Fix incorrect assumptions about the size of size_t

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
This commit is contained in:
Paul Elliott 2021-03-01 17:49:42 +00:00
parent 3e7908189a
commit 717ba77e52

View File

@ -82,7 +82,7 @@ static void mbedtls_base64_cond_assign(unsigned char * dest, const unsigned char
*/
static unsigned char mbedtls_base64_eq(size_t in_a, size_t in_b)
{
uint32_t difference = in_a ^ in_b;
size_t difference = in_a ^ in_b;
/* MSVC has a warning about unary minus on unsigned integer types,
* but this is well-defined and precisely what we want to do here. */
@ -97,7 +97,9 @@ static unsigned char mbedtls_base64_eq(size_t in_a, size_t in_b)
#pragma warning( pop )
#endif
difference >>= 31;
/* cope with the varying size of size_t per platform */
difference >>= ( sizeof( difference ) * 8 - 1 );
return (unsigned char) ( 1 ^ difference );
}