Allocate zend_strings with correct size

For me (32bit) sizeof(zend_string) is 20, which means that the
char[1] array at the end is padded with three bytes. Thus allocating
based on sizeof(zend_string)-1 overallocates by those 3 padding bytes.

This commit fixes the allocation size, by using XtOffsetOf.

Zend/zend.h

@@ -347,6 +347,36 @@ void zend_error_noreturn(int type, const char *format, ...) __attribute__ ((nore
 # define UNEXPECTED(condition) (condition)
 #endif
 
+#ifndef XtOffsetOf
+# if defined(CRAY) || (defined(__ARMCC_VERSION) && !defined(LINUX))
+# ifdef __STDC__
+# define XtOffset(p_type, field) _Offsetof(p_type, field)
+# else
+# ifdef CRAY2
+# define XtOffset(p_type, field) \
+    (sizeof(int)*((unsigned int)&(((p_type)NULL)->field)))
+
+# else /* !CRAY2 */
+
+# define XtOffset(p_type, field) ((unsigned int)&(((p_type)NULL)->field))
+
+# endif /* !CRAY2 */
+# endif /* __STDC__ */
+# else /* ! (CRAY || __arm) */
+
+# define XtOffset(p_type, field) \
+    ((long) (((char *) (&(((p_type)NULL)->field))) - ((char *) NULL)))
+
+# endif /* !CRAY */
+
+# ifdef offsetof
+# define XtOffsetOf(s_type, field) offsetof(s_type, field)
+# else
+# define XtOffsetOf(s_type, field) XtOffset(s_type*, field)
+# endif
+
+#endif
+
 #include "zend_string.h"
 
 static zend_always_inline zend_uint zval_refcount_p(zval* pz) {

Zend/zend_ini.h

@@ -27,36 +27,6 @@
 
 #define ZEND_INI_ALL (ZEND_INI_USER|ZEND_INI_PERDIR|ZEND_INI_SYSTEM)
 
-#ifndef XtOffsetOf
-# if defined(CRAY) || (defined(__ARMCC_VERSION) && !defined(LINUX))
-# ifdef __STDC__
-# define XtOffset(p_type, field) _Offsetof(p_type, field)
-# else
-# ifdef CRAY2
-# define XtOffset(p_type, field) \
-    (sizeof(int)*((unsigned int)&(((p_type)NULL)->field)))
-
-# else /* !CRAY2 */
-
-# define XtOffset(p_type, field) ((unsigned int)&(((p_type)NULL)->field))
-
-# endif /* !CRAY2 */
-# endif /* __STDC__ */
-# else /* ! (CRAY || __arm) */
-
-# define XtOffset(p_type, field) \
-    ((long) (((char *) (&(((p_type)NULL)->field))) - ((char *) NULL)))
-
-# endif /* !CRAY */
-
-# ifdef offsetof
-# define XtOffsetOf(s_type, field) offsetof(s_type, field)
-# else
-# define XtOffsetOf(s_type, field) XtOffset(s_type*, field)
-# endif
-
-#endif
-
 #define ZEND_INI_MH(name) int name(zend_ini_entry *entry, char *new_value, uint new_value_length, void *mh_arg1, void *mh_arg2, void *mh_arg3, int stage TSRMLS_DC)
 #define ZEND_INI_DISP(name) void name(zend_ini_entry *ini_entry, int type)
 

Zend/zend_string.h

@@ -54,6 +54,8 @@ END_EXTERN_C()
 #define STR_RELEASE(s)					zend_str_release(s)
 #define STR_EMPTY_ALLOC()				CG(empty_string)
 
+#define _STR_HEADER_SIZE XtOffsetOf(zend_string, val)
+
 static zend_always_inline zend_ulong zend_str_hash_val(zend_string *s)
 {
 	if (!s->h) {
@@ -93,7 +95,7 @@ static zend_always_inline zend_uint zend_str_delref(zend_string *s)
 
 static zend_always_inline zend_string *zend_str_alloc(int len, int persistent)
 {
-	zend_string *ret = pemalloc(sizeof(zend_string) + len, persistent);
+	zend_string *ret = pemalloc(_STR_HEADER_SIZE + len + 1, persistent);
 
 	GC_REFCOUNT(ret) = 1;
 #if 1
@@ -111,7 +113,7 @@ static zend_always_inline zend_string *zend_str_alloc(int len, int persistent)
 
 static zend_always_inline zend_string *zend_str_safe_alloc(size_t n, size_t m, size_t l, int persistent)
 {
-	zend_string *ret = safe_pemalloc(n, m, sizeof(zend_string) + l - 1, persistent);
+	zend_string *ret = safe_pemalloc(n, m, _STR_HEADER_SIZE + l + 1, persistent);
 
 	GC_REFCOUNT(ret) = 1;
 #if 1
@@ -161,7 +163,7 @@ static zend_always_inline zend_string *zend_str_realloc(zend_string *s, int len,
 		ret = STR_ALLOC(len, persistent);
 		memcpy(ret->val, s->val, (len > s->len ? s->len : len) + 1);
 	} else if (STR_REFCOUNT(s) == 1) {
-		ret = perealloc(s, sizeof(zend_string) + len, persistent);
+		ret = perealloc(s, _STR_HEADER_SIZE + len + 1, persistent);
 		ret->len = len;
 		STR_FORGET_HASH_VAL(ret);
 	} else {
@@ -180,7 +182,7 @@ static zend_always_inline zend_string *zend_str_safe_realloc(zend_string *s, siz
 		ret = STR_SAFE_ALLOC(n, m, l, persistent);
 		memcpy(ret->val, s->val, ((n * m) + l > s->len ? s->len : ((n * m) + l)) + 1);
 	} else if (STR_REFCOUNT(s) == 1) {
-		ret = safe_perealloc(s, n, m, sizeof(zend_string) + l - 1, persistent);
+		ret = safe_perealloc(s, n, m, _STR_HEADER_SIZE + l + 1, persistent);
 		ret->len = (n * m) + l;
 		STR_FORGET_HASH_VAL(ret);
 	} else {

ext/opcache/ZendAccelerator.c

@@ -336,7 +336,7 @@ zend_string *accel_new_interned_string(zend_string *str TSRMLS_DC)
 		idx = Z_NEXT(p->val);
 	}
 
-	if (ZCSG(interned_strings_top) + ZEND_MM_ALIGNED_SIZE(sizeof(zend_string) + str->len) >=
+	if (ZCSG(interned_strings_top) + ZEND_MM_ALIGNED_SIZE(_STR_HEADER_SIZE + str->len + 1) >=
 	    ZCSG(interned_strings_end)) {
 	    /* no memory, return the same non-interned string */
 		return str;
@@ -348,7 +348,7 @@ zend_string *accel_new_interned_string(zend_string *str TSRMLS_DC)
 	ZCSG(interned_strings).nNumOfElements++;
 	p = ZCSG(interned_strings).arData + idx;
 	p->key = (zend_string*) ZCSG(interned_strings_top);
-	ZCSG(interned_strings_top) += ZEND_MM_ALIGNED_SIZE(sizeof(zend_string) + str->len);
+	ZCSG(interned_strings_top) += ZEND_MM_ALIGNED_SIZE(_STR_HEADER_SIZE + str->len + 1);
 	p->h = h;
 	GC_REFCOUNT(p->key) = 1;
 #if 1

ext/opcache/zend_persist.c

@@ -40,13 +40,13 @@
 			STR_RELEASE(str); \
 			str = new_str; \
 		} else { \
-	    	new_str = _zend_shared_memdup((void*)str, sizeof(zend_string) + (str)->len, 0 TSRMLS_CC); \
+	    	new_str = _zend_shared_memdup((void*)str, _STR_HEADER_SIZE + (str)->len + 1, 0 TSRMLS_CC); \
 			STR_RELEASE(str); \
 	    	str = new_str; \
 		} \
     } while (0)
 # define zend_accel_memdup_string(str) \
-	zend_accel_memdup(str, sizeof(zend_string) + (str)->len)
+	zend_accel_memdup(str, _STR_HEADER_SIZE + (str)->len + 1)
 # define zend_accel_store_interned_string(str) do { \
 		if (!IS_ACCEL_INTERNED(str)) { \
 			zend_accel_store_string(str); \

ext/opcache/zend_persist_calc.c

@@ -33,7 +33,7 @@
 
 #if ZEND_EXTENSION_API_NO > PHP_5_3_X_API_NO
 # define ADD_STRING(str) \
-		ADD_DUP_SIZE((str), sizeof(zend_string) + (str)->len)
+		ADD_DUP_SIZE((str), _STR_HEADER_SIZE + (str)->len + 1)
 # define ADD_INTERNED_STRING(str, do_free) do { \
 		if (!IS_ACCEL_INTERNED(str)) { \
 			zend_string *tmp = accel_new_interned_string(str TSRMLS_CC); \

ext/standard/php_smart_str.h

@@ -45,11 +45,11 @@
 
 #ifdef SMART_STR_USE_REALLOC
 #define SMART_STR_DO_REALLOC(b, w) do {								\
-	(b)->s = erealloc((buf)->s, sizeof(zend_string) + (b)->a);		\
+	(b)->s = erealloc((buf)->s, _STR_HEADER_SIZE + (b)->a + 1);		\
 } while (0)
 #else
 #define SMART_STR_DO_REALLOC(b, w) do {								\
-	(b)->s = perealloc((b)->s, sizeof(zend_string) + (b)->a, (w));	\
+	(b)->s = perealloc((b)->s, _STR_HEADER_SIZE + (b)->a + 1, (w));	\
 } while (0)
 #endif
 

ext/standard/tests/strings/str_pad_variation5.phpt

@@ -24,7 +24,7 @@ echo "*** Testing str_pad() function: with large value for for 'pad_length' argu
 
 //defining '$input' argument
 $input = "Test string";
-$pad_length = PHP_INT_MAX - 20 + 1; /* sizeof(zend_string) is 20, 1 character is included in zend_string structure itself */
+$pad_length = PHP_INT_MAX - 16; /* zend_string header is 16 bytes */
 var_dump( str_pad($input, $pad_length) );
 
 ?>