Testing INA values over an OUT pin (in ASM)

peterz

I have·PIN1 that I·set as OUT and change values at certain times in·a Spin·program.
So I do, in Spin:
...
OUTA[noparse][[/noparse]1]:=1
...
OUTA[noparse][[/noparse]1]:=0
...
etc

In the ASM part of the program I test this pin from time to time·to check its value. For this I use an INA instruction.
I use the following code to test this pin:

test··· 2, ina wc
if_c jmp #Is_One

However the condition is·never met, that is, I never detect it as 1, despite I put the PIN1 to 1 many times while in the Spin part.

First question, is it right to perform an INA over an OUT pin ?
Second, is this·the right way to perform the test ?

SSteve

I'm guessing you are wanting to test against the literal value 2 rather than the contents of register 2 which is what you are really doing. Try this:

test pinMask, ina wc
...
pinMask long %0010

You can't use ina as the destination and you can't use a literal value as the destination, so you have to put the test mask in a DAT statement.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
OS-X: because making Unix user-friendly was easier than debugging Windows

links:
My band's website
Our album on the iTunes Music Store

Graham Stabler

If you are trying to check pin 1 you could do:

test ina,#2 wc

Also labels tend to be written as :Is_One so it would be

jmp #:Is_One

But I'm not sure if the : is just a convention, I suspect so.

Graham

SSteve

Graham Stabler said...
If you are trying to check pin 1 you could do:

test ina,#2 wc

That will not work. You can't use ina as a destination register. The compiler won't complain, but the code won't work. Take a look at the second entry on page 6 of the Propeller Tricks & Traps document.

Graham Stabler said...
Also labels tend to be written as :Is_One so it would be

jmp #:Is_One

But I'm not sure if the : is just a convention, I suspect so.

The colon isn't just a convention, it makes a label local. A local label's scope is between two non-local labels. This is to make it possible to use multiple instances of generic labels. Example:

Section1 mov counter, #16
:loop add this, that
    ...
  djnz counter, #:loop
Section2 mov counter, #16
:loop sub theother, thing
    ...
  djnz counter, #:loop

Without the colons, the above code wouldn't compile.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
OS-X: because making Unix user-friendly was easier than debugging Windows

links:
My band's website
Our album on the iTunes Music Store

Graham Stabler

Ah, I assumed that because it wasn't really a destination for the test command it would not mind. I guess that's why its a trap.

The : thing shows how programming languages are like normal langauges, I have just always used the : since the start but never learnt why.

Graham

big_mark

SSteve said...
This is to make it possible to use multiple instances of generic labels.

I am assuming (hoping) that you mean reusing labels in different objects, and not in the same block of assembly, even in different functions? Sounds like a potential debugging nightmare to me.

SSteve

That's what the local label capability is designed for, but just because you can use identical labels in the same block of assembly doesn't mean you have to.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
OS-X: because making Unix user-friendly was easier than debugging Windows

links:
My band's website
Our album on the iTunes Music Store

big_mark

It still sounds a little sloppy to me. I can't think of any good reason to allow multiple uses of the same label in a block of code. If one must use the label "loop", for example, at least make the label unique. So "loop" becomes "Subroutinename_loop" for instance. Its not like it'll take up any extra Propellor RAM...

Mark

Mike Green

Part of the reason for local labels is to avoid the need for other local references like "$+1" or "$-1" where "$" is the current instruction location. Often these are very useful and a complex descriptive label is not helpful in understanding the code.