[vdr] Re: Again: wakeup on VIA EPIA board

[Sergei Haller <Sergei.Haller@math.uni-giessen.de>]

On Sun, 22 Dec 2002, Harald Milz (HM) wrote:

HM> Setting the wakeup time in the BIOS setup results in a correct wakeup -
HM> but:
HM>
HM> - nvram-wakeup doesn't work because although the wakeup time is in bytes
HM>   0x72 and up it seems static, and no wakeup day there.
HM> - the set_timer script from the nvram-wakeup CVS works somehow but is
HM>   utterly confusing because it fiddles with the system time which isn't
HM>   always reset correctly on boot. And it needs to keep RTC wakeup enabled
HM>   permanently, and if the system time isn't corrent the machine will wake
HM>   up inadvertently from time to time.
HM> - the RTC wakeup time (in /proc/drivers/rtc) appears to be a constant
HM>   (23:59:59, and no wakeup day there). This time must be there from an
HM>   earlier manual setting but it doesn't change whatever I do.

the confusing thing is that there are (at least) three different alarms:
(please correct me if I'm wrong)
 * rtc alarm (rises a soft irq at that time)
 * mainboard wakeup for ATX boards (wakes up from the off state, that
   means after you shutdown your machine)
 * acpi alarm (don't really know more about this one)

the first one can be used for a wakeup from Suspend-To-Ram (APM)
the third one can be used for the same in ACPI mode.

the even more confusing thing is that some boards use the same bytes in
NVRAM for the first and the second alarms.

The reason why there's no date in nvram is simple: nvram on modern boards
is at least 256 bytes large (or even larger). The aderssing via IO ports
(the way nvram driver does this) can only address 128 of them -- so if
the manufacturer stores the date and or time of the wakeup in the upper
part of the nvram, we're lost.

--> Does someone know how to read the whole nvram independant of the
    board?


HM> - the SuSE 8.0 and 8.1 kernels do have some ACPI but writing the wakeup
HM>   time to /proc/acpi/alarm always results in a freeze of the machine.

don't know if this is implemented -- try newest kernels with the newest
acpi patch.


HM> - has somebody found out whether and how wake-on-LAN works? There is no
HM>   apparent BIOS setting. My fileserver runs 24x7 so that could be a
HM>   workaround.

I think, this would be the easiest solution for you.

HM> - do I have to write a small program to set the RTC wakeup time ... ?

this would be an 24h alarm -- that means your vdr machine would wake up
once a day


c ya
        Sergei
-- 
--------------------------------------------------------------------  -?)
         eMail:       Sergei.Haller@math.uni-giessen.de               /\\
-------------------------------------------------------------------- _\_V
Be careful of reading health books, you might die of a misprint.
                -- Mark Twain



-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe vdr" as subject.