Is My CPU Toasted?

[pebcak]

I have an Intel Core i7-8565U with 4 cores and 8 threads on this machine (Dell XPS 13 9380).

I am used to see all the 8 threads enabled and working in htop for example.

All of a sudden, just today, 4 threads (or maybe 2 cores plus 2 threads) are showinf as being “offline”:

I haven’t changed any configurations on the system since it was booted the last time before this happened.

CPU:
  Info: model: Intel Core i7-8565U socket: BGA1528 (U3E1) note: check bits: 64
    type: MCP arch: Whiskey Lake gen: core 8 level: v3 note: check built: 2018
    process: Intel 14nm family: 6 model-id: 0x8E (142) stepping: 0xB (11)
    microcode: 0xF4
  Topology: cpus: 1x cores: 4 smt: disabled cache: L1: 256 KiB
    desc: d-4x32 KiB; i-4x32 KiB L2: 1024 KiB desc: 4x256 KiB L3: 8 MiB
    desc: 1x8 MiB
  Speed (MHz): avg: 1176 high: 1900 min/max: 400/4600 base/boost: 4100/4600
    scaling: driver: intel_cpufreq governor: conservative volts: 0.9 V
    ext-clock: 100 MHz cores: 1: 1900 2: 700 3: 400 4: 1707 bogomips: 15999
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx

I would appreciate if anyone could give me a hint on what I am looking at here.

Not sure either what logs to look into. Please tell me and I’ll provide it.

[anthony93]

What does lscpu -e -c say?

[pebcak]

Thanks for the reply @anthony93 !

Here is the otput:

$ lscpu -e -c
CPU NODE SOCKET CORE L1d:L1i:L2:L3 ONLINE MAXMHZ MINMHZ MHZ
  4    -      -    - -                 no      -      -   -
  5    -      -    - -                 no      -      -   -
  6    -      -    - -                 no      -      -   -
  7    -      -    - -                 no      -      -   -

[BendTheKnee]

@pebcak

smt: disabled

You somehow disabled hyperthreading. Did you fool around in your bios?

[pebcak]

I fool around around a bit but not in the BIOS recently.

However I did add mds=full,nosmt to my kernel boot parameters after seeing a vulnerability being mentioned in dmesg and referring to a link at kernel.org documentation.

Even so, the cores/threads were all online until just now.

[BendTheKnee]

I think that if you remove that kernel parameter and reboot you’ll be okay again.

EDIT: or change it to mds=full

mds=full > If the CPU is vulnerable, enable all available mitigations for the MDS vulnerability, CPU buffer clearing on exit to userspace and when entering a VM. Idle transitions are protected as well if SMT is enabled.

It does not automatically disable SMT.

[pebcak]

I’m sure you are right but I wonder why it did work until just now and stopped working all of a sudden?

Here is what I was referring to:

Mitigation control on the kernel command line

The kernel command line allows to control the MDS mitigations at boot time with the option “mds=”. The valid arguments for this option are:

full
	

If the CPU is vulnerable, enable all available mitigations for the MDS vulnerability, CPU buffer clearing on exit to userspace and when entering a VM. Idle transitions are protected as well if SMT is enabled.

It does not automatically disable SMT.

full,nosmt
	

The same as mds=full, with SMT disabled on vulnerable CPUs. This is the complete mitigation.

off
	

Disables MDS mitigations completely.

https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/mds.html

[pebcak]

@BendTheKnee

EDIT: or change it to mds=full

Going to try that.

[BendTheKnee]

@pebcak

I wonder why it did work until just now and stopped working all of a sudden?

No idea why the parameters should load at each boot.

[pebcak]

@BendTheKnee

change it to mds=full

I booted up with only this now and all cores are back online however I get again the dmesg “warning” and reference to the link that I posted above:

[0.802192] MDS CPU bug present and SMT on, data leak possible. See https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/mds.html for more details.
[ 0.802192] MMIO Stale Data CPU bug present and SMT on, data leak possible. See https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/processor_mmio_stale_data.html for more details.

Also from inxi -aC

Type: mds mitigation: Clear CPU buffers; SMT vulnerable
Type: mmio_stale_data mitigation: Clear CPU buffers; SMT vulnerable

[ricklinux]

You should have got an amd Ryzen.

[pebcak]

Yeah… but this has been a nice lil’ machine so far and I still like it a lot.

It was a Developer Edition from the beginning with Ubuntu pre-installed and it has run every GNU-Linux I have thrown at it nicely and flawlessly. I have to add, despite Ubuntu being Ubuntu, it ran like a breeze on this machine.

[nate]

I wonder, is this vulnerability really a problem without physical access to the MB/CPU? I’m tempted to disable all vulnerabilities mitigations on my machine to gain an extra 10-20% performance, as I’m CPU-bound lately.

[pebcak]

You may be right.

I have been wondering about this myself but haven’t really looked into the details of this specific vulnerability.

You are tempted but you are not actually disabling all the mitigations, right?

[nate]

I didn’t have the time to do a proper research.

[pebcak]

Thanks @BendTheKnee for bringing to my attention that the issue was due to disabled SMT.

I am using mds=full kernel parameter now and as mentioned above I get those vulnerability messages in the output of dmesg and inxi. I guess i have to live with it.

Marking it