[vdr] Re: NVRAM WakeUp

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

On Fri, 19 Oct 2001, Arno Esser wrote:

> Yes, and what it takes it worther than that is the fact, that this data
> seems to be changing for every kernel versions. Yesterday I made some tests with
> linux 2.4.4, 2.4.7 and 2.4.10

2.4.4 had just some small part of nvram mapped into /dev/nvram
2.4.6 and newer have the bytes 14-127.
2.4.6 and 2.4.7 (dont know of 2.4.8) had a typo in nvram.c

#define RTC_FIRST_BYTE    14         /* RTC register number of first ... */
#define NVRAM_BYTES       128        /* number of NVRAM bytes */

instead of:

#define RTC_FIRST_BYTE    14         /* RTC register number of first ... */
#define NVRAM_BYTES  128-RTC_FIRST_BYTE /* number of NVRAM bytes */

havent tried 2.4.10 and later.


> (which btw runs very unstable on my machine and
> crashes the computer completely every half on hour). The dumps of all
> versions looks different. Either the complete size of the dump is different or the
> positions where the data seems to be stored. Especially the position ( where
> I presume it ) and value the checksum is complete different every time.

in 2.4.6 and 2.4.7 the last 14 bytes of /dev/nvram are mapped to the
realtime clock bytes. therefore they change every second.


c ya
        Sergei

--------------------------------------------------------------------
         eMail:       Sergei.Haller@math.uni-giessen.de
--------------------------------------------------------------------
Be careful of reading health books, you might die of a misprint.
                -- Mark Twain