1 |
Dave Nebinger wrote: |
2 |
> On Tuesday 11 October 2005 07:37 am, Steve [Gentoo] wrote: |
3 |
> |
4 |
>> I'm also vaguely hopeful that there may |
5 |
>> be a more efficient lower-level solution which wouldn't require the |
6 |
>> overhead of a process to 'pass-on' the tcp data... maybe integrated with |
7 |
>> ipchains or pf or similar? |
8 |
>> |
9 |
> If you choose to roll your own solution, that would be difficult. Youve |
10 |
> already accepted the connection, so the firewall is now configured to allow |
11 |
> the packets back and forth only when related to your connection. |
12 |
> |
13 |
I realise that the idea would necessarily be substantially more |
14 |
challenging than just writing a proxy... but I'm sure it is possible. |
15 |
I'm guessing I'd need to interact at the IP packet level, recognise the |
16 |
start of a TCP stream (buffering packets as necessary) then re-play them |
17 |
to the right port and force the packet filter to re-direct that TCP |
18 |
stream. It would not be worth my time to try and make this work if it |
19 |
isn't already available for me to just compile and use. |
20 |
> Technically the proxy development is not difficult, but for newbies it can be |
21 |
> frustrating working out the nuances of processing asynchronous data arriving |
22 |
> on one pipe let alone two. |
23 |
> |
24 |
I'm confident that I could write a proxy that would do this... as you |
25 |
suggest - it's not rocket science. Conversely, I'm lazy enough to just |
26 |
use one that's already written if one exists... which, I'm guessing, is |
27 |
likely as I doubt I'm the first person to tackle this. |
28 |
|
29 |
Steve |
30 |
|
31 |
-- |
32 |
gentoo-user@g.o mailing list |