Hi!
LKML
I also suffer the same problem with my Slackware 12 installation and generic-smp kernel. (no problems with WinXP)
My laptop has an Hitachi 120GB SATA drive.

I usually hear that noise when the laptop completely freezes and I have to push the power button (why laptops doesn't have reset button?!) or remove the battery. I know that this is dangerous but I do this twice a year.

Well, this bug(#7674) has been fixed in 2.6.22 kernel versions.
Do you notice that noise in your laptops?
Which kernel are you using?
I'm going to compile and install the latest kernel tomorrow or later.

Thanks in advance.
P.D.:Sorry for my bad english

EDIT: Maybe this is not the best section for this thread...sorry , sometimes I forget that other sections exist xD

The only thing I can think of is that you aren't letting Slack completely shut down before turning the power off and thus the emergency park. Do you notice any unusual logging when you restart? Does Slack force you to check the drive for errors upon reboot?

Yes, strange ... I have same problem on my laptop Acer Aspire 5100 , Slackware 12 default kernel and normal shutdown procedure shutdown -h now, I hear that notorious noise click Slack doesn't me force to check HD for errors upon reboot ...

Thekid: I let laptop completely shutdown, the shutdown process is OK, disable pcmcia services, unmount media...then...CLACK.
So If I use Slackware I have to reboot and shutdown in windows.

And as pego said Slack doesn't check any partition for errors (ext3) upon reboot.

Look at this, from the launchpad.net thread:

Quoted from an Hitachi pdf
This is a critical kernel bug that has been solved in recent kernel versions. I haven't got time to update it now.

Well, it seems you know what the solution is, upgrading the kernel. If this is a machine you use regularly, my suggestion would be to make the time for the upgrade before you end up having to buy a new HDD.

Hi,

Most modern hard disk subsystems provide protective head parking as a feature that adds another layer of protection to auto park by safely parking the heads even if other system components lose functionality. This does provide a means by the integrated drive controls to safely park the heads of the subsystem.

In the older drive subsystems(RLL, MFM type drives) we would have too issue a 'park' command which was a small program that would park the heads in a known safe area.

With modern HD we don't really have that issue any longer. With the 2.6 kernel unload problem, yes there are ways to prevent this from occurring. If you are using 'APS' or even some other protective scheme for the laptop then some noise can occur from the park.

You can check the .config of your kernel for the 'HDAPS' by 'cat /usr/src/linux/.config |grep -i aps' to see if there is a module or compiled in.

I personally don't want to introduce another level, since most subsystems already take care of the concern.

This is result from command "cat /usr/src/linux/.config |grep -i aps" on my system:

CONFIG_DM_SNAPSHOT=m
CONFIG_SENSORS_HDAPS=m

I suppose that first line is irrelevant ... sorry onebuck if I misunderstand you but does your post mean that there is no reason for panic? And thanks for your help.

Hi,

I'm not sure which model of SATA drive you are using. If I had the drive model then more could be found out. Generally the SATA has a great control function. The head safety features are about the same.

If you just hear a set or click that is the autopark or some say unload. But if you hear repeated actions then that is when I would worry. You can tell when that occurs.

I'm sure the Hitachi drive that you are using does have this feature. If the the drive subsystems should fail then the feature will perform as designed.

I would not panic. If you look at the data then that should help. The load/unload cycle is specified at 600,000.

You can also look into the kernel options.