[PATCH] Skip down interfaces

Serge E. Hallyn serue at us.ibm.com
Wed Mar 31 09:13:53 PDT 2010


Quoting Dan Smith (danms at us.ibm.com):
> This makes the netns checkpoint code skip interfaces that are not up.
> Later, we want to make it possible to checkpoint down interfaces with
> a flag, but for now this helps prevent people from getting stuck on
> stock kernels with various unsupported and persistent virtual interfaces,
> such as tunnel devices.
> 
> Signed-off-by: Dan Smith <danms at us.ibm.com>

Acked-by: Serge Hallyn <serue at us.ibm.com>
Tested-by: Serge Hallyn <serue at us.ibm.com>

Now this really is only a problem for devices which use
register_pernet_device() right?  Could/should we add a
flag to such devices, and only skip those when they're
down?

Or, can we detect whether the device has a checkpoint
operation, checkpoint it if so regardless, and, if not,
if it is down, skip it?

Mind you if noone minds this patch (for now) then it
works for me.

thanks,
-serge

> Cc: serue at us.ibm.com
> ---
>  net/checkpoint_dev.c |    8 +++++++-
>  1 files changed, 7 insertions(+), 1 deletions(-)
> 
> diff --git a/net/checkpoint_dev.c b/net/checkpoint_dev.c
> index bc0415d..1e053f9 100644
> --- a/net/checkpoint_dev.c
> +++ b/net/checkpoint_dev.c
> @@ -268,7 +268,13 @@ int checkpoint_netns(struct ckpt_ctx *ctx, void *ptr)
>  		goto out;
> 
>  	for_each_netdev(net, dev) {
> -		ret = checkpoint_obj(ctx, dev, CKPT_OBJ_NETDEV);
> +		if (dev->flags & IFF_UP)
> +			ret = checkpoint_obj(ctx, dev, CKPT_OBJ_NETDEV);
> +		else
> +			/* TODO: There should be a flag to enable checkpoint
> +			 * of downed interfaces
> +			 */
> +			ret = 0;
>  		if (ret < 0)
>  			break;
>  	}
> -- 
> 1.6.2.5


More information about the Containers mailing list