Comment 6 for bug 172282

I recently reinstalled my toshiba p100 laptop from Dapper to Hardy... Recently, I also changed my HDD for a WD Scorpio 320Gb.

This configuration was working nicely under dapper, with both original drive (fujitsu 100Gb) and new one I used a few days before reinstall.

Under Hardy, I was surprised to hear a new HDD tick almost every minute, even on AC. smartmon proved I was due to heads parking very often.

Here are the results of my experiences:
Hardy defaults to a -B 128 even on AC. It's nonsense. On some drives (like my WD scorpio), it leads to excessive head parks. On some (my office HP nc6400 Seagate momentus 80Gb for instance), It does not.

So, to have something working with ALL HDD, hdparm sould default to -B 255, which effectively seems to switch off drive power management (with 254, the WD Scorpio still heavily park heads), at least on AC.

On battery, value 128 is also irrelevant as it does not permit spin down to save power! So we have excessive parks without huge spin down power saving!

Proposed fixes (scripts in /etc/acpi/*.d with appropriate hdparm line) work... until you resume from suspend to ram! It seems acpi override script settings after their execution.

This bug lasts since several kernel revisions, even if I discovered it recently. It should really be solved in hurry and considered really critical because a user that will lost his HDD becaus of linux is really not likely to use it again.

Regards.