r1ch.net forums
* Home Help Search Login Register
r1ch.net  |  r1ch.net stuff  |  Anticheat  |  Topic: Anticheat and Linux i386
Pages: 1 [2] 3 4  All
Print
Author Topic: Anticheat and Linux i386  (Read 76472 times)
dk_sn1p3r
Member

Posts: 23


« Reply #15 on: December 07, 2006, 03:49:42 am »

I would love to see a linux flavor  for clients but I could completly understand never seing one for linux clients I love the fact that the anticheat module is able to run on linux servers I might end up hosting a linux server eventually but not atm...
Logged
undercouves
Member

Posts: 1


« Reply #16 on: December 15, 2006, 03:37:05 pm »

They cant release the source code. That would provide coders the necessary information to crack the anticheat. Notice that there is no opensource anticheat due to that problem.

And about portability, you don't need several binaries for different distros, since every distro is a slightly modified linux, mostly the X desktop and its utilities. Almost every distro uses the same kernel version (2.6) with minor changes. What they would need is a version for the most used archs (x86, x86_64, ppc, etc), but then again, only a few users have non-x86 computer.

The code... probably, the anticheat's DLL is based on some language that is not supported in linux (IE: vb). What linux users can do is, to install something like CEDEGA to run win32 based apps. Most linux gamers use it to run unsupported games.

Anyway... as i always say: if you like linux, use it. But if it doesn't fit your needs, you can always use something else.

Btw, suse IS NOT the most used distro, not even by gamers.
Logged
dk_sn1p3r
Member

Posts: 23


« Reply #17 on: December 17, 2006, 06:57:12 am »

Personally I use fedora but sence I dual boot and if i'm going to be using linux for quake i'd rather use linux client because why emulate and use crappier libraries when I can reboot to windows 2k for better performance... I'm kinda disapointed that linux doesn't match up and that everything has to be open source but on the contrary I don't think you'd have to provide the source for anticheat.dll because thats something rich concocted all to him self i'm sure the code thats in r1q2 for the dll would be shown but isn't there a way to clean that up as to not show anything important for the anti.dll besides thats how the windows code is released isn't the latest version of r1q2 source available and not the .dll? couldn't you do the same on linux so long as there is an linux coder on the team or rich has the time to code for linux...

Oh and lastly what do you guys use for a client on linux do you guys emulate or use a native linux client?
Logged
wision
Member

Posts: 237



« Reply #18 on: December 17, 2006, 12:25:26 pm »

afaik there won't be any linux anticheat because of win api..

i tried sdlq2, q2pro and aprq2... with sdlq2 i had to change some variable declarations (removed static declarations of some mouse variables) so i could compile it... but ingame i had bad feeling about sensitivity etc... with q2pro i had bad feeling about sensitivity too so i tried aprq2 and aprq2 works nice (still had to remove some xmms and other useless functions to make it compileable <changes in Makefile didn't work>)... i'm used to 120fps but glue step appears quite often with 120 so i had to decrease it to 110...
anybody knows how to get rid of glue step for whatever maxfps i use?
Logged
R1CH
Administrator
Member

Posts: 2625



« Reply #19 on: December 17, 2006, 01:24:12 pm »

sv_strafejump_hack needs to be enabled on the server you play on.
Logged
dk_sn1p3r
Member

Posts: 23


« Reply #20 on: December 18, 2006, 09:11:00 pm »

when your all done messing around with that can I have a copy of the code I'd like to get a relitivly decant anticheat enabled client running for linux just for kicks....
Logged
TgT
Member

Posts: 103


« Reply #21 on: January 25, 2007, 09:43:44 pm »

So there is no chance of anticheat variant for linux at all?  I mean AT ALL! I bet there is a sollution but you dont have time/motivation right?
Logged
R1CH
Administrator
Member

Posts: 2625



« Reply #22 on: January 26, 2007, 02:42:52 am »

No, since Linux is fundamentally impossible to verify as a trusted environment. You have many hundreds of distributions, each with different binaries depending on CPU type and then of course the thousands of users who build things themselves from source. When you can't even trust that the kernel is legit, trying to create an anticheat solution off such a shaky foundation is pointless. Not to mention binary incompatibility that still plagues Linux.
Logged
zorg
Member

Posts: 49


« Reply #23 on: January 26, 2007, 03:40:52 am »

Just an idea.

What about a versoin of anticheat for linux, which only has some limited functionality? For example, the filechecking.  In this case, linux players could at least verify that they are not using any modified models/skins/whatever.

Still, I don't have any idea, how "deep" each anticheat codepart of all the different functions has to be implemented into the OS. But maybe some "simple" parts are possible.
Logged
TgT
Member

Posts: 103


« Reply #24 on: January 26, 2007, 08:41:14 am »

Hundreds of distros but only few used. Gentoo,Suse,Ubuntu mostly... and I still think that some sort of anticheat like zorg said CAN be done. I wonder how Warsow works then?
Logged
QwazyWabbit
Member

Posts: 402


« Reply #25 on: January 26, 2007, 09:03:09 am »

Just an idea.

What about a versoin of anticheat for linux, which only has some limited functionality? For example, the filechecking.  In this case, linux players could at least verify that they are not using any modified models/skins/whatever.

Still, I don't have any idea, how "deep" each anticheat codepart of all the different functions has to be implemented into the OS. But maybe some "simple" parts are possible.

But this is the heart of the A/C problem. If you can't trust the OS you can't trust results reported from an application. The core problem of any A/C solution is that it can't be hacked and the platform must not be subverted. If you can compile your own kernel you can make it report anything you want to an anti-cheat module. The problem still exists in Windows but the task is more difficult. A filesystem shim might work there but writing that kind of driver is not trivial.

As for simple parts; this was tried and failed with NoCheat.
Logged
R1CH
Administrator
Member

Posts: 2625



« Reply #26 on: January 26, 2007, 09:40:50 am »

I wonder how Warsow works then?
No offense to whoever made it, but Warsow's anticheat is pretty much worthless. It can be subverted in just a few bytes of patches and makes no effort to really stop cheats, last time I looked at it at least all it seemed to do was essentially a checksum of the executable and game data.

A simple version, as QwazyWabbit points out, NoCheat was just that - an attempt at least to verify the executable and DLL and a select number of models. It was easily hacked, in fact, the NoCheat 'code' generation code was all completely open in the Linux binary to anyone with a disassembler. With the Windows DLL there was at least the possibility of using a PE protector, such things in Linux are difficult to implement (see UPX for example) and hard to come by, especially given the wide array of distributions and kernels it must support. Having a Linux version do anything without protecting the code against casual disassembly just hurts the overall security of the entire system.
Logged
darkstalker
Guest
« Reply #27 on: January 27, 2007, 04:07:57 pm »

is it possible to patch anticheat.dll so that it works with wine under linux?
i tried and it didn't work (maybe because some missing api calls in the wine implementation)
Logged
Xtyfe
Guest
« Reply #28 on: February 04, 2007, 06:49:58 am »

this whole discussion is moot, there are no cheats that work under linux or any other non-windows system (correct me if im wrong)
with the exception of anything added to pak files, which shouldnt be a problem either since there are no pak editors for linux (same goes here, i know nothing about this)

the best way to go about this is to exclude people using linux and these other OSs so they are not branded as cheaters
Logged
QwazyWabbit
Member

Posts: 402


« Reply #29 on: February 04, 2007, 03:10:30 pm »

PAK files are portable from Windows to Linux so a pak file that was spiked for Windows will work in Linux clients too. A modified client (if modified and recompiled from source) would work in Linux and Windows so exempting Linux clients is not a solution to the cheating problem. The A/C module is problematic on Linux for many reasons and some of them may simply not be fixable. The WINE solution may be viable but it would take some testing and I am sure there will be people complaining about that restriction too.
Logged
Pages: 1 [2] 3 4  All
Print
r1ch.net  |  r1ch.net stuff  |  Anticheat  |  Topic: Anticheat and Linux i386
Jump to:  

Powered by SMF 1.1.19 | SMF © 2013, Simple Machines