Hello,
I strongly agree with you OCY. But as Hugh said it was a misunderstanding of "+"
and "|" operators. You must be a genuine C programmer to write a code in C.
Thanks all of you for your kind help.
--- In m..., "old_cow_yellow" wrote:
>
> I think a bigger and broader problem is, people do not care about
understanding even basic concepts. They rush into "send me sample code" or even
"send me the code".
>
> This is especially a problem for a lot of self proclaimed c programmers. Yes,
c is "easy to write" and "easy to understand". But it does not mean lazy and
careless people can program correctly in c.
>
> --- In m..., Hugh Molesworth wrote:
> >
> > Also this is not (yet) your error, but keep in
> > mind that you potentially could introduce one of
> > the classic c errors caused by incorrectly using
> > operators. "+" means add two items together,
> > which you don't actually want; "|" means or two
> > items together, which you do want. If you want to
> > combine two bitfields in an "or" fashion, so the
> > result has a "1" in any bit where either of the
> > original operands have a "1". "+" will do this as
> > well, if you can guarantee that "1"s in each
> > operand do not overlap, but in your case they do not.
> >
> > Thus these two produce the same result:
> > 0x0001 + 0x0002 = 0x0003
> > 0x0001 | 0x0002 = 0x0003
> >
> > These two DO NOT produce the same result:
> > 0x0001 + 0x0001 = 0x0002
> > 0x0001 | 0x0001 = 0x0001
> >
> > So, for example, let's assume you forget that
> > WDT_ARST_1000 already encompasses WDTPW,
> > WDTCTL = WDTPW + WDT_ARST_1000;
> > is not the same as
> > WDTCTL = WDTPW | WDT_ARST_1000;
> >
> > Out of interest, TI, and hence IAR et al,
> > propagate this error since they uses "+" in their
> > own header files, but of course IAR et al are
> > careful to only use correct definition
> > combinations so it doesn't matter. Incorrectly
> > combining these declarations however, causes
> > grief, even when the combination looks reasonable.
> >
> > Hugh
> >
> > At 09:11 AM 4/13/2009, you wrote:
> > You have:
> >
> > ---quote---
> > void ResetWatchdog (void)
> > {
> > WDTCTL = WDTPW + WDTCNTCL; //Refresh watchdog counter
> > }
> > ---unquote---
> >
> > What does this procedure do? Is the Watchdog
> > still using ACLK after a call to ResetWatchdog()?
> >
> > You did not say what WDTCNTCL is defined to be.
> > But I suspect that it is 0x0008. In that case, the Watchdog will use
SMCLK.
> >
> > Sorry that I did not give you the fish. And I did
> > not teach you how to fish either. I only tried to
> > remind you to use your head. I tried twice and
> > failed twice. I was not trying to be funny. The world is too funny
already.
> >
> > --- In m..., "ti2tt" wrote:
> > >
> > > Hello,
> > >
> > > Below is the code I am using for watchdog,
> > >
> > > #define WDT_ARST_1000 (WDTPW+WDTCNTCL+WDTSSEL)
> > >
> > > void WatchdogHalt (void)
> > > {
> > > WDTCTL = WDTPW + WDTHOLD; // Stop watchdog timer
> > > }
> > >
> > > void EnableWatchdog (void)
> > > {
> > > WDTCTL = WDT_ARST_1000;
> > > }
> > >
> > > void ResetWatchdog (void)
> > > {
> > > WDTCTL = WDTPW + WDTCNTCL; //Refresh watchdog counter
> > > }
> > >
> > > void main (void)
> > > {
> > > WatchdogHalt(); //Halt internal WDOG
> > > _SelectCrystalOsc(); //Configure XTAL
> > > EnableWatchdog(); //Enable internal WDOG
> > > ResetWatchdog(); //Reset internal WDOG
> > > while(1)
> > > {
> > > ResetWatchdog(); //Reset internal WDOG
> > > }
> > > }
> > >
> > > The bits related to watchdog sfrs are defined
> > in msp430F24x.h file. Please guide me.
> > >
> > > Thanks in advance.
> > > --- In m..., "old_cow_yellow" wrote:
> > > >
> > > > In your original posting, you said"
> > > >
> > > > "The refresh routine for watchdog counter consists as:
> > > > " WDTCTL = WDTPW + WDTCNTCL;"
> > > >
> > > > And you also said:
> > > >
> > > > "I am using ACLK as source for watchdog counter."
> > > >
> > > > I am not sure about that. I can only assume
> > that the #define(s) you used make "ACLK as source for watchdog counter".
> > > >
> > > > --OCY
> > > >
> > > > --- In m..., "ti2tt" wrote:
> > > > >
> > > > > Hi,
> > > > >
> > > > > As you all have seen that the #defines are
> > correct in the code but still my question remains
> > unanswered. I am handling the watchdog correctly
> > but at the same time after periodic refresh, I am
> > getting watchdog reset. Can anyone clarify this?
> > > > >
> > > > > Whether the use of ACLK for watchdog clock
> > source in my code which is 1MHz/32768 is
> > acceptable or not as the MCLK in my case is 250
> > KHz? Can anyone clarify this also?
> > > > >
> > > > > I badly need you people help for this. Thanks in advance.
> > > > > --- In m..., Joe Radomski wrote:
> > > > > >
> > > > > > You are correct.. I looked atthe
> > crossworks header file and it does indeed include
> > the password... so it wount reset, but my other
> > point is valid.. I'm not sure what header I
> > looked at initially that didn't� have the
> > password... I usually explicitly show the bits
> > when I do configurations, I don't make it a habit
> > of using defines like that so that it forces be
> > to make sure I handle things correctly, unless I
> > have to make the code compile for several devices
> > then these definitions work out better (for
> > example alkso having to have the code work on a
> > 5xx series which has an extra bit in the divider)
> > and don't have to add in as many conditional compiles...
> > > > > > �
> > > > > > the stopping of the WDT timer in his
> > first line is unnecessary as crossworks does it in startup.asm by
default.
> > > > > > �
> > > > > >
> > > > > >
> > > > > > --- On Fri, 4/10/09, old_cow_yellow wrote:
> > > > > >
> > > > > > From: old_cow_yellow
> > > > > > Subject: [msp430] Re: Use of internal WATCHDOG for MSP430F2418
> > > > > > To: m...
> > > > > > Date: Friday, April 10, 2009, 8:04 PM
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > Not really. It depends on what are defined as what.
> > > > > >
> > > > > > Many of the header files I have seen has effectively:
> > > > > > #define WDTCTL 0x0120
> > > > > > #define WDT_ARST_1000 0x5A0C
> > > > > >
> > > > > > Using such a header file, the statement:
> > > > > > WDTCTL = WDT_ARST_1000;
> > > > > > will not issue a Watchdog Password Violation reset
> > > > > >
> > > > > > On the other hand, if WDTCTL is defined
> > to be something other than 0x0120, the same
> > statement may have nothing to do with the Watchdog.
> > > > > >
> > > > > > What it appears to be saying is not
> > necessarily what you get. That is the unspecified
> > #define can do for you. And that is why I asked
> > would val=THREE resulted in val being set to 3.
> > > > > >
> > > > > > --- In msp430@yahoogroups. com, Joe Radomski
wrote:
> > > > > > >
> > > > > > > you are correct.. the way hisexample
> > �is currently setting the register will issue
> > an immediate reset.. since no password is
> > provided, but if he does get it set correcty the
> > timeout period is very small and will probably reset anyway..
> > > > > > > �
> > > > > > > �
> > > > > > > �
> > > > > > > �
> > > > > > > �
> > > > > > >
> > > > > > >
> > > > > > > --- On Fri, 4/10/09, citymouse2u wrote:
> > > > > > >
> > > > > > > From: citymouse2u
> > > > > > > Subject: [msp430] Re: Use of internal WATCHDOG for MSP430F2418
> > > > > > > To: msp430@yahoogroups. com
> > > > > > > Date: Friday, April 10, 2009, 1:24 PM
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > I think you need to use:
> > > > > > > WDTCTL = WDTPW + WDT_ARST_1000;
> > > > > > > Changing the WD without the password will cause a watchdog
reset.
> > > > > > >
> > > > > > > --- In msp430@yahoogroups. com, "ti2tt" wrote:
> > > > > > > >
> > > > > > > > Hello Forum members,
> > > > > > > >
> > > > > > > > I am using MSP430F2418 with
> > CrossStudio. I want to use internal watchdog of
> > this controller. I have configured the watchdog as
> > > > > > > >
> > > > > > > > WDTCTL = WDTPW + WDTHOLD;
> > > > > > > > WDTCTL = WDT_ARST_1000;
> > > > > > > >
> > > > > > > > The refresh routine for watchdog counter consists as:
> > > > > > > >
> > > > > > > > WDTCTL = WDTPW + WDTCNTCL;
> > > > > > > >
> > > > > > > > With this configuration, I am
> > getting a watchdog reset after certain intervals.
> > Here I am using a XTAL of 1MHz connected to XT1
> > pin. I have configured MCLK=250KHz and ACLK=1MHz.
> > I am using ACLK as source for watchdog counter.
> > > > > > > > What is the reason for watchdog
> > getting resetted? Is any other initialisation/
> > handling is required for the watchdog?
> > > > > > > >
> > > > > > > >
> > > > > > > > The datasheet for MSP430F2418 says
> > that "The WDT+ counter clock should be slower or
> > equal than the system (MCLK) frequency". Does
> > this mean that my clock source selection will not
> > work for watchdog? What combination I can use for
> > watchdog clock source? Has anyone faced any issue
> > other than this with the internal watchdog of MSP430?
> > > > > > > >
> > > > > > > > Your earliest help in this regard
> > will be highly appreciated. Thanks in advance.
> > > > > > > >
> > > > > > >
>
