Slot1launch: Separate reading 1T-ROM and MROM cartridge headers - Fix address of ModuleParams - Fix loading and setting personalData

[Lorenzooone]

What's changed?

1. Fixed issue preventing MROM DSi cartridges from launching in DSi mode (by splitting the code for reading the header for MROM and 1T-ROM). Also fixes launching cartridges in DSi mode in melonDS, since it assumes all cartridges are MROMs, even for big games like Pokémon Black (1T-ROM on real cartridges). melonDS slot1launch DSi mode support was broken by mistake in commit f181b22 , though since using DSi mode currently requires changing a define and the Makefile, it was not noticed.

2. Fixed the address used for ModuleParams (and adds a special case for SpiderMan 2 - USA). Previously, it wrongly assumed the magic values were at the beginning of ModuleParams, while they are at the end. This also fixes the various isSdk5 checks to work properly. In general, the isSdk5 check being fixed makes it so some DSi games that previously did not work now launch (and improves compatibility with forcing full DSi mode for some DS games, even though that's not a real goal). Finally, also introduces a NULL check to prevent reading bad data.

3. Fixed reading and changing the personalData (like the language) that is passed to the games. The pointers used were swapped. Though many games end up loading the data from the firmware regardless, so... It's more of a "that was wrong, let's fix it" kind of deal.

Ported from https://github.com/Lorenzooone/Simple-DS-Slot-1-Launcher

Where have you tested it?

3DS via TwilightMenu++ application and DSi via unlaunch + TwilightMenu++.
Tested with a retail DS cartridge, a retail DSi cartridge and a dev DSi cartridge with homebrew on it.
Also tested on melonDS, now that it works again.

Pull Request status

• This PR has been tested using the latest version of devkitARM and libnds.