[sdcc-devel] [ sdcc-Bugs-1373846 ] Risky Startup Code

[SourceForge.net <no...@so...>]

Bugs item #1373846, was opened at 2005-12-05 21:56
Message generated for change (Comment added) made by maartenbrock
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1373846&group_id=599

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: msc51(8051) target
>Group: known bugs
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Nobody/Anonymous (nobody)
>Assigned to: Maarten Brock (maartenbrock)
Summary: Risky Startup Code

Initial Comment:
SDCC -- Version 2.5.0 #1020
Command: SDCC ctest.c

I disassembled the ctest.ihx file and found the 
following code fragment in the startup code:

    push  00    ; Push 00 to stack, R0 if in BANK 0
    mov   R0,82 ; move DPL to R0 IN CURRENT REG BANK!
    mov   @R0,A ; Store A where R0 Points
    pop   00    ; Restore 00 from stack

To function correctly, this code and other similar 
pieces of code in the startup code REQUIRE, that the 
register bank select bits in the PSW be zero.  If they 
have any non-zero value then R0 in the corresponding 
register bank will get the value of DPL.

If this piece of code was to be called from an 
interrupt routine using register bank 1, 2 , or 3 then 
the push/pop sequence would be useless and the value 
of R0 would be corrupted.

I really think that, if you know what addressing model 
to use at compile time, then this kind of linear 
search to find the correct addressing mode is pretty 
lame.  That's just my opinion -- I could be wrong.

Bruce Graham
bg...@hu...


----------------------------------------------------------------------

>Comment By: Maarten Brock (maartenbrock)
Date: 2005-12-05 22:15

Message:
Logged In: YES 
user_id=888171

Bruce,
Yes this was a bug in SDCC 2.5.0. Current development is 
at 2.5.4 and the bug has been squashed quite some time ago.
Maarten

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1373846&group_id=599