Files
rclone/backend/sftp/ssh_external_test.go
Dirk Petersen 7181430fa1 sftp: fix data race when using --sftp-ssh
When the sftp backend uses an external ssh program (--sftp-ssh), one
goroutine waits for the ssh program to finish by calling cmd.Wait().
cmd.Wait() sets cmd.ProcessState when the program ends.

At the same time, CanReuse() calls exited(), which reads
cmd.ProcessState without any lock. This is a data race and the Go
race detector reports it. It happens in normal use, for example when
a connection is put back in the pool after a stat.

The race only decides whether a pooled connection can be reused, so
it is unlikely to cause problems in practice, but it makes programs
that use the sftp backend fail when they are tested with -race.

This fix sets an atomic.Bool after cmd.Wait() returns and makes
exited() read that instead of cmd.ProcessState.

It also adds a test which calls exited() while Wait() is running. It
fails with -race before this fix and passes after it.
2026-10-16 17:41:03 +01:00

103 lines
2.5 KiB
Go

//go:build !plan9
package sftp
import (
"testing"
"time"
"github.com/rclone/rclone/fs"
"github.com/stretchr/testify/assert"
)
// TestSSHExternalWaitMultipleCalls verifies that calling Wait() multiple times
// doesn't cause zombie processes
func TestSSHExternalWaitMultipleCalls(t *testing.T) {
// Create a minimal Fs object for testing
opt := &Options{
SSH: fs.SpaceSepList{"echo", "test"},
}
f := &Fs{
opt: *opt,
}
// Create a new SSH session
session := f.newSSHSessionExternal()
// Start a simple command that exits quickly
err := session.Start("exit 0")
assert.NoError(t, err)
// Give the command time to complete
time.Sleep(100 * time.Millisecond)
// Call Wait() multiple times - this should not cause issues
err1 := session.Wait()
err2 := session.Wait()
err3 := session.Wait()
// All calls should return the same result (no error in this case)
assert.NoError(t, err1)
assert.NoError(t, err2)
assert.NoError(t, err3)
// Verify the process has exited
assert.True(t, session.exited())
}
// TestSSHExternalCloseMultipleCalls verifies that calling Close() multiple times
// followed by Wait() calls doesn't cause zombie processes
func TestSSHExternalCloseMultipleCalls(t *testing.T) {
// Create a minimal Fs object for testing
opt := &Options{
SSH: fs.SpaceSepList{"sleep", "10"},
}
f := &Fs{
opt: *opt,
}
// Create a new SSH session
session := f.newSSHSessionExternal()
// Start a long-running command
err := session.Start("sleep 10")
if err != nil {
t.Skip("Cannot start sleep command:", err)
}
// Close should cancel and wait for the process
_ = session.Close()
// Additional Wait() calls should return the same error
err2 := session.Wait()
err3 := session.Wait()
// All should complete without panicking
// err1 could be nil or an error depending on how the process was killed
// err2 and err3 should be the same
assert.Equal(t, err2, err3, "Subsequent Wait() calls should return same result")
// Verify the process has exited
assert.True(t, session.exited())
}
// TestSSHExternalExitedWhileWaiting verifies that exited() can be called
// while another goroutine is waiting for the process. Run with -race.
func TestSSHExternalExitedWhileWaiting(t *testing.T) {
f := &Fs{
opt: Options{
SSH: fs.SpaceSepList{"echo", "test"},
},
}
session := f.newSSHSessionExternal()
assert.NoError(t, session.Start("exit 0"))
go func() {
_ = session.Wait()
}()
assert.Eventually(t, session.exited, 10*time.Second, time.Millisecond)
}