Distribution: SuSE 8.1
Hardware Environment: Laptop Acer TM630
Software Environment: sysvinit-2.82-187
Problem Description: Till 2.4.24, respectively 2.6.2, halt -p -d -f did kill
off the power of the laptop. With 2.4.25, as well as 2.6.3 it no longer does.
If I do an echo 0xffff > /proc/acpi/debug_level before shuting down, I got
lots of messages, afer the halt-command. It seems to be a loop, but it scrolls
so fast, that I could not read any of them.
cow:~ # cat /proc/acpi/info
states: S0 S3 S4 S5
Steps to reproduce: shutdown -h now
After further digging I found out, that starting with a vanilla 2.4.24 kernel
(which by itself works as it should), applying acpi-20031203-2.4.24.diff.bz2
introduces the regression. If I have /proc/acpi/debug_level set to 7, it
reports me after the "power down." message the following:
hwsleep-0265  acpi_enter_sleep_state: Entering sleep state [S5]
After that nothing more happens. The machine has to be cut power by pressing
the power-button >4sec.
looks like the same regression spotted in bug 2109
*** This bug has been marked as a duplicate of 2109 ***