[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH v2 2/2] lxc: don't forbid <interface type='direct'>



On Thu, Dec 06, 2018 at 10:58:23AM -0500, Laine Stump wrote:
Commit 017dfa27d changed a few switch statements in the LXC code to
have all possible enum values, and in the process changed the switch
statement in virLXCControllerGetNICIndexes() to return an error status
for unsupported interface types, but it erroneously put type='direct'
on the list of unsupported types.

type='direct' (implemented with a macvlan interface) is supported on
LXC, but it's interface shouldn't be placed on the list of interfaces
given to CreateMachineWithNetwork() because the interface is put
inside the container, while CreateMachineWithNetwork() only wants to
know about the parent veths of veth pairs (the parent veth remains on
the host side, while the child veth is put into the container).

Resolves: https://bugzilla.redhat.com/1656463
Signed-off-by: Laine Stump <laine laine org>
---
src/lxc/lxc_controller.c | 14 +++++++++++++-
1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c
index 07342cbc77..cff004a034 100644
--- a/src/lxc/lxc_controller.c
+++ b/src/lxc/lxc_controller.c
@@ -364,6 +364,16 @@ static int virLXCControllerGetNICIndexes(virLXCControllerPtr ctrl)
    size_t i;
    int ret = -1;

+    /* Gather the ifindexes of the "parent" veths for all interfaces
+     * implemented with a veth pair. These will be used when calling
+     * virCgroupNewMachine (and eventually the dbus method
+     * CreateMachineWithNetwork). ifindexes for the child veths, and
+     * for macvlan interfaces, *should not* be in this list, as they
+     * will be moved into the container. Only the interfaces that will
+     * remain outside the container, but are used for communication
+     * with the container, should be added to the list.
+     */
+
    VIR_DEBUG("Getting nic indexes");
    for (i = 0; i < ctrl->def->nnets; i++) {
        int nicindex = -1;
@@ -388,6 +398,9 @@ static int virLXCControllerGetNICIndexes(virLXCControllerPtr ctrl)
            ctrl->nicindexes[ctrl->nnicindexes-1] = nicindex;
            break;

+        case VIR_DOMAIN_NET_TYPE_DIRECT:
+           break;
+
        case VIR_DOMAIN_NET_TYPE_USER:
        case VIR_DOMAIN_NET_TYPE_VHOSTUSER:
        case VIR_DOMAIN_NET_TYPE_SERVER:
@@ -395,7 +408,6 @@ static int virLXCControllerGetNICIndexes(virLXCControllerPtr ctrl)
        case VIR_DOMAIN_NET_TYPE_MCAST:
        case VIR_DOMAIN_NET_TYPE_UDP:
        case VIR_DOMAIN_NET_TYPE_INTERNAL:
-        case VIR_DOMAIN_NET_TYPE_DIRECT:
        case VIR_DOMAIN_NET_TYPE_HOSTDEV:
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("Unsupported net type %s"),

Technically, from the point of view of this function whether the net
type is supported or not does not matter (qemuBuildInterfaceCommandLine
uses 'break;' for all the other networks.

But any effort spent on this function would be better spent on removing
the QEMU/LXC duplication and not bothering calling it on non-systemd
systems, so:

Reviewed-by: Ján Tomko <jtomko redhat com>

Jano

Attachment: signature.asc
Description: PGP signature


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]