On 02/08/15 12:14, Jeffrey Walton wrote: > According to -fno-strict-aliasing in the online manual > (https://gcc.gnu.org/onlinedocs/gcc-4.4.2/gcc/Optimize-Options.html): > > Allow the compiler to assume the strictest aliasing rules ... > ... unless the types are almost the same. > > I think "almost the same" applies to sizes and alignments, and not > qualifiers like const and volatile. > > Does GCC consider an 'int' and a 'volatile int' (or 'int*' and a > 'volatile int*') almost the same? It's not symmetrical. The rules are the rules of the C standard. I don't think there's any intention to use GCC-specific rules. So: << ISO/IEC 9899:1999 (E) 6.2.7 Compatible type and composite type 1 Two types have compatible type if their types are the same. Additional rules for determining whether two types are compatible are described in 6.7.2 for type specifiers, in 6.7.3 for type qualifiers, and in 6.7.5 for declarators. ... For two qualified types to be compatible, both shall have the identically qualified version of a compatible type; 6.7.3 Type qualifiers Syntax 1 type-qualifier: const restrict volatile Semantics 3 4 If an attempt is made to refer to an object defined with a volatile-qualified type through use of an lvalue with non-volatile-qualified type, the behavior is undefined. >> It's OK to cast a pointer to int to a pointer to volatile int and then dereference that pointer, but not vice versa. So, GCC must assume that a pointer to volatile int may point to a non-volatile int object. But the reverse is not true: GCC need not assume that a pointer to int may point to a volatile int. Andrew.
