Erik,

On Fri, 2009-08-07 at 15:40 +0200, Erik Nordmark wrote:
> I've looked at uts/common/inet except the iptun* files.

Thanks very much for the review, responses inline:

> 
> ip.h:
> You can remove the definition of TUN_CMD; I don't think it is used any more.

ACCEPT

> 
> ip.c: Shouldn't be bump InDelivers when we call connp->conn_recv to send 
> to iptun?

ACCEPT: Yes, and we'll need to pass in the ill that received the packet
into ip_iptun_input() order to bump the right mib.

> 
> ip6.c: Ditto.

ACCEPT

> 
> ip_ndp.c: You've removed the special handling if IP_ADDR_LEN around line 
> 3682. With this change how does a point-to-multipoint device driver tell 
> the nexthop that the packet should be sent to? Perhaps 6to4 tunnels 
> don't need this because the nexthop is always extracted from the IPv6 
> destination address. But I don't see how other point-to-multipoint 
> devices (IP over X.25, Fibrechannel, etc) can work with this removed.

DISCUSS: 6to4 tunnels would also benefit from this point-to-multipoint
mechanism, but it currently can't because the GLDv3 framework can't
currently cope with processing a DL_UNITDATA_REQ that has a destination
address of a different length than the length of the link's registered
link-layer address.  This could be made to work in the future for 6to4.

For other types of links, I can bring that code back, but it won't work
as-is.  The reason is that this will inhibit tunnel links from
negotiating proper fast-path headers.  What I can do is bring the code
back, but only for link types that are not DL_IPV4, DL_IPV6, and
DL_6TO4.  Does that make sense to you?

> 
> ipclassifier.c:
> The new ipcl_iptun_classify functions do not check tsol_receive_local(). 
> Should they? Currently in onnv-gate we do check since the tunnel fanout 
> happens through ip_fanout_proto which continues looking for a matching 
> conn_t if tsol_receive_local fails.
> (Is this what "XXX KEBE ASKS - What about TX and labels?" is asking about?)

ACCEPT: Yes, that's what KEBE was asking about, and this needs to be
done.  I'll consult with Will Young or Ken Powell about this and fix
this.

> 
> spd.c: Typo at
> 5862                  * NOTE:  Even if our policy is trasnport mode, set the

ACCEPT

> 
> At L5933 what prevents inner_ipv* from pointing past b_wptr? The reason 
> I'm asking is that you have removed the asserts that checked for this.

ACCEPT: It would be good to verify what danmcd intended here since this
was his change (and I don't fully comprehend the intricacies of the
IPsec "frag cache" that these packets come from), but I'll add the
ASSERT() calls back in, they can't possibly hurt.

Thanks!
-Seb


_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to