Move pod jobs to parallel execution

Make Podman pod operations that do not involve starting
containers (which needs to be done in a specific order) use the
same parallel operation code we use to make `podman stop` on
large numbers of containers fast. We were previously stopping
containers in a pod serially, which could take up to the timeout
(default 15 seconds) for each container - stopping 100 containers
that do not respond to SIGTERM would take 25 minutes.

To do this, refactor the parallel operation code a bit to remove
its dependency on libpod (damn circular import restrictions...)
and use parallel functions that just re-use the standard
container API operations - maximizes code reuse (previously each
pod handler had a separate implementation of the container
function it performed).

This is a bit of a palate cleanser after fighting CI for two
days - nice to be able to return to a land of sanity.

Signed-off-by: Matthew Heon <matthew.heon@pm.me>

pkg/parallel/parallel.go

@@ -1,6 +1,7 @@
 package parallel
 
 import (
+	"context"
 	"sync"
 
 	"github.com/pkg/errors"
@@ -42,3 +43,32 @@ func SetMaxThreads(threads uint) error {
 func GetMaxThreads() uint {
 	return numThreads
 }
+
+// Enqueue adds a single function to the parallel jobs queue. This function will
+// be run when an unused thread is available.
+// Returns a receive-only error channel that will return the error (if any) from
+// the provided function fn when fn has finished executing. The channel will be
+// closed after this.
+func Enqueue(ctx context.Context, fn func() error) <-chan error {
+	retChan := make(chan error)
+
+	go func() {
+		jobControlLock.RLock()
+		defer jobControlLock.RUnlock()
+
+		defer close(retChan)
+
+		if err := jobControl.Acquire(ctx, 1); err != nil {
+			retChan <- errors.Wrapf(err, "error acquiring job control semaphore")
+			return
+		}
+
+		err := fn()
+
+		jobControl.Release(1)
+
+		retChan <- err
+	}()
+
+	return retChan
+}