[ECOS] Re: ipv6 question

[cyl cyl <cylecos at gmail dot com>]

I used wireshark . And found the reset bit of the response package was
set 1. May anyone tell me which file or function does this work? So I
can trace it.

Flags: 0x014 (RST, ACK)



2013/1/5 cyl cyl <cylecos@gmail.com>:
> Hello:
>     Is anyone familiar with ipv6? My UDP works well now, but TCP is
> not . My computer connects my ecos equipment fail .
>
> This is ipv4 connect : (IPV4 works well.)
>
>         ISR
>         INT_STATUS: 8a
>
>
>         RPIdx 3 RCIdx 2
>         total_len 78 rx_status[2] = c400004d rx_len 78
>         lpc2xxx_eth_recv: sg_len 2
>         bil: 0 sg_list[0].len: 14 bnl: 14
>         bib: 78 buffer_pos: 0
>         (1) memcpy(0xA0A721DA, 0x7FE00CB8, 14) [0xA0A721DA - 0xA0A721E8]
>         bil: 0 sg_list[1].len: 64 bnl: 64
>         bib: 78 buffer_pos: 14
>         (2) memcpy(0xA0A3BB2C, 0x7FE00CC6, 64)[0xA0A3BB2C - 0xA0A3BB6C]
>         RCIdx++ 3
>         RPIdx 3 RCIdx 3 Iterator 2
>         can_send: 1
>         TPI: 6 TCI: 6
>         TX: sg_len 1, total_len 74, key 0xA0A3BB00
>         TPI: 7 TCI: 6
>         ISR
>         can_send: 0
>         INT_STATUS: ca
>         TX finished - key: 0xA0A3BB00
>
> This is IPV6 connect :
>
>         ISR
>         INT_STATUS: 8a
>
>
>         RPIdx 3 RCIdx 2
>         total_len 98 rx_status[2] = c4000061 rx_len 98
>         lpc2xxx_eth_recv: sg_len 2
>         bil: 0 sg_list[0].len: 14 bnl: 14
>         bib: 98 buffer_pos: 0
>         (1) memcpy(0xA0A721DA, 0x7FE00CB8, 14) [0xA0A721DA - 0xA0A721E8]
>         bil: 0 sg_list[1].len: 84 bnl: 84
>         bib: 98 buffer_pos: 14
>         (2) memcpy(0xA0A3B7AC, 0x7FE00CC6, 84)[0xA0A3B7AC - 0xA0A3B800]
>         RCIdx++ 3
>         RPIdx 3 RCIdx 3 Iterator 2
>         can_send: 1
>         TPI: 6 TCI: 6
>         TX: sg_len 2, total_len 74, key 0xA0A3BA00
>         TPI: 8 TCI: 6
>         ISR
>         can_send: 0
>         INT_STATUS: c2
>         TX finished - key: 0xA0A3BA00
>
> The "tx total_len" of ipv4 and ipv6 both are 74. I'm not familiar with
> TCP/IP but I think They should not be the same. Does anyone hit this
> problem before?
>
> 2012/12/31 cyl cyl <cylecos@gmail.com>:
>> I found it. If the ip I set is a linklocal ip (fe80) , then
>> s6_adr16[1] will change to ifp->if_index.
>>
>>         if (sa6 && sa6->sin6_family == AF_INET6) {
>>                 if (IN6_IS_ADDR_LINKLOCAL(&sa6->sin6_addr)) {
>>                         if (sa6->sin6_addr.s6_addr16[1] == 0) {
>>                                 /* link ID is not embedded by the user */
>>                                 sa6->sin6_addr.s6_addr16[1] =
>>                                         htons(ifp->if_index);
>>                         } else if (sa6->sin6_addr.s6_addr16[1] !=
>>                                     htons(ifp->if_index)) {
>>                                 return(EINVAL); /* link ID contradicts */
>>                         }
>>
>> So I set another IP (3ffe .....) ,and it works OK.
>>
>> 2012/12/24 cyl cyl <cylecos@gmail.com>:
>>> Hello:
>>>
>>>         In function init_net_IPv6, I set "fe80::219:dbff:fea9:f3d5"
>>> for my eth0 ipv6 address. But _show_all_interfaces says my ip is
>>> "fe80:1::219:dbff:fea9:f3d5". I don't know why it adds a "1" in it.
>>>         The prefix I set is "fe80:0:0:0" .
>>>
>>>
>>> ipv6: fe80::219:dbff:fea9:f3d5, in6mask64: ff ff ff ff ff ff ff ff
>>>
>>> 0xa0b53204 - eth0 - 192.168.1.239
>>> 0xa0b53220 - eth0 - fe80:1::200:ff:fe6c:18f0
>>> 0xa0b5323c - eth0 - fe80:1::219:dbff:fea9:f3d5
>>> 0xa0b53274 - lo0 - ::1
>>> 0xa0b53290 can_send: 1

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss