[Coco] nitros9 proposal - cache sector 0?

[Aaron Wolfe]

Hrmmm.. can we not trust data in the system page? Well... I know we
really can't, but it seems this would also be a problem for the
current implementation, since the buffer used currently in normal disk
ops would be similarly at risk (along with the device drive static
storage containing vital things like drive geometry, etc) and sector 0
seems to be in it nearly ever other disk operation.  I don't think any
of the standard modules do anything to ensure their data is intact.  I
agree it could be a concern, but not sure its of any greater risk than
what is already present in the existing implementation, unless I'm
missing something (quite possible!).

Of course this reminds me of how some 6809/OS9 systems had an MMU with
memory protection, I think SSB and GIMIX offered them.  Wouldn't that
be nice? In my own personal "dream coco 4", an upgraded MMU would top
the list :)



On Sat, Jan 25, 2014 at 11:45 AM, Steve <6809er at srbsoftware.com> wrote:
> My problem with cache this sector is corruption.  This a very important bit
> of information of the file system and should the cache buffer get damaged by
> crash code on the CoCo side, you just lost the disk.
>
> A CRC or other check would be needed to verify the data has not been
> corrupted.  But the CoCo doing this type of check could take as long as
> reading it from the server.  Don't forget that writing this sector back to
> the server will take longer because to updating the cache and the CRC.
>
> Good idea, but caching this sector may not save the time that you are hoping
> for.
>
> Steve
>
>
> On 1/25/2014 8:05 AM, Aaron Wolfe wrote:
>>
>> Its often been noticed that sector 0 of an RBF filesystem gets
>> read/written an awful lot when doing nearly any disk I/O.  I've been
>> working on the dw server some this morning and am being reminded of
>> just how much.
>>
>> Would it be practical to cache this sector somewhere?  256 bytes is
>> not tiny, but also not impossible to find in the typical system page.
>> The logic for a simple write-through cache would not be very complex.
>>   I guess I'm looking to the experts for reasons this can't work or
>> should never be done before looking at how to do it :)  Any thoughts?
>>
>> -Aaron
>>
>> --
>> Coco mailing list
>> Coco at maltedmedia.com
>> http://five.pairlist.net/mailman/listinfo/coco
>
>
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco