Description of problem: Debian has announced that the Skylake/Kabylake Intel processors has a bug in the hyperthreading which causes data corruption and crashes. For some Skylake processors an update of the microcode package fixes this. Mageia has released microcode-20170511 for cauldron, which I assume fixes this bug on select Skylake processors, and also has it in nonfree/updates-testing on Mageia 5, but it does not seem to be moving from there to release. Is there some problem?
No problem, I just havent had time to review it and assign it for testing
Ubuntu just announced 20170707 microcode which fixes the bugs on all the Skylake and Kabylake processors. It would be good to get this out since this bug really is horrible if it strikes apparently.
Great, I waited/hoped that Intel would release the Kaby Lake firnwares too (first they stated it would only be provided to hw manufacturers :/)
So here goes...
Updated microcode fixes HT bug in Intel Skylake and Kaby Lake processors
This microcode update fixes a Hyperthreading bug in Intel processor/microcode
based on Skylake and Kaby Lake platforms (includes Xeon v5 and v6 and some
This defect can, when triggered, cause unpredictable system behavior: it
could cause spurious errors, such as application and system misbehavior,
data corruption, and possible data loss.
Note, in order for the firmware to get loaded early in the boot process, we
suggest you recreate the initrd after installing this update.
To do so, open up a terminal window, use the command 'su -' to gain root
privilegies and then use the command 'dracut -f'. Rebooting the system after
that will make the new microcode being used.
Release (media or process) =>
$ uname -r
$ grep name /proc/cpuinfo | sort -u
model name : Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz
$ grep 'microcode' /proc/cpuinfo
microcode : 0x9e
package installed cleanly
After rebuilding initrd and re-booting:
$ dmesg | grep microcode
[ 0.000000] microcode: CPU0 microcode updated early to revision 0xba, date = 2017-04-09
[ 0.476091] microcode: CPU0 sig=0x506e3, pf=0x2, revision=0xba
$ grep 'microcode' /proc/cpuinfo
microcode : 0xba
No regressions noted
OK for mga5-64
Just installed on Mageia 6 (updated to today)dmesg|grep microcode
[ 0.000000] microcode: microcode updated early to revision 0x62, date = 2017-04-27
[ 0.674346] microcode: sig=0x806e9, pf=0x80, revision=0x62
[ 0.674416] microcode: Microcode Update Driver: v2.01 <email@example.com>, Peter Oruba
No idea why the different revision and date to the previous Mageia5 installation.
No real testing yet. Hyperthreading still not re-enabled in bios.
Testing mga6 64 without problems on KBL i5(dual core with HT)
Rebooted after install.
Checked using the "stress" package and top to ensure 4 cores shown and loaded.
# stress -c 4 -t 30
stress: info:  dispatching hogs: 4 cpu, 0 io, 0 vm, 0 hdd
stress: info:  successful run completed in 30s
watched top in another terminal, pressed 1 to show all cores.
Ran several times with much whirring of fans.
Installed on nvidia GTX750 machine with 8x Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz and rebooted.
[ 0.412884] microcode: CPU0 sig=0x306c3, pf=0x2, revision=0x20
30 second stress tests on CPUs and RAM ran fine - some core swapping during the runs.
8x Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
Installed the microcode update and rebooted.
microcode : 0x20
Stress tests on cpu and RAM completed successfully.
In the light of Comment 6, Comment 8, added OK for MGA6.
Doubting that this appliers to 32-bit, validating; advisory uploaded.
MGA5TOO MGA5-64-OK =>
MGA5TOO MGA5-64-OK advisory MGA6-64-OK
An update for this issue has been pushed to the Mageia Updates repository.