At 09:29 AM 07/24/99, you wrote:
>>> This is not feasable. During sector I/O from floppy you can't miss a
>>single
>>> byte. Even a trivial interrupt routine which returns immediately with a
>>reti
>>> instruction takes too long during diskload.
>>
>>Even if you have a good CRC error detection-routine so that you can retry if
>>the data failed???
>
> Isn't CRC a so processor-intensive routine? I think u could use something
>'lighter', like checksum.
CRC is only a little more processor-intensive than a checksum. If my memory
is correct (that's always an uncertainty), CRC can be implemented as a
combination of adding, shifting and XORring with a fixed table.
But CRC is not necessary to check if data was missed, because the FDC
already has an error code for that (at least the one used in the 8250).
And as Alex Wulms pointed out, the time between MSX interrupts (1/60 or
1/50 sec) is too short for a sector to be found in a track, so the
retry-if-interrupted approach will fail anyway.
Bye,
Maarten
****
MSX Mailinglist. To unsubscribe, send an email to [EMAIL PROTECTED] and put
in the body (not subject) "unsubscribe msx [EMAIL PROTECTED]" (without the
quotes :-) Problems? contact [EMAIL PROTECTED] (www.stack.nl/~wiebe/mailinglist/)
****
