Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.arch > #1464

Re: Patent application

Show all headers | View raw

On 5/10/2011 10:22 AM, Robert Myers wrote:
> On May 9, 5:01 pm, "Andy \"Krazy\" Glew"<a...@SPAM.comp-arch.net>
> wrote:
>> Let me mention in passing the following patent application which is
>> publicly visible:
>>
>> http://appft1.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=...
>>
>> 20110078389
>>
>> MANAGING AND IMPLEMENTING METADATA IN CENTRAL PROCESSING UNIT USING
>> REGISTER EXTENSIONS
>>
>> A set of default registers of a processor are expanded into metadata
>> registers on the processor of a computer system. The default registers
>> having stored thereon data, while metadata which is related to the data
>> is stored separately on the metadata registers.
>>
>> Inventors:      Patel; Baiju V.; (Portland, OR) ; Gopalakrishna; Rajeev;
>> (Hillsboro, OR) ; Glew; Andrew F.; (Hillsboro, OR) ; Kushlis; Robert J.;
>> (Worcester, MA) ; Van Dyke; Don Alan; (Rescue, CA) ; Cihula; Joseph
>> Frank; (Hillsboro, OR) ; Mallick; Asit K.; (Saratoga, CA) ; Crossland;
>> James B.; (Banks, OR) ; Neiger; Gilbert; (Portland, OR) ; Rodgers; Scott
>> Dion; (Hillsboro, OR) ; Dixon; Martin Guy; (Portland, OR) ; Charney;
>> Mark Jay; (Lexington, MA) ; Gottlieb; Jocob; (Kiryat Tivon, IL)
>> Serial No.:     571269
>> Series Code:    12
>> Filed:  September 30, 2009
>
> Seems like swatting a mosquito with a sledgehammer?

Amusing:  my present employer shoots mosquitoes with a laser.


> If you're going to put metadata into memory, why limit the advantage
> to bounds checking?

(0) The patent application makes no claims about memory metadata, 
although it does mention it in passing.

(1) I don't want to stop at bounds checking.

(2) However, I am not at all certain that classic data types are a good 
use of such metadata.  Now, the memory mechanisms referred to in the 
patent, but not claimed, allow large amounts of memory metadata to used; 
but, still, there are limits.

(3) As an OS guy, I tend to think more about providing mandatory 
security, not just security that can be turned off via a cast. Although 
I think it is key that there must be a layer of optional security, that 
can be cast around.  I discussed this in comp.arch years ago, before the 
work reflected in this patent application.


Anyway, the answer to any question along the line of "Why stop there?", 
not just for the sort of thing mentioned in the patent but for other 
things, is usually "We thought of going further, we may even have wanted 
to go further, or have plans to go further, but we stopped where we did 
because it was enough for a first step."  Or "because this is as much as 
we thought we might be able to push through".  Or "more than enough".

Dave Sager said something like:  it's not hard to figure out what the 
ultimate, ideal, form of XYZZY computer architecture should be.  It is 
hard to figure out the small incremental steps that can get there.

Back to comp.arch | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Patent application "Andy \"Krazy\" Glew" <andy@SPAM.comp-arch.net> - 2011-05-09 14:01 -0700
  Re: Patent application Edward Feustel <efeustel@hughes.net> - 2011-05-10 06:22 -0400
  Re: Patent application Robert Myers <rbmyersusa@gmail.com> - 2011-05-10 10:22 -0700
    Re: Patent application "Andy \"Krazy\" Glew" <andy@SPAM.comp-arch.net> - 2011-05-11 14:22 -0700
      Re: Patent application Rick Jones <rick.jones2@hp.com> - 2011-05-11 22:51 +0000

csiph-web