Update Python from 3.11.8 to 3.12.2

1 files changed, 56 insertions, 285 deletions

diff --git a/contrib/tools/python3/src/Modules/md5module.c b/contrib/tools/python3/src/Modules/md5module.c
index 48b11e0779..2122f8b18b 100644
--- a/contrib/tools/python3/src/Modules/md5module.c
+++ b/contrib/tools/python3/src/Modules/md5module.c
@@ -22,6 +22,7 @@
 #include "Python.h"
 #include "hashlib.h"
 #include "pycore_strhex.h"        // _Py_strhex()
+#include "pycore_typeobject.h"    // _PyType_GetModuleState()
 
 /*[clinic input]
 module _md5
@@ -43,283 +44,19 @@ typedef long long MD5_INT64; /* 64-bit integer */
 #define MD5_BLOCKSIZE    64
 #define MD5_DIGESTSIZE   16
 
-/* The structure for storing MD5 info */
+#include "_hacl/Hacl_Hash_MD5.h"
 
-struct md5_state {
-    MD5_INT64 length;
-    MD5_INT32 state[4], curlen;
-    unsigned char buf[MD5_BLOCKSIZE];
-};
 
 typedef struct {
     PyObject_HEAD
-
-    struct md5_state hash_state;
+    // Prevents undefined behavior via multiple threads entering the C API.
+    // The lock will be NULL before threaded access has been enabled.
+    PyThread_type_lock lock;
+    Hacl_Streaming_MD5_state *hash_state;
 } MD5object;
 
 #include "clinic/md5module.c.h"
 
-/* ------------------------------------------------------------------------
- *
- * This code for the MD5 algorithm was noted as public domain. The
- * original headers are pasted below.
- *
- * Several changes have been made to make it more compatible with the
- * Python environment and desired interface.
- *
- */
-
-/* LibTomCrypt, modular cryptographic library -- Tom St Denis
- *
- * LibTomCrypt is a library that provides various cryptographic
- * algorithms in a highly modular and flexible manner.
- *
- * The library is free for all purposes without any express
- * guarantee it works.
- *
- * Tom St Denis, tomstdenis@gmail.com, https://www.libtom.net
- */
-
-/* rotate the hard way (platform optimizations could be done) */
-#define ROLc(x, y) ( (((unsigned long)(x)<<(unsigned long)((y)&31)) | (((unsigned long)(x)&0xFFFFFFFFUL)>>(unsigned long)(32-((y)&31)))) & 0xFFFFFFFFUL)
-
-/* Endian Neutral macros that work on all platforms */
-
-#define STORE32L(x, y)                                                                     \
-     { (y)[3] = (unsigned char)(((x)>>24)&255); (y)[2] = (unsigned char)(((x)>>16)&255);   \
-       (y)[1] = (unsigned char)(((x)>>8)&255); (y)[0] = (unsigned char)((x)&255); }
-
-#define LOAD32L(x, y)                            \
-     { x = ((unsigned long)((y)[3] & 255)<<24) | \
-           ((unsigned long)((y)[2] & 255)<<16) | \
-           ((unsigned long)((y)[1] & 255)<<8)  | \
-           ((unsigned long)((y)[0] & 255)); }
-
-#define STORE64L(x, y)                                                                     \
-     { (y)[7] = (unsigned char)(((x)>>56)&255); (y)[6] = (unsigned char)(((x)>>48)&255);   \
-       (y)[5] = (unsigned char)(((x)>>40)&255); (y)[4] = (unsigned char)(((x)>>32)&255);   \
-       (y)[3] = (unsigned char)(((x)>>24)&255); (y)[2] = (unsigned char)(((x)>>16)&255);   \
-       (y)[1] = (unsigned char)(((x)>>8)&255); (y)[0] = (unsigned char)((x)&255); }
-
-
-/* MD5 macros */
-
-#define F(x,y,z)  (z ^ (x & (y ^ z)))
-#define G(x,y,z)  (y ^ (z & (y ^ x)))
-#define H(x,y,z)  (x^y^z)
-#define I(x,y,z)  (y^(x|(~z)))
-
-#define FF(a,b,c,d,M,s,t) \
-    a = (a + F(b,c,d) + M + t); a = ROLc(a, s) + b;
-
-#define GG(a,b,c,d,M,s,t) \
-    a = (a + G(b,c,d) + M + t); a = ROLc(a, s) + b;
-
-#define HH(a,b,c,d,M,s,t) \
-    a = (a + H(b,c,d) + M + t); a = ROLc(a, s) + b;
-
-#define II(a,b,c,d,M,s,t) \
-    a = (a + I(b,c,d) + M + t); a = ROLc(a, s) + b;
-
-
-static void md5_compress(struct md5_state *md5, const unsigned char *buf)
-{
-    MD5_INT32 i, W[16], a, b, c, d;
-
-    assert(md5 != NULL);
-    assert(buf != NULL);
-
-    /* copy the state into 512-bits into W[0..15] */
-    for (i = 0; i < 16; i++) {
-        LOAD32L(W[i], buf + (4*i));
-    }
-
-    /* copy state */
-    a = md5->state[0];
-    b = md5->state[1];
-    c = md5->state[2];
-    d = md5->state[3];
-
-    FF(a,b,c,d,W[0],7,0xd76aa478UL)
-    FF(d,a,b,c,W[1],12,0xe8c7b756UL)
-    FF(c,d,a,b,W[2],17,0x242070dbUL)
-    FF(b,c,d,a,W[3],22,0xc1bdceeeUL)
-    FF(a,b,c,d,W[4],7,0xf57c0fafUL)
-    FF(d,a,b,c,W[5],12,0x4787c62aUL)
-    FF(c,d,a,b,W[6],17,0xa8304613UL)
-    FF(b,c,d,a,W[7],22,0xfd469501UL)
-    FF(a,b,c,d,W[8],7,0x698098d8UL)
-    FF(d,a,b,c,W[9],12,0x8b44f7afUL)
-    FF(c,d,a,b,W[10],17,0xffff5bb1UL)
-    FF(b,c,d,a,W[11],22,0x895cd7beUL)
-    FF(a,b,c,d,W[12],7,0x6b901122UL)
-    FF(d,a,b,c,W[13],12,0xfd987193UL)
-    FF(c,d,a,b,W[14],17,0xa679438eUL)
-    FF(b,c,d,a,W[15],22,0x49b40821UL)
-    GG(a,b,c,d,W[1],5,0xf61e2562UL)
-    GG(d,a,b,c,W[6],9,0xc040b340UL)
-    GG(c,d,a,b,W[11],14,0x265e5a51UL)
-    GG(b,c,d,a,W[0],20,0xe9b6c7aaUL)
-    GG(a,b,c,d,W[5],5,0xd62f105dUL)
-    GG(d,a,b,c,W[10],9,0x02441453UL)
-    GG(c,d,a,b,W[15],14,0xd8a1e681UL)
-    GG(b,c,d,a,W[4],20,0xe7d3fbc8UL)
-    GG(a,b,c,d,W[9],5,0x21e1cde6UL)
-    GG(d,a,b,c,W[14],9,0xc33707d6UL)
-    GG(c,d,a,b,W[3],14,0xf4d50d87UL)
-    GG(b,c,d,a,W[8],20,0x455a14edUL)
-    GG(a,b,c,d,W[13],5,0xa9e3e905UL)
-    GG(d,a,b,c,W[2],9,0xfcefa3f8UL)
-    GG(c,d,a,b,W[7],14,0x676f02d9UL)
-    GG(b,c,d,a,W[12],20,0x8d2a4c8aUL)
-    HH(a,b,c,d,W[5],4,0xfffa3942UL)
-    HH(d,a,b,c,W[8],11,0x8771f681UL)
-    HH(c,d,a,b,W[11],16,0x6d9d6122UL)
-    HH(b,c,d,a,W[14],23,0xfde5380cUL)
-    HH(a,b,c,d,W[1],4,0xa4beea44UL)
-    HH(d,a,b,c,W[4],11,0x4bdecfa9UL)
-    HH(c,d,a,b,W[7],16,0xf6bb4b60UL)
-    HH(b,c,d,a,W[10],23,0xbebfbc70UL)
-    HH(a,b,c,d,W[13],4,0x289b7ec6UL)
-    HH(d,a,b,c,W[0],11,0xeaa127faUL)
-    HH(c,d,a,b,W[3],16,0xd4ef3085UL)
-    HH(b,c,d,a,W[6],23,0x04881d05UL)
-    HH(a,b,c,d,W[9],4,0xd9d4d039UL)
-    HH(d,a,b,c,W[12],11,0xe6db99e5UL)
-    HH(c,d,a,b,W[15],16,0x1fa27cf8UL)
-    HH(b,c,d,a,W[2],23,0xc4ac5665UL)
-    II(a,b,c,d,W[0],6,0xf4292244UL)
-    II(d,a,b,c,W[7],10,0x432aff97UL)
-    II(c,d,a,b,W[14],15,0xab9423a7UL)
-    II(b,c,d,a,W[5],21,0xfc93a039UL)
-    II(a,b,c,d,W[12],6,0x655b59c3UL)
-    II(d,a,b,c,W[3],10,0x8f0ccc92UL)
-    II(c,d,a,b,W[10],15,0xffeff47dUL)
-    II(b,c,d,a,W[1],21,0x85845dd1UL)
-    II(a,b,c,d,W[8],6,0x6fa87e4fUL)
-    II(d,a,b,c,W[15],10,0xfe2ce6e0UL)
-    II(c,d,a,b,W[6],15,0xa3014314UL)
-    II(b,c,d,a,W[13],21,0x4e0811a1UL)
-    II(a,b,c,d,W[4],6,0xf7537e82UL)
-    II(d,a,b,c,W[11],10,0xbd3af235UL)
-    II(c,d,a,b,W[2],15,0x2ad7d2bbUL)
-    II(b,c,d,a,W[9],21,0xeb86d391UL)
-
-    md5->state[0] = md5->state[0] + a;
-    md5->state[1] = md5->state[1] + b;
-    md5->state[2] = md5->state[2] + c;
-    md5->state[3] = md5->state[3] + d;
-}
-
-
-/**
-   Initialize the hash state
-   @param md5   The hash state you wish to initialize
-*/
-static void
-md5_init(struct md5_state *md5)
-{
-    assert(md5 != NULL);
-    md5->state[0] = 0x67452301UL;
-    md5->state[1] = 0xefcdab89UL;
-    md5->state[2] = 0x98badcfeUL;
-    md5->state[3] = 0x10325476UL;
-    md5->curlen = 0;
-    md5->length = 0;
-}
-
-/**
-   Process a block of memory though the hash
-   @param md5   The hash state
-   @param in     The data to hash
-   @param inlen  The length of the data (octets)
-*/
-static void
-md5_process(struct md5_state *md5, const unsigned char *in, Py_ssize_t inlen)
-{
-    Py_ssize_t n;
-
-    assert(md5 != NULL);
-    assert(in != NULL);
-    assert(md5->curlen <= sizeof(md5->buf));
-
-    while (inlen > 0) {
-        if (md5->curlen == 0 && inlen >= MD5_BLOCKSIZE) {
-           md5_compress(md5, in);
-           md5->length    += MD5_BLOCKSIZE * 8;
-           in             += MD5_BLOCKSIZE;
-           inlen          -= MD5_BLOCKSIZE;
-        } else {
-           n = Py_MIN(inlen, (Py_ssize_t)(MD5_BLOCKSIZE - md5->curlen));
-           memcpy(md5->buf + md5->curlen, in, (size_t)n);
-           md5->curlen    += (MD5_INT32)n;
-           in             += n;
-           inlen          -= n;
-           if (md5->curlen == MD5_BLOCKSIZE) {
-              md5_compress(md5, md5->buf);
-              md5->length += 8*MD5_BLOCKSIZE;
-              md5->curlen = 0;
-           }
-       }
-    }
-}
-
-/**
-   Terminate the hash to get the digest
-   @param md5  The hash state
-   @param out [out] The destination of the hash (16 bytes)
-*/
-static void
-md5_done(struct md5_state *md5, unsigned char *out)
-{
-    int i;
-
-    assert(md5 != NULL);
-    assert(out != NULL);
-    assert(md5->curlen < sizeof(md5->buf));
-
-    /* increase the length of the message */
-    md5->length += md5->curlen * 8;
-
-    /* append the '1' bit */
-    md5->buf[md5->curlen++] = (unsigned char)0x80;
-
-    /* if the length is currently above 56 bytes we append zeros
-     * then compress.  Then we can fall back to padding zeros and length
-     * encoding like normal.
-     */
-    if (md5->curlen > 56) {
-        while (md5->curlen < 64) {
-            md5->buf[md5->curlen++] = (unsigned char)0;
-        }
-        md5_compress(md5, md5->buf);
-        md5->curlen = 0;
-    }
-
-    /* pad up to 56 bytes of zeroes */
-    while (md5->curlen < 56) {
-        md5->buf[md5->curlen++] = (unsigned char)0;
-    }
-
-    /* store length */
-    STORE64L(md5->length, md5->buf+56);
-    md5_compress(md5, md5->buf);
-
-    /* copy output */
-    for (i = 0; i < 4; i++) {
-        STORE32L(md5->state[i], out+(4*i));
-    }
-}
-
-/* .Source: /cvs/libtom/libtomcrypt/src/hashes/md5.c,v $ */
-/* .Revision: 1.10 $ */
-/* .Date: 2007/05/12 14:25:28 $ */
-
-/*
- * End of copied MD5 code.
- *
- * ------------------------------------------------------------------------
- */
 
 typedef struct {
     PyTypeObject* md5_type;
@@ -337,6 +74,7 @@ static MD5object *
 newMD5object(MD5State * st)
 {
     MD5object *md5 = (MD5object *)PyObject_GC_New(MD5object, st->md5_type);
+    md5->lock = NULL;
     PyObject_GC_Track(md5);
     return md5;
 }
@@ -350,8 +88,12 @@ MD5_traverse(PyObject *ptr, visitproc visit, void *arg)
 }
 
 static void
-MD5_dealloc(PyObject *ptr)
+MD5_dealloc(MD5object *ptr)
 {
+    Hacl_Streaming_MD5_legacy_free(ptr->hash_state);
+    if (ptr->lock != NULL) {
+        PyThread_free_lock(ptr->lock);
+    }
     PyTypeObject *tp = Py_TYPE(ptr);
     PyObject_GC_UnTrack(ptr);
     PyObject_GC_Del(ptr);
@@ -373,13 +115,15 @@ static PyObject *
 MD5Type_copy_impl(MD5object *self, PyTypeObject *cls)
 /*[clinic end generated code: output=bf055e08244bf5ee input=d89087dcfb2a8620]*/
 {
-    MD5State *st = PyType_GetModuleState(cls);
+    MD5State *st = _PyType_GetModuleState(cls);
 
     MD5object *newobj;
     if ((newobj = newMD5object(st))==NULL)
         return NULL;
 
-    newobj->hash_state = self->hash_state;
+    ENTER_HASHLIB(self);
+    newobj->hash_state = Hacl_Streaming_MD5_legacy_copy(self->hash_state);
+    LEAVE_HASHLIB(self);
     return (PyObject *)newobj;
 }
 
@@ -394,10 +138,9 @@ MD5Type_digest_impl(MD5object *self)
 /*[clinic end generated code: output=eb691dc4190a07ec input=bc0c4397c2994be6]*/
 {
     unsigned char digest[MD5_DIGESTSIZE];
-    struct md5_state temp;
-
-    temp = self->hash_state;
-    md5_done(&temp, digest);
+    ENTER_HASHLIB(self);
+    Hacl_Streaming_MD5_legacy_finish(self->hash_state, digest);
+    LEAVE_HASHLIB(self);
     return PyBytes_FromStringAndSize((const char *)digest, MD5_DIGESTSIZE);
 }
 
@@ -412,15 +155,23 @@ MD5Type_hexdigest_impl(MD5object *self)
 /*[clinic end generated code: output=17badced1f3ac932 input=b60b19de644798dd]*/
 {
     unsigned char digest[MD5_DIGESTSIZE];
-    struct md5_state temp;
-
-    /* Get the raw (binary) digest value */
-    temp = self->hash_state;
-    md5_done(&temp, digest);
-
+    ENTER_HASHLIB(self);
+    Hacl_Streaming_MD5_legacy_finish(self->hash_state, digest);
+    LEAVE_HASHLIB(self);
     return _Py_strhex((const char*)digest, MD5_DIGESTSIZE);
 }
 
+static void update(Hacl_Streaming_MD5_state *state, uint8_t *buf, Py_ssize_t len) {
+#if PY_SSIZE_T_MAX > UINT32_MAX
+  while (len > UINT32_MAX) {
+    Hacl_Streaming_MD5_legacy_update(state, buf, UINT32_MAX);
+    len -= UINT32_MAX;
+    buf += UINT32_MAX;
+  }
+#endif
+  Hacl_Streaming_MD5_legacy_update(state, buf, (uint32_t) len);
+}
+
 /*[clinic input]
 MD5Type.update
 
@@ -438,7 +189,18 @@ MD5Type_update(MD5object *self, PyObject *obj)
 
     GET_BUFFER_VIEW_OR_ERROUT(obj, &buf);
 
-    md5_process(&self->hash_state, buf.buf, buf.len);
+    if (self->lock == NULL && buf.len >= HASHLIB_GIL_MINSIZE) {
+        self->lock = PyThread_allocate_lock();
+    }
+    if (self->lock != NULL) {
+        Py_BEGIN_ALLOW_THREADS
+        PyThread_acquire_lock(self->lock, 1);
+        update(self->hash_state, buf.buf, buf.len);
+        PyThread_release_lock(self->lock);
+        Py_END_ALLOW_THREADS
+    } else {
+        update(self->hash_state, buf.buf, buf.len);
+    }
 
     PyBuffer_Release(&buf);
     Py_RETURN_NONE;
@@ -531,7 +293,7 @@ _md5_md5_impl(PyObject *module, PyObject *string, int usedforsecurity)
         return NULL;
     }
 
-    md5_init(&new->hash_state);
+    new->hash_state = Hacl_Streaming_MD5_legacy_create_in();
 
     if (PyErr_Occurred()) {
         Py_DECREF(new);
@@ -540,7 +302,15 @@ _md5_md5_impl(PyObject *module, PyObject *string, int usedforsecurity)
         return NULL;
     }
     if (string) {
-        md5_process(&new->hash_state, buf.buf, buf.len);
+        if (buf.len >= HASHLIB_GIL_MINSIZE) {
+            /* We do not initialize self->lock here as this is the constructor
+             * where it is not yet possible to have concurrent access. */
+            Py_BEGIN_ALLOW_THREADS
+            update(new->hash_state, buf.buf, buf.len);
+            Py_END_ALLOW_THREADS
+        } else {
+            update(new->hash_state, buf.buf, buf.len);
+        }
         PyBuffer_Release(&buf);
     }
 
@@ -601,6 +371,7 @@ md5_exec(PyObject *m)
 
 static PyModuleDef_Slot _md5_slots[] = {
     {Py_mod_exec, md5_exec},
+    {Py_mod_multiple_interpreters, Py_MOD_PER_INTERPRETER_GIL_SUPPORTED},
     {0, NULL}
 };