Bay Trail C-State Bug - is it now fixed and in which kernel version/patch level?

English Hardware
#1

Hello,

Can someone point me to what version of the openSUSE kernel fixed the Baytrail C-State problem?

https://bugzilla.kernel.org/show_bug.cgi?id=109051#c854

Basically I have around 500 units in the wild with J1900 processors and they will occasionally freeze randomly - sometimes after a week, sometimes after a month, and sometimes after 6 months. I get no syslog info to point me in the right direction, the logs just stop. Power is still on but the device can’t do anything.

I did some more investigation, and I believe I’m running into the Bay Trail C-State problem, so I’ll need to update all OSes in the wild, but I want to make sure I am not just reintroducing the same problem. I went through the kernel changelog in 5.3.18-59.37-default, but I can’t tell if any of the Bay Trail or C-state mentions fixed this problem (that people still currently run into apparently). I could build a new image and just add the " intel_idle.max_cstate=1" kernel parameter, but I’d like to find out first which kernel this was supposed to be fixed in, as I may have a newer build that already contains the fix.

Thanks for any help.

#2

Ask developers about that by creating bug report. They may add needed patches to Leap 15.3 kernel (if it not done yet).

#3

“intel_idle.max_cstate=<n>” and “processor.max_cstate=<n>” are both Kernel Command Line options – <https://www.kernel.org/doc/html/latest/admin-guide/pm/cpuidle.html?highlight=max_cstate#> –

  • Do you really have to build a new image to get this option into the systems you support or, can you simply add the option to the GRUB2 script on each system?
#4

I could add it to GRUB but we try to avoid making one-off changes like that.