Official Release 1 ( 17-Dec-2002) does not generate this problem.
This is a bug when using patches from 10 Feb 2003.
The first scenario (right shift 3) generates incorrect code. it should be
bit.b #llo(8), &0x0030
c-code :
/* Initialise key matrix variables */
KP_CurKeyVal = (P5IN & 0x08) >> 3;
list file output:
246:src/keys.c **** /* Initialise key matrix variables */
247:src/keys.c **** KP_CurKeyVal = (P5IN & 0x08) >> 3;
223 .stabn 68,0,247,.LM16-KEYS_InitSW
224 .LM16:
225 0054 F2B0 0300 bit.b #llo(3), &0x0030
225 3000
226 005a C243 0000 clr.b &KP_CurKeyVal
227 005e C263 0000 adc.b &KP_CurKeyVal
Correct code is generated for a right shift of 2
c-code:
/* Initialise key matrix variables */
KP_CurKeyVal = (P5IN & 0x08) >> 2;
list file:
246:src/keys.c **** /* Initialise key matrix variables */
247:src/keys.c **** KP_CurKeyVal = (P5IN & 0x08) >> 2;
223 .stabn 68,0,247,.LM16-KEYS_InitSW
224 .LM16:
225 0054 5F42 3000 mov.b &0x0030, r15
226 0058 7FF2 and.b #llo(8), r15
227 005a 12C3 clrc
228 005c 4F10 rrc.b r15
229 005e 4F11 rra.b r15
230 0060 C24F 0000 mov.b r15, &KP_CurKeyVal
Regards,
Phil Hancock
Design Engineer
Tekelek Pty Ltd
7 Comley St
Export Park
South Australia 5950
Ph 61 8 8234 3011
Fx 61 8 8234 3242
