root / target-mips / TODO @ 1e3d0552
History | View | Annotate | Download (1.1 kB)
| 1 |
Unsolved issues/bugs in the mips/mipsel backend |
|---|---|
| 2 |
----------------------------------------------- |
| 3 |
|
| 4 |
General |
| 5 |
------- |
| 6 |
- [ls][dw][lr] report broken (aligned) BadVAddr |
| 7 |
- Missing per-CPU instruction decoding, currently all implemented |
| 8 |
instructions are regarded as valid |
| 9 |
- Applications running on top of a emulated Linux segfault sometimes |
| 10 |
when the Qemu FPU emulation is disabled, the tb misses a flush |
| 11 |
in that case. |
| 12 |
- gdb breakpoints inside the emulated system work only due to a hack |
| 13 |
which disassembles over the end of the current tb. |
| 14 |
- Floating point results of ceil/floor/round are wrong for IEEE cornercases. |
| 15 |
- recip/rsqrt FPU instructions are not implemented |
| 16 |
|
| 17 |
MIPS64 |
| 18 |
------ |
| 19 |
- Only lighly tested but apparently functional as of 2007-05-31. |
| 20 |
|
| 21 |
"Generic" 4Kc system emulation |
| 22 |
------------------------------ |
| 23 |
- Doesn't correspond to any real hardware. |
| 24 |
|
| 25 |
PICA 61 system emulation |
| 26 |
------------------------ |
| 27 |
- No framebuffer support yet. |
| 28 |
|
| 29 |
MALTA system emulation |
| 30 |
---------------------- |
| 31 |
- We fake firmware support instead of doing the real thing |
| 32 |
- Real firmware falls over when trying to init RAM, presumably due |
| 33 |
to lacking system controller emulation. |