You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Harmless PoC: ExecuTorch MethodMeta memory_planned_buffers vector length OOB

MethodMeta::num_memory_planned_buffers / memory_planned_buffer_size index ExecutionPlan.non_const_buffer_sizes[index+1] using a forged FlatBuffer vector length with no bound against the program buffer. Under default Program::Verification::Minimal, loading a malicious .pte and calling memory_planned_buffer_size(0) triggers ASan heap-buffer-overflow READ at method_meta.cpp:357.

Distinct from submitted MethodMeta Tensor.sizes, packaged MethodMeta I/O index (inputs/outputs), Program execution_plan / backend_delegate_data, and Method::init values/operators/delegates/list-items.

Files

poc_method_meta_memory_planned_asan.cpp, asan.log
method_load_stub.cpp (link stub for unused Method::load)
NOTES.md

./poc_method_meta_memory_planned_asan

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support