mpm said:
But creates another problem.
Now you can't jump past the INTO interrupt vector off-page. (located
at 0003H)
You'll have to tell the Assembler / Compiler to do an in-page jump to
avoid servicing the interrupt.
No big deal, I guess.
Typically you don't have RAM at bottom of memory. So there is no
problem! But the ' Intel feature' repeats on every memory chip in the
system!
We had an EEPROM at 8000h in XDATA space. Sometimes the first byte of
this EEPROM was overwritten. It depends on ramp-up speed of VCC and Q
(=delay) of the oscillator.
Some memoriy devices restrict the write access in some way - mostly on
VCC value. As long as the memory device blocks the write all is ok.
Thereafter it depends...
To be save: Don't use the first byte of EVERY writable memory in the system.
A small risk scenario persists for corruption of another byte address
somewhere. Do you know when which output line in the controller goes
active??
At least for us the problem was completely solved.
Or go the WR-blocking way with for example a MAX691.
- Henry