[Adhoc] Partially fixes multiplayer regression on GTA games.

[anr2me]

Partially fixes #20224

[hrydgard]

Sure. But what was the reason this was added?

[anr2me]

Was added on #19310 to prevent the connection to be in TIME_WAIT and might still existed on the router, was suspected to be the cause of port remapping (since the port mapping might still existed, thus attempting to create the same connection on a new socket might be remapped to a different port), but unconfirmed, since i don't have internet with working port forwarding to test it,

[hrydgard]

Ah ok. Reverting speculative changes seems very reasonable :)

(hm, that's a strange build error on alpine, I think we can ignore it)

[anr2me]

I will need to test the artifact first before we can merge this.

[anr2me]

Okay, the data no longer lost and can be received successfully.

Btw, is this a valid usage? return hleLogDebug(Log::sceNet, i + 1);
I'm seeing E[HLE]: HLE\HLE.cpp:1115 HLE function stack mismatch (D:\a\ppsspp\ppsspp\Core\HLE\sceNetAdhoc.cpp:3474)! stackSize = 0 error message on that return line.
Do i need to add a string as the 3rd argument to fix this may be?

[hrydgard]

It's correct but the "hleLog*" family should only be used at the very end of a syscall, since it logs all the params and return value and stuff. Other logging should be done with DEBUG/INFO/ERROR_LOG and friends.

[anr2me]

Well that return hleLogDebug(Log::sceNet, i + 1); used to print the socket id, along with the syscall like 1 = sceNetAdhocPtpOpen(args), but not sure why it got stacksize error.

PS: was accidentally clicking the close button while trying to comment using my phone 😅

[hrydgard]

You can ignore the stack size errors for now, I'll figure it out later.

[anr2me]

OK, we can merge this now.
But the GTA issue shouldn't be closed yet, since there is still an issue with group scanning.