I have some troubles with this commit.

I tested improxy five months ago and reverted this commit because I don't need to forward packets from one downstream to other interfaces. After that my router started to forward IPv6 multicast from wan to lan correctly.

I decided to check it again.
In this commit you disabled k_mcast_msfilter.
As a result, Linux Kernel sends MLD2 reports (leaves) with incorect record type.
Attachment contains additional information:

2018-07-02 13:06:00.058 INFO(wan): Expecting MLDv2 record type CHANGE_TO_INCLUDE_MODE for group ff0e::1, but found CHANGE_TO_EXCLUDE_MODE

Hi,
I need to forward IGMPv3 messages from enp0s8 to enp0s9, my .conf is this:

igmp enable version 3
mld enable
upstream enp0s8
downstream enp0s9

But I get the following error:

DEBUG[main@580]: logLevel = 5
DEBUG[load_config@76]: buffer = igmp enable version 3
DEBUG[load_config@94]: token is igmp
DEBUG[load_config@76]: buffer = mld enable
DEBUG[load_config@94]: token is mld
DEBUG[load_config@76]: buffer = upstream enp0s8
DEBUG[load_config@94]: token is upstream
DEBUG[load_config@185]: upstream interface is enp0s8
DEBUG[load_config@76]: buffer = downstream enp0s9
DEBUG[load_config@94]: token is downstream
DEBUG[load_config@194]: downstream interface is enp0s9
DEBUG[init_interface@189]: enp0s9's ipv6 address is fe80::a00:27ff:feff:4040
DEBUG[init_interface@195]: if_index = 4
DEBUG[init_interface@208]: if_mtu = 1500
DEBUG[init_interface@214]: if_falgs = 4163
DEBUG[init_interface@189]: enp0s8's ipv6 address is fe80::a00:27ff:fe89:fa25
DEBUG[init_interface@195]: if_index = 3
DEBUG[init_interface@208]: if_mtu = 1500
DEBUG[init_interface@214]: if_falgs = 4163
DEBUG[init_interface@178]: enp0s9's ipv4 address is 192.168.2.2
DEBUG[init_interface@195]: if_index = 4
DEBUG[init_interface@208]: if_mtu = 1500
DEBUG[init_interface@214]: if_falgs = 4163
DEBUG[init_interface@178]: enp0s8's ipv4 address is 192.168.1.1
DEBUG[init_interface@195]: if_index = 3
DEBUG[init_interface@208]: if_mtu = 1500
DEBUG[init_interface@214]: if_falgs = 4163
ERROR[init_mproxy4@332]: can't create igmp socket
ERROR[main@619]: exiting.........

Do you have an idea why igmp socket can be created?

Running on Ubuntu VM:

description: Computer
width: 64 bits
capabilities: vsyscall32
*-core
description: Motherboard
physical id: 0
*-memory
description: System memory
physical id: 0
size: 16GiB
*-cpu
product: Intel(R) Core(TM) i7-10850H CPU @ 2.70GHz
vendor: Intel Corp.
physical id: 1
bus info: cpu@0
width: 64 bits

Thanks

First of all, thanks you a lot for this package.
It does help me solve the problem. The only one workable for IGMP/MLD proxy I can find.
Thank You.

I try this software on a ARM little endian and got some warning message as title.
I fix it and attach the changes. Hope this help someone.

input.c, the "ntohl" part is not existed in original code. I add this to fix the problem.

} else if(p_pia->ss.ss_family == AF_INET) {

    if (!IN_MULTICAST(ntohl(p_pia->v4.sin_addr.s_addr))) {
        IMP_LOG_ERROR("group address %s isn't multicast adddress\n",
                      imp_pi_ntoa(p_pia));
        return -1;
    }

    /*
     *    Address Range                 Size       Designation
     *    -------------                 ----       -----------
     *    224.0.0.0 - 224.0.0.255       (/24)      Local Network Control Block
     *
     *    Addresses in the Local Network Control Block are used for protocol
     *    control traffic that is not forwarded off link.
     *    [RFC 5771 section 4]
     */

Hello and thank you for this igmp/mld proxy code
I run some static code analysis on it and I would like to share somes results with you

1. input.c mcast_recv_igmp() => unreachable code : else if (ip->protocol == 0)
2. proxy.c main() => unreachable code : free_resource();
   I think the proxy need to handle system signal in order to cleanup ressources
3. output.c send_igmp_mld_query() => unreachable code : if(p_if == NULL) (line 84)
4. proxy.c load_config() => resource lost : need to close fd
5. output.c send_igmp_mld_query() => possible memory leak : p = malloc(max_len)
   if family is neither AF_INET nor AF_INET6

I hope it help
Best Reagrds

Function imp_get_mfcc_ttls() checks that

p_if->if_index < length

where length is MAXVIFS (32 for linux).

But interface index may be greater than this value.
SOHO routers may have 40 interfaces (e.g. Realtek-based routers with real and virtual wifi interfaces, wds etc). As a result, eth1 (wan) and br0 (lan) have indexes greater than 32 and improxy doesn't call
imp_membership_db_mfc_add() because imp_get_mfcc_ttls returns 0.

I don't quite understand this limitation. I removed it for my device. But as a generic solution, maybe you should check the number of interfaces for which you call k_add_ip6_mif or k_add_ip4_vif.

Dear author:
I would like to report a problem (bug).

at source_timer_handler() in data.c, the code segment could have crash risk.

The imp_group_cleanup(p_gp) will free (p_gp) at end of that function.
After return, the last line "imp_membership_db_update(&p_gp->group_addr);"

reference it again. But p_gp is freed. So, that could have problem with some compiler or system..

if(p_gp->type == GROUP_INCLUDE && imp_source_exist_allow(p_gp) == 0){

    /* If there are no more source records for the group, delete group record.
     * rfc 3376 [6.3]
     */
    imp_group_cleanup(p_gp);
}
imp_membership_db_update(&p_gp->group_addr);