| --- Log opened Fri Mar 13 00:00:19 2015 | ||
| -!- n0bawk` is now known as n0bawk | 10:41 | |
| -!- n0bawk` is now known as n0bawk | 12:42 | |
| sheridp | Hi all, I've run into a bug using gcc with -msoft-mul: the compiler is still trying to use l.mul instructions causing the cpu to error. I saw on some forums that msoft-mul might have some problems; does anyone have any updates on this? | 18:13 |
|---|---|---|
| stekern | sheridp: if you have a reasonable testcase, I can take a look at it | 19:48 |
| sheridp | stekern: You know, I'm only finding the problem when using floats/doubles. Maybe when I compiled gcc/newlib I forgot to use --with-float=soft | 20:14 |
| sheridp | Sorry, I think I'm getting a bit mixed up. I'm finding the l.mul within the __mulsf3 function which get's included when I execute a multiplication involving a float | 20:23 |
| sheridp | it looks like __mulsf3 is defined in mulsf3.o which is archived in soft-float/libgcc.a | 20:26 |
| sheridp | So I think i need to add an -msoft-mul when compiling mulsf3.o but I'm not exactly sure where to put that in the configuration options / scripts | 21:05 |
| sheridp | OK, I fixed the issue by doing the following: in /opt/or1k-gcc/gcc/config/or1k/t-or1knd I added msoft-mul to MULTILIB_OPTIONS and soft-mul to MULTILIB_DIRNAMES | 22:04 |
| sheridp | this gives versions of libgcc.a that do not use the l.mul instruction | 22:05 |
| sheridp | I'm hitting an undefined reference to memcpy as a result of the following line: int colList [4] = {0,1,2,3}; | 23:27 |
| sheridp | does anyone know where to grab memcpy from? | 23:27 |
| --- Log closed Sat Mar 14 00:00:21 2015 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!