| --- Log opened Thu Sep 17 00:00:47 2015 | ||
| andrzejr | olofk, getting there. As usual, there were more bugs but now I can boot newlib code in RTL simulation. HW still fails (works with asm code, though) but now I know what parts are OK. | 00:24 |
|---|---|---|
| dalias | does the or1k gcc target actually work? i'm looking at its codegen and it seems bogus... | 04:11 |
| dalias | it stores values at negative offsets from the stack pointer _then_ decrements the stack pointer | 04:12 |
| dalias | unless there is some redzone contract i'm not aware of, this will result in horrible stack clobbering by signal handlers | 04:13 |
| dalias | and simply 'return x;' compiles to (pic) 17 instructions | 04:18 |
| dalias | ah ok there is a 128-byte redzone | 04:19 |
| dalias | so i don't get why it's adjusting the stack pointer at all | 04:19 |
| dalias | http://sprunge.us/FJGe | 04:20 |
| dalias | the stack pointer adjustment is immediately undone and never used :-p | 04:20 |
| wbx | is qemu-system-or32 really really slow? no command works after first start with f.e. sash as init. | 18:30 |
| --- Log closed Fri Sep 18 00:00:48 2015 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!