Analysis Task: Shared Return Calls - Index 0 out of bounds for lenght 0 about ghidra-switch-loader

[TSRBerry]

I don't see any reference to SwitchLoader in the stacktrace. Next time please, copy and paste them into your issue in a code block, that makes it a lot easier to work with them.

It definitely looks odd, could you provide more information? What binary are you analyzing? Which Switch-Loader version are you using? Was this binary analyzed using an older Ghidra version before?

[StevensND]

I don't see any reference to SwitchLoader in the stacktrace. Next time please, copy and paste them into your issue in a code block, that makes it a lot easier to work with them.

It definitely looks odd, could you provide more information? What binary are you analyzing? Which Switch-Loader version are you using? Was this binary analyzed using an older Ghidra version before?

Right now I'm analyzing another game. I will copy the errors into a .txt and upload it later.

I'm analyzing a main file extracted from the game. Switch-Loader version is the one compatible with Ghidra 11.0 (I just merged the changes you did and compile it by myself).

When I used Ghidra 10.4 I didn't have these issues. So could be an issue with Ghidra 11.0?. Could you give me your Switch-Loader 11.0 version compiled to test?. Thanks.

[StevensND]

Here's the pastebin with the errors that I'm getting.

[TSRBerry]

What game are you analyzing? Could you use hactool to get the SDK version of the the nca file containing the main nso that you are analyzing?

[TSRBerry]

Nevermind, I'm able to reproduce this when I try to analyze the main nso of "Super Mario Bros. Wonder".
I'll try to figure it out, but I'm currently pretty busy, so it might take some time.

[TSRBerry]

This is a Ghidra issue and was fixed just 2 weeks ago: https://github.com/NationalSecurityAgency/ghidra/blob/85d276bce0bb2055c5556825afc810d3163ed15f/Ghidra/Framework/SoftwareModeling/src/main/java/ghidra/program/model/lang/protorules/HomogeneousAggregate.java#L71-L74

The current release does not check if the result is empty and tries to access index 0 unconditionally:

if (!DatatypeFilter.extractPrimitives(dt, MAX_PRIMITIVES, null, res)) {
    return false;
}
DataType base = res.get(0);

I assume the next release will contain this fix.

[StevensND]

What game are you analyzing? Could you use hactool to get the SDK version of the the nca file containing the main nso that you are analyzing?

It happens with all games that I analyzed using Ghidra 11.

So it seems to be a Ghidra 11 issue then?.

Then I guess I have to stay at 10.4 for now.