Bug 2534

Summary: new acpi_wake_gpes_always_on kernel option didn't work
Product: ACPI Reporter: Hermann Lauer (Hermann.Lauer)
Component: Config-InterruptsAssignee: Len Brown (lenb)
Status: CLOSED PATCH_ALREADY_AVAILABLE    
Severity: normal CC: acpi-bugzilla
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.4.26 Subsystem:
Regression: --- Bisected commit-id:
Attachments: dmesg output (via syslog)

Description Hermann Lauer 2004-04-17 16:40:08 UTC
Distribution: Debian testing
Hardware Environment:
Software Environment:
Problem Description:

The new "acpi_wake_gpes_always_on" kernel option didn't work. Without these
the LID switch on my laptop did not work any more (has worked with
2.4.25 and apci patches)

As you can see from the attached dmesg output, the kernel option
is recognized:
Apr 17 01:14:53 localhost kernel: ACPI: wake GPEs not disabled

But later on, there are wake GPEs found:

Apr 17 01:14:53 localhost kernel: evgpeblk-0926 [06] ev_create_gpe_block   :
Found 2 Wake, Enabled 1 Runtime GPEs in this block

If I understand linux-2.4.26/drivers/acpi/events/evgpeblk.c correct,
alls gpe are initialised as ACPI_GPE_TYPE_RUNTIME and in line 879ff only if the
"acpi_wake_gpes_always_on" kernel option is not set the GPEs are checked if they
have to be changed to ACPI_GPE_TYPE_WAKE -- so with this option set there should
be no wake GPEs.

Just guessing: maybe acpi_ut_init_globals() is called after 
acpi_wake_gpes_always_on_setup() ?

Thanks for all the work on acpi linux code.
Comment 1 Hermann Lauer 2004-04-17 16:42:48 UTC
Created attachment 2617 [details]
dmesg output (via syslog)
Comment 2 Luming Yu 2004-04-26 20:49:54 UTC
We will release a improved wakeup/runtime GPE implementation soon. 
Thanks,
Luming
Comment 3 Len Brown 2004-11-14 20:46:41 UTC
This was fixed early on in 2.4.28: 
http://linux.bkbits.net:8080/linux-2.4/cset@411d887a5dK6KUykwxHohbB0vxJssw 
and is also available in the 24-stable-release ACPI patch for 2.4.27. 
 
closing.