Setting output bits in Pasm

AV_Tech

Hello readers of the forum.

I have a question about setting output bits in Pasm. When i set or reset 1 bit I use the example in Part_1 of my code snippet below. But when I want to set more bit's, but stil not touch other bit's, I use the code snippet of Part_2. Are I am right, is this the right way or is there an other way to set/reset for example a nibble to the output pin's in Pasm.

Greeting Abraham

dat

'example data = %00000000_00000000_00001010_00000000 or %00000000_00000000_00001101_00000000 

Start         mov       dira,   Pin_out         ' Set the output's   

Part_1        or        outa,   Clk_2           ' Set a bit
              andn      outa,   Clk_2           ' Reset a bit

Part_2        andn      outa,   mask_nibb       ' Clear a nibble
              or        outa,   data            ' Set data nibble

Pin_out    long         %00000000_00011111_11111111_00000000 
Clk_2      long         %00000000_00010000_00000000_00000000     
mask_nibb  long         %00000000_00000000_00001111_00000000

Electrodude

That's the usual way to do it on the P1.

You can also use "or" and "andn" to modify dira if you need to change pin directions at runtime, but a "mov dira, whatever" is fine for setup.

Phil Pilgrim (PhiPi)

Your Part_2 code will cause a glitch on the pins you're trying to change. If that's not acceptable, you can do this instead:

Part_2        mov       temp,outa
              andn      temp,mask_nibb       ' Clear a nibble
              or        temp,data            ' Set data nibble
              mov       outa,temp

-Phil

AV_Tech

Phil Pilgrim (PhiPi) wrote: »

Your Part_2 code will cause a glitch on the pins you're trying to change. If that's not acceptable, you can do this instead:

Part_2        mov       temp,outa
              andn      temp,mask_nibb       ' Clear a nibble
              or        temp,data            ' Set data nibble
              mov       outa,temp

-Phil

Phil.

Thanks for this suggestion. I go use it.

Abraham.

ersmith

Phil:

Thanks for pointing out the glitching problem with the original Part_2 code. That pointed out a fastspin bug; the optimizer wasn't special casing OUTA or DIRA, and so the code generated did look like that original one. I've added a special case for OUTA[a..b] and DIRA[a..b] to use the second form, so it's fixed in fastspin 3.1.2.

Eric