The RTC chip is a Xicor X1205 Real Time Clock/Calendar which is connected externally to the CPU via I2C. (see LinksToHardwareComponentDocumentation for data sheet)
There has been an unconfirmed report that the Xicor X1205 RTC chip has been replaced with a ST Microelectronics M41T11 RTC chip.
Kas11 has a working kernel module for the Xicor X1205 in the AnyU2 CVS repo.
The I2C address of the chip is 0x6F. The base register is 0x30. The status register is at 0x3F.
If reading the time via hwclock always returns the same date/time, it's possible your RTC clock is stalled. As suggested by Kas11, removing the battery for a few minutes and reinstalling it fixes the problem in some cases.
bbo: Error: root@LKG62CC02:~# date
Fri Dec 31 00:00:41 UTC 1999 root@LKG62CC02:~# hwclock -lw hwclock: Could not set the RTC time: Invalid argument root@LKG62CC02:~# hwclock Fri Dec 31 00:00:00 1999 0.000000 seconds root@LKG62CC02:~# hwclock Fri Dec 31 00:00:00 1999 0.000000 seconds
Solution: set date ! and write to RTC
root@LKG62CC02:~# date 011317402006 Fri Jan 13 17:40:00 UTC 2006 root@LKG62CC02:~# hwclock -lw root@LKG62CC02:~# hwclock Fri Jan 13 17:43:45 2006 0.000000 seconds root@LKG62CC02:~# hwclock Fri Jan 13 17:43:47 2006 0.000000 seconds
The Chip also has an /INT interrupt pin, which does not seem to be connected to other Hardware on the board. Maybe an alarm time could be set by software so that with hardware modification the slug could be woken up.
The RTC is capable of waking up the NSLU2, see http://johnarthur.wordpress.com/2008/06/26/a-battery-powered-time-lapse-camera/. According to the comments a Linux kernel from 2.6.26rc-9 on should contain the patch correcting an issue preventing setting the alarm. Hardware-modifications are shown there as well. But it seems to be that you need a special program to set the alarm clock, which you can find on the pade as well.
Has anybody tried to connect the /INT pin to the ForcePowerAlwaysOn-hack?