On Tue, Jan 24, 2012 at 4:27 PM, PICdude <picdude3@narwani.org> wrote:
> Hmmm... still tinkering with C18 and ran into this oddity...
>
> #define S2_MIN 90
> #define S2_MAX 125
> #define S2_STEP 5
> ...
> void main(void)
> {
> =A0 =A0 =A0 =A0...
> =A0 =A0 =A0 =A0S2 =3D S2_MIN; =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0// Initialize
> =A0 =A0 =A0 =A0...
> =A0 =A0 =A0 =A0// (In button responder routine...)
> =A0 =A0 =A0 =A0S2 +=3D S2_STEP; =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0// Increment to next value
> =A0 =A0 =A0 =A0if (S2 >=3D (S2_MAX + S2_STEP)) =A0 =A0 =A0 =A0 =A0 // If =
max passed...
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0S2 =3D S2_MIN; =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0// ...cycle back to start
> =A0 =A0 =A0 =A0...
> }
>
> Every time S2 is incremented, the comparison checks to see if the max
> value has been exceeded, and if so, it will reset it to the min value.
> =A0But it fails.
>
> If I change the comparison to this, it works...
>
> =A0 =A0 =A0 =A0if (S2 > S2_MAX)
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0S2 =3D S2_MIN; =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0// Cycle back around to start
>
> Disassembly of the non-working code (with comments I've now added).
> 0xc0 and 0xbf are the upper and lower bytes of S2 resp. =A0Computations
> in comments are at S2=3D130.
>
> 845: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0if (S2 >=3D (S2_MAX + S2_STEP))
> =A0 0B16 =A0 =A00E82 =A0 =A0 MOVLW 0x82 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0// 130, correct.
> =A0 0B18 =A0 =A05DBF =A0 =A0 SUBWF 0xbf, W, BANKED =A0 =A0 =A0 // Result=
=3D0, so C=3D1, Z=3D1
> =A0 0B1A =A0 =A00EFF =A0 =A0 MOVLW 0xff
> =A0 0B1C =A0 =A059C0 =A0 =A0 SUBWFB 0xc0, W, BANKED =A0 =A0 =A0// 511-0 i=
s +ve, so C=3D1, Z=3D0.
> =A0 0B1E =A0 =A0E303 =A0 =A0 BNC 0xb26 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 // So this should not branch...?
> 847: =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0 =A0 =A0S2 =3D S2_MIN;
> =A0 0B20 =A0 =A00E5A =A0 =A0 MOVLW 0x5a
> =A0 0B22 =A0 =A06FBF =A0 =A0 MOVWF 0xbf, BANKED
> =A0 0B24 =A0 =A06BC0 =A0 =A0 CLRF 0xc0, BANKED
>
>
> If I understand what's happening here, the lower bytes are subtracted
> first, and C is carried over into the next operation where the upper
> byte is subtracted. =A0Not sure why 255 is being used for that though.
> But at S2=3D130, C should be 1 when it reaches the BNC instruction.
> But's it's skipping the code after that branch. =A0What's really
> confusing me here is the SUBWFB instruction. =A0I've looked at the
> datasheet for it and the examples provided for its use, and can't see
> how they get those results.
>
> Huh?
>

How is "S2" declared?  If it is an int8_t (aka char) then 125+5 =3D
-126, which is always going to be less than S2_MIN, so S2 will remain
at 90.

--=20
Regards,
Mark
markrages@gmail

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.