diff options
author | Jamie Liu <jamieliu@google.com> | 2020-09-15 15:57:38 -0700 |
---|---|---|
committer | gVisor bot <gvisor-bot@google.com> | 2020-09-15 15:59:25 -0700 |
commit | 8b15effd9ec6091feb9ec554ebe38cc02e765560 (patch) | |
tree | 323785581a4830744fdb2753063b5031cd7216db /pkg/sentry/kernel/task_test.go | |
parent | 456c6c33e104e7421a5a815bba69d46b289ae724 (diff) |
Improve syserror_test.
- It's very difficult to prevent returnErrnoAsError and returnError from being
optimized out. Instead, replace BenchmarkReturn* with BenchmarkAssign*, which
store to globalError.
- Compare to a non-nil globalError in BenchmarkCompare* and BenchmarkSwitch*.
New results:
BenchmarkAssignErrno
BenchmarkAssignErrno-12 1000000000 0.615 ns/op
BenchmarkAssignError
BenchmarkAssignError-12 1000000000 0.626 ns/op
BenchmarkCompareErrno
BenchmarkCompareErrno-12 1000000000 0.522 ns/op
BenchmarkCompareError
BenchmarkCompareError-12 1000000000 3.54 ns/op
BenchmarkSwitchErrno
BenchmarkSwitchErrno-12 1000000000 1.45 ns/op
BenchmarkSwitchError
BenchmarkSwitchError-12 536315757 10.9 ns/op
PiperOrigin-RevId: 331875387
Diffstat (limited to 'pkg/sentry/kernel/task_test.go')
0 files changed, 0 insertions, 0 deletions