPU that has its own RAM, and if that RAM is being maintained. Again, this is all speculation.

The other culprit is going to be your USB bus. During sleep, devices on that bus are still provided 5V and they are told to behave themselves. This is done so that the bus does not need to be re enumerated after wakeup. It also provides for wake-on LAN, or other real world events that need be monitored. Not all USB devices are well behaved -- especially phones, MP3 players, or other battery powered devices that charge their internal batteries from the host. Also, remember that not all USB devices are external. Check what is on your system with usb -vv and see what they report in terms of low power state.

WLANs might consume more than one would expect during a sleep, especially if they are using some sort of WoL. You might check your logs and see what the system does as it comes out of sleep; for example, does it have to reassociate with an AP, or was it already there?

OdinEidolon · 2017-02-19 16:16:43

ewaller wrote:

Seems a bit high. Let's explore some worse case numbers.
I would have expected a one, may two hundred mW. You do have a fair amount of RAM, so the time the processor will need to be awake to refresh that RAM will be a bit longer than laptops that are more memory challenged. I would expect the refresh on a typical machine to take on the order of 250nS, and that the machine need to do that every 64mS (roughly 16 times per second). Your RAM might take twice that long. We still should be talking on the order of 300mW.
I honestly have no set-of-the-pants feeling for PCI-e bus power use. I have a hunch that there is some overhead there in order to maintain state. Especially if there is a GPU that has its own RAM, and if that RAM is being maintained. Again, this is all speculation.
The other culprit is going to be your USB bus. During sleep, devices on that bus are still provided 5V and they are told to behave themselves. This is done so that the bus does not need to be re enumerated after wakeup. It also provides for wake-on LAN, or other real world events that need be monitored. Not all USB devices are well behaved -- especially phones, MP3 players, or other battery powered devices that charge their internal batteries from the host. Also, remember that not all USB devices are external. Check what is on your system with usb -vv and see what they report in terms of low power state.
WLANs might consume more than one would expect during a sleep, especially if they are using some sort of WoL. You might check your logs and see what the system does as it comes out of sleep; for example, does it have to reassociate with an AP, or was it already there?

Thanks for the very detailed answer! I love to be able to see the calculations.
WoL/WoWlan is disabled. I'll do a test without USB devices (I usually have a small USB drive plugged in) and with PCI-e power management set to low power. The GPU is disabled, so I don't think that might be a culprit. When waking for sleep, the Wlan card must reassociate, and it takes a few seconds to do so, which is something I expect to be normal.

Unfortunately I do not know whether this happened also before the RAM upgrade (from 8 to 32GB). I did not notice it.

OdinEidolon · 2017-02-20 07:43:05

OdinEidolon wrote:

ewaller wrote:
Seems a bit high. Let's explore some worse case numbers.
I would have expected a one, may two hundred mW. You do have a fair amount of RAM, so the time the processor will need to be awake to refresh that RAM will be a bit longer than laptops that are more memory challenged. I would expect the refresh on a typical machine to take on the order of 250nS, and that the machine need to do that every 64mS (roughly 16 times per second). Your RAM might take twice that long. We still should be talking on the order of 300mW.
I honestly have no set-of-the-pants feeling for PCI-e bus power use. I have a hunch that there is some overhead there in order to maintain state. Especially if there is a GPU that has its own RAM, and if that RAM is being maintained. Again, this is all speculation.
The other culprit is going to be your USB bus. During sleep, devices on that bus are still provided 5V and they are told to behave themselves. This is done so that the bus does not need to be re enumerated after wakeup. It also provides for wake-on LAN, or other real world events that need be monitored. Not all USB devices are well behaved -- especially phones, MP3 players, or other battery powered devices that charge their internal batteries from the host. Also, remember that not all USB devices are external. Check what is on your system with usb -vv and see what they report in terms of low power state.
WLANs might consume more than one would expect during a sleep, especially if they are using some sort of WoL. You might check your logs and see what the system does as it comes out of sleep; for example, does it have to reassociate with an AP, or was it already there?
Thanks for the very detailed answer! I love to be able to see the calculations.
WoL/WoWlan is disabled. I'll do a test without USB devices (I usually have a small USB drive plugged in) and with PCI-e power management set to low power. The GPU is disabled, so I don't think that might be a culprit. When waking for sleep, the Wlan card must reassociate, and it takes a few seconds to do so, which is something I expect to be normal.
Unfortunately I do not know whether this happened also before the RAM upgrade (from 8 to 32GB). I did not notice it.

Ok, did a test tonight. With all powersaving activated and no USB devices connected, the laptop used 10.53Wh of battery, as reported by upower, in exactly 9 hours and 40 minutes. That's more than a Watt! Something must be wrong...

moesi · 2017-04-01 20:02:10

Hi!

I have some problems with the tempreture when just surfing with Firefox or Chromium. I did a clean install some weeks ago, and i maybe missed some configuration (compared to installation before)... well the battery-drain is also very high. I have bumblee-setup as in the wiki explained. I also tried acpi call for disabling the gpu. I also use the bootparamters from the wiki site. https://wiki.archlinux.org/index.php/Dell_XPS_15_(9550) But nothing helped.

It is really annoying. Maybe someone can help me out a bit?

FHD-i7

greeting
Moesi

OdinEidolon · 2017-04-01 20:32:13

moesi wrote:

Hi!
I have some problems with the tempreture when just surfing with Firefox or Chromium. I did a clean install some weeks ago, and i maybe missed some configuration (compared to installation before)... well the battery-drain is also very high. I have bumblee-setup as in the wiki explained. I also tried acpi call for disabling the gpu. I also use the bootparamters from the wiki site. https://wiki.archlinux.org/index.php/Dell_XPS_15_(9550) But nothing helped.
It is really annoying. Maybe someone can help me out a bit?
FHD-i7

greeting
Moesi

What is the cause of the high power usage, according to Powertop?
Are you sure the GPU is off? Check /proc/acpi/bbswitch

Arch Linux

#276 2017-01-02 14:22:18

Re: Arch on New XPS 15(Late 2015)

#277 2017-01-15 11:34:01

Re: Arch on New XPS 15(Late 2015)

#278 2017-01-16 08:36:44

Re: Arch on New XPS 15(Late 2015)

#279 2017-01-17 10:02:08

Re: Arch on New XPS 15(Late 2015)

#280 2017-01-17 10:13:29

Re: Arch on New XPS 15(Late 2015)

#281 2017-01-17 10:34:21

Re: Arch on New XPS 15(Late 2015)

#282 2017-01-17 10:35:56

Re: Arch on New XPS 15(Late 2015)

#283 2017-01-17 10:48:42

Re: Arch on New XPS 15(Late 2015)

#284 2017-01-17 10:52:59

Re: Arch on New XPS 15(Late 2015)

#285 2017-01-17 11:02:41

Re: Arch on New XPS 15(Late 2015)

#286 2017-01-17 11:06:40

Re: Arch on New XPS 15(Late 2015)

#287 2017-01-17 11:08:40

Re: Arch on New XPS 15(Late 2015)

#288 2017-01-17 11:36:46

Re: Arch on New XPS 15(Late 2015)

#289 2017-01-17 12:17:03

Re: Arch on New XPS 15(Late 2015)

#290 2017-01-17 12:27:44

Re: Arch on New XPS 15(Late 2015)

#291 2017-01-19 08:48:29

Re: Arch on New XPS 15(Late 2015)

#292 2017-01-19 08:53:51

Re: Arch on New XPS 15(Late 2015)

#293 2017-01-19 09:38:45

Re: Arch on New XPS 15(Late 2015)

#294 2017-01-30 13:14:22

Re: Arch on New XPS 15(Late 2015)

#295 2017-02-19 11:47:14

Re: Arch on New XPS 15(Late 2015)

#296 2017-02-19 16:08:10

Re: Arch on New XPS 15(Late 2015)

#297 2017-02-19 16:16:43

Re: Arch on New XPS 15(Late 2015)

#298 2017-02-20 07:43:05

Re: Arch on New XPS 15(Late 2015)

#299 2017-04-01 20:02:10

Re: Arch on New XPS 15(Late 2015)

#300 2017-04-01 20:32:13

Re: Arch on New XPS 15(Late 2015)

Board footer