Add Test for linksInContainer = false

This commit introduces a new test case that tests the code flow when the linksInContainer parameter is set to false, indicating that links start in a different namespace instead of the pod namespace. This scenario requires a different setup than usual, involving the addition of dummy links to a different namespace rather than the pod namespace.

1. Created a  function (addLinksInNS) responsible for adding links during namespace setup for each test. This function is needed for the new test, where dummy links are added to the init namespace.

2. Added an option to modify the linksInContainer setting in the string config used across all tests. This change allows each test to set the linksInContainer option to true or false as needed.

3. Introduced a test case that verifies the correct addition and deletion of the plugin while links are active in the host network namespace. To avoid redundancy, new functions (validateBondIFConf, validateBondSlavesConf) were added.

4. Included a missing netNs.Close() statement in the bond code's setLinksinNetNs function to properly close the pod namespace.

5. Also changed the setLinksinNetNs function. Previous versions didn't include transfer of the namespace back to the current namespace after function execution. In this commit instead of using ns.netNs.Set function, ns.netNs.Do was used, including the transfer to the initial namespace. This change also included creation of a move links from namespace function for a more readable implementation. 

Signed-off-by: Alina Sudakov <asudakov@redhat.com>

bond-cni

@@ -0,0 +1 @@
+/home/asudakov/go/src/github.com/k8snetworkplumbingwg/bond-cni

bond/bond.go

@@ -173,8 +173,8 @@ func deattachLinksFromBond(linkObjectsToDeattach []netlink.Link, netNsHandle *ne
 	return nil
 }
 
-func setLinksinNetNs(bondConf *bondingConfig, nspath string, releaseLinks bool) error {
-	var netNs, curnetNs ns.NetNS
+func setLinksInNetNs(bondConf *bondingConfig, nspath string, releaseLinks bool) error {
+	var podNs, hostNS ns.NetNS
 	var err error
 
 	linkNames := []string{}
@@ -186,49 +186,48 @@ func setLinksinNetNs(bondConf *bondingConfig, nspath string, releaseLinks bool)
 		linkNames = append(linkNames, s)
 	}
 
-	if netNs, err = ns.GetNS(nspath); err != nil {
+	if podNs, err = ns.GetNS(nspath); err != nil {
 		return fmt.Errorf("failed to open netns %q: %v", nspath, err)
 	}
+	defer podNs.Close()
 
-	if curnetNs, err = ns.GetCurrentNS(); err != nil {
+	if hostNS, err = ns.GetCurrentNS(); err != nil {
 		return fmt.Errorf("failed to get init netns: %v", err)
 	}
 
 	if releaseLinks {
-		if err := netNs.Set(); err != nil {
-			return fmt.Errorf("failed to enter netns %q: %v", netNs, err)
-		}
+		err = moveLinksBetweenNs(linkNames, podNs, hostNS, "host")
+	} else {
+		err = moveLinksBetweenNs(linkNames, hostNS, podNs, "container")
 	}
+	return err
+}
 
-	if len(linkNames) >= 2 { // currently supporting two or more links to one bond
-		for _, linkName := range linkNames {
-			// get interface link in the network namespace
-			link, err := netlink.LinkByName(linkName)
-			if err != nil {
-				return fmt.Errorf("failed to lookup link interface %q: %v", linkName, err)
-			}
-
-			// set link interface down
-			if err = netlink.LinkSetDown(link); err != nil {
-				return fmt.Errorf("failed to down link interface %q: %v", linkName, err)
-			}
+func moveLinksBetweenNs(links []string, from ns.NetNS, to ns.NetNS, toNsName string) error {
+	return from.Do(func(ns.NetNS) error {
+		if len(links) < 2 { // currently supporting two or more links to one bond
+			return fmt.Errorf("Bonding requires at least two links, we have %+v", len(links))
+		} else {
+			for _, linkName := range links {
+				// get interface link in the network namespace
+				link, err := netlink.LinkByName(linkName)
+				if err != nil {
+					return fmt.Errorf("failed to lookup link interface %q: %v", linkName, err)
+				}
 
-			if releaseLinks { // move link inteface to network netns
-				if err = netlink.LinkSetNsFd(link, int(curnetNs.Fd())); err != nil {
-					return fmt.Errorf("failed to move link interface to host netns %q: %v", linkName, err)
+				// set link interface down
+				if err = netlink.LinkSetDown(link); err != nil {
+					return fmt.Errorf("failed to down link interface %q: %v", linkName, err)
 				}
-			} else {
-				if err = netlink.LinkSetNsFd(link, int(netNs.Fd())); err != nil {
-					return fmt.Errorf("failed to move link interface to container netns %q: %v", linkName, err)
+
+				// move link interface to network netns
+				if err = netlink.LinkSetNsFd(link, int(to.Fd())); err != nil {
+					return fmt.Errorf("failed to move link interface to %s netns %q: %v", toNsName, linkName, err)
 				}
 			}
-
 		}
-	} else {
-		return fmt.Errorf("Bonding requires at least two links, we have %+v", len(linkNames))
-	}
-
-	return nil
+		return nil
+	})
 }
 
 func createBond(bondName string, bondConf *bondingConfig, nspath string, ns ns.NetNS) (*current.Interface, error) {
@@ -249,7 +248,7 @@ func createBond(bondName string, bondConf *bondingConfig, nspath string, ns ns.N
 	defer netNsHandle.Close()
 
 	if !bondConf.LinksContNs {
-		if err := setLinksinNetNs(bondConf, nspath, false); err != nil {
+		if err := setLinksInNetNs(bondConf, nspath, false); err != nil {
 			return nil, fmt.Errorf("Failed to move the links (%+v) in container network namespace, error: %+v", bondConf.Links, err)
 		}
 	}
@@ -423,7 +422,7 @@ func cmdDel(args *skel.CmdArgs) error {
 	}
 
 	if !bondConf.LinksContNs {
-		if err := setLinksinNetNs(bondConf, args.Netns, true); err != nil {
+		if err := setLinksInNetNs(bondConf, args.Netns, true); err != nil {
 			return fmt.Errorf("Failed set links (%+v) in host network namespace, error: %+v", bondConf.Links, err)
 		}
 	}