> Moore's Law caught up with us; when doing speed tests on new platforms
> (for AES-NI) we found that the calculation of the number of operations
> to do was overflowing a 31-bit 'long'.
> 
> By switching to 'unsigned long' and re-ordering the calculations a bit,
> we can postpone that overflow for a year or two. We may eventually need
> to use a 64-bit datatype, or impose an upper limit of 4 milliard loops.

Or two milliard, in which case one doesn't even have to switch to 
unsigned. Please test http://cvs.openssl.org/chngview?cn=19095. I opted 
for limiting loops at 2^31, because extending to 64-bit is problematic, 
as it implies different types on different platforms [not to mention 
not-really-portable conversion specifiers for 64-bit types]. I didn't 
opt for 2^32, because 2^31 provides as adequate accuracy in lesser 
changes. A.


______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [email protected]
Automated List Manager                           [email protected]

Reply via email to