Optimize set_syscall_retval for riscv by predicting the return
value to be zero.
Details
- Reviewers
mhorne kp - Group Reviewers
riscv - Commits
- rS367138: Optimize set_syscall_retval for riscv by predicting the return
Diff Detail
- Repository
- rS FreeBSD src repository - subversion
- Lint
Lint Passed - Unit
No Test Coverage - Build Status
Buildable 34448 Build 31550: arc lint + arc unit
Event Timeline
Do we see measurable performance improvements with this?
(Looking at the generated code it does seem to be better, but measurements would make me feel better.)
I wouldn’t expect any visible results. The syscall path overhead is still huge, and there’s not a single reason for it - it’s a “death by thousand cuts” situation, which needs to be resolved in small chunks, like this one.
Relative imports are better, assuming this is a (namespace )? package. It makes it easier to lint items like this without installing everything beforehand.