Linux 5.19 Frowns On x86/x86_64 Late Microcode Loading - "It's Just Lottery & Broken"

[NobodyXu]

I'm not sure if you mean having a single file or having a generic build that works everywhere, but in the first case I think you can bundle the initrd in your regular kernel, and for UEFI you can make a PE that includes both the kernel and initrd, there's a guide in Lennart's blog (he promotes it for systemd-boot, but should work regardless).

If you build your own kernel, then you can put the initramfs directly into the kernel in `make menuconfig`.

The real challenge here lies on how you compile libc for your initramfs, as it depends on kernel headers generated by the compilation of the kernel.
So you would have to compile your kernel first, then build your initramfs, then compile your kernel again with the updated initramfs.

[sinepgib]

If you build your own kernel, then you can put the initramfs directly into the kernel in `make menuconfig`.

The real challenge here lies on how you compile libc for your initramfs, as it depends on kernel headers generated by the compilation of the kernel.
So you would have to compile your kernel first, then build your initramfs, then compile your kernel again with the updated initramfs.

Shouln't the "don't break userspace" rule be enough as long as you keep the same .config?
Then it's not as much compiling as it is buildintg the image AFAICT.

[carewolf]

I'm not sure if you mean having a single file or having a generic build that works everywhere, but in the first case I think you can bundle the initrd in your regular kernel, and for UEFI you can make a PE that includes both the kernel and initrd, there's a guide in Lennart's blog (he promotes it for systemd-boot, but should work regardless).

I meant a single image file. It is easier to test things that way, but also forces you to make sure you have all the right modules integrated into the image. If you want something that works everywhere initrd is much easier.

[NobodyXu]

Shouln't the "don't break userspace" rule be enough as long as you keep the same .config?
Then it's not as much compiling as it is buildintg the image AFAICT.

Yeah, but on the first run, you would have to build the kernel twice, so it indeed makes thing a bit harder.

[gigaplex]

Yeah, but on the first run, you would have to build the kernel twice, so it indeed makes thing a bit harder.

If the initramfs just needs the headers, you don't need to do a full compile the first time.

[NobodyXu]

If the initramfs just needs the headers, you don't need to do a full compile the first time.

How do I only generate and install the headers? Is there a special target in the linux Makefile just for this?

[gigaplex]

How do I only generate and install the headers? Is there a special target in the linux Makefile just for this?

I think it's

make headers_install

[NobodyXu]

I think it's

make headers_install

Thanks

[Hafting]

I can't imagine why you'd want to load the proper microcode after your CPU starts doing stuff...

Well, load microcode to make the cpu work better? Surely, the cpu is not so broken that it cannot boot without microcode - because then it couldn't even boot up and get to the microcode loading.

If I get new improved microcode, why should I need to boot the thing? Just load the code. Perhaps the cpu needs to go single-core for a while - but just do that then?