emoose / re4-research

Various tools & modding research for Resident Evil 4
BSD 3-Clause "New" or "Revised" License
35 stars 4 forks source link

IDA 7.6 database #1

Open emoose opened 2 years ago

emoose commented 2 years ago

(E: IDA 7.7 database at #3, will probably only update that from now on)

License: The annotations included in the IDA database, such as names, structures, and comments, are made available to you under the following terms:

Please note that this license only applies to the information contained in the provided files, and does not necessarily apply to any derived works or modifications that you create using this information.

These license terms can be modified upon request, please feel free to get in contact.


Figured it'd be better to link new IDBs at one place instead of posting different comments everywhere.

IDB can be found here: pc-1.1.0_enum - Copy.7z.zip (old version see posts below for latest) (remove .zip extension so that it ends with .7z, as github doesn't allow 7z uploads for some reason...)

This names ~60% of functions in the bio4.exe, and I also made a start on a struct for pG / Globals (seems that struct is 0x8000+ bytes long though so would take a lot of work to label everything there... right now I'm just trying to get the correct field types inside it, haven't really named many fields yet)

(if anyone has done any work on naming any vars between bio4.exe+0x85A760 - bio4.exe+0x862E00 (1.1.0), or made any trainer that acts on that range or anything like that, please consider posting what you have, that range holds the GLOBALS struct, so any info would be appreciated!)

Unfortunately there are some things that are currently unnamable:

emoose commented 2 years ago

Named a few more funcs, filled out GLOBALS struct types quite a bit, & added EM_LIST struct:

Download: pc-1.1.0_enum - Copy.7z.zip

Still quite a lot to do inside GLOBALS though, need to find accesses to gapXXXX fields & name/type them properly (most of the remaining ones seem to be accessing structs instead of raw fields though, maybe PS2/VR version can give us names of those structs)

emoose commented 2 years ago

Had a small breakthrough: turns out the Steam release actually contains .rel files for a debug GC/Wii build for some reason!

Sadly these aren't that useful since we don't have the main executable to go with them, but it does give us debug st3_X.rel files to look at, which are pretty close to the st3_X.SYM symbols files that were included with the GC debug build (that for some reason didn't include the st3_X.rel files itself)

The SYM doesn't match up exactly, but comparing sizes of functions in the SYM against the REL does let us name some things, then it's pretty simple to port the names over to PC :) Will hopefully have an updated IDB to post here soon.

E: ahh, just realized that the st3_X files are on a second disc on GC, sadly doesn't seem that the released debug build included disc 2 with it, too bad. Maybe the GC retail second disc will match up closer to the .sym though, hm...

emoose commented 2 years ago

Updated with st3_X names (R3XXInit etc), bringing it up to ~68% named 😄

Download: pc-1.1.0_enum - Copy.7z.zip (remove .zip extension)

Seems some of the R3XX funcs call into some main.dol funcs that currently aren't named here, and since we don't have symbols for the main.dol for non-debug GC it'd be hard to work out what the name for those funcs are :/

Maybe PS2 version could give some info there, need to add support for imports to the PS2 rel loader, but not really sure how they work yet.

(would be great if debug disc 2 appears some day, would make things so much easier...)

nipkownix commented 2 years ago

Ah, just saw this! Nice work with those IDB files.

emoose commented 2 years ago

Named a bunch of Criware funcs (still a ton to go though >.>), fixed up the cSofdec class (also made a cSofdec struct with some fields inside, might be useful at some point), named some more memory/heap related stuff, and probably a bunch more stuff I forgot about too: pc-1.1.0_enum_latest.7z.zip (remove .zip extension)

emoose commented 2 years ago

Added some extra structs & names around the CameraBinocular/CameraScope/FocusAnimation etc classes, named some of the gamepad input fields (need to try looking into how M+KB is translated though), and some more Criware crap: pc-1.1.0_enum - Copy_latest.7z.zip

E: some more adds, started filling out PAD struct, added PAD_BTN enum, named some DInput8 GUIDs & related funcs: pc-1.1.0_enum_newest.7z.zip

emoose commented 2 years ago

Oh whoops, messed something up with the Joy / Key structs...

So there's (AFAIK) an array of 4 JOY structs (0x270 bytes each) at 0xC63008, some symbols call that array Joy, and game code seems to reach into Joy[0] to check for inputs (I guess maybe debug build might check the other Joy indexes)

However just before the Joy[4] array, there's also some more fields related to inputs at 0xC62EF0, RE4VR seems to call this area Bio4Key, and GC just calls it Key - I thought this was just another JOY struct, but it's actually too small for that, difference between start of Key & Joy[0] is only 0x118 bytes - in the IDBs above I just had it set as JOY, meaning the Joy[4] array ended up being swallowed up into the Key field...

Weird thing is that even with it swallowed up like this, the fields didn't seem that messed up - eg. accesses to Joy[0].buttonsOn_XXX just ended up being Key.pc_buttonsOn_XXX, since the stuff after 0x118 in the JOY struct seems to mostly just be a repeat (pc_ is just a guess, seemed like those fields are only set if using mouse/keyboard)

Anyway for now I've just made a separate KEY struct with only the fields in the first 0x118 of JOY, seems fine like this, but still not sure what's going on here... maybe Joy is actually meant to be part of the KEY struct (not sure why Joy would have been exposed in symbols if so though), or maybe JOY struct is actually something like

struct JOY {
  KEY controller; // 0x118 bytes
  KEY pc; // repeat, only set for lastDevice = 2?
  KEY wiipad; // maybe not full KEY struct, but seems to have some similarities
}

instead, not sure yet though.

Here's fixed IDB anyway: pc-1.1.0_enum - Copy_latest.7z.zip

E: will probably start a new IDB soon with IDA 7.7, maybe has better analysis than 7.6 did.

emoose commented 2 years ago

Posted IDA 7.7 db here: https://github.com/emoose/re4-research/issues/3