1 |
On 12/19/08, Jakub Ladman <ladmanj@×××××.cz> wrote: |
2 |
> Because of reading this message, i am trying to create toolchain (with |
3 |
> crossdev) for the maverick crunch again. |
4 |
> I made a try with gcc-4.3.2 and 4.1.2, but both failed with this messages: |
5 |
> |
6 |
> collect2: ld terminated with signal 8 [Výjimka matematického koprocesoru |
7 |
> (SIGFPE)] during the last stage - compiling C++ compiler. |
8 |
> |
9 |
> I have czech localization, so: |
10 |
> Výjimka matematického koprocesoru = FPU exception |
11 |
> |
12 |
> What you think about? |
13 |
|
14 |
I've never had that failure. You need to say exactly what tools you |
15 |
are using to create the cross toolchain, what exact command lines you |
16 |
issue, what special lines you have in config files. |
17 |
|
18 |
Like I say, I failed to find a combination of options that will set |
19 |
Maverick hardfloat as the default, so I just create a regular ARM |
20 |
toolchain, then pass |
21 |
-mcpu=ep9312 -mfpu=maverick -mfloat-abi=softfp |
22 |
when using it to compile binaries. Yes, it's a pain. |
23 |
|
24 |
If you want to compile for Maverick FP, check out |
25 |
martinwguy.co.uk/martin/crunch (published yesterday!) where there is a |
26 |
set of patches to make gcc-4.3.2 generate working code for Maverick |
27 |
Crunch. Otherwise GCC producing Maverick instrutions will definitely |
28 |
output bad code that gives the wrong results - this may be the root |
29 |
cause of the problem you describe. |
30 |
|
31 |
Good luck, let us know how it goes. |
32 |
|
33 |
M |