1 |
To reactivate this old 486 laptop that's been sitting in my basement, I |
2 |
set out to install it with a tiny Gentoo system and use it as a DSL |
3 |
router. The HD is 1.3 GB, so a full glibc-based system wouldn't be much |
4 |
of a problem, but I wanted to experiment with embedded stuff anyway, |
5 |
so... |
6 |
Well, I've never sone a Stage1 install. Upon my first try with Gentoo I |
7 |
ran into some problem and thought WTH, I'll just go with Stage3. But |
8 |
now, following the HOWTO at http://www.bulah.com/embeddedgentoo.html, I |
9 |
have to do it. |
10 |
All is fine up to the bootstrapping. I have a P4 Gentoo machine, trying |
11 |
to compile for i486. My short make.conf: |
12 |
|
13 |
CFLAGS="-Os -march=i486 -pipe -fomit-frame-pointer" |
14 |
CHOST="i486-gentoo-linux-uclibc" |
15 |
CXXFLAGS="${CFLAGS}" |
16 |
FEATURES="-sandbox buildpkg" |
17 |
UCLIBC_CPU="486" |
18 |
USE="bitmap-fonts minimal truetype-fonts" |
19 |
|
20 |
Trying to boostrap gcc fails with a segfault: |
21 |
|
22 |
gengtype-yacc.c: In function `yydestruct': |
23 |
gengtype-yacc.c:725: warning: traditional C rejects ISO C style function definitions |
24 |
stage1/xgcc -Bstage1/ -B/usr/i486-gentoo-linux-uclibc/bin/ -DEFAULT_PIE_SSP -DEFAULT_RELRO -DEFAULT_BIND_NOW -DUSE_UCLIBC -march=i486 -pipe -O2 -DIN_GCC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long -DHAVE_CONFIG_H -DGENERATOR_FILE -o gengtype \ gengtype.o gengtype-lex.o gengtype-yacc.o ../libiberty/libiberty.a |
25 |
/usr/i486-gentoo-linux-uclibc/bin/ld: warning: creating a DT_TEXTREL in object. |
26 |
./gengtype |
27 |
make[2]: *** [s-gtype] Segmentation fault |
28 |
make[2]: Leaving directory `/var/tmp/portage/gcc-3.3.5.20050130-r1/work/build/gcc' |
29 |
make[1]: *** [stage2_build] Error 2 |
30 |
make[1]: Leaving directory `/var/tmp/portage/gcc-3.3.5.20050130-r1/work/build/gcc' |
31 |
make: *** [bootstrap-lean] Error 2 |
32 |
|
33 |
I found that if I hack the Makefile to link all of those helper programs |
34 |
in gcc-3.3.5.20050130-r1/work/build/gcc statically, they won't segfault |
35 |
and even produce something that will compile. But then I get another |
36 |
segfault when the resulting "xgcc" binary is run for the first time, so |
37 |
there is probably a systematic problem. |
38 |
Any ideas on what might be going wrong would be highly appreciated. |
39 |
|
40 |
cheers! |
41 |
Matthias |
42 |
|
43 |
-- |
44 |
I prefer encrypted and signed messages. KeyID: 90CF8389 |
45 |
Fingerprint: 8E 1F 10 81 A4 66 29 46 B9 8A B9 E2 09 9F 3B 91 |