[PATCH 4/4] dhcp6c: DHCPv6 client

[Leonid Lisovskiy]

First of all, thank you for code review. I'll try to follow all your
recommendations and will prepare next version.
Below is few open questions:

>On Tue, Oct 18, 2011 at 3:06 AM, Denys Vlasenko <vda.linux at googlemail.com> wrote:
>dhcp6c_script.c:
>                               /* since we count total length above, it is safely to use strcat() */
>                               switch (v->lvtype) {
>                                   case DHCP6_LISTVAL_VBUF:
>                                           strcat(*curr, v->val_vbuf.dv_buf);
>                                           strcat(*curr, " ");
>strcat is inefficient. strcat(" ") doubly so:
>can use *p++ = ' ' instead.

Are you sure that following code will be much efficient:

	len = strlen(*curr);
	*(*curr + len++) = ' ';
	*(*curr + len++) = '\0';

Or I miss something?
Moreover, I saw similar code in editors/vi.c: strcat(readbuffer, " ");


>On Tue, Oct 18, 2011 at 5:04 PM, Denys Vlasenko <vda.linux at googlemail.com> wrote:
>> Are use of getifaddrs/freeifaddrs prohibited in bb?
>I guess it's not a very good idea to use them, they are not portable.
>Why dhch6 even needs to know some intimate things about interfaces?
>I only understand the need to know MAC, nothing else.

There are two places of getifaddrs() dhcp6c use:

1) if6.c:if6init() - it tries to determine existent IPv6 address on
listen interface.
	Used in server only, so must be moved to server code.

2) if6.c:gethwid() - determine MAC, will be replaced to SIOCGIFCONF +
udhcp_read_interface().
	Iteration through interfaces needed since ppp0, for example, hasn't MAC.


>BTW, udhcpc is handling just one iface, which simplifies logic a lot.
>What do you think about using the same approach for DHCPv6?

I apologize that it is definitely acceptable for DHCPv6 client, but I'm not
sure for server.


>+       /* find the corresponding event based on the received xid */
>+       ev = find_event_withid(ifp->ifid, ntohl(dh6->dh6_xid) & DH6_XIDMASK);
>+       if (ev == NULL) {
>+               bb_error_msg("XID mismatch");
>+               goto fail;
>+       }

>Why we emit a message? It's just a packet to somebody else.
>I think it needs to be log1()...

We receive a valid DHCPv6 message on our interface, correct multicast address
and message type, but XID isn't our. This can be occur either if we have another
DHCPv6 client/relay/server on same host/interface, or it is a bug.
Not enough reason for error message?