(Q) - freeRTOS context switch qualification about romwbw

[wwarthen]

Odd. Sorry I don't have any immediate thoughts.

-Wayne

[wwarthen]

Q

So, I'm not totally sure that the BBR for user TPA is actually 0xF0.
Could you please confirm that matching specifically 0xF0 is the most effective method to match only the user TPA?

For a Z180 with 512K ROM & 512K RAM (normal setup), then BBR=0xF0 is synonymous with user TPA being mapped to low 32K.

Based on your description, it sounds like interrupts might be suppressed while waiting for the character. Is that possible?

Thanks,

Wayne

[feilipu]

Thanks. Helps so much to get clarification.

It sounds like interrupts might be suppressed while waiting for the character. Is that possible?

Things should be running normally, unless a call is inside HBIOS, then the System Tick ISR will recognise that a) the BBR is not in USR TPA, or b) the SP is set within the HBIOS call stack area (top page). If either of these things happen, then the System Tick will reset the timer, and exit. Of course the System Tick will only be entered if interrupts are enabled (which they should be). Otherwise in the normal case a Tick increment will happen and a context switch opportunity happens.

I have just resolved one issue, where the call to increment the System Tick was not properly protected by the SP location check. I just fixed that, and at the same time simplified things a little.

This has made things incrementally better. But, I'm still fishing to find the underlying problem with getchar().

[feilipu]

I'm still looking, but will close this as it's sure to be on my code base. Thanks again.