[Bug libstdc++/100444] std::random_device isn't random on AMD

["ecree429 at virginmedia dot com" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100444

--- Comment #7 from Edward Cree <ecree429 at virginmedia dot com> ---
(In reply to Richard Biener from comment #4)
> people that
> cannot be bothered to update their ucode or the kernel are not likely
> bothered to update libstdc++ either.

Fwiw, my kernel is fairly up-to-date (5.9.0-3-amd64); it and my libstdc++ get
updates pretty regularly from the distro, whereas (like most normal folks) I
don't tend to reflash my BIOS without a good reason.  (Which this is, I grant
you, but I only found out recently that it was causing me any issues, and I'm
holding off on the update in case affected projects want me to test mitigations
/ run additional experiments on the buggy hw.)  And while Debian can apply
updated ucode at boot-time, it's not enabled by default because the ucode
binaries are non-free.

So a user of a freedom-respecting OS, following normal routine update
procedures, will automatically get libstdc++ updates but not ucode updates.

But if, having considered it, you decide it's not worth it for libstdc++ to
mitigate this, that's fine by me — after all, now that I know about the issue,
I can fix my system.  Question is whether you want to try to protect other
users from hitting the same thing.