Group the load/store and msr/mrs instructions together in swtch.S and other places. This may be useful for efficiency.
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Skipped - Unit
Tests Skipped
Event Timeline
| sys/arm64/arm64/swtch.S | ||
|---|---|---|
| 100–101 | Is there a reason you changed this to ldr, ldp and not keep it as ldp, ldr? | |
| sys/arm64/arm64/swtch.S | ||
|---|---|---|
| 100–101 | This is mostly stylistic. ldr, ldp and ldp, ldr should be equally fine in practice. | |
| sys/arm64/arm64/swtch.S | ||
|---|---|---|
| 100–101 | Ok, it looks like there is a penalty on some CPUs for non-64bit aligned ldp, however as these are all 64bit aligned it should be ok. | |
Comment Actions
So where does this stand? My arm assembler fu is weak, but I know that at least the pcb.h bits are good. I don't know how to read Andrew's comments on this, though.