1 |
Hello, |
2 |
|
3 |
I have a bit more brain resources to continue the integration of Dracut |
4 |
with Genkernel, at last. Dracut was meant to replace internal Genkernel |
5 |
initramfs generator, but recently Genkernel is being developed quite |
6 |
actively, therefore I think it would be a waste to just replace it. |
7 |
Both can coexist. Although there are a lot of differences between |
8 |
Genkernel and Dracut, therefore there's no much sense to continue the |
9 |
integration the way I have started. I have tried to replace initramfs |
10 |
generator transparently, translating command line options and |
11 |
configuration to Dracut options, to be fully compatible. Quickly it has |
12 |
become a burden, because Genkernel code has to be updated according to |
13 |
changes in Dracut. Just imagine you would have to update Genkernel for |
14 |
every kernel release. |
15 |
|
16 |
I propose to extract the initramfs generator from Genkernel and make it |
17 |
a standalone project. The rest of the Genkernel could be rewritten to |
18 |
support both Gentoo's initramfs generator, Dracut and maybe some other |
19 |
in the future. Of course this would be my job to do so and I'm only |
20 |
asking for acceptance from you. |
21 |
|
22 |
Another problem arisen is the name and versioning. Solutions: |
23 |
|
24 |
1) The framework integrating old Genkernel and Dracut could be just |
25 |
bumped to: |
26 |
|
27 |
a) 4 - because there's no official 4.x release, |
28 |
b) 5 - because there's an unofficial and dead 4.x: |
29 |
https://github.com/eedgar/genkernel4 |
30 |
|
31 |
2) The framework could be named genboot, because it actually takes care |
32 |
of kernel, initramfs and boot manager entries. Although we're losing |
33 |
old good name. |
34 |
|
35 |
I am definitely for (1). The extracted old internal Genkernel initramfs |
36 |
generator would need to be named differently for either case, for |
37 |
example: geninitrd, geninitramfs, genifs, genird or geninitfs. |
38 |
|
39 |
|
40 |
-- |
41 |
Amadeusz Żołnowski |