You are not logged in.
I am running a custom built kernel 4.11 (acpi_rev_override=1) with a Samsung SSD and power consumption is about 4-5 W with screen dimmed to 5%.
Things are running quite smooth, but I am experiencing screen flickering from time to time. Has this been observed and solved before? I think its not related to the kernel as it has been there before my recent updates, but it is starting to annoy me.
Edit: It is not the 4K panel and setting brightness to 100% doesn't seem to fix it.
I had the same problem with screen-flickering. It appeared to be caused by
enable_fbc=1 enable_psr=1 disable_power_well=0
removing those parameters fixed the flickering for me. I'm not having any issues with power consumption - battery lasts a good 12+ hours with the nvidia card off, 1080p screen
Offline
Kernel 4.11 and 4.10.13 prevent my system from booting. 4.10.10 is okay.
I'm running:
Systemd boot
NVE harddrives
Xorg
GDM
Gnome 3.xxWhen I run these kernels GDM does start. For a few seconds the computer will take keyboard input. I can change my terminal with Ctrl+Alt+F2 and get into a terminal, from there everything will work fine in that terminal. If I don't change into a terminal, after a few seconds mouse clicks and all keyboard input is ignored. The Mouse will continue to move just fine, the backlight on the keyboard acts correctly.
I don't even know where to start debugging. As a temp "fix" I'm running linux-lts kernel, which runs fine.
Here is my boot config:
https://pastebin.com/4Rn5nkbhHere is a a journalctl -b of a failed boot (Gdm debug is on):
https://pastebin.com/ykuzkpP8Here is a journalctl -b of a successful boot:
https://pastebin.com/pb1TXhbXAnyone else hit by this? Any other information I can provide? Thanks.
Sounds like an X problem, you might have more luck looking at the Xorg log. The kernel logs show that nouveau is loading and resulting in lots of errors. Unless you intend to use nouveau its probably best to blacklist it.
Offline
Hi Daniel, yes from the outputs of the nvme-cli seems the samsung ssd goes to the deeper states very fast.
However my power consumption on idle does never go below 9.5 W (when I am really idle) what numbers do you now report with this new kernel? Also I just installed archlinux on the ssd and I found out that the power consumption is the same as the case arch is installed in the usb stick.. still get 9.5 W.
Also do you have any idea or clue for what we can do to further optimize this power discharge? in particular to solve this problem of the PCI devices stuck at 100%?
Mmm yes here's mine. Maybe in a year or two I'll upgrade the SSD
nvme get-feature -f 0x0c -H /dev/nvme0
get-feature:0xc (Autonomous Power State Transition), Current value:0x000001
Autonomous Power State Transition Enable (APSTE): Enabled
Auto PST Entries .................
Entry[ 0]
.................
Idle Time Prior to Transition (ITPT): 1500 ms
Idle Transition Power State (ITPS): 3
.................
Entry[ 1]
.................
Idle Time Prior to Transition (ITPT): 1500 ms
Idle Transition Power State (ITPS): 3
.................
Entry[ 2]
.................
Idle Time Prior to Transition (ITPT): 1500 ms
Idle Transition Power State (ITPS): 3
.................
Entry[ 3]
.................
Idle Time Prior to Transition (ITPT): 8500 ms
Idle Transition Power State (ITPS): 4
.................
Entry[ 4]
.................
Idle Time Prior to Transition (ITPT): 0 ms
Idle Transition Power State (ITPS): 0
.................
...
I'm not convinced that those 100% device usages are a problem - It was the same for me on 4.10.x when my idle power consumption was 5-6W and from what I understand that's comparable to Windows and quite good for a 4K 15' quad core laptop. I don't really know what the percentage means - Does it mean the device is "on" or that the pci link is "active" etc? Certainly the card reader should be as asleep as possible given that its empty.
Is 9.5W with screen brightness at minimum? I always turn it down to minimum to test. 8-9 W is common for me on 4.11, although rarely I get ~7W. On 4.10 it was 5-6W consistently and the only difference I found in powertop was the increased time in CPU package C3 state vs C7.
Offline
Hello, is anyone having issues with the frequency scaling after resuming from suspend? I get very high CPU frequencies after resuming using the intel_pstate, if I disable it via kernel command line, the acpi module can't scale up the frequency and it stays at 800Mhz forever, and of course I lose the turbo boost. Any ideas?
Offline
I'm not convinced that those 100% device usages are a problem - It was the same for me on 4.10.x when my idle power consumption was 5-6W and from what I understand that's comparable to Windows and quite good for a 4K 15' quad core laptop. I don't really know what the percentage means - Does it mean the device is "on" or that the pci link is "active" etc? Certainly the card reader should be as asleep as possible given that its empty.
Is 9.5W with screen brightness at minimum? I always turn it down to minimum to test. 8-9 W is common for me on 4.11, although rarely I get ~7W. On 4.10 it was 5-6W consistently and the only difference I found in powertop was the increased time in CPU package C3 state vs C7.
Its 7.07 W on idle with the screen at minimum brightness. I also would like to understand how to get where you was with 4.10, it would be amazing and definitive to have a power consumption of 5-6 W on idle. From the current situation seems that we are losing 3 W somewhere.
Offline
@Naveg Thanks, I think your solution has helped. I have not observed any more screen flickering since I have removed parameters
enable_fbc=1 enable_psr=1 disable_power_well=0
Last edited by zzzad (2017-05-29 20:31:51)
Offline
Just an update about current state for newcomers
Dell XPS 9560
00:00.0 Host bridge: Intel Corporation Device 5910 (rev 05)
00:01.0 PCI bridge: Intel Corporation Skylake PCIe Controller (x16) (rev 05)
00:02.0 VGA compatible controller: Intel Corporation Device 591b (rev 04)
00:04.0 Signal processing controller: Intel Corporation Skylake Processor Thermal Subsystem (rev 05)
00:14.0 USB controller: Intel Corporation Sunrise Point-H USB 3.0 xHCI Controller (rev 31)
00:14.2 Signal processing controller: Intel Corporation Sunrise Point-H Thermal subsystem (rev 31)
00:15.0 Signal processing controller: Intel Corporation Sunrise Point-H Serial IO I2C Controller #0 (rev 31)
00:15.1 Signal processing controller: Intel Corporation Sunrise Point-H Serial IO I2C Controller #1 (rev 31)
00:16.0 Communication controller: Intel Corporation Sunrise Point-H CSME HECI #1 (rev 31)
00:17.0 SATA controller: Intel Corporation Sunrise Point-H SATA Controller [AHCI mode] (rev 31)
00:1c.0 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #1 (rev f1)
00:1c.1 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #2 (rev f1)
00:1d.0 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #9 (rev f1)
00:1d.4 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #13 (rev f1)
00:1d.6 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #15 (rev f1)
00:1f.0 ISA bridge: Intel Corporation Sunrise Point-H LPC Controller (rev 31)
00:1f.2 Memory controller: Intel Corporation Sunrise Point-H PMC (rev 31)
00:1f.3 Audio device: Intel Corporation Device a171 (rev 31)
00:1f.4 SMBus: Intel Corporation Sunrise Point-H SMBus (rev 31)
01:00.0 3D controller: NVIDIA Corporation GP107M [GeForce GTX 1050 Mobile] (rev ff)
02:00.0 Network controller: Qualcomm Atheros QCA6174 802.11ac Wireless Network Adapter (rev 32)
03:00.0 Unassigned class [ff00]: Realtek Semiconductor Co., Ltd. RTS525A PCI Express Card Reader (rev 01)
04:00.0 Non-Volatile memory controller: Toshiba America Info Systems Device 0115 (rev 01)
HDD: Toshiba SSD
4k Monitor
Power usage: Average 15W with brightness set to 80-90
With brightness set to 10% usage around 10W
Customizations:
Compiled kernel with
apci_override
enabled and kernel option set
Bumblebee + bbswitch
Latest stable nvidia from extra
TLP service enabled v1.0-1
Xorg with AwesomeWM
Nothing else touched
Touch screen enabled but autosuspend enabled by
echo 'auto' > '/sys/bus/usb/devices/1-9/power/control';
Everything woks perfectly
Last edited by Osleg (2017-05-30 18:22:51)
Offline
Hello, is anyone having issues with the frequency scaling after resuming from suspend? I get very high CPU frequencies after resuming using the intel_pstate, if I disable it via kernel command line, the acpi module can't scale up the frequency and it stays at 800Mhz forever, and of course I lose the turbo boost. Any ideas?
I installed the linux-lts kernel and I don't have this issue running it, only thing I lose is the nvme power saving settings, which with the toshiba drive I didn't notice too much improvements.
Offline
Well one thing is not working perfectly... expected
While running
optirun/primusrun glxspeheres64
i'm getting around 30FPS while with mesa around 60
osleg in ~ λ glxspheres64
Polygons in scene: 62464 (61 spheres * 1024 polys/spheres)
Visual ID of window: 0xba
Context is Direct
OpenGL Renderer: Mesa DRI Intel(R) HD Graphics 630 (Kaby Lake GT2)
59.053016 frames/sec - 236.330703 Mpixels/sec
56.065285 frames/sec - 224.373775 Mpixels/sec
59.980887 frames/sec - 240.044051 Mpixels/sec
^Cosleg in ~ λ vblank_mode=0 primusrun glxspheres64
Polygons in scene: 62464 (61 spheres * 1024 polys/spheres)
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
Visual ID of window: 0xba
Context is Direct
OpenGL Renderer: GeForce GTX 1050/PCIe/SSE2
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
39.476164 frames/sec - 316.964152 Mpixels/sec
39.053911 frames/sec - 313.573769 Mpixels/sec
38.041862 frames/sec - 305.447772 Mpixels/sec
Any ideas what could be the issue?
Last edited by Osleg (2017-05-30 20:46:41)
Offline
Hey everyone,
first time arch linux user, since this was the only system where I there was an awesome WIKI article for the laptop and I got Power Consumption down to level that I can work for a day on battery. Thanks for the WIKI article.
Only problem I had was the Wireless controller:
Qualcomm Atheros QCA6174 802.11ac Wireless Network Adapter (rev 32)
/
168c:003e (rev 32)
. My connections speed was about 15 to 25 Mbps, instead of the full 100Mbps my link gives me. It seemed to be mostly uploads, but overall performance was not good. It took ages to connect to the WIFI, sometimes pages would not load immediately and so on.
I tried everything, mostly all the tips about which firmware to use, but nothing changed the connection speed. The system reported a 6MBps connection, although it was definitely faster than 6, but not much more.
Connected to 82:2a:a8:12:db:9c (on wlp2s0)
SSID: xxx
freq: 5180
RX: 323225533 bytes (3650809 packets)
TX: 2121267279 bytes (6985132 packets)
signal: -42 dBm
tx bitrate: 6.0 MBit/s
bss flags: short-slot-time
dtim period: 1
beacon int: 100
After hours of searching I found this thread https://bbs.archlinux.org/viewtopic.php … 0#p1689990 and since I already had a custom kernel running for bbswitch, I made the according changes. Voila, system still reports 6MBps but now I have the bandwidth I am used to with my old laptop. Connecting to a network is done in seconds.
Is there a way to get this into the WIKI, for anyone else facing the problem, or at least so that I may remember the next time I reinstall the laptop It seems to be a bit of luck whether or not the installed card is performing as expected, but I guess I am not the only one. And after spending 2k on a laptop I don't want to spend another buck for a new WIFI card.
Offline
Hey everyone,
first time arch linux user, since this was the only system where I there was an awesome WIKI article for the laptop and I got Power Consumption down to level that I can work for a day on battery. Thanks for the WIKI article.
Only problem I had was the Wireless controller:
Qualcomm Atheros QCA6174 802.11ac Wireless Network Adapter (rev 32)
/
168c:003e (rev 32)
. My connections speed was about 15 to 25 Mbps, instead of the full 100Mbps my link gives me. It seemed to be mostly uploads, but overall performance was not good. It took ages to connect to the WIFI, sometimes pages would not load immediately and so on.
I tried everything, mostly all the tips about which firmware to use, but nothing changed the connection speed. The system reported a 6MBps connection, although it was definitely faster than 6, but not much more.
Connected to 82:2a:a8:12:db:9c (on wlp2s0) SSID: xxx freq: 5180 RX: 323225533 bytes (3650809 packets) TX: 2121267279 bytes (6985132 packets) signal: -42 dBm tx bitrate: 6.0 MBit/s bss flags: short-slot-time dtim period: 1 beacon int: 100
After hours of searching I found this thread https://bbs.archlinux.org/viewtopic.php … 0#p1689990 and since I already had a custom kernel running for bbswitch, I made the according changes. Voila, system still reports 6MBps but now I have the bandwidth I am used to with my old laptop. Connecting to a network is done in seconds.
Is there a way to get this into the WIKI, for anyone else facing the problem, or at least so that I may remember the next time I reinstall the laptop It seems to be a bit of luck whether or not the installed card is performing as expected, but I guess I am not the only one. And after spending 2k on a laptop I don't want to spend another buck for a new WIFI card.
Thank you! this is a really useful discovery can you just illustrate here the steps you did to apply the patch ? So I can also test it to check the difference. Then I think Daniel could insert it in the wiki.
Last edited by simo8989 (2017-06-02 09:24:41)
Offline
Has anyone had/solved the issue that I am currently seeing.
Using bumblebee with bbswitch and the gpu in an off state, my touchpad starts freezing up. It does work, but it will randomly freeze for a few seconds at a time mid-use. I tried using synaptics instead of libinput, but the problem still persists. If I remove acpi_osi=! acpi_osi="Windows 2009" from the grub config, my touchpad returns to a normal state, however my computer stops booting properly. It will only boot about 1 out of every 10 times and the message it is usually stuck on is "Started TLP system startup/shutdown."
In summary:
GPU in on state: touchpad and boot work normally but runs hot and bad battery life
GPU with bumblebee and bbswitch currently in off state: touchpad suffers constant freezing with acpi_osi=! acpi_osi="Windows 2009"
Removing acpi_osi=! acpi_osi="Windows 2009" causes computer to not boot properly
kernel: linux 4.11.2-1
Offline
Just an update about current state for newcomers
Dell XPS 9560
HDD: Toshiba SSD
4k MonitorPower usage: Average 15W with brightness set to 80-90
With brightness set to 10% usage around 10W
Customizations:
Compiled kernel withapci_override
enabled and kernel option set
Bumblebee + bbswitch
Latest stable nvidia from extra
TLP service enabled v1.0-1
Xorg with AwesomeWM
Nothing else touched
Touch screen enabled but autosuspend enabled byecho 'auto' > '/sys/bus/usb/devices/1-9/power/control';
Everything woks perfectly
To add to your contribution,
My setup is XPS 15 9560
4k Display
Toshiba SSD
Stock 4.11.3 kernel
acpi_osi=! acpi_osi="Windows 2009" boot parameters, I added these to fix hangs when running lspci and nvidia-smi
Bumblebee + bbswitch
Latest nvidia from pacman
TLP caused me issues (broke suspend, random hangs), However I am getting similar results using just powertop applying all fixes except PCI Nvidia one.
Touchscreen enabled
Min brightness I can get down to almost 7-8W however, my polybar setup is quite power hungry, uses about 2-3W on top of that (need to fix that).
Max brightness gives around 13-15W, more if under load.
I run i3wm with Compton (alot of transparency) to give you an idea of how that would affect things.
I've been using this setup for a few days now and seems very stable. Haven't noticed any issues.
This is my first post here. I felt compelled to contribute after all the great work everyone here has done, so I hope someone can gain something from this.
Thanks again all for the great work.
Last edited by DevTony72 (2017-06-06 16:20:16)
Offline
Thank you! this is a really useful discovery can you just illustrate here the steps you did to apply the patch ? So I can also test it to check the difference. Then I think Daniel could insert it in the wiki.
Yeah sure, I simply followed the steps for creating a custom kernel here: https://wiki.archlinux.org/index.php/Ke … ild_System . All following commands take linux/trunk as the working directory.
Optional: I edited config.i686 and config.x64_64 and enabled CONFIG_ACPI_REV_OVERRIDE_POSSIBLE to get bumblebee working correctly.
Then I had to manually created a patch file, since the patch file from the linked post had incorrect whitespaces: https://gist.githubusercontent.com/DaCo … tch-ath10k
Updated the checksums:
updpkgsums
Setup PKGBUILD to correct settings:
- renamed the package to linux-custom
- added patch-ath10k to the sources
- told prepare() to apply the patch at the end.
pkgbase=linux-custom
...
source=(....
'patch-ath10k'
...
prepare() {
...
patch -p1 -i "${srcdir}/patch-ath10k"
...
}
...
After that I simply ran makepkg -s and waited for my custom kernel to be done, installed created linux-custom and linux-custom-header packages via pacman and reconfigured the bootloader and the result is awesome. Not sure if I did everything right with applying the patch, but again I am an arch linux newbie and this worked for me
Last edited by cookiemonster (2017-06-02 16:34:14)
Offline
FWIW, I have raised the following: https://bugs.archlinux.org/task/54304
Offline
After hours of searching I found this thread https://bbs.archlinux.org/viewtopic.php … 0#p1689990 and since I already had a custom kernel running for bbswitch, I made the according changes. Voila, system still reports 6MBps but now I have the bandwidth I am used to with my old laptop. Connecting to a network is done in seconds.
Nice find! To be honest, I hadn't really even realised I had a problem - it wasn't bad enough, I'd just put it down to "wireless". But it's definitely a lot better now!
Offline
FWIW, I have raised the following: https://bugs.archlinux.org/task/54304
Which was promptly closed as duplicate of #53281. I have asked for it to be reopened noting that: "Yes, this is duplicate of #53281; however, that issue was closed as for the XPS13, it was found to be a rare hardware bug and thus considered a won't fix. This is NOT the case for the XPS15 (9560) whereby this Kernel option is essential and always will be (as per the text in the issue). Please consider re-opening this."
I'm not hopeful to be honest. Looks like we might be stuck compiling a custom kernel forever.
Offline
You might want to look at #53281 before making wild assumptions. It's not for the XPS13 and it's not closed.
Offline
patrickb wrote:FWIW, I have raised the following: https://bugs.archlinux.org/task/54304
Which was promptly closed as duplicate of #53281. I have asked for it to be reopened noting that: "Yes, this is duplicate of #53281; however, that issue was closed as for the XPS13, it was found to be a rare hardware bug and thus considered a won't fix. This is NOT the case for the XPS15 (9560) whereby this Kernel option is essential and always will be (as per the text in the issue). Please consider re-opening this."
I'm not hopeful to be honest. Looks like we might be stuck compiling a custom kernel forever.
Turns out I was looking at the wrong ticket. The #53281 issue is definitely still open. Phew. Ooops. (Sorry Kernel package team.)
Offline
You might want to look at #53281 before making wild assumptions. It's not for the XPS13 and it's not closed.
Ah, I'm glad you popped into this thread actually. I do apologise. I was confusing 53281 and 47989 (which is closed and relates to the XPS 13).
(So there's an hour of my online life that would probably be better deleted.)
Last edited by patrickb (2017-06-05 02:03:25)
Offline
Is anybody using i8kutils to control the fans speed? I found that when using this laptop on AC it almost never stops running the fans even if the temperatures are really low, like 35 C. Also, when the fans are on, they run at a minimum speed that is higher than the minimum speed on windows 10.
If somebody has already tried or can provide useful information on how to better manage the speed of the fans without harming the system it would be really helpful.
Thanks
Last edited by simo8989 (2017-06-07 17:21:30)
Offline
Has anyone had/solved the issue that I am currently seeing.
Using bumblebee with bbswitch and the gpu in an off state, my touchpad starts freezing up. It does work, but it will randomly freeze for a few seconds at a time mid-use. I tried using synaptics instead of libinput, but the problem still persists. If I remove acpi_osi=! acpi_osi="Windows 2009" from the grub config, my touchpad returns to a normal state, however my computer stops booting properly. It will only boot about 1 out of every 10 times and the message it is usually stuck on is "Started TLP system startup/shutdown."
In summary:
GPU in on state: touchpad and boot work normally but runs hot and bad battery life
GPU with bumblebee and bbswitch currently in off state: touchpad suffers constant freezing with acpi_osi=! acpi_osi="Windows 2009"
Removing acpi_osi=! acpi_osi="Windows 2009" causes computer to not boot properlykernel: linux 4.11.2-1
If you follow the original post and the wiki you'll notice this is very much expected behavior. Explanations and solutions are provided.
In short: In your case the system freezes because TLP tries to change GPU power state, which is the known problem with stock settings. Using the acpi_osi flags fixes this but also messes up the touchpad as described in original post. Compiling your kernel with ACPI_REV_OVERRIDE_POSSIBLE option set and using acpi_rev_override boot flag circumvents GPU power issues without affecting touchpad.
Offline
Is anybody using i8kutils to control the fans speed? I found that when using this laptop on AC it almost never stops running the fans even if the temperatures are really low, like 35 C. Also, when the fans are on, they run at a minimum speed that is higher than the minimum speed on windows 10.
If somebody has already tried or can provide useful information on how to better manage the speed of the fans without harming the system it would be really helpful.
Thanks
I tried it, but as well as being difficult to set up it doesn't work so well imo. You can make the fans turn off more often but instead of the stock speeds (0,2400,3270,3900,4100,4800,5100) you can only manually set (0,2400,4800): https://github.com/vitorafsr/i8kutils/issues/5. Also you have to compile smm.c and write something to ensure that that smm command happens before i8kutils sets the fan speed (and after if you want to revert to automatic control)
https://www.reddit.com/r/Dell/comments/ … n_and_fan/
My problem was that when the fans do need to go faster than 2400, 4800 is much much louder than 3200 and I prefer to deal with 2400 instead of 0 most of the time. I found that putting a heat conducting pad on the southbridge and SSD made the BIOS fan control much more quiet. The southbridge especially heated up to 70 when I plugged in an external screen.
Last edited by hedgepigdaniel (2017-06-08 22:50:10)
Offline
One thing I've probably forgot to mention before is that I've not found a solution for turning off the screen when closing the lid. It does not seem to do it on the hardware side like with many devices, and systemd/etc. actions for lid events only seem to cover stuff like suspend/hibernate. You could probably plug something custom into it, but it's not been a big enough deal for me to look into further. I've just plugged the search function key to run "xset dpms force off" in case I want to turn off the screen manually.
If anyone has figured something out for this, I'd like to hear about it! I often need to run around with the laptop lid closed but without suspending so it would be great to save some battery while I do.
Offline
I spent great part of a day trying to figure out why the headphone jack in the TB16 dock didn't work in my laptop, it turns out it was muted, but since you can't select the USB Audio card in alsamixer because it breaks, the solution is to boot into windows, un-mute it (and set the volume to 100% just in case) and go back to Linux, making the TB16 work 100%
Offline