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

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>

diff --git a/src/net/timeout_test.go b/src/net/timeout_test.go
index 3c6aa27cc1f3353b55eb19c62a53fa98b4f3ad06..d1cfbf853cbb4a77b7860b0897e58366ee5ffa07 100644 (file)
--- a/src/net/timeout_test.go
+++ b/src/net/timeout_test.go
@@ -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