Hi Harald and list!

I finished my patch to add an ip_conntrack_expect_alloc(...) function. I 
also patched all the current conntrack helpers available in the stable 
branch and patch-o-matic-ng.

Some considerations:

- I tested the irc, tftp and ftp helpers, they work fine for me. I 
haven't tested the others since I don't have access to all those 
protocols (I mean, I would need more time to install some programs and 
do some testing), but I have no problem about fixing problems related to 
those patches.

- I saw that Patrick submitted recently a patch to clean up the amanda 
helper (BTW, nice work!), I tried to patch my kernel with his patch 
available in patch-o-matic but it failed. So I patched the current. I 
also have no problem about patching the new one.

BTW, I know that there's a skeleton of a helper in the documention, due 
to they mostly look the same but I think that I could add it more details.

- There's a memset in expect_alloc (see expect.patch) after the 
allocation is done, I don't like it anyway but I just added to keep 
backward compatibility because some helper usually do it. I think that 
it's better setting all the fields of the new expectation in the helper 
(this implies being a bit more restrictive with the programmer but well, 
this is kernel programming anyway, it's restrictive :-)).

I could delete that memset and have a look at all the helpers to make 
sure that all the fields are correctly set.

- I didn't exported the expect_insert function, but you could do it 
later to use it in the ct_sync code.

- I found a condition where expect_related could succesfully insert an 
expectaction when removing an old one and return EPERM (see 
expect_related-newnat14.patch). I think that this doesn't make sense. Am 
I forgetting something? That patch fix that problem.

- I also have no problem about updating the documentation to add info 
about expect_alloc.

cheers,
Pablo