code question
rjo__
Posts: 2,114
in Propeller 2
There has to be a structural issue that I am confused about here, but I have been fiddling for the most part of a day and it eludes me.
This code runs in its own cog.
It copies a small region of an image from HUB Ram into LUT Ram. It then copies the LUT Ram contents back to a different region of the image in Hub Ram. Simple and it works fine.
However, when I started to add call subroutines, I ran into an issue.
If you look at the subroutines section(line 51), you'll see a subroutine that is not actually called, does absolutely nothing, but does corrupt the data I am passing from the LUT to the Hub.
If I comment out the subroutine, everything is fine.
How can a subroutine, which does nothing and is never called, have any impact on a program?
This code runs in its own cog.
It copies a small region of an image from HUB Ram into LUT Ram. It then copies the LUT Ram contents back to a different region of the image in Hub Ram. Simple and it works fine.
However, when I started to add call subroutines, I ran into an issue.
If you look at the subroutines section(line 51), you'll see a subroutine that is not actually called, does absolutely nothing, but does corrupt the data I am passing from the LUT to the Hub.
If I comment out the subroutine, everything is fine.
How can a subroutine, which does nothing and is never called, have any impact on a program?
dat org
area_test
mov at_testpoint, ##32064
mov at_testPointaddr, at_testpoint
add at_testPointaddr, ##_SCREEN_BUFFER
mov at_inc2output,##129_408 'offset from input point output point (640*202)+128
mov at_rgnwidth,#21
mov at_rgnheight,#21
mov at_rgnwm1div2,#10
mov at_rgnhm1div2,#10
mov at_rgnbase,at_testPointaddr
sub at_rgnbase,#at_rgnwm1div2
mul at_rgnhm1div2,##640
sub at_rgnbase,at_rgnhm1div2
mov at_rgnlinept0,at_rgnbase
mov at_temp1,at_rgnwidth
mov at_lutaddr,#0
mov at_temp2,at_rgnheight
'***
at_region2lut
mov at_temp1,at_rgnwidth
rdfast #0,at_rgnlinept0
rep @at_hline,#20
rfbyte at_pixval
wrlut at_pixval,at_lutaddr
add at_lutaddr,#1
at_hline
add at_rgnlinept0,##640
djnz at_temp2,#at_region2lut
'***
mov at_lutaddr,#0
mov at_rgnlinept0,##_SCREEN_BUFFER
add at_rgnlinept0,##1040
mov at_temp2,#21
at_lut2region
mov at_targetpt, at_rgnlinept0
mov at_temp1,#21
wrfast #0,at_rgnlinept0
rep @at_hline2,#20
rdlut at_pixval, at_lutaddr
add at_lutaddr,#1
wfbyte at_pixval
at_hline2
add at_rgnlinept0,##640
djnz at_temp2, #at_lut2region
jmp #area_test 'back to start
'__________________________________________________________________________________________________________
'**************
'Subroutines
at_neibrhood '<----------------???????
ret
'************* 'start address of output picture (640*202)+128
at_bool res 1
at_lutaddr res 1
at_lutpt0 res 1
at_temp1 res 1
at_temp2 res 1
at_pixval res 1
at_outputLinept res 1
at_testline res 1
at_testval res 1
at_testPoint res 1
at_testPointaddr res 1
at_imageBase res 1
at_imageEnd res 1
at_outputBase res 1
at_outputaddr res 1
at_rgnbase res 1
at_rgnwidth res 1
at_rgnheight res 1
at_rgnwm1div2 res 1
at_rgnhm1div2 res 1
at_inc2output res 1
at_rgnlinept0 res 1
at_targetpt res 1
at_testpix res 1
fit $1F0
Comments
A 21x21 area is being captured from the first camera on the left and displayed
at the top of the frame toward the right edge.
Without the inclusion of an uncalled, do nothing subroutine, the image capture
is fine. But with the inclusion of the do-nothing, uncalled subroutine, the image capture appears to include part of the white frame from the left side of the original under-laid image.
How can including an uncalled, do-nothing subroutine possibly do anything?
Have you ever tried to replace the 'ret' instruction with a 'res 1' statement, or even another instruction, such as a 'nop'?
If the problem has to do with the address of any / all other labels, it will continue to interfere with your code.
Otherwise, it will magically disappear, and you may start to blame the compiler.
Another useful test might be to check the $ hex image of the code output to detect any incorrect compiler behavior.
Hope that helps a bit
Henrique
But it is on my list!!!
I actually looked for it and didn't see it... until the next day after begging for help:)
It is a short piece of code... how I did not see it is beyond me. I was so busy systematically looking in all the wrong places...for an entire day!!!
My bigly problem is that I can't help trying to optimize every little task... I write it the slow way to get it working and then play with faster and faster methods.
I have now decided to completely re-organize my entire demo...
It might never get done:)
I love it!!!
Very addictive.