Problem in XScale exiting from sleep mode after a fault signal assert

Hi all,

I'm working on a DBPXA255 based platform and Windows CE 4.1.

When I assert one of the BATT_FAULT or VDD_FAULT signals, the
microprocessor enters in sleep mode correctly. But when I deassert
that fault signal and use a wake up source (GPIO1 for instance) the
power consumption goes a bit high and it seems that is trying to wake
up from sleep, but I cannot see the Windows CE image working another
time. It has to be a software problem, but my doubt is:

if the software was working properly, would the execution continue in
the same point that it was when the fault signal was asserted???

What could be this problem's source???

Thanks in advance,

Lur

Jim
Mon Feb 09 13:52:28 CST 2004

Any time the PXA255 enters sleep, the processor itself is basically powered
down, and begins execution from the reset vector when it is brought out of
sleep. Early, within the boot code, the Reset Controller Status Register
(RCSR @ 0x40F00030) is examined to determine the type of reset. After that
point, alternate paths can be followed. A cold boot (HWR bit set) requires
full initialization of the platform. If the SMR bit is set, a full register
restore is required from the register save area. (This assumes all the
registers were saved and the SDRAM was put into self refresh mode prior to
entering sleep.)

This whole process is not for the faint of heart......

Jim Merkle


"lur" <lur444@hotmail.com> wrote in message
news:55dc9062.0402090902.24ba0822@posting.google.com...
> Hi all,
>
> I'm working on a DBPXA255 based platform and Windows CE 4.1.
>
> When I assert one of the BATT_FAULT or VDD_FAULT signals, the
> microprocessor enters in sleep mode correctly. But when I deassert
> that fault signal and use a wake up source (GPIO1 for instance) the
> power consumption goes a bit high and it seems that is trying to wake
> up from sleep, but I cannot see the Windows CE image working another
> time. It has to be a software problem, but my doubt is:
>
> if the software was working properly, would the execution continue in
> the same point that it was when the fault signal was asserted???
>
> What could be this problem's source???
>
> Thanks in advance,
>
> Lur