Gentoo Archives: gentoo-user

From: "J. Roeleveld" <joost@××××××××.org>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] Re: Question about genkernel's default kernel config
Date: Mon, 15 Aug 2016 08:12:31
Message-Id: 2060260.7un78O2rPg@andromeda
In Reply to: Re: [gentoo-user] Re: Question about genkernel's default kernel config by Jeroen Mathon
1 On Monday, August 15, 2016 09:48:14 AM Jeroen Mathon wrote:
2 > I have used genkernel for my initramfs.
3 >
4 > The dracut one seems to hang on a kernel panic(Implying that it did not
5 > load the correct modules for luks and lvm).
6
7 In my experience, normal for both...
8
9 > Besides the size of your initramfs should not matter since its only used
10 > to mount the partitions and do the pre-kernel loading.
11
12 Agreed, except for the next part:
13
14 > The only thing that you will win by making your own initramfs would be
15 > size and boot speed.
16
17 If done sanely, the reduced complexity makes debugging and maintenance a lot
18 easier.
19
20 > Correct me if im wrong btw.
21 >
22 > On 15-08-16 02:45, Rich Freeman wrote:
23 > > On Sun, Aug 14, 2016 at 7:38 PM, Mike Gilbert <floppym@g.o> wrote:
24 > >> On Sun, Aug 14, 2016 at 4:29 PM, J. Roeleveld <joost@××××××××.org> wrote:
25 > >>> On August 14, 2016 8:55:32 PM GMT+02:00, Neil Bothwick
26 <neil@××××××××××.uk> wrote:
27 > >>>> On Sun, 14 Aug 2016 11:48:08 -0700, Ian Zimmerman wrote:
28 > >>>>> Don't you still need genkernel if you want to build an initramfs?
29 > >>>>
30 > >>>> No, dracut.
31 > >>>>
32 > >>>>> The handbook (amd64) seems to imply you do, and I don't know of an
33 > >>>>
34 > >>>> easy
35 > >>>>
36 > >>>>> way to build an initramfs just with the bare kernel source.
37 > >>>>
38 > >>>> Dracut.
39 > >>>
40 > >>> Dracut if you just want something quick and easy.
41 > >>> If you want something small and reliable, build your own.
42 > >>
43 > >> A homegrown initramfs created by a novice is going to be more reliable
44 > >> than one created by dracut or genkernel? Seems unlikely.
45 > >
46 > > This is a silly argument. Everybody knows that anybody who wants
47 > > something reliable just writes their own kernel in the first place.