Gigabyte and AsRock have rolled out new BIOS for their AM4 motherboards which include AGESA 1.0.0.4 firmware from AMD. This version of the AGESA *will* advertise that it has SEV support even if it DOES NOT support it.
The Linux kernel WILL absolutely try to use the SEV support and the code as of right now DOES NOT have any kind of timeout when it, in reality, isn't present.
The result of this is that IF YOU UPGRADE TO THE LATEST AM4 BIOS YOU WILL END UP WITH A SYSTEM THAT SIMPLY HANGS ON BOOT. I put this in big capital letters because it's a rather important point. You won't be able to boot live USB sticks either since all kernels except very old ones without SEV support will just hang forever.
AMD has submitted a patch to the kernel team which adds a timeout when SEV is advertised but not present. It is a bit strange that they did this instead of fixing their firmware bug. This patch is not yet accepted.
There you have it, I have warned you, NO-NO upgrading to the latest AM4 BIOSes or you will not be able to use your computer! It will be a brick and you will have to either downgrade your BIOS so you can boot GNU/Linux or switch to Windows.
The fuck? 1.0.0.6 was released like more than a year ago?
Brandon Martin
OP is a fag, and most likely a shilltel
Josiah Jenkins
Not entirely an accurate description, the kernel patch was written and submitted by an AMD employee. marc.info/?l=linux-crypto-vger&m=153436754612783 probably the same guy responsible for the firmware bug in the first place
See picture related, AMD AGESA versions are absolute nonsense. >1004a April 2017 > 1.0.0.6a June 2017 >1.0.7.2a December 2017 >1.0.0.0a February 2018 (???) >1.0.0.1a March 2018 > 1.0.0.4 August 18th, 2018 not sure what they are thinking
If you are running linux and you happen to do the bios update and hit this issue, simply reboot and downgrade properly back to a stable BIOS.
Nothing of value lost/gained. It does surprise me though that they didn't catch this shit. I've encountered it myself and had to downgrade. Nothing gets bricked in BIOS so you simply revert back to prior version
Luis Sanchez
>most likely a shilltel I can assure you that I am, in fact, not the shilltel. I ran into this when I updated my BIOS and found that I could no longer boot my computer. After having wasted time tried to see if it was causes by some BIOS setting - which it wasn't, I found that the only solution was to downgrade my BIOS.
This thread is simply a way of letting others know so they don't make the mistake of "upgrading" AM4 Linux systems to a bad BIOS.
Still gonna require work from kernel devs to ensure the patch doesn't break things. And patching a kernel non-bug to accommodate AMD's bug? Disgusting. Hopefully the kernel devs drag their feet on this so people will blame AMD for this shit. THEN maybe they'll fix their own bug.
>rolled out new BIOS for their AM4 motherboards which include AGESA 1.0.0.4 wasn't it like 40 days back?
David Miller
The Gigabyte BIOS I got burned with for my board is dated 2018/08/08. ASUS hasn't yet pushed out BIOSes with the new AGESA for their boards (perhaps because they did internal testing and found that it broke GNU/LInux systems?). Perhaps it's not that new but it is the latest.
Joseph Rogers
>all of these cucks complaining that their botnet doesn't work
do share how well libreboot works on not ancient outdated hardware. yeah, I know it's great for freedom and not botnet. A typewriter is also not botnet but it doesn't really compare to a modern computer.
Illustrate my point, tell us all about what fancy hardware you're booting with libreboot
John Taylor
Gentoo doesn't have this problem.
Luke Ward
>Buying into the new technology Jew
Literally don't need more than two cores.
Oliver Cruz
Gentoo doesn't have this problem if you compile your kernel with CONFIG_CRYPTO_DEV_SP_PSP=n or use openrc instead of systemd-udev, that's .. true.
It's more fair to say that some Gentoo users don't have this problem and some do.
Yeah but compiling a new kernel on a live cd takes nothing, remember you can easily pass parameters to the installation live cd before booting.
Robert Cruz
Wendell from Level1Techs brought up the issue you fuckign wangblow
and wendel is the biggest AMD fanboy. he just want what's good for AMD
Colton King
what ethnicity is this creature?
Xavier Johnson
>all these SJW cucks working for years to make LibreBoot work but it turns out the hardware (Intel processors) have a fuckhuge amount of hardware security vulnerabilities
it's like putting make-up on a pig. Sure, the pig will look prettier, but it will still be a pig
It must be difficult differentiating your RAM when it all comes from the same forge. Pretty girls and bright colors are pretty standard ways of distracting the buyer.
Eli Russell
AYY LMAO Y Y
Lincoln Barnes
Manly Tears transition went very well, good for her.
Christian Rogers
It's also worth noting that AGESA 1.0.0.4 seems to cause suspend/resume issues for those few that have decided to use that BIOS with various kernel workarounds. Seems like they made more than one mistake with this one, might as well wait for a fixed version since previous BIOS works just fine. Lacks support for new chips like the Athlon 200GE, though, but who's got one of those anyway
