Implement UKI and CPU microcode hooks - #49
Conversation
|
Thanks, this looks good. I can test intel ucode. For uki, I have been running similar code for a while. It is perhaps good to note that for hooks (and late hooks) order matters (e.g. device managers first, and here ucode before uki). |
|
I've updated some of the manpage (I realized there were some more changes I wanted to make), but have not made a commit yet; are there any more order dependent hooks? The order in which hooks appear is significant; a device manager hook (e.g. the 'eudev' hook) should be the first element of the list, the 'ucode' hook should be placed before the 'uki' hook, if both are used. Edit: AFAICT there shouldn't be ordering issues with other hooks so I've pushed my changes. If there are any, let me know and I'll change it and force push the changes to keep the commit log clean. |
|
Also, unrelated to this pull request, I implemented a plymouth hook, seems to work including with LUKS but it still has a few issues I need to figure out. |
|
Looks good to me, tested with intel and uki. I was using |
|
(I do not use zfs at the moment, but if you end up opening another pr maybe someone else can help.) |
|
@illiliti Any comments, before I merge this? |
I decided to create a GitHub account, this is an updated version of #48.
Pretty sure in this version all issues have been fixed, but I still have only tested with an AMD machine so it might not work for Intel CPUs.
This pull request, aside from implementing the aforementioned hooks, also allows arbitrary generation of temporary files which will not be included in the final initramfs but will be deleted using the same mechanism other temporary files are deleted, and also adds support for a new type of hook which runs after the initramfs has been generated (required to implement both hooks added by this set of patches).
I have been using this version for a while with no issues.