[Coco] Is the 6809 flawed?

[Dave Philipsen]

Of course adding can end up with an incorrect result too. That’s why we have flags to make things right. 

Dave

> On Sep 7, 2018, at 5:34 PM, Richard Cavell <richardcavell at mail.com> wrote:
> 
> If I understand it, ASL can result in an overflow such that the result is mathematically incorrect. 
> 
> But LSL doesn’t.  The bits are just bits. Pixels, perhaps.  If one falls off the left hand side, that’s what was expected and its replacement is don’t care.
> 
>> On 07/09/2018 at 23:18, linville wrote:
>> 
>> Not a bug. That's pretty much the whole point.
>> What a dumb controversy...
>> John
>> -------- Original message --------From: Arthur Flexser <flexser at fiu.edu> Date: 9/7/18  5:24 PM  (GMT-05:00) To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com> Subject: Re: [Coco] Is the 6809 flawed? 
>> John Linville, who wrote the article you linked to, clearly does not
>> consider this a bug:
>> 
>> " Did the CPU architects at Motorola get it wrong? *Spoiler alert:*
>> *No!* Neither
>> did architects for Intel, MIPS, Sparc, ARM, or any of the other CPUs that
>> similarly implement a single instruction for both operations. The fact is
>> that when operating on two's-complement
>> <https://en.wikipedia.org/wiki/Two%27s_complement> numbers,
>> logical-shift-left and arithmetic-shift-left are, in fact, exactly the same
>> operation."
>> 
>> Art
>> 
>>> On Fri, Sep 7, 2018 at 4:47 PM, Dave Philipsen <dave at davebiz.com> wrote:
>>> 
>>> Search for the 6809 ASL bug using Google. No one has ever reported such a
>>> thing as far as I can tell.
>>> 
>>> Dave
>>> 
>>> 
>>>> On 9/7/2018 3:29 PM, Salvador Garcia via Coco wrote:
>>>> 
>>>>    I am not familiar with 6809 assembler, but for this bug to have gotten
>>>> past Motorola testing is, as Vizzini would say, Inconceivable!
>>>> 
>>>> Salvador
>>>> 
>>>> 
>>>> 
>>>> 
>>>>       On Friday, September 7, 2018, 1:47:13 PM CDT, Dave Philipsen <
>>>> dave at davebiz.com> wrote:
>>>>      If you haven’t watched episode 73 of CoCoTALK!, you should. A
>>>> contention is brought up that claims the 6809 ASL instruction “does not
>>>> work for signed math”.  Is it true?  After 40 years are we just now finding
>>>> out that the 6809 has a bug and that Motorola engineers “messed up here”?
>>>> John Linville was the first to notice the flawed tutorial and comment about
>>>> it here: https://retrotinker.blogspot.com/2018/09/yes-lsl-and-asl-rea
>>>> lly-are-same.html
>>>> 
>>>> You can see for yourself by visiting the CoCoTALK! Webpage here:
>>>> http://cocotalk.live/video-episodes/ and viewing episode 73.  You’ll
>>>> find the interesting part around 1:23:00.  Hear for yourself and determine
>>>> whether our favorite CoCo processor, the 6809 is flawed and in regard to
>>>> the Arithmetic Shift Left if you should “just never use that in
>>>> life...never use that command and you’ll be fine.”
>>>> 
>>>> 
>>>> 
>>>> Dave
>>>> 
>>>> 
>>> 
>>> --
>>> Coco mailing list
>>> Coco at maltedmedia.com
>>> https://pairlist5.pair.net/mailman/listinfo/coco
>>> 
>> 
>> -- 
>> Coco mailing list
>> Coco at maltedmedia.com
>> https://pairlist5.pair.net/mailman/listinfo/coco
>> 
>> -- 
>> Coco mailing list
>> Coco at maltedmedia.com
>> https://pairlist5.pair.net/mailman/listinfo/coco
> 
> -- 
> Coco mailing list
> Coco at maltedmedia.com
> https://pairlist5.pair.net/mailman/listinfo/coco