Real-Time Clock problem

During CE boot, I found DoGeneralInit function in filesys module try to set
the RTC
to 6/1/1999, 12:00. However, I did not find the source file fsmain.c
required by debugger for source level debugging.
How can I prevent filesys from setting RTC?

Paul
Mon Aug 29 15:54:59 CDT 2005

That's not what you want to do. What processor are you working with? Are
you implementing handling of the INIT RTC command that happens during kernel
startup? That's where you'd really want to do something different, if
you're running, say on a PXA255, which doesn't have a battery-backed
real-time clock, but on a platform where you *do* have one. You'd also have
to implement OEMSetRealTime, OEMGetRealTime, and OEMSetAlarmTime, maybe, to
handle that case.

It would be worth telling us what version of the OS you're talking about,
too!

Paul T.

"Ben" <Ben@discussions.microsoft.com> wrote in message
news:B65049A3-51AF-417A-BEA2-72F1316FBED5@microsoft.com...
> During CE boot, I found DoGeneralInit function in filesys module try to
> set
> the RTC
> to 6/1/1999, 12:00. However, I did not find the source file fsmain.c
> required by debugger for source level debugging.
> How can I prevent filesys from setting RTC?

Ben
Tue Aug 30 12:14:02 CDT 2005

Hi Paul,

The processor I am using is Cirrus EP9315. It has RTC w/o battery backup.
So, I have a battery backed dallas rtc chip on board. The OEM functions you
mentioned are came with board support package. I modified them so that
OEMGetRealTime gets time from rtc chip on boot, and OEMSetRealTime updates
the time of rtc chip. The only problem is DoGeneralInit function in filesys
module try to set the RTC to 6/1/1999, 12:00 when filesys module is loaded on
boot. I am using CE 4.2 with all updates installed.

Thanks

"Paul G. Tobey [eMVP]" wrote:

> That's not what you want to do. What processor are you working with? Are
> you implementing handling of the INIT RTC command that happens during kernel
> startup? That's where you'd really want to do something different, if
> you're running, say on a PXA255, which doesn't have a battery-backed
> real-time clock, but on a platform where you *do* have one. You'd also have
> to implement OEMSetRealTime, OEMGetRealTime, and OEMSetAlarmTime, maybe, to
> handle that case.
>
> It would be worth telling us what version of the OS you're talking about,
> too!
>
> Paul T.
>
> "Ben" <Ben@discussions.microsoft.com> wrote in message
> news:B65049A3-51AF-417A-BEA2-72F1316FBED5@microsoft.com...
> > During CE boot, I found DoGeneralInit function in filesys module try to
> > set
> > the RTC
> > to 6/1/1999, 12:00. However, I did not find the source file fsmain.c
> > required by debugger for source level debugging.
> > How can I prevent filesys from setting RTC?
>
>
>

Paul
Tue Aug 30 12:19:15 CDT 2005

Your own BSP must be triggering that behavior. Add debug messages to the
OEMGet and OEMSet functions, as well as to the OEMIoControl handler and see
what's happening and when.

By the way, unless you also handle the set-alarm-time function and its
interrupts, daylight saving time transitions, as well as other time-based
things that the system is expecting to have work, won't. That's something
that you might not think of during testing, but you should fix it now,
either by arranging for the OEMSetAlarmTime to use the RTC chip alarms, or
by trying to keep the processor RTC synced with the external RTC chip and
using the processor's RTC alarms.

Paul T.

"Ben" <Ben@discussions.microsoft.com> wrote in message
news:0DDB46B2-8655-4FD4-939F-6D80C205E7AD@microsoft.com...
> Hi Paul,
>
> The processor I am using is Cirrus EP9315. It has RTC w/o battery backup.
> So, I have a battery backed dallas rtc chip on board. The OEM functions
> you
> mentioned are came with board support package. I modified them so that
> OEMGetRealTime gets time from rtc chip on boot, and OEMSetRealTime updates
> the time of rtc chip. The only problem is DoGeneralInit function in
> filesys
> module try to set the RTC to 6/1/1999, 12:00 when filesys module is loaded
> on
> boot. I am using CE 4.2 with all updates installed.
>
> Thanks
>
> "Paul G. Tobey [eMVP]" wrote:
>
>> That's not what you want to do. What processor are you working with?
>> Are
>> you implementing handling of the INIT RTC command that happens during
>> kernel
>> startup? That's where you'd really want to do something different, if
>> you're running, say on a PXA255, which doesn't have a battery-backed
>> real-time clock, but on a platform where you *do* have one. You'd also
>> have
>> to implement OEMSetRealTime, OEMGetRealTime, and OEMSetAlarmTime, maybe,
>> to
>> handle that case.
>>
>> It would be worth telling us what version of the OS you're talking about,
>> too!
>>
>> Paul T.
>>
>> "Ben" <Ben@discussions.microsoft.com> wrote in message
>> news:B65049A3-51AF-417A-BEA2-72F1316FBED5@microsoft.com...
>> > During CE boot, I found DoGeneralInit function in filesys module try to
>> > set
>> > the RTC
>> > to 6/1/1999, 12:00. However, I did not find the source file fsmain.c
>> > required by debugger for source level debugging.
>> > How can I prevent filesys from setting RTC?
>>
>>
>>

Bruce
Wed Aug 31 07:50:19 CDT 2005


Are you actually starting the RTC, they don't necessarily do that
automatically. Are you reading from it successfully? Are you resetting the
internal RTC based on the date read from the external RTC?

--
Bruce Eitman (eMVP)
Senior Engineer
beitman AT applieddata DOT net

Applied Data Systems
www.applieddata.net
An ISO 9001:2000 Registered Company
Microsoft WEP Gold-level Member


"Ben" <Ben@discussions.microsoft.com> wrote in message
news:0DDB46B2-8655-4FD4-939F-6D80C205E7AD@microsoft.com...
> Hi Paul,
>
> The processor I am using is Cirrus EP9315. It has RTC w/o battery backup.
> So, I have a battery backed dallas rtc chip on board. The OEM functions
> you
> mentioned are came with board support package. I modified them so that
> OEMGetRealTime gets time from rtc chip on boot, and OEMSetRealTime updates
> the time of rtc chip. The only problem is DoGeneralInit function in
> filesys
> module try to set the RTC to 6/1/1999, 12:00 when filesys module is loaded
> on
> boot. I am using CE 4.2 with all updates installed.
>
> Thanks
>
> "Paul G. Tobey [eMVP]" wrote:
>
>> That's not what you want to do. What processor are you working with?
>> Are
>> you implementing handling of the INIT RTC command that happens during
>> kernel
>> startup? That's where you'd really want to do something different, if
>> you're running, say on a PXA255, which doesn't have a battery-backed
>> real-time clock, but on a platform where you *do* have one. You'd also
>> have
>> to implement OEMSetRealTime, OEMGetRealTime, and OEMSetAlarmTime, maybe,
>> to
>> handle that case.
>>
>> It would be worth telling us what version of the OS you're talking about,
>> too!
>>
>> Paul T.
>>
>> "Ben" <Ben@discussions.microsoft.com> wrote in message
>> news:B65049A3-51AF-417A-BEA2-72F1316FBED5@microsoft.com...
>> > During CE boot, I found DoGeneralInit function in filesys module try to
>> > set
>> > the RTC
>> > to 6/1/1999, 12:00. However, I did not find the source file fsmain.c
>> > required by debugger for source level debugging.
>> > How can I prevent filesys from setting RTC?
>>
>>
>>