| Jia | qemu can run or-linux now, with some littlle bugs. | 03:13 |
|---|---|---|
| gxti | neat | 03:34 |
| michaelliu | please, How to compile a 32-bit binary on a 64-bit linux machine with bld-all.sh | 06:16 |
| michaelliu | file /openrisc/bin/or32-linux-gcc | 06:17 |
| michaelliu | or32-linux-gcc: ELF 64-bit LSB executable | 06:18 |
| stekern | is it legal to place a load to r9 in a l.jr r9 delay slot? i.e.: l.jr r9; l.lwz r9, -4(r1) | 06:39 |
| juliusb | stekern: good question. I would say it is, as the jump has been calculated by the time you're processing the delay slot | 12:57 |
| juliusb | but... this isn't mentioned in the architecture spec is it? | 12:58 |
| juliusb | maybe it's something for the spec revision | 12:58 |
| juliusb | spec revision update | 12:58 |
| stekern | hmm, are you saying you think it should be legal or not? | 13:06 |
| stekern | arch spec only mentions putting r9 in l.jal(r) delay slots | 13:07 |
| jemarch | hi | 13:22 |
| stekern | btw, by legal, I mean the jump should go to what's in -4(r1) | 13:22 |
| stekern | it could of course be any register, and it doesn't necessarily need to be a load, any l.jr rX with an instruction that uses rX as dest in the delay slot would do | 13:28 |
| stekern | but yeah, you're probably right, it's implied that the branch address is calculated before the the delay slot | 13:38 |
| juliusb | i think it _should_ be legal | 13:38 |
| stekern | yes, legal it _should_ be, but I realized that I should probably have questioned what the expected outcome would be ;) | 13:41 |
| stekern | because I realised to things; my definition of legal was probably wrong and I write sentences that sound like Yoda | 14:15 |
| stekern | s/to/two | 14:15 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!