Hi Andy,

I'm sorry for the confusion. The tip I learned from you many years ago was
referring to the #define directive for I/O, not the DQ line assumption in
the code snippet. My apologies.

I modified his code to show him a different way of doings things with the
code. The detailed points you made are excellent, as you pointed out the
hardware ramifications of both methods.

Regards,

Ken Pergola



-----Original Message-----
From: pic microcontroller discussion list
[mailto:PICLIST@MITVMA.MIT.EDU]On Behalf Of Andrew Warren
Sent: Thursday, August 14, 2003 12:28 AM
To: PICLIST@MITVMA.MIT.EDU
Subject: [PIC]: Re: a better "if .. then .. else" in assembly?


Ken Pergola <PICLIST@mitvma.mit.edu> wrote:

> Andy Warren taught me this many years ago (thanks Andy!)
>
>     BCF     CLK        ; Bring CLK low
>     BSF     DQ         ; Assume DQ line will be a '1'
>     BTFSS   ReadCmd,0  ; If it IS a '1', skip ahead.
>     BCF     DQ         ; Otherwise, force DQ line to 0
>     BSF     CLK        ; Bring CLK high to tell temp. sensor

    Hmm... I hope he also pointed out that that method is ONLY
    acceptable in cases where a brief pulse on the DQ pin before
    every "0" bit won't cause problems.

    He should also have mumbled something about possible
    read-modify-write issues with such fast consecutive writes to the
    same port, and he should CERTAINLY have reminded you to check the
    setup-time specs for the device to which you're talking, to see
    whether a NOP or two is needed between BCF DQ and BSF CLK.

    By the way... The original method Sabachka showed,

        BTFSC   SomeBit
        BSF     Pin
        BTFSS   SomeBit
        BCF     Pin

    eliminates the little glitch before "0" bits (at the expense of
    one more instruction and one more cycle), but it has a subtle
    failure mode:  If "SomeBit" can change between the BTFSC and the
    BTFSS (e.g., if SomeBit is an I/O pin, or a bit in a timer
    register or something, rather than a general-purpose register
    bit), then BOTH or NEITHER of the branches will be taken.
    Often, this isn't a problem, but it's something to keep in mind.

    -Andy

    P.S.  My personal syntax preference for register-oriented
          instructions that place their result back in the register
          is neither ",1" nor ",F"; I simply omit the suffix
          entirely.  Makes the code much more readable, in my
          opinion:

              DECF    REG     ;Decrement REG.
              DECF    REG,W   ;W = (REG) - 1, REG unchanged.

    P.P.S.  Oh, one more:

            Andrew Kieran <PICLIST@mitvma.mit.edu> wrote:

            > I prefer to replace each argument (in this case, PORTx
            > and BITx) with separate equivalents [rather than] using
            > one word (e.g. CLK) to represent two items
            >
            > When I'm debugging, this helps me to check each part
            > of the instruction for accuracy.

                Yes, but if you use one symbol to represent the
                whole quantity, you don't HAVE to check each part of
                the instruction for accuracy.

=== Andrew Warren -- aiw@cypress.com
=== Principal Design Engineer
=== Cypress Semiconductor Corporation
===
=== Opinions expressed above do not
=== necessarily represent those of
=== Cypress Semiconductor Corporation

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu