Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
From: Peter Enderborg <peterend@axis.com>
To: "FFmpeg development discussions and patches"
	<ffmpeg-devel@ffmpeg.org>,
	"Rémi Denis-Courmont" <remi@remlab.net>
Subject: Re: [FFmpeg-devel] [PATCH] avformat/udp: Select output interfaces for ipv6 multicast
Date: Tue, 12 Aug 2025 12:59:32 +0200
Message-ID: <e21b0817-b6da-4e46-8390-f0b51da87362@axis.com> (raw)
In-Reply-To: <47824AC2-40F4-4FA4-B9A6-0F71646F4117@remlab.net>


On 8/6/25 17:28, Rémi Denis-Courmont wrote:
>
> Le 5 août 2025 14:56:18 GMT+07:00, Peter Enderborg <peterend@axis.com> a écrit :
>> This fixes two old TODO's in ipv6 multicast handling.
>> A helper function to iterate over all interfaces added to
>> help join the multicast group on interface if approperite.
>>
>> The default value is IN6ADDR_ANY (::) and it then joins all interfaces
>> that are up and have multicast support. Local address can be
>> specified and then it join ALL interfaces that use that specific
>> local address.
>>
>> Limitations (TODO's)
>> Handling when network configuration is changed. if up/down etc.
>>
>> Signed-off-by: Peter Enderborg <peterend@axis.com>
>> ---
>> libavformat/udp.c | 108 ++++++++++++++++++++++++++++++++++++++--------
>> 1 file changed, 90 insertions(+), 18 deletions(-)
>>
>> diff --git a/libavformat/udp.c b/libavformat/udp.c
>> index 035db785c2..0a981447a5 100644
>> --- a/libavformat/udp.c
>> +++ b/libavformat/udp.c
>> @@ -68,6 +68,13 @@
>> #include "libavutil/thread.h"
>> #endif
>>
>> +#if HAVE_STRUCT_IPV6_MREQ && defined(IPPROTO_IPV6)
>> +#include <sys/types.h>
>> +#include <net/if.h>
>> +#include <sys/ioctl.h>
>> +#include <ifaddrs.h>
>> +#endif
>> +
>> #ifndef IPV6_ADD_MEMBERSHIP
>> #define IPV6_ADD_MEMBERSHIP IPV6_JOIN_GROUP
>> #define IPV6_DROP_MEMBERSHIP IPV6_LEAVE_GROUP
>> @@ -205,6 +212,79 @@ static int udp_set_multicast_ttl(int sockfd, int mcastTTL,
>>
>>      return 0;
>> }
>> +#if HAVE_STRUCT_IPV6_MREQ && defined(IPPROTO_IPV6)
>> +static int udp_ipv6_multicast_iterate(int sockfd, struct sockaddr_in6 *addr,
>> +                                      struct sockaddr_in6 *local_addr, int mop, void *logctx)
>> +{
>> +    struct in6_addr anyipv6 = IN6ADDR_ANY_INIT;
>> +    struct ifaddrs *ifal=NULL,*ife=NULL;
>> +    int iindex_fd = socket(AF_INET, SOCK_STREAM, IPPROTO_IP);
>> +    int membership_changed = 0;
>> +
>> +    if (iindex_fd >= 0 && !getifaddrs(&ifal)) {
>> +        for(ife=ifal; ife!=NULL; ife=ife->ifa_next) {
>> +            if (ife->ifa_addr &&
>> +                (ife->ifa_addr->sa_family == AF_INET6) &&
>> +                (ife->ifa_flags & IFF_MULTICAST) &&
>> +                (ife->ifa_flags & IFF_UP)) {
>> +                if ((!memcmp(&local_addr->sin6_addr, &anyipv6, sizeof(struct in6_addr))) ||
>> +                    (!memcmp(&local_addr->sin6_addr, &((struct sockaddr_in6 *)ife->ifa_addr)->sin6_addr, sizeof(struct in6_addr)))) {
>> +                    struct ifreq if_req;
>> +                    strncpy(if_req.ifr_name, ife->ifa_name, IFNAMSIZ);
>> +                    if (ioctl(iindex_fd, SIOCGIFINDEX, &if_req) != -1) {
>> +                        struct ipv6_mreq mreq6;
>> +                        memcpy(&mreq6.ipv6mr_multiaddr, &addr->sin6_addr, sizeof(struct in6_addr));
>> +                        mreq6.ipv6mr_interface = if_req.ifr_ifindex;
>> +                        switch (mop) {
>> +                        case IPV6_JOIN_GROUP:
>> +                            if (setsockopt(sockfd, IPPROTO_IPV6, IPV6_ADD_MEMBERSHIP, &mreq6, sizeof(mreq6)) == 0) {
>> +                                membership_changed = 1;
>> +                            } else {
>> +                                /* Subscribe for all ipv6 addresses belong to a
>> +				   interface.It can generate EADDRINUSE but
>> +				   is accaptable. */
>> +                                if (errno != EADDRINUSE) {
>> +                                    ff_log_net_error(logctx, AV_LOG_ERROR, "setsockopt(IPV6_ADD_MEMBERSHIP)");
>> +                                    goto error;
>> +                                }
>> +                            }
>> +                            break;
>> +                        case IPV6_LEAVE_GROUP:
>> +                            if (setsockopt(sockfd, IPPROTO_IPV6, IPV6_DROP_MEMBERSHIP, &mreq6, sizeof(mreq6)) == 0) {
>> +                                membership_changed = 1;
> You can't leave a group you didn't join and you can't assume that the set of interfaces is constant over time.
Did you read the commit message?
> Besides joining a group on more than one interface is generally a terrible idea, as you would receive the same data multiple time (+ reordering).

It's not a problem. At least not in linux. You will get an EADDRINUSE if 
you have another request for the same interface. The problem
occurs when the network topology is changed. Still, it's a better 
solution to try with the topology that exist when the application tries
to join.  You will get duplicates if multiple network interfaces are 
connected to the same network segment, but not in a more
reasonable setup, and you can still narrow it down by using a specific 
interface address in that case.

> It makes sense to track interfaces and subscribe to all of them for discovery protocols, but it's way more involved (and less portable) than this.

It make sense to try all interfaces when the application uses ANY 
address as input parameter. Without the patch it uses
interface 0, then the linux kernel will fallback to do a rt6_lookup. The 
function rt6_lookup fallback selects one interface
based on the routing table. However, the routing table is a function for 
sending, not for receiving, and to change routing
tables it requires admin privileges. This will result in that users 
without this admin privilege can not access the multicast stream.

>
>> +                            } else {
>> +                                /* Unsubscribe for all ipv6 addresses belong to a
>> +				   interface. Ignore EADDRNOTAVAIL. */
>> +                                if (errno != EADDRNOTAVAIL) {
>> +                                    ff_log_net_error(logctx, AV_LOG_ERROR, "setsockopt(IPV6_DROP_MEMBERSHIP)");
>> +                                    goto error;
>> +                                }
>> +                            }
>> +                            break;
>> +                        default:
>> +                            av_log(logctx, AV_LOG_ERROR, "unknown multicast operation\n");
>> +                            goto error;
>> +                        }
>> +                    }
>> +                }
>> +            }
>> +        }
>> +    }
>> +    freeifaddrs(ifal);
>> +    close(iindex_fd);
>> +
>> +    if (!membership_changed) {
>> +        av_log(logctx, AV_LOG_ERROR, "no valid interfaces found\n");
>> +        return AVERROR(EINVAL);
>> +    }
>> +    return 0;
>> +error:
>> +    freeifaddrs(ifal);
>> +    close(iindex_fd);
>> +    av_log(logctx, AV_LOG_ERROR, "udp_ipv6_multicast_iterate failed\n");
>> +    return AVERROR(EINVAL);
>> +}
>> +#endif
>>
>> static int udp_join_multicast_group(int sockfd, struct sockaddr *addr,
>>                                      struct sockaddr *local_addr, void *logctx)
>> @@ -226,15 +306,11 @@ static int udp_join_multicast_group(int sockfd, struct sockaddr *addr,
>> #endif
>> #if HAVE_STRUCT_IPV6_MREQ && defined(IPPROTO_IPV6)
>>      if (addr->sa_family == AF_INET6) {
>> -        struct ipv6_mreq mreq6;
>> -
>> -        memcpy(&mreq6.ipv6mr_multiaddr, &(((struct sockaddr_in6 *)addr)->sin6_addr), sizeof(struct in6_addr));
>> -        //TODO: Interface index should be looked up from local_addr
>> -        mreq6.ipv6mr_interface = 0;
>> -        if (setsockopt(sockfd, IPPROTO_IPV6, IPV6_ADD_MEMBERSHIP, &mreq6, sizeof(mreq6)) < 0) {
>> -            ff_log_net_error(logctx, AV_LOG_ERROR, "setsockopt(IPV6_ADD_MEMBERSHIP)");
>> -            return ff_neterrno();
>> -        }
>> +      return udp_ipv6_multicast_iterate(sockfd,
>> +					(struct sockaddr_in6 *)addr,
>> +					(struct sockaddr_in6 *)local_addr,
>> +					IPV6_JOIN_GROUP,
>> +					logctx);
>>      }
>> #endif
>>      return 0;
>> @@ -260,15 +336,11 @@ static int udp_leave_multicast_group(int sockfd, struct sockaddr *addr,
>> #endif
>> #if HAVE_STRUCT_IPV6_MREQ && defined(IPPROTO_IPV6)
>>      if (addr->sa_family == AF_INET6) {
>> -        struct ipv6_mreq mreq6;
>> -
>> -        memcpy(&mreq6.ipv6mr_multiaddr, &(((struct sockaddr_in6 *)addr)->sin6_addr), sizeof(struct in6_addr));
>> -        //TODO: Interface index should be looked up from local_addr
>> -        mreq6.ipv6mr_interface = 0;
>> -        if (setsockopt(sockfd, IPPROTO_IPV6, IPV6_DROP_MEMBERSHIP, &mreq6, sizeof(mreq6)) < 0) {
>> -            ff_log_net_error(logctx, AV_LOG_ERROR, "setsockopt(IPV6_DROP_MEMBERSHIP)");
>> -            return -1;
>> -        }
>> +	return udp_ipv6_multicast_iterate(sockfd,
>> +					  (struct sockaddr_in6 *)addr,
>> +					  (struct sockaddr_in6 *)local_addr,
>> +					  IPV6_LEAVE_GROUP,
>> +					  logctx);
>>      }
>> #endif
>>      return 0;
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

  reply	other threads:[~2025-08-12 10:59 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-08-05  7:56 Peter Enderborg
2025-08-06 15:28 ` Rémi Denis-Courmont
2025-08-12 10:59   ` Peter Enderborg [this message]
2025-08-07 10:49 ` [FFmpeg-devel] [PATCH v2 0/2] " Peter Enderborg
2025-08-07 10:49 ` [FFmpeg-devel] [PATCH v2 1/2] configure: Add test_ioctl and test for SIOCGIFINDEX Peter Enderborg
2025-08-07 10:49 ` [FFmpeg-devel] [PATCH v2 2/2] avformat/udp: Select output interfaces for ipv6 multicast Peter Enderborg

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=e21b0817-b6da-4e46-8390-f0b51da87362@axis.com \
    --to=peterend@axis.com \
    --cc=ffmpeg-devel@ffmpeg.org \
    --cc=remi@remlab.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
		ffmpegdev@gitmailbox.com
	public-inbox-index ffmpegdev

Example config snippet for mirrors.


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git