1995-12-28 - Re: fun with rsaref and 64-bit machines

Header Data

From: Jeff Weinstein <jsw@netscape.com>
To: sameer <sameer@c2.org>
Message Hash: 668c358536468f3c462aa238a88e8e53608d990a90113160a846539732caa5a2
Message ID: <30E12C02.54F7@netscape.com>
Reply To: <199512270120.RAA19559@infinity.c2.org>
UTC Datetime: 1995-12-28 04:32:14 UTC
Raw Date: Thu, 28 Dec 1995 12:32:14 +0800

Raw message

From: Jeff Weinstein <jsw@netscape.com>
Date: Thu, 28 Dec 1995 12:32:14 +0800
To: sameer <sameer@c2.org>
Subject: Re: fun with rsaref and 64-bit machines
In-Reply-To: <199512270120.RAA19559@infinity.c2.org>
Message-ID: <30E12C02.54F7@netscape.com>
MIME-Version: 1.0
Content-Type: text/plain


sameer wrote:
> 
>         Does anyone know if RSAREF works on 64 bit CPUs? It seems
> that RSAREF's NN_Mod is getting stuck in digit.c at:
> 
> 99         */
> 100       while ((t[1] > 0) || ((t[1] == 0) && t[0] >= c)) {
> 101         if ((t[0] -= c) > (MAX_NN_DIGIT - c))
> 102           t[1]--;
> 103         aLow++;
> 104       }

  In global.h, UINT4 is typedef'd to be unsigned long int, which is a 64 bit
value on the DEC alpha system.  The type UINT4 is supposed to be a 32-bit
value.  On the DEC system you need to make that be an unsigned int, which is
32-bits.  I've complained to RSA before about this problem, as it bit me
when I was porting SSLREF to the DEC machine.  They mumbled something about
people using the pre-compiled libraries, and ignored me.  It would be
so easy for them to just ifdef this one line...

	--Jeff

-- 
Jeff Weinstein - Electronic Munitions Specialist
Netscape Communication Corporation
jsw@netscape.com - http://home.netscape.com/people/jsw
Any opinions expressed above are mine.





Thread