Visitar URL original
gh-96735: Fix undefined behaviour in struct unpacking functions by mdickinson · Pull Request #96739 · python/cpython · GitHub
Skip to content

gh-96735: Fix undefined behaviour in struct unpacking functions #96739

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mdickinson merged 10 commits into from
Sep 25, 2022
Merged

gh-96735: Fix undefined behaviour in struct unpacking functions #96739

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
22fdce6
Fix undefined behaviour in bu_ulonglong
mdickinson Sep 10, 2022
e12943b
Add regression test
mdickinson Sep 10, 2022
8efefd8
Also apply the fix to lu_longlong
mdickinson Sep 10, 2022
cba4b8b
Merge branch 'main' into fix-struct-unpack-undefined-behaviour
mdickinson Sep 10, 2022
0d49e73
Fix similar undefined behaviour issues in bu_int and lu_int
mdickinson Sep 10, 2022
bfd714f
Merge remote-tracking branch 'mdickinson/fix-struct-unpack-undefined-…
mdickinson Sep 10, 2022
5aeef09
Remove redundant test
mdickinson Sep 10, 2022
204e924
Fix accumulator type
mdickinson Sep 10, 2022
994703e
Branch-free sign-extension
mdickinson Sep 13, 2022
1294440
Fix unconventional spacing
mdickinson Sep 13, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions Misc/NEWS.d/next/Library/2022-09-10-16-46-16.gh-issue-96735.0YzJuG.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
Fix undefined behaviour in :func:`struct.unpack`.
96 changes: 70 additions & 26 deletions Modules/_struct.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -805,19 +805,38 @@ static const formatdef native_table[] = {

/* Big-endian routines. *****************************************************/

static PyObject *
bu_short(_structmodulestate *state, const char *p, const formatdef *f)
{
unsigned long x = 0;

/* This function is only ever used in the case f->size == 2. */
assert(f->size == 2);
Py_ssize_t i = 2;
const unsigned char *bytes = (const unsigned char *)p;
do {
x = (x<<8) | *bytes++;
} while (--i > 0);
/* Extend sign, avoiding implementation-defined or undefined behaviour. */
x = (x ^ 0x8000U) - 0x8000U;
return PyLong_FromLong(x & 0x8000U ? -1 - (long)(~x) : (long)x);
}

static PyObject *
bu_int(_structmodulestate *state, const char *p, const formatdef *f)
{
long x = 0;
Py_ssize_t i = f->size;
unsigned long x = 0;

/* This function is only ever used in the case f->size == 4. */
assert(f->size == 4);
Py_ssize_t i = 4;
const unsigned char *bytes = (const unsigned char *)p;
do {
x = (x<<8) | *bytes++;
} while (--i > 0);
/* Extend the sign bit. */
if (SIZEOF_LONG > f->size)
x |= -(x & (1L << ((8 * f->size) - 1)));
return PyLong_FromLong(x);
/* Extend sign, avoiding implementation-defined or undefined behaviour. */
x = (x ^ 0x80000000U) - 0x80000000U;
return PyLong_FromLong(x & 0x80000000U ? -1 - (long)(~x) : (long)x);
}

static PyObject *
Expand All @@ -835,16 +854,19 @@ bu_uint(_structmodulestate *state, const char *p, const formatdef *f)
static PyObject *
bu_longlong(_structmodulestate *state, const char *p, const formatdef *f)
{
long long x = 0;
Py_ssize_t i = f->size;
unsigned long long x = 0;

/* This function is only ever used in the case f->size == 8. */
assert(f->size == 8);
Py_ssize_t i = 8;
const unsigned char *bytes = (const unsigned char *)p;
do {
x = (x<<8) | *bytes++;
} while (--i > 0);
/* Extend the sign bit. */
if (SIZEOF_LONG_LONG > f->size)
x |= -(x & ((long long)1 << ((8 * f->size) - 1)));
return PyLong_FromLongLong(x);
/* Extend sign, avoiding implementation-defined or undefined behaviour. */
x = (x ^ 0x8000000000000000U) - 0x8000000000000000U;
return PyLong_FromLongLong(
x & 0x8000000000000000U ? -1 - (long long)(~x) : (long long)x);
}

static PyObject *
Expand Down Expand Up @@ -1009,7 +1031,7 @@ static formatdef bigendian_table[] = {
{'c', 1, 0, nu_char, np_char},
{'s', 1, 0, NULL},
{'p', 1, 0, NULL},
{'h', 2, 0, bu_int, bp_int},
{'h', 2, 0, bu_short, bp_int},
{'H', 2, 0, bu_uint, bp_uint},
{'i', 4, 0, bu_int, bp_int},
{'I', 4, 0, bu_uint, bp_uint},
Expand All @@ -1026,19 +1048,38 @@ static formatdef bigendian_table[] = {

/* Little-endian routines. *****************************************************/

static PyObject *
lu_short(_structmodulestate *state, const char *p, const formatdef *f)
{
unsigned long x = 0;

/* This function is only ever used in the case f->size == 2. */
assert(f->size == 2);
Py_ssize_t i = 2;
const unsigned char *bytes = (const unsigned char *)p;
do {
x = (x<<8) | bytes[--i];
} while (i > 0);
/* Extend sign, avoiding implementation-defined or undefined behaviour. */
x = (x ^ 0x8000U) - 0x8000U;
return PyLong_FromLong(x & 0x8000U ? -1 - (long)(~x) : (long)x);
}

static PyObject *
lu_int(_structmodulestate *state, const char *p, const formatdef *f)
{
long x = 0;
Py_ssize_t i = f->size;
unsigned long x = 0;

/* This function is only ever used in the case f->size == 4. */
assert(f->size == 4);
Py_ssize_t i = 4;
const unsigned char *bytes = (const unsigned char *)p;
do {
x = (x<<8) | bytes[--i];
} while (i > 0);
/* Extend the sign bit. */
if (SIZEOF_LONG > f->size)
x |= -(x & (1L << ((8 * f->size) - 1)));
return PyLong_FromLong(x);
/* Extend sign, avoiding implementation-defined or undefined behaviour. */
x = (x ^ 0x80000000U) - 0x80000000U;
return PyLong_FromLong(x & 0x80000000U ? -1 - (long)(~x) : (long)x);
}

static PyObject *
Expand All @@ -1056,16 +1097,19 @@ lu_uint(_structmodulestate *state, const char *p, const formatdef *f)
static PyObject *
lu_longlong(_structmodulestate *state, const char *p, const formatdef *f)
{
long long x = 0;
Py_ssize_t i = f->size;
unsigned long long x = 0;

/* This function is only ever used in the case f->size == 8. */
assert(f->size == 8);
Py_ssize_t i = 8;
const unsigned char *bytes = (const unsigned char *)p;
do {
x = (x<<8) | bytes[--i];
} while (i > 0);
/* Extend the sign bit. */
if (SIZEOF_LONG_LONG > f->size)
x |= -(x & ((long long)1 << ((8 * f->size) - 1)));
return PyLong_FromLongLong(x);
/* Extend sign, avoiding implementation-defined or undefined behaviour. */
x = (x ^ 0x8000000000000000U) - 0x8000000000000000U;
return PyLong_FromLongLong(
x & 0x8000000000000000U ? -1 - (long long)(~x) : (long long)x);
}

static PyObject *
Expand Down Expand Up @@ -1213,7 +1257,7 @@ static formatdef lilendian_table[] = {
{'c', 1, 0, nu_char, np_char},
{'s', 1, 0, NULL},
{'p', 1, 0, NULL},
{'h', 2, 0, lu_int, lp_int},
{'h', 2, 0, lu_short, lp_int},
{'H', 2, 0, lu_uint, lp_uint},
{'i', 4, 0, lu_int, lp_int},
{'I', 4, 0, lu_uint, lp_uint},
Expand Down