From: swarnerx3 AT acadia DOT net (Scott Warner)
Subject: Re: WARNING: Serious Pentium Bug
10 Nov 1997 22:32:04 -0800
Message-ID: <199711102337.SAA11324.cygnus.gnu-win32@p2.acadia.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
To: "Bryon Roche" <broche AT bensonlaw DOT com>, <gnu-win32 AT cygnus DOT com>

So, the Pentium *should* recognize the instruction as illegal and crash the
program, but instead the program crashes the OS?  However did this get
discovered, and are there more instances?  It occurs to me that no
processor is safe from these kinds of things.

----------
> From: Bryon Roche <broche AT bensonlaw DOT com>
> To: Scott Warner <swarnerx3 AT acadia DOT net>
> Subject: Re: WARNING: Serious Pentium Bug
> Date: Monday, November 10, 1997 4:54 PM
> 
> The contents of char *x is just raw, assembled x86 code, so it assigns
> the address of local function pointer *f the address of x, the ASM code,
> then calls it. Apparently, the code crashes an _Intel_ _Pentium_ (NOT
> P6/PII)
> because of incorrect implementation of exception trapping. (i.e. this
> bug
> does not exist on non-Intel CPUs.
> 
> -- 
> /---------------------------\
> |Bryon Roche                |
> |mailto:broche AT bensonlaw DOT com|
> \---------------------------/
> 

-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".