Gentoo Archives: gnap-dev

From: Philipp Riegger <lists@××××××××××××.de>
To: gentoo-soc@l.g.o, gnap-dev@l.g.o, Christel Dahlskjaer <christel@g.o>, g2boojum@g.o
Subject: [gnap-dev] GNAP Cross Compile Support: Status Report
Date: Mon, 09 Jul 2007 22:10:57
Message-Id: 1184018652.32228.58.camel@b136a
1 Hi.
2
3 I want to tell you about the status of my project.
4
5 My plan was like this:
6
7 Part 1: 2 weeks: Getting to know all the things.
8 Part 2: 4 weeks: Integrate cross compile support into gnap/catalyst.
9 Part 3: 2 weeks: Make gnap working on arches other than x86/amd64.
10 Part 4: 3 weeks: Proof of concept, support one board.
11 1 week of buffer.
12
13 If you want more details... well, its in the project description. Is
14 that pblicly available? I'm not sure. If not and you want to read it,
15 write a mail.
16
17 At the moment i'm slightle behind schedule and things also have changed
18 a little.
19
20 6 weeks are gone, i almost finished part 1. From my point of view, the
21 gnap_* scripts were in quite a bad state. I cleaned them up a bit, did
22 some refacturing and some improvements. Some small things still have to
23 be done, and i have to create the patchset, but i'm almost finished. The
24 plan is also to provide more than just one example configurations. Bass
25 was working on a new portage snapshot, a new seedstage and also a new
26 kernel. I tested this intensively, built several images and extensions
27 and looked into the kernel patches used (src.rpm from redhat/fedora).
28
29 For part 2 i'm still waiting for vapiers patches. I emailed him when the
30 project started, theese mails seem to have been lost. I pinged him on
31 IRC several times and finally reached him one week ago, he said he would
32 fiel a bug with the patches and inform me. Some people told me i should
33 start from scratch because vapier is quite busy and does not have time
34 to send the patches. But since he's much more into this and i really
35 want to see what he did, i have been waiting until now. It's not that i
36 have done nothing, i built some cross toolchains, looked into cross
37 building stage1 image building and played around a bit. I understood
38 most of how this stuff works (i hope) and i know how i would do this.
39 I'll remind vapier and if i don't hear anything positive i'll start from
40 scratch tomorrow. Time is running out.
41
42 I started on part 3 and it works better than i thought. The hardest part
43 was getting access to non-x64, non-amd64 machines and to get the rights
44 i need (root, at least in a chroot). Since half a week i have what i
45 need on ppc(64), i'm very happy with what i reached so far. Building
46 with glibc works great, i just have some problems getting started with
47 uclibc.
48
49 Part 4 is a little difficult. I could choose a ppc board and build
50 images on the ppc64 machines not using cross compilers, or i focuse more
51 on the other tasks and maybe leave that aside. I'll have to look into
52 that as soon as get further with the other stuff.
53
54 Ok, that's what i have. A little behind, but i'm working. There are some
55 problems i faced so far which surely made things harder for me:
56
57 - The last semester used more time than was planed. I did not have the
58 time to look into the things i am working with until summer of code
59 started, and this started for me about 1 1/2 week late because i still
60 had to finnish university stuff.
61 - My Laptop decided to not cooperate anymore with me and i had to get
62 an old PC from a friend and get it running/configure it so that i can
63 work with. Not that easy, since i'm abroad, just finished my exchange
64 studies, and i have nothing here.
65 - Well, i had to get used to a proper version control (the cmmands are
66 easy, it's harder to get as much out of it as possible).
67
68 But all of the above mentioned things are working now, i'm very happy
69 with all of them and i finally have my focus where it should be, on
70 GSoC.
71
72 Something on what i plan for the future:
73
74 Today/tomorrow i will sort all the things out i did as improvement for
75 the gnap_* scripts, finnish the work and clean things up. I'll rerun the
76 test builds over night to make (at least to some point) sure, that i did
77 not break anything.
78
79 I'll ping vapier again and again and like mentioned above, tomorrow i'll
80 look into that stuff myself. I should also make the decision about part
81 4, i'll discuss that in the next meeting with my mentors next monday and
82 make a decision next week.
83
84 Some things that need to be sorted out:
85
86 I read that kingtaco wanted to ask christel about misinforming me about
87 the soc box gentoo bought. Nobody ever told me what the misinformation
88 was. I also heard that it is not ready. Quite good, i'm not the only one
89 behind schedule. I heard that it is quite powerful (which might prove
90 very usefull for me) but it seems not to be possible to get the level of
91 access i need, which would be sudo for a special chroot script. The
92 usual hardened setup gentoo servers have and the script would make this
93 safe (at leat me and the admin of a box i got that level of access
94 thought that). Ok, 4 chroot restrictions from grsec need to be disabled
95 for catalyst, but it should still be quite safe. I never got a reply to
96 the mail i sent.
97
98 So much from me, if you want to know more, i read my mail.
99
100 Philipp
101
102 --
103 gnap-dev@g.o mailing list