In message <[EMAIL PROTECTED]> on Mon, 02 Dec 2002 10:51:30 +0100, Andy
Polyakov <[EMAIL PROTECTED]> said:
appro> > levitte> bn_div_words(0xC383,0x838B4B53,0x8000)
appro> >
appro> > Hmm, a call like that gave me an aruthmetic error on Linux...
appro>
appro> According to bc 0xC383838
> levitte> bn_div_words(0xC383,0x838B4B53,0x8000)
>
> Hmm, a call like that gave me an aruthmetic error on Linux...
According to bc 0xC383838B4B53 / 0x8000 = 0x18707. The
result is 33 bits or in other words the operation *overflows*. According
to IA-32 manual "overflow is indi
In message <[EMAIL PROTECTED]> on Mon, 02 Dec 2002
01:19:21 +0100 (CET), Richard Levitte - VMS Whacker <[EMAIL PROTECTED]> said:
levitte> bn_div_words(0xC383,0x838B4B53,0x8000)
Hmm, a call like that gave me an aruthmetic error on Linux...
--
Richard Levitte \ Spannvägen 38, II \ [EMA
I don't understand what BN_div() expects from bn_div_words(). I see
calls like this:
bn_div_words(0xC383,0x838B4B53,0x8000)
and I wonder what the hell is expected to come out of that. The
result of dividing 0xC383838B4B53 with 0x8000 is 0x18707,
which can't be correctly repr