gadiluna / safe Goto Github PK

I would like to provide information about how to download the trained SAFE model for ARM architecture

Please, could you help to handle this error.

When I wanted to do a search evaluation, I run python EvaluateSearchEngine.py and found the following error reported.

File "EvaluateSearchEngine.py", line 54, in find_target_fcn
q = cur.execute("SELECT num FROM count_func WHERE file_name='{}' and function_name='{}'".format(fi_name,f_name))
sqlite3.OperationalError: no such table: count_func

Then I go into AMD64PostgreSQL.db and check it, I find that there is no count_func table in the database, but only the following tables

sqlite> .tables
filtered_functions safe_embeddings train             
functions test validation

I want to compare two functions from different architecture, e.g., x86, ARM. I do not know how to train the model for cross-architecture. Can you provide a trained model for cross-architecture?
Thank you.

Hi,
I need some help regarding running of your solution. It is successfully installed and running on Google CoLab but i could not understand how to check function semantic as you mentioned i your paper. Like function have encryption or math operation etc.

Can you guide please

hi, I want to know how to generate the word2id.json file? This is causing me significant confusion, I hope to revecive your answer. Thanks!

Hi,

We couldn't find any field corresponding to labels in the OpenSSL dataset (as mentioned by you in your publication). Can you please explain how to extract/compute them? We were looking at the AMD64MultipleCompilers.db dataset.

We are planning to perform supervised learning along with similarity measures (similar to what you did in the paper).

My colleague has mailed Luca Massarelli regarding the same.

Thanks in advance!

I compiled hellow word using gcc in my 64-bit host machine and used objdump utility to restore the address of main function that I would to embed.

I typed the following commands:

python3.7 safe.py -m data/safe_trained_X86.pb -i a.out -a 0000000000400526
python3.7 safe.py -m data/safe_trained_X86.pb -i a.out -a 400526

and set the base to 16, 32, 36 ... but I got the same error "function not found"