[coreboot] Flashing Coreboot on Lenovo G505s

[awokd]

Mike Banon:
> Problem here is that the latest AMD microcodes officially released by
> AMD at linux-firmware.git are far from being the latest... AMD's f15h
> recent May 2018 commit contains a version 0x6001119 [2012-07-13]
> microcode for TN (CPU ID 0x610F01). Although thats slightly more
> recent than what coreboot has at the moment inside
> F15TnMicrocodePatch0600110F_Enc.c file, I don't understand why AMD
> didn't include 0x600111F [2018-03-05] despite it has been already
> available as a part of UEFI updates as early as March 2018.
> 
> Please take a look at these two changes:
> https://review.coreboot.org/c/coreboot/+/28273/ -
> src/vendorcode/amd/agesa/f15tn: Update microcode to version 0x600111F
> 2018-03-05
> https://review.coreboot.org/c/coreboot/+/28370/ -
> src/vendorcode/amd/agesa/f16kb: Update microcode to version 0x7000110
> 2018-02-09
> Sadly it seems they couldn't be merged to coreboot until AMD releases
> these microcodes "officially" to linux-firmware.git . To avoid the
> manual patching work I wrote some scripts to apply these unofficial
> patches (with SHA256 checksum verification) :
> https://review.coreboot.org/c/coreboot/+/28425/ - AMD microcodes:
> scripts for applying the unofficial (not-merged-yet) updates
> 
> Although its' possible to manually upgrade that binary inside
> linux-firmware.git to the truly latest version (with the help of
> hexedit / Okteta) - its' not very convenient, and your microcode
> version will always depend on your OS configuration. In addition,
> there are some OS like KolibriOS which don't have any microcode
> updating mechanism at all, so have to rely on coreboot in these cases
> On Sun, Sep 23, 2018 at 8:57 PM Nico Huber <nico.h at gmx.de> wrote:
>>
>> On 9/23/18 7:26 PM, Matt B wrote:
>>> Speaking of microcode, I've seen on other threads that the microcode has to
>>> be manually updated for some boards, and for the g505s this is especially
>>> complicate with many manual steps. Can anyone provide a more explicit
>>> series of steps than the below? Is this something being worked on?
>>
>> I might be wrong on this (not experienced with AMD) but usually you can
>> apply microcode updates much easier from the OS (e.g. firmware updates
>> to a state sufficient to boot; OS updates to the final microcode with
>> all current fixes).

Plus from what I've gathered in those discussions, linux-firmware.git is 
lacking the appropriate equivalence table for these AMDs so even if the 
microcode gets updated in the file, the OS will ignore it. I noticed 
that same behaviour when I was hacking on it so approached it from the 
Coreboot side instead, but I think there's enough info now I (or anyone) 
could try testing and submitting a patch against linux-firmware.git 
adding that table. Hopefully AMD releases the updated microcode soon.