1 |
Jean-Marc Beaune wrote: |
2 |
|
3 |
> I tried several times to start with embedded linux but can't really |
4 |
> figure out how to start in a proper way. |
5 |
|
6 |
It's really quite simple. |
7 |
|
8 |
Choose a simple problem to solve that requires an embedded system. |
9 |
Running a large database and sending out packets, is not a valid |
10 |
reason. Why? Because you can use a server for that purpose. |
11 |
Now minimizing that server to fit into an airplane, so it is |
12 |
as light-weight and low power as possible, is a valid reason |
13 |
to build an embedded system. |
14 |
|
15 |
|
16 |
Choose an OS (embedded linux cause it's the coolest and has too many |
17 |
readily available software resources). |
18 |
|
19 |
Embedded linux is not the same thing as embedded (gentoo) linux! |
20 |
Caveat Emptor! |
21 |
|
22 |
Take one sheet of paper and specific the hardware resources you |
23 |
need to solve that problem: |
24 |
|
25 |
Micro controller (decision making capability based on software) |
26 |
|
27 |
A/D analog (input/output) digital converters ? |
28 |
DIO digital input/output |
29 |
usb |
30 |
special sensors such as temperature, pressure, acceleration..... |
31 |
flash |
32 |
amount/types of memory |
33 |
ethernet (or 2) |
34 |
etc etc |
35 |
(you make the list) |
36 |
|
37 |
|
38 |
Folks on this list will help you with identifying and |
39 |
solving your problem. |
40 |
|
41 |
Choose a processor/arch that readily solves your design constraints |
42 |
but has enough computation resources to run embedded linux |
43 |
and to solve your problem, too. |
44 |
|
45 |
|
46 |
Here is where you ask the list for architecture recommendations |
47 |
for example. If your problem requires low power, so as to |
48 |
run off of batteries or a solar panel, then that will |
49 |
affect your selection of processor. Different processors |
50 |
use vastly different amounts of energy. Low power is but |
51 |
one dimension of processor selection. |
52 |
|
53 |
If you just want to noodle in the office, you may want to |
54 |
choose a processor that comes on a SBC (single board |
55 |
computer) such as mini-itx form factor or one of the epia |
56 |
boards and an x86 arch, so it is closest to what you want, |
57 |
and most familiar (arch) to what you have used. |
58 |
|
59 |
You can look at the various arch's supported by gentoo embedded, |
60 |
as that will quickly 'skinny down' your selection. |
61 |
|
62 |
You can even visit various vendor web sites to look through |
63 |
a myriad of boards (SBC) that the various vendors offer. |
64 |
|
65 |
Then ask on this list to see what others have done, so you |
66 |
can leverage their knowledge and get a SBC (embedded gentoo linux |
67 |
board) that is familiar to persons on this list. That way |
68 |
help is easy to find. |
69 |
|
70 |
|
71 |
Maybe you should start by reading the list archives to see |
72 |
what other (embedded) noobs have chosen to begin their |
73 |
quest? |
74 |
|
75 |
|
76 |
Alternatively, maybe GNAP is ready for noobs to once again build a |
77 |
simple firewall on an old x(4)86 box as their first exposure to |
78 |
embedded (gentoo) linux. Make sure there is a 'body of work' |
79 |
within embedded gentoo for the processor/SBC you choose, |
80 |
if you want the least painful path towards your quest(goal) |
81 |
of solving the problem you defined above. A firewall is a nice |
82 |
project, because it's very likely to be something that you |
83 |
use and maintain. Choose your problem, wisely. GNAP is a very |
84 |
inexpensive introduction to embedded gentoo, as most folks |
85 |
have old pc hardware just laying around. |
86 |
|
87 |
|
88 |
Building an embedded system that is part of an aimless wandering |
89 |
experience, most often leads to failure and abandonment. If you |
90 |
build something that you want and will use to solve a problem, then |
91 |
you will be hooked and will be well on your way to becoming an |
92 |
embedded systems develop, imho. |
93 |
|
94 |
|
95 |
Embedded systems are spartan (hard and minimalistic) ventures that |
96 |
are best employed to solve narrowly focused problems. That's the |
97 |
justification for not using a server or workstation to solve |
98 |
the same problem. |
99 |
|
100 |
ymmv, |
101 |
|
102 |
James |