A first look at Ghidra’s Debugger – Game Boy Advance Edition

[mr_golyadkin]

https://wrongbaud.github.io/posts/ghidra-debugger/

[oritsnile]

Look pretty exciting, debugging is the feature I'm always missing when working with Ghidra.

Will be intersting if ida also adds the debugger to there free version.

[JohnCurran]

When I try and connect to gdb (In-VM GNU gdb local debugger, /usr/local/bin/gdb) on MacOS (Big Sur, 11.0.1) I get the following exception:

java.util.concurrent.CompletionException: java.io.IOException: Could not detect GDB's interpreter mode --------------------------------------------------- Build Date: 2020-Dec-27 1528 CST Ghidra Version: 9.3 Java Home: /Library/Java/JavaVirtualMachines/jdk-11.0.2.jdk/Contents/Home JVM Version: Oracle Corporation 11.0.2 OS: Mac OS X 10.16 x86_64

[Shared404]

Off topic but I have to ask. Does anyone else wonder if there are backdoors in NSA projects like this? It seems like a way you could get a foothold into lots of interesting places.

Edit: Spelling.

Also, I recognize this particular program may not be the best target, but SELinux for example, or any of their projects. As for the "It's open source argument", if they provide binaries, do they have instructions for reproducible builds?

[cjbprime → Shared404]

An open source project used by people capable of analyzing the network and disk I/O of binaries they run doesn't sound like a good target -- if something was found it would have massive consequences.

[dboreham → cjbprime]

Backdoors from TLAs don't look like backdoors.

[cjbprime → dboreham]

People who use Ghidra know that backdoors from TLAs don't look like backdoors.

[_mlxl → cjbprime]

I wouldn't say this is true at all, there are lots of kids out there who will cut their teeth with this seeing as it is entirely free.

Also, it is as if people forget the launch of Ghidra; https://github.com/NationalSecurityAgency/ghidra/issues/6

[cjbprime → _mlxl]

Hiding a backdoor requires hiding it from every user who might publicize it, not some users.

Your issue (which is barely a vulnerability at all, in my personal opinion) demonstrates a failure to hide a vulnerability, if you think it was intentionally-placed (which I don't), in that you quickly learned about it.

[fuyu → Shared404]

I can't imaging a project aimed at reverse engineers would be the best place to try to hide that sort of thing. Ghidra is open source, you are free to audit and build it yourself if you are concerned.

[userbinator → fuyu]

Not to take away from your point, but a lot of "career" reverse-engineers are no more knowledgeable about the details of their tools or the low-level than developers.

[FentanylFloyd → fuyu]

> Ghidra is open source, you are free to audit and build it yourself if you are concerned.

^ the manyeyeballs fallacy

[travmatt → FentanylFloyd]

I don’t think the many eyes fallacy states that code audits are impossible or useless.

[FentanylFloyd → travmatt]

but it states that nobody (as in nearly not enough people) does it

[Pompidou → Shared404]

There was one possible backdoor in the past :

https://latesthackingnews.com/2019/03/24/critical-vulnerabil...