Skip to content

bytes.fromhex() reads uninitialized memory #158583

Description

@vstinner

Example:

import array
res = bytes.fromhex(array.array('B', b' 1A 2B 30 '))
try:
    bytes.fromhex(array.array('B', b' 1A 2B 3'))
except ValueError:
    pass

Valgrind reports two "Use of uninitialised value of size 8" errors:

$ PYTHONMALLOC=malloc valgrind ./python x.py 
==863634== Memcheck, a memory error detector
==863634== Copyright (C) 2002-2026, and GNU GPL'd, by Julian Seward et al.
==863634== Using Valgrind-3.27.1 and LibVEX; rerun with -h for copyright info
==863634== Command: ./python x.py
==863634== 
==863634== Use of uninitialised value of size 8
==863634==    at 0x497C07: _PyBytes_FromHex (bytesobject.c:2688)
==863634==    by 0x497CC2: bytes_fromhex_impl (bytesobject.c:2618)
==863634==    by 0x497D07: bytes_fromhex (bytesobject.c.h:1261)
==863634==    by 0x4F491A: cfunction_vectorcall_O (methodobject.c:535)
==863634==    by 0x49BDAF: _PyObject_VectorcallTstate (pycore_call.h:149)
==863634==    by 0x49BE77: PyObject_Vectorcall (call.c:327)
==863634==    by 0x5B0C90: _Py_VectorCallInstrumentation_StackRefSteal (ceval.c:770)
==863634==    by 0x5B7E1C: _PyEval_EvalFrameDefault (generated_cases.c.h:1906)
==863634==    by 0x5CFACD: _PyEval_EvalFrame (pycore_ceval.h:122)
==863634==    by 0x5CFC98: _PyEval_Vector (ceval.c:2176)
==863634==    by 0x5CFD7C: PyEval_EvalCode (ceval.c:681)
==863634==    by 0x649C6B: run_eval_code_obj (pythonrun.c:1406)
==863634== 
==863634== Use of uninitialised value of size 8
==863634==    at 0x497BC1: _PyBytes_FromHex (bytesobject.c:2700)
==863634==    by 0x497CC2: bytes_fromhex_impl (bytesobject.c:2618)
==863634==    by 0x497D07: bytes_fromhex (bytesobject.c.h:1261)
==863634==    by 0x4F491A: cfunction_vectorcall_O (methodobject.c:535)
==863634==    by 0x49BDAF: _PyObject_VectorcallTstate (pycore_call.h:149)
==863634==    by 0x49BE77: PyObject_Vectorcall (call.c:327)
==863634==    by 0x5B0C90: _Py_VectorCallInstrumentation_StackRefSteal (ceval.c:770)
==863634==    by 0x5B7E1C: _PyEval_EvalFrameDefault (generated_cases.c.h:1906)
==863634==    by 0x5CFACD: _PyEval_EvalFrame (pycore_ceval.h:122)
==863634==    by 0x5CFC98: _PyEval_Vector (ceval.c:2176)
==863634==    by 0x5CFD7C: PyEval_EvalCode (ceval.c:681)
==863634==    by 0x649C6B: run_eval_code_obj (pythonrun.c:1406)
==863634== 
==863634== 
==863634== HEAP SUMMARY:
==863634==     in use at exit: 1,618 bytes in 4 blocks
==863634==   total heap usage: 36,249 allocs, 36,245 frees, 9,219,652 bytes allocated
==863634== 
==863634== LEAK SUMMARY:
==863634==    definitely lost: 0 bytes in 0 blocks
==863634==    indirectly lost: 0 bytes in 0 blocks
==863634==      possibly lost: 0 bytes in 0 blocks
==863634==    still reachable: 1,618 bytes in 4 blocks
==863634==         suppressed: 0 bytes in 0 blocks
==863634== Rerun with --leak-check=full to see details of leaked memory
==863634== 
==863634== Use --track-origins=yes to see where uninitialised values come from
==863634== For lists of detected and suppressed errors, rerun with: -s
==863634== ERROR SUMMARY: 2 errors from 2 contexts (suppressed: 0 from 0)

The problem is in _PyBytes_FromHex(): str++ doesn't check if we reached end.

For example, for a buffer of 10 bytes, it can read the 11th: outside the buffer!

Linked PRs

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    interpreter-core(Objects, Python, Grammar, and Parser dirs)type-bugAn unexpected behavior, bug, or error

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions