Most recent kernel where this bug did *NOT* occur:
HP Compaq nc6320
Ubuntu, boot time
When booting using only battery power, the laptop often expereinces a softlock
on CPU#0. Plugging AC in and rebooting seems to be the only cure.
Steps to reproduce:
No AC adapter plugged in
Power up computer.
Bug seems to still occur if the AC is plugged in after initial powerup, but
before the softlock.
Does not appear to occur with either acpi=off, or apm=off.
Booting with acpi=off causes other problems though - see dmesg output for an
IRQ11 which noone cared about. Also, the sound card went slightly barmy and
repeated the ubuntu startup sound sample over and over.
Created attachment 10101 [details]
dmesg output from normal good boot
Created attachment 10102 [details]
dmesg output with acpi=off
Created attachment 10103 [details]
dmesg output with apm=off
Created attachment 10104 [details]
acpidump output from this laptop with a clean boot on ac.
Bug is actually still present with apm=off.
Please replace "quiet" on the cmdline with "debug"
boot on battery to reproduce the failure
and take a screen shot with a digital camera
and attach the .JPG
Also, Please try to reproduce the failure with CONFIG_CPU_FREQ=n
It turns out that this softlock is caused by the Intel Wireless driver, and the
bug is known:
If powered up with the Wireless radio switched off, the driver would spin trying
to get a response from some thermal sensor on the card, which appears to be
powered down with the radio off, and the laptop in battery mode.
Seems odd that it powers down different bits / behaves differently in battery
mode, but thats life.
How do I / we close this bug?