Re: [RFC bluetooth-next 13/19] addrconf: add 802.15.4 short addr slaac

[Alexander Aring <aar@pengutronix.de>]

Hi Hannes,

thanks for your answer.

Am 03/23/2016 um 11:50 AM schrieb Hannes Frederic Sowa:
> On 22.03.2016 14:15, Alexander Aring wrote:
>> This patch adds the autoconfiguration if a valid 802.15.4 short address
>> is available for 802.15.4 6LoWPAN interfaces.
>>
>> Signed-off-by: Alexander Aring <aar@pengutronix.de>
>> ---
>>   net/ipv6/addrconf.c | 40 ++++++++++++++++++++++++++++++++++++++++
>>   1 file changed, 40 insertions(+)
>>
>> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
>> index 27aed1a..eb3e3fb 100644
>> --- a/net/ipv6/addrconf.c
>> +++ b/net/ipv6/addrconf.c
>> @@ -3058,6 +3058,37 @@ static void ipv6_gen_mode_random_init(struct inet6_dev *idev)
>>       s->initialized = true;
>>   }
>>
>> +#ifdef CONFIG_IEEE802154_6LOWPAN
>> +static int addrconf_ifid_ieee802154(u8 *eui, struct net_device *dev)
>> +{
>> +    struct wpan_dev *wpan_dev = lowpan_802154_dev(dev)->wdev->ieee802154_ptr;
>> +
>> +    /* Set short_addr autoconfiguration if short_addr is present only */
>> +    if (!ieee802154_is_valid_src_short_addr(wpan_dev->short_addr))
>> +        return -1;
>> +
>> +    /* For either address format, all zero addresses MUST NOT be used */
>> +    if (wpan_dev->pan_id == cpu_to_le16(0x0000) &&
>> +        wpan_dev->short_addr == cpu_to_le16(0x0000))
>> +        return -1;
>> +
>> +    /* Alternatively, if no PAN ID is known, 16 zero bits may be used */
>> +    if (wpan_dev->pan_id == cpu_to_le16(IEEE802154_PAN_ID_BROADCAST))
>> +        memset(eui, 0, 2);
>> +    else
>> +        ieee802154_le16_to_be16(eui, &wpan_dev->pan_id);
>> +
>> +    /* The "Universal/Local" (U/L) bit shall be set to zero */
>> +    eui[0] &= ~2;
>> +    eui[2] = 0;
>> +    eui[3] = 0xFF;
>> +    eui[4] = 0xFE;
>> +    eui[5] = 0;
>> +    ieee802154_le16_to_be16(&eui[6], &wpan_dev->short_addr);
>> +    return 0;
>> +}
>
> static void configure lowpan_addrconf(idev, addr)
> {
> > +            if (addrconf_ifid_ieee802154(addr.s6_addr + 8,
> > +                             idev->dev) == 0)
> > +                addrconf_add_linklocal(idev, &addr, 0);
> }
>
> #else
>
> static void configure lowpan_addrconf(idev, addr)
> {
> }
>
>> +#endif
>
> So we don't need the ifdef in addrconf_addr_gen. I think it pretty much reflects what upstream wants.
>
>> +
>>   static void addrconf_addr_gen(struct inet6_dev *idev, bool prefix_route)
>>   {
>>       struct in6_addr addr;
>> @@ -3088,6 +3119,15 @@ static void addrconf_addr_gen(struct inet6_dev *idev, bool prefix_route)
>>               addrconf_add_linklocal(idev, &addr, 0);
>>           else if (prefix_route)
>>               addrconf_prefix_route(&addr, 64, idev->dev, 0, 0);
>> +
>> +        /* 802.15.4 6LoWPAN short address slaac handling */
>> +        if (lowpan_is_ll(idev->dev, LOWPAN_LLTYPE_IEEE802154)) {
>> +        }
>
> Do you think it makes sense to put this under the label IN_ADDR_GEN_MODE_EUI64? I don't see a real issue with it, maybe it conflicts a little bit with the naming.
>
> Otherwise I don't have any issue with the change.

I moved the complete part of generating such address into net/6lowpan/, there we have another
netdev notifier functionality for 6lowpan interfaces only.

I think this stuff is very 802.15.4 specific and it's good when we can move it outside the net/ipv6
implementation. I just need to add a "EXPORT_SYMBOL(addrconf_add_linklocal);", so I can
access that function from 6lowpan module.

For the ndisc parts, I added a ndisc_ops structure and add a net/6lowpan/ndisc.c file.
I will send a RFCv2 soon.

- Alex