1 |
Mike Frysinger wrote: |
2 |
> On Thursday 26 April 2007, Jason wrote: |
3 |
>> This sounds like the solution to my problem, although I'm a bit hesitant |
4 |
>> to muddy my ${SYSROOT}... |
5 |
> |
6 |
> i dont understand what you mean by "muddy" ... it's managed by portage, so you |
7 |
> could just as easily do: |
8 |
> ROOT=/usr/$CTARGET emerge -C zlib ... |
9 |
|
10 |
yeah, I think the big fat red warning in the guide led me to assume that |
11 |
I should only let crossdev populate $SYSROOT... I'm over it now. :-) |
12 |
|
13 |
> |
14 |
>> Is this what most people do (libs/headers to SYSROOT)? |
15 |
> |
16 |
> it's what i do ;) |
17 |
|
18 |
Good to know I'm on the right track... |
19 |
|
20 |
So, small update. I added the serial port to the nslu, and cut the |
21 |
resistor (change processor speed from 133MHz to 266MHz). |
22 |
|
23 |
I just uploaded my first kernel (2.6.20-gentoo-r6) and ramdisk via tftp |
24 |
using this guide[1]. A few false starts with the kernel (serial port |
25 |
shouldn't be a module, need ramdisk, etc), and it's looking good. It |
26 |
halts now with two errors, one seems to be a size mismatch between the |
27 |
kernel's config and the physical size of the ramdisk fs. The other is |
28 |
asking for a pointer to init... neat. |
29 |
|
30 |
After doing this: |
31 |
|
32 |
# xmerge --root /rootfs uclibc busybox baselayout-lite dropbear |
33 |
|
34 |
# cd /rootfs |
35 |
|
36 |
# du --max-depth=1 |
37 |
12 ./lost+found |
38 |
1 ./tmp |
39 |
292 ./var |
40 |
4394 ./lib |
41 |
3406 ./usr |
42 |
57 ./sbin |
43 |
21 ./etc |
44 |
825 ./bin |
45 |
1 ./home |
46 |
1 ./root |
47 |
1 ./mnt |
48 |
1 ./proc |
49 |
2 ./dev |
50 |
9015 ./ |
51 |
|
52 |
Not bad. I haven't even cleaned it out yet (I have a whole bunch of |
53 |
unneeded modules in /lib) |
54 |
|
55 |
|
56 |
Jason. |
57 |
|
58 |
[1] - http://www.nslu2-linux.org/wiki/HowTo/TestAnImageInRamUsingRedBoot |
59 |
-- |
60 |
gentoo-embedded@g.o mailing list |