kernel 2.4.20 & acpi-20021212-2.4.20.diff => no /proc/acpi

Tinkster · 03-23-2003, 07:09 PM

Hi Guys!

I was just wondering whether anyone here
managed not only to patch that acpi patch
against a stock 2.4.20 kernel (the patch worked
well for me, too) but could make any use of
it afterwards? To my greatest surprise the
result was that the subdirectories that used
to reside in /proc/acpi (e.g. /proc/acpi/battery)
ended up in /proc (e.g. /proc/battery) afterwards,
and the acpi-daemon refused to run, stating (and
rightfully so) that there's no /proc/acpi ...

If you succeeded, please tell me how :)

Cheers,
Tink

mcleodnine · 03-23-2003, 08:15 PM

If you're using acpid you can tel it to use a different event path. From the man page...

Quote:

-e, --eventfile filename
This option changes the event file from which acpid reads events. Default is
/proc/acpi/event.

Tinkster · 03-23-2003, 08:30 PM

Thanks for the response!

I don't utilize /proc/acpi/event (not interested in
suspending the box, or putting it to sleep), and am
primarily after reading battery states and thermal
information .... but I'll give the patched kernel
another try later :} and check whether tools for
gathering acpi information can happily live without
/proc/acpi ...

Cheers,
Tink

mcleodnine · 03-23-2003, 08:51 PM

Oh... battery info. Tried that on two notebooks so far and was unable to make it go. I could monitor all the buttons and timer events, but not the battery.

I think the way acpid works is that it _needs_ to read the /proc/acpi/battery as a change in battery status generates an event.

Tinkster · 03-23-2003, 08:59 PM

I decided to be brave, patched my kernel
to 2.4.21pre5 and applied the most recent
available acpi-patch for that one ...
make dep && make bzImage running as I type :)
this message ...

Cheers,
Tink

Tinkster · 03-24-2003, 01:56 PM

DOH ... same result. I wonder whether any
of the ACPI developers are going to react
to my question as to why this happens this
time around :(

Cheers,
Tink