Return to Level1Techs.com

Laptop Ryzen 3750H dropping to 400MHz in linux and staying there

Hello!

I purchased an Asus ROG ga502 (Ryzen 7 3750h, 1660ti) and I’m having a really weird issue where my CPU will drop it’s pstate down to 400mhz and stick there until I reboot. I’ve found some workarounds by using zenpstates from aur, but I wanted to reach out and see if anyone else had seen this issue, or if it’s a known issue.

Thanks guys!

P.S. I’d post logs, but sadly I had to throw windows back on the laptop for now as I’m having to temporarily use it for work.

1 Like

Well… This is weird.

Could you please check that it is a general Linux issue by booting your laptop from a LiveUSB, preferably Ubuntu 19.04 Disco Dingo? If the issue does not persist, then it is something to do with Arch. If the issue does persist, however, please share the dmesg --human output after the Pstate drops down.
Also, could you try, if willing to install Ubuntu on some boot drive (which drive is up to you), sudo apt-get install linux-tools-common linux-tools-generic, reboot and then sudo cpupower frequency-set --governor performance && cpupower frequency-info to set the CPU governor to performance, which should keep it at the highest possible clock rate (Pstate) which the current temperature can allow. cpupower frequency-info is just to verify that the current governor was set to performance. Once you get the necessary packages installed on Arch, you can use the exact same second command, the one after the reboot above, without any issues.

Good luck!

1 Like

Thanks @elsandosgrande I’ll give it a go this weekend! Weird thing is, it happened in windows until I installed the asus armoury crate software. It’s the weirdest thing, but several people have reported it under a few flavors of linux (OpenSuse, Ubuntu, Manjaro(arch)). I’ll let you know what happens as soon as I work on it though!

Hmm…

Well, that’s… interesting. Do you have a link to the software? I am really curious as to what it even does.

Sadly it’s a windows app, from the store. Also here are the drivers you have to install with it.

(under utilities Armoury crate service)

I know that it is a Windows program. I asked for the link in order to see the description of the program that is, or should be, present on the download page.

I meant that it’s a store app, it’s the only way to get it. I knew you knew it was a windows program, lol.

Heres the Dmesg Output after it dropped.

dmesg400.txt (80.4 KB)

[[email protected] ~]$ sudo cpupower frequency-set --governor performance && cpupower frequency-info
Setting cpu: 0
Setting cpu: 1
Setting cpu: 2
Setting cpu: 3
Setting cpu: 4
Setting cpu: 5
Setting cpu: 6
Setting cpu: 7
analyzing CPU 0:
  driver: acpi-cpufreq
  CPUs which run at the same hardware frequency: 0
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency:  Cannot determine or is not supported.
  hardware limits: 1.40 GHz - 2.30 GHz
  available frequency steps:  2.30 GHz, 1.70 GHz, 1.40 GHz
  available cpufreq governors: performance schedutil
  current policy: frequency should be within 1.40 GHz and 2.30 GHz.
                  The governor "performance" may decide which speed to use
                  within this range.
  current CPU frequency: Unable to call hardware
  current CPU frequency: 399 MHz (asserted by call to kernel)
  boost state support:
    Supported: no
    Active: no

My guess is that you have tried to change the governor only after the Pstate drop. Try doing so before the drop if you have not already and check if it helps. If you already have and it does not, please do clarify.

As far as the dmesg output goes, I am still looking things up online, but I have found this when searching for “ucsi_ccg”: https://chat.stackexchange.com/rooms/97546/discussion-on-question-by-tram98-fresh-ubuntu-19-04-installation-only-booting-in
The guy never followed up, I presume failure, but it might provide some hint, possibly? The kernel element seems to be related to USB, so…

Also, did the drop occur after 13h or 14h? I ask because of this segment of the output is not entirely chronologically clear to me:

[Oct17 13:32] ucsi_ccg 3-0008: failed to reset PPM!
[  +0.000005] ucsi_ccg 3-0008: PPM init failed (-110)
[Oct17 13:44] audit: type=1130 audit(1571319883.451:66): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=systemd-tmpfiles-clean comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[  +0.000003] audit: type=1131 audit(1571319883.451:67): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=systemd-tmpfiles-clean comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[Oct17 13:53] audit: type=1130 audit(1571320394.917:68): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=systemd-timedated comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[ +14.576538] audit: type=1130 audit(1571320409.494:69): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=systemd-timesyncd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[ +30.033448] audit: type=1131 audit(1571338435.375:70): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=systemd-timedated comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[Oct17 14:09] input: Arc Touch BT Mouse as /devices/virtual/misc/uhid/0005:045E:0804.0006/input/input26
[  +0.000095] input: Arc Touch BT Mouse Keyboard as /devices/virtual/misc/uhid/0005:045E:0804.0006/input/input28
[  +0.000062] hid-generic 0005:045E:0804.0006: input,hidraw4: BLUETOOTH HID v0.01 Mouse [Arc Touch BT Mouse] on DC:F5:05:9B:69:8C
[Oct17 14:31] audit: type=1130 audit(1571340689.472:71): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=NetworkManager-dispatcher comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[ +10.207377] audit: type=1131 audit(1571340699.678:72): pid=1 uid=0 auid=4294967295 ses=4294967295 subj==unconfined msg='unit=NetworkManager-dispatcher comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[Oct17 14:34] ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.SBRG.EC0._QE8.TEMF], AE_NOT_FOUND (20190703/psargs-330)
[  +0.000010] ACPI Error: Aborting method \_SB.PCI0.SBRG.EC0._QE8 due to previous error (AE_NOT_FOUND) (20190703/psparse-529)

Aaaand I’m off to bed now. Good night!

I actually found out this is a known issue with AMD ASEGA on these newer laptops, looks like it even happens in windows still. When the laptop hits 90c it drops to that speed but isn’t be reset as it should. ASUS is treating it as an RMA, but none of the RMA’d laptops are any different.

I tried changing the governor before and after and the same thing happens. Though in performance mode it does a better job of not hitting 90c so I can atleast run it for a while.

For the ucsi_ccg I believe thats an issue with the switchable graphics, it goes away if I remove nouveau and throw on nvidia, but throwing on nvidia breaks everything since optimus switching isn’t set up.

The drop happened at 14h. Those ACPI errors look interesting though.

\_SB.PCI0.SBRG.EC0._QE8.TEMF - Is power management stuff. I found something similar in https://github.com/RehabMan/Laptop-DSDT-Patch/issues/14 Which is for DSDT patching for hackintoshes.

Also here is an arch linux post about the same issue:

https://bbs.archlinux.org/viewtopic.php?id=248477

  1. What is RMA?
  2. The performance governor should make your temperatures higher, not lower. I am confused by your laptop.
  3. My laptop has similar ACPI errors, but they refer to some buffer, as opposed to an unresolvable symbol.
[  +0,000993] ACPI BIOS Error (bug): AE_AML_BUFFER_LIMIT, Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20190703/dsopcode-198)
[  +0,000900] ACPI Error: Aborting method \HWMC due to previous error (AE_AML_BUFFER_LIMIT) (20190703/psparse-529)
[  +0,000909] ACPI Error: Aborting method \_SB.WMID.WMAA due to previous error (AE_AML_BUFFER_LIMIT) (20190703/psparse-529)
[  +0,000983] ACPI BIOS Error (bug): AE_AML_BUFFER_LIMIT, Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20190703/dsopcode-198)
[  +0,000944] ACPI Error: Aborting method \HWMC due to previous error (AE_AML_BUFFER_LIMIT) (20190703/psparse-529)
[  +0,000951] ACPI Error: Aborting method \_SB.WMID.WMAA due to previous error (AE_AML_BUFFER_LIMIT) (20190703/psparse-529)
[  +0,000985] ACPI BIOS Error (bug): AE_AML_BUFFER_LIMIT, Field [D128] at bit offset/length 128/1024 exceeds size of target Buffer (160 bits) (20190703/dsopcode-198)
[  +0,000986] ACPI Error: Aborting method \HWMC due to previous error (AE_AML_BUFFER_LIMIT) (20190703/psparse-529)
[  +0,001013] ACPI Error: Aborting method \_SB.WMID.WMAA due to previous error (AE_AML_BUFFER_LIMIT) (20190703/psparse-529)

RMA - return merchandise authorization (Meaning returning for repair)

I thought it would too, but setting it to performance also kicks up my fan curve, it’s really weird.

Yeah, I’m not really sure on the ACPI stuff. It’s a little above my knowledge level.

I verified this happening on Ubuntu, Pop!_OS, Fedora, Manjaro and Windows. Happens every time the CPU hits 90c (which apparently they say is normal) and the bios can’t reset it back up to normal speeds after cooling down. Using the nvidia card helps keep everything cooler and it doesn’t hit 90c, but until Asus fixes the issue it doesn’t seem like there is going to be a way to fix this short of workarounds.

This isn’t a thermal flaw. It is a BIOS design flaw for this laptop. I’m on RMA #2 for this exact issue, with no end in sight.

Yeah. The BIOS issue is when it hits 90c it drops to 400MHz and stays there. But it absolutely is a BIOS issue. no RMA will fix it, it has to be a BIOS update from ASUS. They fixed it somehow in windows with the stupid armory crate crap, not sure how though.

Armoury Crate usually installs ITSELF on ASUS BIOSes. Z390 BIOSes and newer has the program IN THE UEFI that autostarts and installs itself even from clean Windows USB installs. You have to disable it in the UEFI.

Weird. I sadly can’t find that option on this laptop. I was reading that a future bios update that is supposed to fix this is in the works on the ASUS side. Here’s Hopinh.

1 Like

Theres been an update and a kernel fix for this issue: https://www.phoronix.com/scan.php?page=news_item&px=ASUS-TUF-Ryzen-Thermal-Fix

I just patched 5.4.12-1 manjaro kernel and am testing now. I’ll let y’all know what happens. so far so good though!

Problem solved on ubuntu 19.

I put the instructions on pastebin with id 1rZyKJmB since I can’t include links here.

Hope that helps. It looks like it will be built into 5.6.

Ryan

Hello. I wanted to shed some light on this. I patched Kernel 5.4 (Ubuntu 20.04) by following these instructions and my laptop still throttles after the ACPI reports the temp to be over 90 dC and while newer bios lets the system get hot it throttles AFTER cooling down! Its like this, On BIOS 207 and earlier, the CPU would throttle when the system got too hot. On BIOS 208 and newer, the CPU would throttle when it got too cool. There seems to be no middle to it! I am currently running kernel 5.4 (Patched) and BIOS 207. And no, the newer BIOS does not work. Has anyone got this working on Ubuntu and if so please help me! I wrote a tutorial on Ubuntu Forums called “Setup Guide 2020: Asus ROG Zephyrus GA502 Ryzen 7 3750H + Nvidia GTX 1660 Ti” Please take a read when you have the chance and leave any info you know that I did not include already so I can improve my guide and be a happy person. Thank You!