1 |
In all of the uclibc ebuilds (all in the current portage) I find the |
2 |
following lines: |
3 |
|
4 |
if [[ $(tc-is-softfloat) != "yes" ]] ; then |
5 |
sed -i -e '/^HAS_FPU=y$/d' .config |
6 |
echo 'HAS_FPU=n' >> .config |
7 |
fi |
8 |
|
9 |
According to the documentation of tc-is-softfloat the return is: |
10 |
# - only: the target is always softfloat (never had fpu) |
11 |
# - yes: the target should support softfloat |
12 |
# - no: the target should support hardfloat |
13 |
|
14 |
So, if the target should support hardfloat it forces the FPU off? I may |
15 |
be just tired, but that seems backwards to me. |
16 |
|
17 |
I have an i586 target, so it has an FPU. If I comment out those lines |
18 |
above in the ebuild uclibc compiles fine and runs fine. If I leave it |
19 |
in I get an error when it tries to compile that the softfloat routines |
20 |
are not found. |
21 |
|
22 |
Am I confused? Am I doing something else wrong and this is just a symptom? |
23 |
|
24 |
Thanks in advance for the help! |
25 |
|
26 |
Scott =) |
27 |
|
28 |
-- |
29 |
Electrical Engineering/Web Development |
30 |
Hunt Utilities Group LLC http://www.hugllc.com |
31 |
Pine River, MN (218) 587-5001 |
32 |
Hopkins, MN (952) 935-2418 |
33 |
|
34 |
-- |
35 |
gentoo-embedded@g.o mailing list |