|
|
Log in / Subscribe / Register

Intel Skylake/Kaby Lake processors: broken hyper-threading

Intel Skylake/Kaby Lake processors: broken hyper-threading

Posted Jun 26, 2017 22:18 UTC (Mon) by SEJeff (guest, #51588)
Parent article: Intel Skylake/Kaby Lake processors: broken hyper-threading

I wonder if this is a problem in RHEL / Fedora that has been already addressed, or still needs to be.


to post comments

Intel Skylake/Kaby Lake processors: broken hyper-threading

Posted Jun 27, 2017 7:57 UTC (Tue) by TimSmall (guest, #96681) [Link]

No fix on CentOS 7 as of the time of writing. A quick hack is to copy over /lib/firmware/intel-ucode/ from a Debian box, followed by:

# verify processors and microcode versions
grep -E 'model|stepping|code' /proc/cpuinfo | sort -u
# request microcode reload
echo 1 > /sys/devices/system/cpu/microcode/reload
# verify upgrade
dmesg | grep microcode

... this works with the stock CentOS 7 3.10 kernel, but may not work on newer kernels depending on kernel options IIRC. In general there is a small risk to doing this on a booted system, because CPU features which running software are using could suddenly disappear if the microcode update disables them. This is why this is normally done very early in the boot sequence by the initramfs, assuming this is the case on CentOS 7, then the initramfs will probably need to be rebuilt with "dracut -f" to get this fix to persist across reboots (and/or apply it in a safer way in general).

Intel Skylake/Kaby Lake processors: broken hyper-threading

Posted Jun 27, 2017 13:41 UTC (Tue) by tdz (subscriber, #58733) [Link]

I checked yesterday on my Skylake-based system with Fedora 25. The system had the patched microcode and the BIOS was from Nov last year. So I guess the BIOS didn't fix is yet and chances are that Fedora has the patch already.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds