New: amd_pstate=active reset computer
https://bugzilla.kernel.org/show_bug.cgi?id=219051
Bug ID: 219051
Summary: amd_pstate=active reset computer
Product: Power Management
Version: 2.5
Hardware: AMD
OS: Linux
Status: NEW
Severity: normal
Priority: P3
Component: cpufreq
Assignee: [email protected]
Reporter: [email protected]
Regression: No
Hello!
I have a ASUS TUF A17 FA706QR laptop. Ryzen 5800H, nvidia dgpu.
I have installed Ubuntu Desktop 24.04 LTS with kernel 6.8 and the laptop started to reset shortly after booting: from a few minutes to 30 minutes, nothing in the logs.
Before I had Ubuntu 20.04 with 5.15 kernel, no problems, but amd_pstate was not active.
Reset occurs by leaving the laptop at idle or doing something like playing an youtube video.
I also tried kdump without getting any dump files.
I noticed that the processor runs at lower frequencies than 20.04, so do the fans, most of the time at 0 rpm. Temps 50-70 degrees.
I ran stress on one core, youtube play in the same time, cpu temp ~90 degrees, no reset!
I switched the power profile to Performance and it didn't reset. It only happens on Balanced or Power Saver profiles.
Next step was amd_pstate=passive, reset only on Power Saver.
The fix was amd_pstate=disable. No reset at all
Not happening on other Asus laptops with Ryzen 6800H, Lenovo (no Nvidia) 5800H, 6800H, 7840HS with amd_pstate=active, Ubuntu 22/24, 6.5 and 6.8 kernels.
Thanks!
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #1 from Artem S. Tashkinov ([email protected]) --- Are you running the latest BIOS version? If not, please flash.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #2 from Artem S. Tashkinov ([email protected]) ---
Also, it's worth noting that if your system resets itself (and you get a confirmation about that on next boot where dmesg says your partitions have been recovered), it usually indicates a HW issue, rather than a software one.
Unfortunately I've no clue what could be wrong and how to test it.
Some AMD users have been playing with undervolting their CPUs using e.g. Smokeless_UMAF or similar things, and I wanna hope it's not what you've done. Undervolting may make your system unstable.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #3 from Catalin ([email protected]) --- I already have the latest version of bios.
It's a laptop with very limited bios settings, undervolting is not one of them. Pretty much is default, secureboot activated, disk encrypted
I got some errors in dmesg:
[ 864.970948] asus_wmi: fan_curve_get_factory_default (0x00110024) failed: -61 [ 864.972365] asus_wmi: fan_curve_get_factory_default (0x00110025) failed: -61 [ 864.973533] asus_wmi: fan_curve_get_factory_default (0x00110032) failed: -19
I have them on other laptop with 24.04 and no issues
and
[ 864.511925] systemd-journald[638]: File
/var/log/journal/9ce66a6c7ebf406aa908624aafc66e86/system.journal corrupted or uncleanly shut down, renaming and replacing.
I agree it could be a HW error, but with amd_pstate disabled I have no resets .. at least for 10 hours..
If I can give you more information I will.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
Artem S. Tashkinov ([email protected]) changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]
--- Comment #4 from Artem S. Tashkinov ([email protected]) --- Casting Mario.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #5 from Artem S. Tashkinov ([email protected]) --- Mario, any ideas what is going on here and how it can be fixed/debugged?
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
Perry Yuan(AMD) ([email protected]) changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]
--- Comment #6 from Perry Yuan(AMD) ([email protected]) --- Mario is OOO lately,
[ 864.970948] asus_wmi: fan_curve_get_factory_default (0x00110024) failed: -61 [ 864.972365] asus_wmi: fan_curve_get_factory_default (0x00110025) failed: -61 [ 864.973533] asus_wmi: fan_curve_get_factory_default (0x00110032) failed: -19
The errors show the BIOS broken to get factory default fan curve data. I would like sugguest to submit support case to ASUS for further help.
The system reset is triggered when system reach critical temperature.
"I noticed that the processor runs at lower frequencies than 20.04, so do the fans, most of the time at 0 rpm. Temps 50-70 degrees."
Like you said, the frequencies are lower than 20.04 which have no amd-pstate driver loaded, so the system temperature increasing slower than 24.04 kernel which has amd-pstate-epp driver loaded.
I would guess the root cause is system fan error instead of amd-pstate driver. After you resolve the Fan issue, system will not reset any more.
static int fan_curve_get_factory_default(struct asus_wmi *asus, u32 fan_dev)
{
struct fan_curve_data *curves;
u8 buf[FAN_CURVE_BUF_LEN];
int err, fan_idx;
u8 mode = 0;
if (asus->throttle_thermal_policy_available)
mode = asus->throttle_thermal_policy_mode;
/* DEVID_<C/G>PU_FAN_CURVE is switched for OVERBOOST vs SILENT */
if (mode == 2)
mode = 1;
else if (mode == 1)
mode = 2;
// asus_wmi_evaluate_method_buf failed here, it is a broken bios issue.
err = asus_wmi_evaluate_method_buf(asus->dsts_id, fan_dev, mode, buf,
FAN_CURVE_BUF_LEN);
if (err) {
pr_warn("%s (0x%08x) failed: %d\n", __func__, fan_dev, err);
return err;
}.....
}
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #7 from Catalin ([email protected]) --- Update:
It crashed with amd_pstate=disable after 15+ hours.
Also I tried Windows with crash in 4 hours, so it has nothing to to with amd_pstate. You can close it. Most likely faulty motherboard. Thanks!
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #8 from Perry Yuan(AMD) ([email protected]) --- (In reply to Catalin from comment #7)
Quoted 5 lines
Update:
It crashed with amd_pstate=disable after 15+ hours.
Also I tried Windows with crash in 4 hours, so it has nothing to to withamd_pstate. You can close it. Most likely faulty motherboard. Thanks!
Thanks to hear the feedback.
BTW, You can check if the system fan is spinning normally in case the system CPU die temperature reach critical point(+84.8 degree).
sensors utility can help to check the die temperature like below. #sensors
k10temp-pci-00c3
Adapter: PCI adapter
Tctl: +37.2°C
On some AMD CPUs, there is a difference between the die temperature (Tdie) and the reported temperature (Tctl). Tdie is the real measured temperature, and Tctl is used for fan control. While Tctl is always available as temp1_input, the driver exports Tdie temperature as temp2_input for those CPUs which support it. (Documentation/hwmon/k10temp.rst)
Looks like your system HW/BIOS broken, vendors maybe have solution for your case.
Perry.
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
Artem S. Tashkinov ([email protected]) changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |INVALIDSignature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.
[Bug 219051] amd_pstate=active reset computer
In reply to [email protected]https://bugzilla.kernel.org/show_bug.cgi?id=219051
--- Comment #9 from Perry Yuan(AMD) ([email protected]) ---
Hello, Thank you for your message. I am currently out of the office for Labor
Day Holiday. If you have any urgent issues or need CPPC, Hetero Core related urgent assistance, please reach out to [Limonciello, Mario]. For non-urgent matters, I will respond to your email as soon as possible upon my return on [5/6]. Best wishes, Perry Yuan
Signature · 4 lines
You may reply to this email to add a comment. You are receiving this mail because: You are the assignee for the bug.