How to use failpoint in go test

[root@xxx]# cat main_test.go
package main

import (
        "testing"
"github.com/pingcap/failpoint"
)

func TestGetValue(t *testing.T) {
    failpoint.Enable("main/failpoint-name","return(4)")
        t.Log("result:", GetValue())
}
[root@xxx]# cat main.go
package main

import (
        "fmt"
        "github.com/pingcap/failpoint"
)

func GetValue() int {
        if val, _err_ := failpoint.Eval(_curpkg_("failpoint-name")); _err_ == nil {
                return val.(int)
        }

        return 1
}
func main() {

        fmt.Println("Hello world ", GetValue())
}
[root@xxx]# GOCACHE=/tmp/failpoint-cache  GO_FAILPOINTS="main/failpoint-name=return(4)" CGO_ENABLED=1 GO111MODULE=on go test-v -toolexec /usr/bin/failpoint-toolexec main_test.go main.go binding__failpoint_binding__.go
=== RUN   TestGetValue
    main_test.go:10: result: 1
--- PASS: TestGetValue (0.00s)
PASS
ok      command-line-arguments  0.006s

expect:

[root@xxx]# GOCACHE=/tmp/failpoint-cache  GO_FAILPOINTS="main/failpoint-name=return(4)" CGO_ENABLED=1 GO111MODULE=on go test-v -toolexec /usr/bin/failpoint-toolexec main_test.go main.go binding__failpoint_binding__.go
=== RUN   TestGetValue
    main_test.go:10: result: 4
--- PASS: TestGetValue (0.00s)
PASS

Reproduce the report with main_test.go, main.go, binding__failpoint_binding__.go, and the shown go test command. Compare the observed result with the expected result of 4, then inspect the failpoint-toolexec and GO_FAILPOINTS integration to identify where the configuration is lost. Done means the supplied test command produces result: 4 or the usage limitation is documented.