* Henrique de Moraes Holschuh <[email protected]> [240604 02:51]: > Hello Chris, > > Did you test the result? Does the system boots with an updated microcode > after the change ?
I do not have access to x86 hardware with an intel cpu. > It has to be tested with mkinitramfs and Dracut, the resulting > early-initramfs must be "almost" byte-equal (timestamps can change, but not > the paths/filenames inside, nor the *offset* where the data starts relative > to start-of-file. THIS IS IMPORTANT. > > I believe it should really just work with mkinitramfs, since we ask > iucode_tool to generate the early-initramfs in that case, and iucode_tool > doesn't care where the source data is, as long as it can read it. One should > still test it to be sure. > > Dracut, I haven't looked into what it is doing. Hopefully it also uses > iucode_tool nowadays... Generally I think we should be good, because the files are in the same place *on the installed filesystem* as before. Many moons ago I checked that the initramfs output was the same, but I cannot remember if it was bit identical or not. I'd appreciate it if anyone could do a verification test with the expected hardware and/or provide an autopkgtest. Chris
