Brenda*n,*
thanks. It's really good to here that. Even if 99.999% of the problems
comes from the code we write ourselves it is still nice to take away
that 0.0001% that comes from problems in hardware.
I'm in the process of doing a new test project at the moment to try to
trace this down under a more controlled environment.
Thanks for taking the time to sharing this info. Much appreciated!
Cheers
/Ake
brendanmurphy37 wrote:
>
> Ake,
>
> Thanks for your comments.
>
> I'm not sure if you want to hear this or not, but we've just put in
> support for the watchdog. It worked first time, with no problems.
> Code is below: two functions, one called as part of startup sequence,
> the other called from a watchdog task that's scheduled every second.
> If we spin in a loop locking out all other tasks, the system is reset
> as expected.
>
> Our system has several sources of interrupt enabled (both UARTs, I2C
> etc.): it works regardless of what's happeneing elsewhere.
>
> By the way, we use GCC 3.3.1 with -O2 to compile.
>
> Regards
> Brendan
>
> =========================================
> CODE STARTS HERE:
>
> #define REG(addr) (*(volatile unsigned int *)(addr))
>
> #define WDOG_BASE (0xE0000000)
> #define WDOG_WDMOD (0xE0000000)
> #define WDOG_WDTC (0xE0000004)
> #define WDOG_WDFEED (0xE0000008)
> #define WDOG_WDTV (0xE000000C)
>
> /* define basic watchdog timeout value as 100 ms */
>
> #define WATCH_TIMER_100MS (250000) /* 10 MHz / 4 is i/p clock */
>
> /*
> * Configure and start watchdog
> */
>
> static
> void watchdog_init(void)
>
> {
> /* set the watchdog timer constant */
>
> REG(WDOG_WDTC) = WATCH_TIMER_100MS * (50); /* 5 seconds for now */
>
> /* set mode to reset on underflow */
>
> REG(WDOG_WDMOD) = 0x3;
>
> /* feed watchdog to start it off */
>
> REG(WDOG_WDFEED) = 0xaa;
> REG(WDOG_WDFEED) = 0x55;
>
> return;
> }
>
> /*
> * Feed the watchdog to prevent us from being reset
> */
>
> void HwWatchdogFeed(void)
>
> {
> /* write the magic sequence to keep the dog happy */
>
> REG(WDOG_WDFEED) = 0xaa;
> REG(WDOG_WDFEED) = 0x55;
>
> return;
> }
>
> END OF CODE
> ========================================
>
> --- In lpc2000@yahoogroups.com, "Ake Hedman, eurosource" <akhe@b...>
> wrote:
> >
> > Thanks Brendan,
> >
> > This is the right approach surely. I'm at that state now and have
> no
> > other alternative then to back a bit and go through this "the right
> > way". I just have myself to blame. Have used watchdogs for many
> years on
> > PIC, AVR and other processors and expected no problem there except
> for
> > the usual forgetting to feed it in certain places. Oh well it's my
> first
> > project on this processor family so this makes me learn useful
> stuff
> > just as you write (just a bit hard to appreciate that right now ;-
> ) ).
> > Really need the watchdog to work in this application though so
> there is
> > no shortcut.
> >
> > I just debug using GPIO pins and one of the serial ports at the
> moment
> > so there is no emulator connected now.
> >
> > Thanks for your tips, suggestions and moral support. I report my
> > findings back.
> > /Ake
> >
> >
> > brendanmurphy37 wrote:
> >
> > >
> > > Ake,
> > >
> > > I can almost feel your frustration: we've all been there!
> > >
> > > I'm very interested in the outcome of this, as our own task for
> next
> > > week is to get the watchdog working on our own LPC2134-based
> system.
> > >
> > > I'm a bit surprised that no one has volunteered some working code.
> > > Surely someone out there is using this?
> > >
> > > Unfortunately, as we haven't got there yet, I've nothing specific
> to
> > > suggest, other than a general strategy of getting something simple
> > > working first, and then building up to where you are. For example:
> > >
> > > 1. As simple as possible startup, initialisation (all peripherals
> > > disabled, all interrupts off etc.) and an application that
> > > initialises the watchdog and uses GPIO to signal what it's doing
> (and
> > > maybe reads to indicate whether or not the watchdog should be
> fed).
> > > In other words, the simplest possible program with a working
> > > watchdog.
> > >
> > > 2. Same program, but with your normal initialisation and setup
> code,
> > > added incrementally. Still working?
> > >
> > > 3. Compare and contrast with your real application, particularly
> in
> > > how peripherals, interrupts etc. are managed.
> > >
> > > In other words, get something simple working first, and build up
> from
> > > there. It'll take time, but maybe less time then starting from a
> > > larger system that doesn't work.
> > >
> > > Hopefully, somewhere along the line, all will become obvious.
> > >
> > > By the way, are you running your code using an emulator connected?
> > > I've seen strange behaviour in the past with watchdogs enabled.
> Try
> > > running stand-alone, if you haven't done so already.
> > >
> > > As a final suggestion, based on what you say below, I'd look very
> > > carefully at every location interrupts are enabled/disabled (at
> the
> > > peripheral level, at the VIC and at the CPSR). Also at your
> startup
> > > and IRQ dispatch code. Maybe the processor is resetting and re-
> > > initialising stuff without you realising? maybe throwing an
> exception
> > > you haven't noticed?
> > >
> > > As a final comment: 'though no doubt it feels like it, it's not
> time
> > > wasted: you'll know way more about the watchdog and processor at
> the
> > > end of all this then when you started.
> > >
> > > Good luck!
> > >
> > > Brendan
> > >
> > >
> > > --- In lpc2000@yahoogroups.com, "Ake Hedman, eurosource"
> <akhe@b...>
> > > wrote:
> > > >
> > > > WD story update
> > > >
> > > > I changed the code to have the watchdog raising an IRQ instead
> of a
> > > > RESET. I initialize stuff and then go in in a while loop.
> > > >
> > > > 1.) With interrupts enabled but no interrupt channel active this
> > > will
> > > > not trigger the watchdog. Why?
> > > >
> > > > 2.) With one of the interrupt channels active ( UART0 writing
> some
> > > > characters ) the watchdog works. Without a dog feed the
> watchdog is
> > > > triggered and with a feed it runs fine. However what I
> initialize
> > > WDTC
> > > > with does to seem to matter. I always get about 4 ms period
> when I
> > > > toggle a pin in the WD interrupt (60MHz clock).
> > > >
> > > > Is this behaviour recognized by anyone? Please!
> > > >
> > > > BTW the chip is LPC2138
> > > >
> > > > /Ake
> > > >
> > > >
> > > > dr_danish_ali wrote:
> > > >
> > > > > Ok Ake, here's my dumb suggestion:
> > > > >
> > > > > Try enabling the watchdog, but just to interrupt not reset?
> > > > > You'll need an interrupt handler / VIC channel to do it - can
> you
> > > > > spare one?
> > > > > That ISR need just raise a flag for now.
> > > > >
> > > > > Once the watchdog is fed and running happily, you might then
> be
> > > able
> > > > > to enable resets on
> > > > > it.
> > > > >
> > > > > Hope this helps,
> > > > > Danish
> > > > >
> > > > > --- In lpc2000@yahoogroups.com, "Ake Hedman, eurosource"
> > > <akhe@b...>
> > > > > wrote:
> > > > > >
> > > > > > I am still trying to solve my problems with the watchdog.
> > > > > >
> > > > > > I have now scaled away most stuff of my application and can
> see
> > > that
> > > > > the
> > > > > > problem occurs when either of three interrupts occur
> > > (UART0/UART1/I2C0)
> > > > > > in the system. Without the watchdog everything works fine
> but
> > > if I
> > > > > > enable the watchdog everything crashes. I have tried to just
> > > enable and
> > > > > > trig one of the interrupts in turn but the situation is the
> > > same.
> > > > > >
> > > > > > I'm totally out of clues at the moment so any (also things
> that
> > > > > might be
> > > > > > considered dumb) are very welcome.
> > > > > >
> > > > > > Regards
> > > > > > /Ake
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > SPONSORED LINKS
> > > > > Microprocessor
> > > > > <http://groups.yahoo.com/gads?
> > >
> t=ms&k=Microprocessor&w1=Microprocessor&w2=Microcontrollers&w3=Pic+mic
> > > rocontrollers&w4=8051+microprocessor&c=4&s=93&.sig=tsVC-
> > > J9hJ5qyXg0WPR0l6g>
> > > > > Microcontrollers
> > > > > <http://groups.yahoo.com/gads?
> > >
> t=ms&k=Microcontrollers&w1=Microprocessor&w2=Microcontrollers&w3=Pic+m
> > >
> icrocontrollers&w4=8051+microprocessor&c=4&s=93&.sig=DvJVNqC_pqRTm8Xq0
> > > 1nxwg>
> > > > > Pic microcontrollers
> > > > > <http://groups.yahoo.com/gads?
> > >
> t=ms&k=Pic+microcontrollers&w1=Microprocessor&w2=Microcontrollers&w3=P
> > >
> ic+microcontrollers&w4=8051+microprocessor&c=4&s=93&.sig=TpkoX4KofDJ7c
> > > 6LyBvUqVQ>
> > > > >
> > > > > 8051 microprocessor
> > > > > <http://groups.yahoo.com/gads?
> > >
> t=ms&k=8051+microprocessor&w1=Microprocessor&w2=Microcontrollers&w3=Pi
> > >
> c+microcontrollers&w4=8051+microprocessor&c=4&s=93&.sig=1Ipf1Fjfbd_HVI
> > > lekkDP-A>
> > > > >
> > > > >
> > > > >
> > > > > --------------------------------------------------------------
> ----
> > > ------
> > > > > YAHOO! GROUPS LINKS
> > > > >
> > > > > * Visit your group "lpc2000
> > > > > <http://groups.yahoo.com/group/lpc2000>" on the web.
> > > > >
> > > > > * To unsubscribe from this group, send an email to:
> > > > > lpc2000-unsubscribe@yahoogroups.com
> > > > > <mailto:lpc2000-unsubscribe@yahoogroups.com?
> > > subject=Unsubscribe>
> > > > >
> > > > > * Your use of Yahoo! Groups is subject to the Yahoo!
> Terms of
> > > > > Service <http://docs.yahoo.com/info/terms/>.
> > > > >
> > > > >
> > > > > --------------------------------------------------------------
> ----
> > > ------
> > > > >
> > > >
> > > >
> > > > --
> > > > ---
> > > > Ake Hedman (YAP - Yet Another Programmer)
> > > > eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
> > > > Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
> > > > Company home: http://www.eurosource.se
> > > > Kryddor/Te/Kaffe: http://www.brattberg.com
> > > > Personal homepage: http://www.eurosource.se/akhe
> > > > Automated home: http://www.vscp.org
> > > >
> > > >
> > > >
> > > > [Non-text portions of this message have been removed]
> > > >
> > >
> > >
> > >
> > >
> > >
> > > ------------------------------------------------------------------
> ------
> > > YAHOO! GROUPS LINKS
> > >
> > > * Visit your group "lpc2000
> > > <http://groups.yahoo.com/group/lpc2000>" on the web.
> > >
> > > * To unsubscribe from this group, send an email to:
> > > lpc2000-unsubscribe@yahoogroups.com
> > > <mailto:lpc2000-unsubscribe@yahoogroups.com?
> subject=Unsubscribe>
> > >
> > > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
> > > Service <http://docs.yahoo.com/info/terms/>.
> > >
> > >
> > > ------------------------------------------------------------------
> ------
> > >
> >
> >
> > --
> > ---
> > Ake Hedman (YAP - Yet Another Programmer)
> > eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
> > Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
> > Company home: http://www.eurosource.se
> > Kryddor/Te/Kaffe: http://www.brattberg.com
> > Personal homepage: http://www.eurosource.se/akhe
> > Automated home: http://www.vscp.org
> >
> >
> >
> > [Non-text portions of this message have been removed]
> >
>
>
>
>
>
> ------------------------------------------------------------------------
> YAHOO! GROUPS LINKS
>
> * Visit your group "lpc2000
> <http://groups.yahoo.com/group/lpc2000>" on the web.
>
> * To unsubscribe from this group, send an email to:
> lpc2000-unsubscribe@yahoogroups.com
> <mailto:lpc2000-unsubscribe@yahoogroups.com?subject=Unsubscribe>
>
> * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
> Service <http://docs.yahoo.com/info/terms/>.
>
>
> ------------------------------------------------------------------------
>
--
---
Ake Hedman (YAP - Yet Another Programmer)
eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
Company home: http://www.eurosource.se
Kryddor/Te/Kaffe: http://www.brattberg.com
Personal homepage: http://www.eurosource.se/akhe
Automated home: http://www.vscp.org
[Non-text portions of this message have been removed]Message
Re: [lpc2000] Re: Problem with watchdog
2005-12-08 by Ake Hedman, eurosource