Re: [Sdcc-user] weird user of the volatile keyword.

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

Hi again Rod,

Ok, if this is the case, then we definitely need to know if the latest 
snapshot gives you the same problems. If so, then we have a new 
bug, if not than this was fixed already and nothing stands in the way of 
a new release.

> I forgot to place the initialisation lines like this:
> 
>     wCurTmr = ReadTimerTic();
>     wOldTmr = wCurTmr;
> 
> 
> These lines are just before the while 1....
> 
> I do not beleive that the bug report you quoted is relevant but I did not supply
> enough infomation sorry about that.
> 
> Rod
> 
> On Fri Apr  1  1:08 , 'Maarten Brock' <sou...@ds...> sent:
> 
> >
> >
> >
> >
> >
> >Rod,
> >
> >
> >This is a known bug (895992). The problem lies in the uninitialized 
> >variables. The liverange analysis doesn't correctly detect these 
> >variables to be alive at the start of the loop. Initialize them and try again 
> >please.
> >
> >
> >> All,
> >> 
> >> I have been trying to get my main loop to run correctly.  I have been 
> >> having trouble with the following code:
> >> 
> >> void main( void )
> >> {
> >>     UINT16 wOldTmr;
> >>     UINT16 wCurTmr;
> >> 
> >>     while( 1 )
> >>     {
> >>         while( wOldTmr == wCurTmr )
> >>         {
> >>             wCurTmr = ReadTimerTic();
> >>         }
> >>         wOldTmr = wCurTmr;
> >> 
> >>     /* Some more code here */
> >> }
> >> 
> >> The function ReadTimerTic returns a 16-bit value my problem was that 
> >> this loop at the top of main was supposed to delay 1mS so my main loop 
> >> would run every 1mS.  This is not the case it seems that the optomizer 
> >> is allowing the loop to exit immediately without delaying 1mS.  It was 
> >> not until I set both variables to volatile that the loop run correctly.  
> >> I have no problem with this but the definition of volatile is to tell 
> >> the compiler to reload a variable on each reference to the variable in 
> >> case of a change in value outside the main program flow.  In this case 
> >> the variable wCurTmr will change but that change is in sync with the 
> >> normal program flow.  So my question is whey do I have to have the 
> >> volatile declaration in any case?  This would seem to me to be a case of 
> >> an over aggressive optomizer.  I have never had to do this with the Keil 
> >> compiler so I was surprised when I did this with SDCC.
> >> 
> >> My code is all working now and I have very happy but I would clasify 
> >> this as unexpected and nothing else.
> >> My version of sdcc is:
> >> SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 
> >> 2.4.0 (Mar 22 2005) (UNIX)
> >> 
> >> Regards,
> >> Rod Boyce
> >
> >
> >
> >Greets,
> >Maarten
> >
> >
> >
> >
> >
> >-------------------------------------------------------
> >This SF.net email is sponsored by Demarc:
> >A global provider of Threat Management Solutions.
> >Download our HomeAdmin security software for free today!
> >http://www.demarc.com/info/Sentarus/hamr30
> >_______________________________________________
> >Sdcc-user mailing list
> >Sdc...@li...
> >https://lists.sourceforge.net/lists/listinfo/sdcc-user
> 
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by Demarc:
> A global provider of Threat Management Solutions.
> Download our HomeAdmin security software for free today!
> http://www.demarc.com/info/Sentarus/hamr30
> _______________________________________________
> Sdcc-user mailing list
> Sdc...@li...
> https://lists.sourceforge.net/lists/listinfo/sdcc-user
> 
>