[SOLVED] Broken hyperthreading in Skylake and Kaby Lake

mcloaked · 2017-06-27 20:55:51

In the last few days there has been a number of articles published and discussions on other forums about the problem of broken hyperthreading in some recent processors needing a patched microcode as well as updated BIOS/EFI. See http://www.mail-archive.com/debian-deve … 47747.html

It isn't clear if there are instances where an arch linux installation in the affected processors might trigger a problem when running particular applications, but it would be useful to know what the current state of the linux microcode is in regard to this particular issue. Of course any required BIOS update is dependent on the supplier of the affected machine. In my case I have a NUC6i5SYK which "may" be affected. In that case the latest BIOS update is from a month ago so presumably does not address this problem, and conceivably a BIOS update will be forthcoming from Intel before long.

Has anyone else come across any issues running arch on affected machines? If so has hyperthreading been disabled - and if that has been done, has this impacted system performance to a noticeable degree?

Last edited by mcloaked (2017-07-03 18:51:01)

Scimmia · 2017-06-27 20:58:36

I think the Debian announcement covers it pretty well. The intel-ucode package in Arch is always recommended if you're running an Intel processor.

Last edited by Scimmia (2017-06-27 20:58:53)

mcloaked · 2017-06-27 21:06:39

The current intel-ucode package in arch has version as per:

$ pacman -Qi intel-ucode
Name            : intel-ucode
Version         : 20170511-1

which does indeed match the version listed in the Debian article as fixing processors with signatures[1] 0x406E3 and 0x506E3 but it also says "This will fix only Skylake processors with model 78 stepping 3, and model 94 stepping 3. The fixed microcode for these two processor models reports revision 0xb9/0xba, or higher." Is it possible that other processors may also be affected that require additional patches? I don't have access to my Skylake machine right now but will check it next weekend for versions.

Scimmia · 2017-06-27 21:08:30

The announcement says that other processors affected don't have publicly released microcode files, nothing we can do about that until Intel releases them to the public.

Last edited by Scimmia (2017-06-27 21:08:42)

LaurentvdB · 2017-06-28 12:42:31

I do have a Skylake processor which has a fixed microcode:

$ lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                4
On-line CPU(s) list:   0-3
Thread(s) per core:    2
Core(s) per socket:    2
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 78
Model name:            Intel(R) Core(TM) i7-6560U CPU @ 2.20GHz
Stepping:              3
CPU MHz:               609.619
CPU max MHz:           3200,0000
CPU min MHz:           400,0000
BogoMIPS:              4416.00
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              4096K
NUMA node0 CPU(s):     0-3
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch epb intel_pt tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp

This processor should suffer from the mentioned bug, but I can verify that the microcode that is mentioned in the Debian post is picked up correctly in Arch:

$ dmesg | grep microcode
[    0.000000] microcode: microcode updated early to revision 0xba, date = 2017-04-09
[    1.083611] microcode: sig=0x406e3, pf=0x40, revision=0xba
[    1.083781] microcode: Microcode Update Driver: v2.01 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

So essentially everything that is mentioned by the Debian mail also holds for Arch.

mcloaked · 2017-07-03 18:50:44

I was able to test my machine, and it is one of those affected:

[mike@home2 ~]$ grep name /proc/cpuinfo | sort -u
model name      : Intel(R) Core(TM) i5-6260U CPU @ 1.80GHz
[mike@home2 ~]$ grep -q '^flags.*[[:space:]]ht[[:space:]]' /proc/cpuinfo && \
> echo "Hyper-threading is supported"
Hyper-threading is supported
[mike@home2 ~]$ grep -E 'model|stepping' /proc/cpuinfo | sort -u
model           : 78
model name      : Intel(R) Core(TM) i5-6260U CPU @ 1.80GHz
stepping        : 3
[mike@home2 ~]$

But the microcode is up to date in the current latest BIOS (The BIOS fix came in April I believe):

sudo journalctl -b | egrep microcode
Jul 03 19:33:02 home2 kernel: microcode: sig=0x406e3, pf=0x40, revision=0xba
Jul 03 19:33:02 home2 kernel: microcode: Microcode Update Driver: v2.2.

So in my case I have the fix. Hopefully any others with this issue will have a fix soon too, though recent posts in the tech media in the last few days suggest that the problem is likely to manifest itself only on very rate occasions with unusual conditions the the processor would be dealing with.

Last edited by mcloaked (2017-07-03 19:13:17)

Arch Linux

#1 2017-06-27 20:55:51

[SOLVED] Broken hyperthreading in Skylake and Kaby Lake

#2 2017-06-27 20:58:36

Re: [SOLVED] Broken hyperthreading in Skylake and Kaby Lake

#3 2017-06-27 21:06:39

Re: [SOLVED] Broken hyperthreading in Skylake and Kaby Lake

#4 2017-06-27 21:08:30

Re: [SOLVED] Broken hyperthreading in Skylake and Kaby Lake

#5 2017-06-28 12:42:31

Re: [SOLVED] Broken hyperthreading in Skylake and Kaby Lake

#6 2017-07-03 18:50:44

Re: [SOLVED] Broken hyperthreading in Skylake and Kaby Lake

Board footer