Linux 5.10.20 Released - Fixes The Erroneous Record-Breaking AMD Clock Frequencies

[phoronix]

Phoronix: Linux 5.10.20 Released - Fixes The Erroneous Record-Breaking AMD Clock Frequencies

Recent kernel point releases have reported erroneous maximum frequencies on AMD Zen 2 / Zen 3 CPUs in the area of 6GHz+ while now with the latest stable releases that is being fixed...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-5.10.20-Fixes-Freq-Report

[birdie]

Speaking of so called "stable" point releases.

With kernel 5.10.13 my laptop goes to sleep and resumes just fine as many times as I want to.

With kernel > 5.10.13 it resumes just once, then dies on consequent resumes (the screen doesn't turn on, the CPU fan starts spinning, Alt + SysRQ does nothing, I cannot ping it, so it's dead).

[Leopard]

Speaking of so called "stable" point releases.

With kernel 5.10.13 my laptop goes to sleep and resumes just fine as many times as I want to.

With kernel > 5.10.13 it resumes just once, then dies on consequent resumes (the screen doesn't turn on, the CPU fan starts spinning, Alt + SysRQ does nothing, I cannot ping it, so it's dead).

Did you report it?

[juxuanu]

I hope this fixes schedutil performing worse than powersave + boosting off in my 3700U laptop. At the moment ondemand gets the job done.

[kripteks]

Speaking of so called "stable" point releases.

With kernel 5.10.13 my laptop goes to sleep and resumes just fine as many times as I want to.

With kernel > 5.10.13 it resumes just once, then dies on consequent resumes (the screen doesn't turn on, the CPU fan starts spinning, Alt + SysRQ does nothing, I cannot ping it, so it's dead).

I have similair thing (deskstop pc), my kernel dont have hibernation/suspend to ram.
In 5.11 and maybe lower idk.
When i need do something outside, i keep computer on but on tty2 (tty1 is graphical), and turn off monitor (button).
After some minutes when i open monitor, tty1 dont respond nothing, tty2 work, i cant have/show tty1/graphical(openbox) so i need to reboot.
I dont know if it kernel/system/nvidia.

[juxuanu]

https://gitlab.freedesktop.org/drm/amd/-/issues/1230

Might be this.

[Volta]

Speaking of so called "stable" point releases.

With kernel 5.10.13 my laptop goes to sleep and resumes just fine as many times as I want to.

With kernel > 5.10.13 it resumes just once, then dies on consequent resumes (the screen doesn't turn on, the CPU fan starts spinning, Alt + SysRQ does nothing, I cannot ping it, so it's dead).

Speaking of stability my laptop died second time in few months from 'stable' windows 10 update. Reverting the update helped, so you should have a clue. PS. unlike your problem the laptop was dying after few minutes after system loaded, so it was hard to revert this 'UPDATE'.

[birdie]

https://gitlab.freedesktop.org/drm/amd/-/issues/1230
Might be this.

I've got an Intel iGPU. I thought about filing a bug report against i915 but decided against it as I'm not sure this driver is the culprit.

[tildearrow]

What's up with the sensors? Why do they always fluctuate on every new processor generation?

Are these changes done for no actual reason other than messing with developers or are the sensors that unstable?

[monty11ez]

I think this is the biggest point release I've seen since I switched to Linux full time early last year.