Re: [sdcc-devel] [ sdcc-Bugs-1715020 ] Cast volatile over literal constant address is ignored.

[Maarten B. <sou...@ds...>]

Philipp,

Have you checked this only applies to hc08? Are at least mcs51 and z80 ok?

Greets,
Maarten

> Bugs item #1715020, was opened at 2007-05-08 15:47
> Message generated for change (Comment added) made by spth
> You can respond by visiting:
> https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1715020&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: hc08 port
> Group: known bugs
>>Status: Closed
>>Resolution: Fixed
> Priority: 5
> Private: No
> Submitted By: Cassius Resende Duarte (cassius_duarte)
>>Assigned to: Philipp Klaus Krause (spth)
> Summary: Cast volatile over literal constant address is ignored.
>
> Initial Comment:
> Cast volatile over literal constant address is ignored, resulting in wrong
> instruction elimination done by the peephole optimizer. See the code
> bellow:
>
> char set_and_read(void) {
>   (*(volatile char*)256) = 0x11;
>   return  (*(volatile char*)256);
> }
>
> At the return value, an extra load from is generated, but eliminated by
> the
> peephole optimizer.
>
> _set_and_read:
> ;hi_page.c:10: (*(volatile char*)256) = 0x11;
>     lda    #0x11
> ;hi_page.c:11: return  (*(volatile char*)256);
> ; Peephole 8b  - removed redundant lda
>     sta    0x0100
> 00101$:
>     rts
>
> WorkAround(s)
> =============
>
> 1 - Disable the Peephole optimzer: sdcc -mhc08 --no-peep <file>
> 2 - Employ a const pointer or global variable.
>
> Additional Information
> ======================
> Command line for running the example: sdcc -mhc08 hi_page.c
> SDCC Version: 2.6.0
>
> ----------------------------------------------------------------------
>
>>Comment By: Philipp Klaus Krause (spth)
> Date: 2010-09-26 14:52
>
> Message:
> Fixed in sdcc 2.9.8 #5995, by disabling peepholes 8a and 8b.
>
> Philipp
>
>
> ----------------------------------------------------------------------
>
> Comment By: Cassius Resende Duarte (cassius_duarte)
> Date: 2007-05-08 15:49
>
> Message:
> Logged In: YES
> user_id=1787323
> Originator: YES
>
> Futher Analisis
> ===============
>
> The rule 8b (from /src/hc08/peeph.def) also invokes the routine
> notVolatileVariable (from /src/SDCCpeeph.c) to check if the optimization
> can be
> done. However, according to the documentation, the function will return
> TRUE for
> constants. Maybe the paser process may generate a special symbol for
> volatile
> constants?
>
>
> Related Bugs
> ============
> None that I could find. The closer one was "[ 663539 ] volatile does not
> work",
> but the issue is with global variables, and the bug is already closed.
>
> ----------------------------------------------------------------------
>
> Comment By: Cassius Resende Duarte (cassius_duarte)
> Date: 2007-05-08 15:47
>
> Message:
> Logged In: YES
> user_id=1787323
> Originator: YES
>
> File Added: hi_page.asm
>
> ----------------------------------------------------------------------
>
> You can respond by visiting:
> https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1715020&group_id=599
>
> ------------------------------------------------------------------------------
> Start uncovering the many advantages of virtual appliances
> and start using them to simplify application deployment and
> accelerate your shift to cloud computing.
> http://p.sf.net/sfu/novell-sfdev2dev
> _______________________________________________
> sdcc-devel mailing list
> sdc...@li...
> https://lists.sourceforge.net/lists/listinfo/sdcc-devel
>
>