Buffered TX and RX (SX/B + Assembly)

JonnyMac

My purpose in sharing this is two-fold:

1) I needed it; someone else may as well.

2) To demonstrate that with very little effort we mortals can take code from gurus like Guenter, Peter, etc., and fold it into our SX/B programs. Terry has done a great job constructing SX/B to allow the smooth incorporation of Assembly segments and I think this is a pretty good demonstration of that.

Notes:
-- TX and RX [noparse][[/noparse]circular] buffers are each eight bytes
-- TX and RX vars and code are independent of each other and can be pulled out and used separately
-- code prevents overrun of buffers when full
-- DELAY_MS used to replace PAUSE -- allows RX and TX while performing the delay

I've tested at the limits: 1200 and 38400 baud. Please let me know if you find any issues as this is getting folded into the latter chapters of my book. Thanks.

T&E Engineer

JonnyMac,

When attempting to compile I got an error. See attached jpg screenshot.

Bean

You are using the BETA compiler.

You'll have to add a "RETURN __PARAM1" before the ENDFUNC to satisfy the compiler.

Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Teacher: What is the difference between ignorance and apathy ?
Student: I don't know and I don't care
Teacher: Correct !
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
www.hittconsulting.com
·

JonnyMac

Terry,

Is that because the compiler expects the FUNC keyword to have an associated RETURN? I'm using 1.51.03 so I have not see the problem reported by T&E.

T&E Engineer

That did it for me:

' -------------------------------------------------
' Subroutine / Function Code
' -------------------------------------------------
' Use: aByte = RX_BYTE
' -- returns "aByte" from 8-byte circular buffer
' -- will wait if buffer is presently empty
' -- rxBufCnt holds byte count of receive buffer (0 to 8)
FUNC RX_BYTE
  ASM
    BANK  rxSerial
    TEST  rxBufCnt    ' check buffer count
    JZ    @RX_BYTE    ' wait if empty
    MOV   W, #rxBuf    ' point to tail
    ADD   W, rxTail
    MOV   FSR, W
    MOV   __PARAM1, IND    ' get byte at tail
    INC   rxTail    ' update tail
    CLRB  rxTail.3    ' keep 0 to 7    
    DEC   rxBufCnt    ' update buffer count
    BANK  0
  ENDASM
  [b]RETURN __PARAM1[/b]
  ENDFUNC

Bean

JonnyMac said...
Terry,

Is that because the compiler expects the FUNC keyword to have an associated RETURN? I'm using 1.51.03 so I have not see the problem reported by T&E.

Jon,
· Yes, it is expecting to see a RETURN somewhere within FUNC...ENDFUNC (I'll have to change the wording of the error message since it references ENDSUB which CAN be used without a return).

· It probably should be a warning instead of an error, since it doesn't prevent anything from compiling.

Bean.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Teacher: What is the difference between ignorance and apathy ?
Student: I don't know and I don't care
Teacher: Correct !
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
www.hittconsulting.com
·

JonnyMac

I agree that it should just be a warning; in cases like this program where the FUNC is all assembly (and __PARAMx is being dealt with manually), there is no need for the RETURN keyword.