Re: [PATCH nft] evaluate: add range specified flag setting (missing NF_NAT_RANGE_PROTO_SPECIFIED)

["Сергей Маринкевич" <s@marinkevich.ru>]

I sorry, my e-mail client goes bananaz.

 > Sergey reports:
 >
 > With nf_tables it is not possible to use port range for masquerading.
 > Masquerade statement has option "to [:port-port]" which give no effect
 > to translation behavior. But it must change source port of packet to
 > one from ":port-port" range.
 >
 > My network:
 >
 >         +-----------------------------+
 >         |   ROUTER                    |
 >         |                             |
 >         |                   Masquerade|
 >         | 10.0.0.1            1.1.1.1 |
 >         | +------+           +------+ |
 >         | | eth1 |           | eth2 | |
 >         +-+--^---+-----------+---^--+-+
 >              |                   |
 >              |                   |
 >         +----v------+     +------v----+
 >         |           |     |           |
 >         | 10.0.0.2  |     |  1.1.1.2  |
 >         |           |     |           |
 >         |PC1        |     |PC2        |
 >         +-----------+     +-----------+
 >
 > For testing i used rule like this:
 >
 >         rule ip nat POSTROUTING oifname eth2 masquerade to :666
 >
 > Run netcat for 1.1.1.2 667(UDP) and get dump from PC2:
 >
 >         15:22:25.591567 a8:f9:4b:aa:08:44 > a8:f9:4b:ac:e7:8f, 
ethertype IPv4 (0x0800), length 60: 1.1.1.1.34466 > 1.1.1.2.667: UDP, 
length 1
 >
 > Address translation works fine, but source port are not belongs to
 > specified range.
 >
 > I see in similar source code (i.e. nft_redir.c, nft_nat.c) that
 > there is setting NF_NAT_RANGE_PROTO_SPECIFIED flag. After adding this,
 > repeat test for kernel with this patch, and get dump:
 >
 >         16:16:22.324710 a8:f9:4b:aa:08:44 > a8:f9:4b:ac:e7:8f, 
ethertype IPv4 (0x0800), length 60: 1.1.1.1.666 > 1.1.1.2.667: UDP, length 1
 >
 > Now it is works fine.
 >
 > Reported-by: Sergey Marinkevich <s@marinkevich.ru>
 > Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
 > ---
 > Sergey, could you try this userspace patch instead? Thanks.
I tried this patch with equal environment but another

net(192.168.122.0/24). Router uses vanilla kernel v5.4.19.
Translation is the same.

     12:59:11.599887 08:00:27:ec:9c:b3 > 52:54:00:57:d2:7d, ethertype 
IPv4 (0x0800), length 60: 192.168.122.38.666 > 192.168.122.1.667: UDP, 
length 4

I think i have to add this tag:

Tested-by: Sergey Marinkevich <s@marinkevich.ru>

 >
 >  src/evaluate.c | 3 +++
 >  1 file changed, 3 insertions(+)
 >
 > diff --git a/src/evaluate.c b/src/evaluate.c
 > index 4a23b231c74d..d0e712dc02f0 100644
 > --- a/src/evaluate.c
 > +++ b/src/evaluate.c
 > @@ -18,6 +18,7 @@
 >  #include <linux/netfilter_arp.h>
 >  #include <linux/netfilter/nf_tables.h>
 >  #include <linux/netfilter/nf_synproxy.h>
 > +#include <linux/netfilter/nf_nat.h>
 >  #include <linux/netfilter_ipv4.h>
 >  #include <netinet/ip_icmp.h>
 >  #include <netinet/icmp6.h>
 > @@ -2950,6 +2951,8 @@ static int stmt_evaluate_nat(struct eval_ctx 
*ctx, struct stmt *stmt)
 >          err = nat_evaluate_transport(ctx, stmt, &stmt->nat.proto);
 >          if (err < 0)
 >              return err;
 > +
 > +        stmt->nat.flags |= NF_NAT_RANGE_PROTO_SPECIFIED;
 >      }
 >
 >      stmt->flags |= STMT_F_TERMINAL;
 > --
 > 2.11.0