Looks good to me. My only concern is the additional gamedata but its minimal and very easy to find.

This changes the max MTU size and when we talked about this when the situation was more "emotional", lets say, this is changing limits the game has defined. For whatever reason Valve still believes limiting the maximum MTU is a good idea, we should respect that (our positions have essentially inversed ;) ).

The solution at that time was to scrap querying servers, and instead do a periodic upload of server data back to an AM WebAPI to do stats. This is a lot more work as there's backend changes, but it was the path forward however many years ago we discussed this.

I'm flexible, I see some SW.ext code in there and that's totally fine, but we're not only changing their MTU on rules inquiries, but on all network traffic.

but we're not only changing their MTU on rules inquiries, but on all network traffic.

That is not true.

This isn't changing it on all network traffic. The check only affects outbound, connectionless packets. No netchannel traffic is affected. Also, CS:GO is the only engine branch with this check, despite most of the other logic around it being identical to others.

You could add a hook on host_rules_show and only change the maxMTU and set the keyvalues on steam if it's enabled.

I had considered that. There's just minimal gain if it isn't "fixed" to match all other games by default, still requiring manual action similar to how it does as-is.

I just realized that I may have misunderstood what you were saying. You didn't mention changing the default, just not making the changes - doing anything, if it is disabled. That's a good idea!

Can we merge this?

I want to do what @peace-maker suggested first, only having it enabled when host_rules_show is true, but haven't had the time yet.

I had a look at this today, the original value should probably be compared to make sure its the value it should be before we start patching incorrect bytes. This way we can identify if an offset changed easier besides a random crash.

Not tested but implemented the change. Compiles locally but fails with ambuild atm needs to be looked at. Regarding my previous crash i totally forgot the signature contained the actual patched bytes so we should be fine without checking. The gamedata also still seems to be good.

@psychonic @Drifter321 what's the skinny on this one?