Re: [PATCH v2 ipsec-next] xfrm: Add possibility to set the default to block if we have no policy

[Antony Antony <antony.antony@secunet.com>]

On Wed, Aug 11, 2021 at 18:14:08 +0200, Nicolas Dichtel wrote:
> Le 18/07/2021 à 09:11, Antony Antony a écrit :
> > From: Steffen Klassert <steffen.klassert@secunet.com>
> Sorry for my late reply, I was off.
> 
> > 
> > As the default we assume the traffic to pass, if we have no
> > matching IPsec policy. With this patch, we have a possibility to
> > change this default from allow to block. It can be configured
> > via netlink. Each direction (input/output/forward) can be
> > configured separately. With the default to block configuered,
> > we need allow policies for all packet flows we accept.
> > We do not use default policy lookup for the loopback device.
> > 
> 
> [snip]
> 
> > diff --git a/include/net/netns/xfrm.h b/include/net/netns/xfrm.h
> > index e946366e8ba5..88c647302977 100644
> > --- a/include/net/netns/xfrm.h
> > +++ b/include/net/netns/xfrm.h
> > @@ -65,6 +65,13 @@ struct netns_xfrm {
> >  	u32			sysctl_aevent_rseqth;
> >  	int			sysctl_larval_drop;
> >  	u32			sysctl_acq_expires;
> > +
> > +	u8			policy_default;
> > +#define XFRM_POL_DEFAULT_IN	1
> > +#define XFRM_POL_DEFAULT_OUT	2
> > +#define XFRM_POL_DEFAULT_FWD	4
> > +#define XFRM_POL_DEFAULT_MASK	7
> > +
> 
> [snip]
> 
> > diff --git a/include/uapi/linux/xfrm.h b/include/uapi/linux/xfrm.h
> > index ffc6a5391bb7..6e8095106192 100644
> > --- a/include/uapi/linux/xfrm.h
> > +++ b/include/uapi/linux/xfrm.h
> > @@ -213,6 +213,11 @@ enum {
> >  	XFRM_MSG_GETSPDINFO,
> >  #define XFRM_MSG_GETSPDINFO XFRM_MSG_GETSPDINFO
> > 
> > +	XFRM_MSG_SETDEFAULT,
> > +#define XFRM_MSG_SETDEFAULT XFRM_MSG_SETDEFAULT
> > +	XFRM_MSG_GETDEFAULT,
> > +#define XFRM_MSG_GETDEFAULT XFRM_MSG_GETDEFAULT
> > +
> >  	XFRM_MSG_MAPPING,
> >  #define XFRM_MSG_MAPPING XFRM_MSG_MAPPING
> >  	__XFRM_MSG_MAX
> > @@ -508,6 +513,11 @@ struct xfrm_user_offload {
> >  #define XFRM_OFFLOAD_IPV6	1
> >  #define XFRM_OFFLOAD_INBOUND	2
> > 
> > +struct xfrm_userpolicy_default {
> > +	__u8				dirmask;
> > +	__u8				action;
> > +};
> > +
> Should XFRM_POL_DEFAULT_* be moved in the uapi?

It is good point. Thanks for the feedback.

> How can a user knows what value is expected in dirmask?
> 
> Same question for action. We should avoid magic values. 0 means drop or accept?

I have an iproute2 patch I want to sent out, moving to uapi would avoid using
hardcoded magic values there.


> Maybe renaming this field to 'drop' is enough.

action is a bitwise flag, one direction it may drop and ther other might
be allow.