[PATCH v2 0/2] Allow users to give source address with --gateway also for IPv6
...just like it happens with IPv4. This offers a solution for https://bugs.passt.top/show_bug.cgi?id=217, as Podman can specify --address and --gateway for IPv6, and provide a unique local address, without the need to introduce other options. Further options might be more appropriate, strictly speaking, but as the future perspective is to eventually generalise NAT tables, it would be nice to avoid them right now. v2: In 2/2, use the address passed as --gateway as source for inbound NAT from the host also when we're not in local mode Stefano Brivio (2): conf: Honour --address, --gateway, --netmask in local mode as well conf, fwd: Prefer same-scope address as inbound source address from host conf.c | 25 +++++++++++++++++++++---- fwd.c | 6 +++++- passt.h | 2 ++ 3 files changed, 28 insertions(+), 5 deletions(-) -- 2.43.0
When I implemented local mode in 14b84a7f077e ("treewide: Introduce
'local mode' for disconnected setups"), I didn't consider the
possibility that, also in that case, the user might want to override
addresses, default gateway or netmask, even though I expressly
mentioned this in the man page:
In this case, **unless configured otherwise**, they will assign the
IPv4 link-local address 169.254.2.1 to the guest or target
namespace, and no IPv6 address.
Fix this by checking if an address, gateway, or netmask length was
explicitly set by the user, before overriding them with the default
parameters for local mode.
This might lead to invalid configurations where we won't be able to
set the default gateway passed by the user, but we print a warning
message, and we assume users know what they're doing in that case.
Link: https://bugs.passt.top/show_bug.cgi?id=217
Fixes: 14b84a7f077e ("treewide: Introduce 'local mode' for disconnected setups")
Signed-off-by: Stefano Brivio
We might have situations, such as the one described in
https://bugs.passt.top/show_bug.cgi?id=217, where using a link-local
address as source in a given namespace doesn't guarantee that we can
reach the intended destination, because, for instance, the inbound
traffic we forward is in turn forwarded to a different interface, such
as a bridge.
In that case, the assumption from 9618d247006a ("ndp, dhcpv6, tcp,
udp: Always use link-local as source if gateway isn't") isn't a safe
one: the user might have specified a valid gateway address, matching
the scope of the destination address, but we won't use it as address
of last resort, and prefer a link-local address with a mismatch in
scope instead.
So, if the user specifies a given default gateway address for IPv6,
note that as 'our_tap_addr', like we would do with with IPv4, and
stick to Rule 2 of RFC 6724, Section 5, when selecting a source
address, by preferring an address with the same scope, if available.
Note that this preference will be applied only if, at the point of
the source address selection, the destination address is already
known, which, in general, only happens if there's an explicit mapping
rule specifying the target destination address.
Otherwise, if there's no explicit rule picking the destination
address, we select the destination address after selecting the source
address. This should probably be improved at some point, to select
matching address pairs (when a compatible one is available) instead of
sequentially picking source and destination addresses. That appears to
be beyond the scope of this patch, though.
Reported-by: Paul Holzinger
On Thu, Jul 23, 2026 at 05:16:03PM +0200, Stefano Brivio wrote:
We might have situations, such as the one described in https://bugs.passt.top/show_bug.cgi?id=217, where using a link-local address as source in a given namespace doesn't guarantee that we can reach the intended destination, because, for instance, the inbound traffic we forward is in turn forwarded to a different interface, such as a bridge.
In that case, the assumption from 9618d247006a ("ndp, dhcpv6, tcp, udp: Always use link-local as source if gateway isn't") isn't a safe one: the user might have specified a valid gateway address, matching the scope of the destination address, but we won't use it as address of last resort, and prefer a link-local address with a mismatch in scope instead.
So, if the user specifies a given default gateway address for IPv6, note that as 'our_tap_addr', like we would do with with IPv4, and stick to Rule 2 of RFC 6724, Section 5, when selecting a source address, by preferring an address with the same scope, if available.
Note that this preference will be applied only if, at the point of the source address selection, the destination address is already known, which, in general, only happens if there's an explicit mapping rule specifying the target destination address.
Otherwise, if there's no explicit rule picking the destination address, we select the destination address after selecting the source address. This should probably be improved at some point, to select matching address pairs (when a compatible one is available) instead of sequentially picking source and destination addresses. That appears to be beyond the scope of this patch, though.
Reported-by: Paul Holzinger
Link: https://bugs.passt.top/show_bug.cgi?id=217 Signed-off-by: Stefano Brivio
Reviewed-by: David Gibson
--- v2: Use the address passed as --gateway as source for inbound NAT from the host regardless of local mode, see:
https://bugs.passt.top/show_bug.cgi?id=217#c5
conf.c | 8 ++++++-- fwd.c | 6 +++++- passt.h | 2 ++ 3 files changed, 13 insertions(+), 3 deletions(-)
diff --git a/conf.c b/conf.c index 8205dbf..faf2681 100644 --- a/conf.c +++ b/conf.c @@ -490,6 +490,8 @@ static unsigned int conf_ip6(unsigned int ifi, struct ip6_ctx *ip6)
if (IN6_IS_ADDR_LINKLOCAL(&ip6->guest_gw)) ip6->our_tap_ll = ip6->guest_gw; + else + ip6->our_tap_addr = ip6->guest_gw;
if (IN6_IS_ADDR_UNSPECIFIED(&ip6->addr) || IN6_IS_ADDR_UNSPECIFIED(&ip6->our_tap_ll)) @@ -507,10 +509,12 @@ static void conf_ip6_local(struct ip6_ctx *ip6) if (IN6_IS_ADDR_UNSPECIFIED(&ip6->guest_gw)) ip6->guest_gw = IP6_LL_GUEST_GW;
- if (IN6_IS_ADDR_LINKLOCAL(&ip6->guest_gw)) + if (IN6_IS_ADDR_LINKLOCAL(&ip6->guest_gw)) { ip6->our_tap_ll = ip6->guest_gw; - else + } else { + ip6->our_tap_addr = ip6->guest_gw; ip6->our_tap_ll = IP6_LL_GUEST_GW; + }
ip6->no_copy_addrs = ip6->no_copy_routes = true; } diff --git a/fwd.c b/fwd.c index 7152169..4ba0af3 100644 --- a/fwd.c +++ b/fwd.c @@ -1090,7 +1090,11 @@ uint8_t fwd_nat_from_host(const struct ctx *c, return PIF_NONE; tgt->oaddr = inany_from_v4(c->ip4.our_tap_addr); } else { - tgt->oaddr.a6 = c->ip6.our_tap_ll; + if (inany_is_linklocal6(&tgt->eaddr) || + IN6_IS_ADDR_UNSPECIFIED(&c->ip6.our_tap_addr)) + tgt->oaddr.a6 = c->ip6.our_tap_ll; + else + tgt->oaddr.a6 = c->ip6.our_tap_addr; } } tgt->oport = ini->eport; diff --git a/passt.h b/passt.h index a61baca..51ccd4f 100644 --- a/passt.h +++ b/passt.h @@ -121,6 +121,7 @@ struct ip4_ctx { * @dns: DNS addresses for DHCPv6 and NDP * @dns_match: Forward DNS query if sent to this address * @our_tap_ll: Link-local IPv6 address for passt's use on tap + * @our_tap_addr: Non-LL IPv6 address for passt's use on tap (if any) * @dns_host: Use this DNS on the host for forwarding * @addr_out: Optional source address for outbound traffic * @ifname_out: Optional interface name to bind outbound sockets to @@ -140,6 +141,7 @@ struct ip6_ctx { struct in6_addr dns[MAXNS]; struct in6_addr dns_match; struct in6_addr our_tap_ll; + struct in6_addr our_tap_addr;
/* PIF_HOST addresses */ struct in6_addr dns_host; -- 2.43.0
-- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibson
participants (2)
-
David Gibson
-
Stefano Brivio