cybercyst/go
1
0
Fork 0
mirror of https://github.com/golang/go.git synced 2025-05-30 19:52:53 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

net: increase timing slop in TimeoutFluctuation tests on NetBSD and OpenBSD

Browse Source 
Decrease the slop everywhere else, since NetBSD and OpenBSD seem to be
the only ones that miss by that much.

For #50189
Updates #36108

Change-Id: I22ac39cc7c254e40358fcd933b5a6016629602c3
Reviewed-on: https://go-review.googlesource.com/c/go/+/372215
Trust: Bryan Mills <bcmills@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
This commit is contained in:
Bryan C. Mills 2021-12-15 10:46:11 -05:00 committed by Bryan Mills
parent 58e8e26924
commit c9ffcca784
1 changed files with 14 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
src/net/timeout_test.go
View File

@ -644,10 +644,20 @@ const (
// timeoutUpperBound returns the maximum time that we expect a timeout of
// duration d to take to return the caller.
func timeoutUpperBound(d time.Duration) time.Duration {
// In https://storage.googleapis.com/go-build-log/1e637cd3/openbsd-amd64-68_3585d3e7.log,
// we observed that an openbsd-amd64-68 builder took 636ms for a 512ms timeout
// (24.2% overhead).
return d * 4 / 3
switch runtime.GOOS {
case "openbsd", "netbsd":
// NetBSD and OpenBSD seem to be unable to reliably hit deadlines even when
// the absolute durations are long.
// In https://build.golang.org/log/c34f8685d020b98377dd4988cd38f0c5bd72267e,
// we observed that an openbsd-amd64-68 builder took 4.090948779s for a
// 2.983020682s timeout (37.1% overhead).
// (See https://go.dev/issue/50189 for further detail.)
// Give them lots of slop to compensate.
return d * 3 / 2
}
// Other platforms seem to hit their deadlines more reliably,
// at least when they are long enough to cover scheduling jitter.
return d * 11 / 10
}
// nextTimeout returns the next timeout to try after an operation took the given