ip_nat_amanda unresolved symbol

ip_nat_amanda unresolved symbol

[Jan Srzednicki]

Hello there,

I've just cvsupped new version of netfilter stuff (16/10/2002). I wanted
to try some of the new modules in patch-o-matic, one of which is the
Amanda connection tracking helper. However, this module contains an
unresolved symbol:

find kernel -path '*/pcmcia/*' -name '*.o' | xargs -i -r ln -sf ../{} pcmcia
if [ -r System.map ]; then /sbin/depmod -ae -F System.map 2.4.19; fi
depmod: *** Unresolved symbols in
/lib/modules/2.4.19/kernel/net/ipv4/netfilter/ip_nat_amanda.o
depmod:         ip_nat_mangle_udp_packet

I grepped all kernel sources and netfilter patches -
ip_nat_mangle_udp_packet is not defined anywhere! I looked on the code of
similar function - ip_nat_mangle_tcp_packet(); it is defined and exported
correctly. Yet, ip_nat_mangle_udp_packet() clearly is missing. I use the
"new" nat headers and schemes, if that matters in any way.

Here's the list of patches I have installed:

Already applied: submitted/2.4.18
                 submitted/DSCP
                 submitted/ECN
                 submitted/REJECT-dont_fragment
                 submitted/REJECT_mark
                 submitted/TOS-oops-fix
                 submitted/ahesp-static
                 submitted/arptables
                 submitted/conntrack+nat-helper-unregister
                 submitted/conntrack
                 submitted/dscp
                 submitted/ecn
                 submitted/helper
                 submitted/ip6tables-export-symbols
                 submitted/ip_nat_irc-srcaddr-fix
                 submitted/ipt_MIRROR-ttl
                 submitted/ipt_REJECT-checkentry
                 submitted/ipt_unclean-ecn
                 submitted/irc-dcc-mask
                 submitted/length-ipv6
                 submitted/local-nat
                 submitted/macro-trailing-semicolon-fix
                 submitted/mangle5hooks
                 submitted/nat-export_symbols
                 submitted/nat-memoryleak-fix
                 submitted/netfilter-arp
                 submitted/skb_clone_copy
                 submitted/ulog-module-unload
                 submitted/ulog-nlgroup-shift-fix
                 submitted/unclean-udpchecksum
                 submitted/z-newnat16
                 base/IPV4OPTSSTRIP
                 base/NETMAP
                 base/TTL
                 base/fuzzy
                 base/iplimit
                 base/ipt_unclean-ubit
                 base/ipv4options
                 base/mport
                 base/nth
                 base/quota
                 base/random
                 base/time
                 extra/CONNMARK
                 extra/amanda-conntrack-nat
                 extra/h323-conntrack-nat
                 extra/ip_tables-proc
                 extra/pptp-conntrack-nat
                 extra/recent
                 extra/string
                 extra/tcp-window-tracking
                 userspace/ipt_REJECT-fake-source


greetz,
-- 
Jan 'Winfried' Srzednicki
w@expro.pl

RE: ip_nat_amanda unresolved symbol

[Sneppe Filip]

[-- Attachment #1: Type: text/plain, Size: 1859 bytes --]

Hi,

Jan Srzednicki [mailto:winfried@expro.pl] wrote:
>
>Hello there,
>
>I've just cvsupped new version of netfilter stuff (16/10/2002). I wanted
>to try some of the new modules in patch-o-matic, one of which is the
>Amanda connection tracking helper. However, this module contains an
>unresolved symbol:
>
>find kernel -path '*/pcmcia/*' -name '*.o' | xargs -i -r ln -sf ../{} pcmcia
>if [ -r System.map ]; then /sbin/depmod -ae -F System.map 2.4.19; fi
>depmod: *** Unresolved symbols in
>/lib/modules/2.4.19/kernel/net/ipv4/netfilter/ip_nat_amanda.o
>depmod:         ip_nat_mangle_udp_packet
>
>I grepped all kernel sources and netfilter patches -
>ip_nat_mangle_udp_packet is not defined anywhere! I looked on the code of
>similar function - ip_nat_mangle_tcp_packet(); it is defined and exported
>correctly. Yet, ip_nat_mangle_udp_packet() clearly is missing. I use the
>"new" nat headers and schemes, if that matters in any way.
>

It seems like the amanda conntracker/nat module depends on some
earlier code written by the same author to mangle udp packets.
However, that earlier code did not get accepted; see the following
mails on netfilter-devel:

http://lists.netfilter.org/pipermail/netfilter-devel/2002-August/009012.html
http://lists.netfilter.org/pipermail/netfilter-devel/2002-September/009371.html
http://lists.netfilter.org/pipermail/netfilter-devel/2002-August/008916.html

So what you can do is:
- if you need to get this to work quickly, you can apply the patch
  that adds ip_nat_mangle_udp_packet() and recompile. This is more of 
  a quickfix solution, though.
- get in touch with the author of the amanda conntracker/nat module,
  because obviously the patch in p-o-m is of little use to anyone in its
  current state. (I've put him email address in the mail loop)

Regards,
Filip Sneppe


[-- Attachment #2: Type: text/html, Size: 2852 bytes --]

Re: ip_nat_amanda unresolved symbol

[netfilter]

[-- Attachment #1: Type: text/plain, Size: 1871 bytes --]

On Thu, Oct 17, 2002 at 01:37:02AM +0200, Sneppe Filip wrote:
>    It seems like the amanda conntracker/nat module depends on some
>    earlier code written by the same author to mangle udp packets.
>    However, that earlier code did not get accepted; see the following
>    mails on netfilter-devel:
>    [2]http://lists.netfilter.org/pipermail/netfilter-devel/2002-August/00
>    9012.html
>    [3]http://lists.netfilter.org/pipermail/netfilter-devel/2002-September
>    /009371.html
>    [4]http://lists.netfilter.org/pipermail/netfilter-devel/2002-August/00
>    8916.html

This is correct.  Harald did not like my implementation, and we came
to an agreement on how to correct it.  Unfortunately, making the
correction has not bubbled to the top of my TODO list (as much as
getting a job has!  :-)

[ Harald, please feel free to correct anything I get wrong here ]

There were two issues.  First, Harald did not like the "multi-mangle"
function that I wrote for UDP.  It was/is at this point an unused
function and a) Harald did not want to bloat the code with an unused
function and b) it would only be used in the rarer of cases and
mulitple calls to the single mangle function would be sufficient.  I
don't necessarily disagree.

The other issue was that my UDP mangle function _always_ calculated a
checksum on the mangled packet, even when it did not have a checksum
prior to mangling.  I have corrected the function to only re-calculate
the checksum when the checksum is 0 (indicated no UDP checksum on the
packet).

>    - get in touch with the author of the amanda conntracker/nat module,
>       because  obviously the patch in p-o-m is of little use to anyone in
>    its
>      current state.

Right.  Stay tuned for a patch to directly follow this e-mail with an
updated patch in it.

b.

-- 
Brian J. Murrell

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]