runc exec: report possible OOM kill

An exec may fail due to memory shortage (cgroup memory limits being too
tight), and an error message provided in this case is clueless:

> $ sudo ../runc/runc exec xx56 top
> ERRO[0000] exec failed: container_linux.go:367: starting container process caused: read init-p: connection reset by peer

Same as the previous commit for run/start, check the OOM kill counter
and report an OOM kill.

The differences from run are

1. The container is already running and OOM kill counter might not be
   zero.  This is why we have to read the counter before exec and after
   it failed.

2. An unrelated OOM kill event might occur in parallel with our exec
   (and I see no way to find out which process was killed, except to
   parse kernel logs which seems excessive and not very reliable).
   This is why we report _possible_ OOM kill.

Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>

libcontainer/container_linux.go

@@ -570,6 +570,7 @@ func (c *linuxContainer) newSetnsProcess(p *Process, cmd *exec.Cmd, messageSockP
 		intelRdtPath:    state.IntelRdtPath,
 		messageSockPair: messageSockPair,
 		logFilePair:     logFilePair,
+		manager:         c.cgroupManager,
 		config:          c.newInitConfig(p),
 		process:         p,
 		bootstrapData:   data,

libcontainer/process_linux.go

@@ -65,6 +65,7 @@ type setnsProcess struct {
 	logFilePair     filePair
 	cgroupPaths     map[string]string
 	rootlessCgroups bool
+	manager         cgroups.Manager
 	intelRdtPath    string
 	config          *initConfig
 	fds             []string
@@ -88,6 +89,8 @@ func (p *setnsProcess) signal(sig os.Signal) error {
 
 func (p *setnsProcess) start() (retErr error) {
 	defer p.messageSockPair.parent.Close()
+	// get the "before" value of oom kill count
+	oom, _ := p.manager.OOMKill()
 	err := p.cmd.Start()
 	// close the write-side of the pipes (controlled by child)
 	p.messageSockPair.child.Close()
@@ -97,6 +100,10 @@ func (p *setnsProcess) start() (retErr error) {
 	}
 	defer func() {
 		if retErr != nil {
+			if newOom, err := p.manager.OOMKill(); err == nil && newOom != oom {
+				// Someone in this cgroup was killed, this _might_ be us.
+				retErr = newSystemErrorWithCause(retErr, "possibly OOM-killed")
+			}
 			err := ignoreTerminateErrors(p.terminate())
 			if err != nil {
 				logrus.WithError(err).Warn("unable to terminate setnsProcess")