GARTHWILSON wrote:
Bit 6 might be extra interesting since the BIT instruction puts it in the C flag
A related detail... one that's quirky and obscure, but heck -- if you're reading a thread about bit-banging then I assume you've gone pretty far down the rabbit hole already.
The BIT instruction certainly has its place, but in certain cases you'll want to
avoid BIT and the special status it brings to bit6. I'm referring to cases where
- utmost speed is required, and
- the code is doing input as well as output.
The latter point rules out half-duplex async. What's eligible is async full-duplex as well as SPI in full-duplex or SPI input-only (which still requires that Clk be output). Here's the key point:
If bit0 of a port is assigned as an output and bit7 is assigned as an input, then INC Port --
a single instruction --
can do input and output simultaneously. It can output a 1 to bit0 and also input bit7 and copy it to the N flag. (We assume bit0 is clear beforehand. And
DEC Port with bit0
set beforehand can output a 0 while also inputting bit7.)
To see this used "in real life" check out my SPI-related post
here. The choice of
bit7 and bit0 is crucial to the very high speed of the routines. ( <=19 cycles/bit for SPI input, <=18 cycles/bit for SPI output )
PS and FWIW: Garth & I seem to have independently discovered this point about INC/DEC outputting while also testing bit7. Somewhere on his web site -- a section on programming tips? -- I stumbled across a reference to it. But I'm not able to find it now.
-- Jeff
_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html
Login
Register
FAQ
Search
