Gentoo Archives: gentoo-dev

From: Ben Kohler <bkohler@g.o>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] Adding USE=udev to linux profiles
Date: Thu, 26 Jul 2018 13:07:12
Message-Id: 8bfd2538-7d42-ce4f-c083-4c5afffa2831@gentoo.org
In Reply to: Re: [gentoo-dev] Adding USE=udev to linux profiles by Andrew Savchenko
1 On 07/26/2018 02:59 AM, Andrew Savchenko wrote:
2 > Hi!
3 >
4 > On Thu, 19 Jul 2018 16:51:17 -0500 Ben Kohler wrote:
5 >> I'd like to propose adding USE=udev to our linux profiles (in
6 >> profiles/default/linux/make.defaults probably). This flag is already
7 >> enabled on desktop profiles but it also affects quite a few packages
8 >> used on non-desktop linux systems.
9 >>
10 >> This flag provides useful functionality that most linux users will want.
11 >> I'm a bit surprised that we still don't have it in all linux profiles,
12 >> but I think we've worked around this in the past by adding IUSE=+udev to
13 >> quite a few of those packages (33 packages, 116 ebuilds, by my count).
14 >>
15 >> This missing flag came to my attention again on bug 661584 where lvm2
16 >> has IUSE=+udev but cryptsetup has only IUSE=udev, so non-desktop users
17 >> have a bit of a mismatch between the 2 and get ugly errors on cryptsetup.
18 >>
19 >> Since this flag only affects linux, I think it makes more sense to set
20 >> it in linux profiles than to use IUSE defaults.
21 >>
22 >> Any objections to this idea?
23 >
24 > A user had contacted me with his input from the HPC world, I'm
25 > redirecting his e-mail here. James is whitelisted now so he can
26 > further participate in this discussion himself if necessary.
27 >
28 > As an HPC user of Gentoo I agree that minimal and highly optimized
29 > Gentoo setups are indeed very useful and must stay that way.
30 >
31 > Begin forwarded message:
32 >
33 > Date: Wed, 25 Jul 2018 13:31:59 -0400
34 > From: james <garftd@×××××××.net>
35 > To: bircoph@g.o
36 > Subject: udev's future
37 >
38 >
39 > Hello Andrew,
40 >
41 >
42 > Sorry, I do not have direct posting ability to gentoo-dev, so in
43 > hopes of finding a dev-sponsor, I hope you will paraphrase this
44 > email to you for the sake of preventing 'dev as a default' or base
45 > setting of any sort.
46 >
47 >
48 > My research and testing for new HPC configurations, has systemd
49 > and udev at the heart of codes to avoid to optimize the
50 > heterogeneous nature of the clusters I'm building. In fact my
51 > development work, although delayed due to transient-illness, is
52 > more of a gentoo-centric convergence of embedded-gentoo, minimal
53 > (server) gentoo, gentoo-hpc clusters and unikernels. As far as
54 > performance and security are concerned 'less' is always better.
55 > Those codes and ebuild that are desired are to added in a higher
56 > level; hoping to continue the leverage the portage tree of
57 > applications, only as dynamically required.
58 >
59 >
60 > Avoidance of setting udev or in any form mandating any part of
61 > systemd will have dire consequences and cost me months, if not
62 > years to find a way to 'totally undo' the ravages of udev.
63 > Minimized kernels are also fundamental to my loosely-coupled
64 > (gentoo) HPC development. Even tiny Rtos based embedded linux
65 > systems are in the process of being included in a loosely-coupled
66 > gentoo centric heterogeneous HPC cluster. I would 'beg' against
67 > making udev primary under any circumstance.
68 >
69 >
70 > Gentoo has a unique and powerful position, just for it's position of
71 > choice and minimizational features; After 15 years, I'd hate to have
72 > to work in another distro, as gentoo has served me extraordinarily
73 > well over the decades.
74 >
75 >
76 > sincerely,
77 > James Horton, PE
78 >
79 > Best regards,
80 > Andrew Savchenko
81 >
82 No one was ever talking about forcing udev usage, just default-enabling
83 support on a few MORE packages than we already do. Our standard linux
84 stage3's are already udev-enabled. But not udev-forced, anywhere.
85
86 Nothing about my proposal was going to force udev on people who don't
87 want udev at all-- let's not even go down that rabbit hole of discussion.
88
89 I was only pushing for more consistency-- either your system would be
90 fully udev-enabled, or not at all.
91
92 -Ben