[SOLVED] How to set up early load of Intel firmware

Tonus · 06-05-2019, 08:08 PM

Thanks abga, I will try that if it's the sure path!

And thanks @rworkman for being the only one I found to state :

Quote:

MICROCODE_ARCH This option specifies a cpio archive containing updated microcode for your CPU. CPU manufacturers occasionally release such updates to fix bugs in the microcode currently embedded in the CPU. The microcode archive will be prepended to the output initrd, where the kernel will find it for early patching. .br Examples:
MICROCODE_ARCH="/boot/intel-ucode.cpio"

abga · 06-05-2019, 08:26 PM

On Haswell (and maybe others) I had to use the early loading method with an initrd boot loader call (check the Gentoo link):
https://www.linuxquestions.org/quest...ml#post5995482

Lysander666 · 07-29-2019, 02:52 PM

There's new microcode at Slackbuilds, so I built and installed it. Built a new initrd for 4.4.186 and now I get this

Quote:

root@psychopig-xxxiv:/var/log# grep microcode dmesg
[ 0.000000] microcode: CPU0 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 0.008671] MDS: Vulnerable: Clear CPU buffers attempted, no microcode
[ 0.002000] microcode: CPU1 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 0.002000] microcode: CPU2 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 0.002000] microcode: CPU3 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 1.595494] microcode: CPU0 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.595561] microcode: CPU1 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.595635] microcode: CPU2 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.595709] microcode: CPU3 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.595832] microcode: Microcode Update Driver: v2.01 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

So it seems to be loading it but it's an old version?

EDIT: an interesting difference in huge kernel for 4.4.153:

Quote:

root@psychopig-xxxiv:/var/log# grep microcode dmesg
[ 0.000000] microcode: CPU0 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 0.002000] microcode: CPU1 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 0.002000] microcode: CPU2 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 0.002000] microcode: CPU3 microcode updated early to revision 0xa0b, date = 2010-09-28
[ 1.471549] microcode: CPU0 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.471624] microcode: CPU1 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.471702] microcode: CPU2 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.471774] microcode: CPU3 sig=0x1067a, pf=0x10, revision=0xa0b
[ 1.471862] microcode: Microcode Update Driver: v2.01 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

i.e. no inclusion of this line

Quote:

[ 0.008671] MDS: Vulnerable: Clear CPU buffers attempted, no microcode

I'm thinking that there's something about the firmware for 4.4.186 that isn't compatible with the newer microcode for systems like mine.

magicm · 07-29-2019, 05:04 PM

That's odd. I get this:

Code:

root@igloo: Mon Jul 29 16:41:46 : /var/log
# grep microcode dmesg
[    0.000000] microcode: CPU0 microcode updated early to revision 0x2f, date = 2019-02-17
[    0.057917] microcode: CPU2 microcode updated early to revision 0x2f, date = 2019-02-17
[    4.480656] microcode: CPU0 sig=0x206a7, pf=0x10, revision=0x2f
[    4.500930] microcode: CPU1 sig=0x206a7, pf=0x10, revision=0x2f
[    4.519903] microcode: CPU2 sig=0x206a7, pf=0x10, revision=0x2f
[    4.537017] microcode: CPU3 sig=0x206a7, pf=0x10, revision=0x2f
[    4.552220] microcode: Microcode Update Driver: v2.01 <tigran@aivazian.fsnet.co.uk>, Peter Oruba

My hardware: Thinkpad T420s

Code:

# inxi
CPU: Dual Core Intel Core i5-2540M (-MT MCP-) speed/min/max: 1017/800/3300 MHz Kernel: 4.4.186 x86_64 Up: 25m

How is it that you told your initrd to load the new microcode? I use a mkinitrd.conf with line

Code:

MICROCODE_ARCH="/boot/intel-ucode.cpio"

Lysander666 · 07-29-2019, 05:44 PM

Quote:

Originally Posted by magicm

How is it that you told your initrd to load the new microcode? I use a mkinitrd.conf with line

Code:

MICROCODE_ARCH="/boot/intel-ucode.cpio"

For generic, the initrd is made with -P /boot/intel-ucode.cpio

For huge,

Code:

initrd = /boot/intel-ucode.cpio

is added to lilo.conf.

Working fine in -current:

Code:

root@lysultra-vi:/var/log# grep microcode dmesg
[    0.000000] microcode: microcode updated early to revision 0xcc, date = 2019-04-01
[    4.310155] microcode: sig=0x406e3, pf=0x80, revision=0xcc
[    4.310432] microcode: Microcode Update Driver: v2.2.
root@lysultra-vi:/var/log# inxi
CPU: Dual Core Intel Core i5-6300U (-MT MCP-) speed/min/max: 2963/400/3000 MHz Kernel: 4.19.58 x86_64 Up: 11h 37m 
Mem: 1082.0/3609.0 MiB (30.0%) Storage: 148.97 GiB (37.5% used) Procs: 199 Shell: bash 5.0.7 inxi: 3.0.34

But 14.2 is giving me the issues I described above. Which is weird, esp in huge since it's the same process - just one line appended to lilo.conf.

Lysander666 · 08-03-2019, 02:35 PM

OK I solved this, according to this link:

https://www.intel.com/content/dam/ww...e_05132019.pdf

my desktop processor, a Q8400 [Yorkfield], no longer receives microcode updates. And this is because Intel have admitted they can't, and won't, fix certain families

https://www.theregister.co.uk/2018/0...ocode_updates/