Don't kill container when volume is unmounted

The gofer session is killed when a gofer backed volume is unmounted. The
gofer monitor catches the disconnect and kills the container. This changes
the gofer monitor to only care about the rootfs connections, which cannot
be unmounted.

Fixes #6259

PiperOrigin-RevId: 385929039

1 files changed, 17 insertions, 10 deletions

diff --git a/runsc/boot/loader.go b/runsc/boot/loader.go
index 898692219..ec9188021 100644
--- a/runsc/boot/loader.go
+++ b/runsc/boot/loader.go
@@ -742,8 +742,11 @@ func (l *Loader) createContainerProcess(root bool, cid string, info *containerIn
 	// ours either way.
 	info.procArgs.FDTable = fdTable
 
-	// Setup the child container file system.
-	l.startGoferMonitor(cid, info.goferFDs)
+	// Gofer FDs must be ordered and the first FD is always the rootfs.
+	if len(info.goferFDs) < 1 {
+		return nil, nil, nil, fmt.Errorf("rootfs gofer FD not found")
+	}
+	l.startGoferMonitor(cid, int32(info.goferFDs[0].FD()))
 
 	mntr := newContainerMounter(info, l.k, l.mountHints, kernel.VFS2Enabled)
 	if root {
@@ -816,17 +819,21 @@ func (l *Loader) createContainerProcess(root bool, cid string, info *containerIn
 }
 
 // startGoferMonitor runs a goroutine to monitor gofer's health. It polls on
-// the gofer FDs looking for disconnects, and kills the container processes if a
-// disconnect occurs in any of the gofer FDs.
-func (l *Loader) startGoferMonitor(cid string, goferFDs []*fd.FD) {
+// the gofer FD looking for disconnects, and kills the container processes if
+// the rootfs FD disconnects.
+//
+// Note that other gofer mounts are allowed to be unmounted and disconnected.
+func (l *Loader) startGoferMonitor(cid string, rootfsGoferFD int32) {
+	if rootfsGoferFD < 0 {
+		panic(fmt.Sprintf("invalid FD: %d", rootfsGoferFD))
+	}
 	go func() {
 		log.Debugf("Monitoring gofer health for container %q", cid)
-		var events []unix.PollFd
-		for _, goferFD := range goferFDs {
-			events = append(events, unix.PollFd{
-				Fd:     int32(goferFD.FD()),
+		events := []unix.PollFd{
+			{
+				Fd:     rootfsGoferFD,
 				Events: unix.POLLHUP | unix.POLLRDHUP,
-			})
+			},
 		}
 		_, _, err := specutils.RetryEintr(func() (uintptr, uintptr, error) {
 			// Use ppoll instead of poll because it's already whilelisted in seccomp.