Bugzilla Bug 148423

Kernel patch for 2.6.17 in app-laptop/hdaps doesn't apply to 2.6.18; ebuild
--config doesn't try to apply it.

Created an attachment (id=97581) [details]
hdaps_protect-2.6.18.patch

Current patch massaged to apply to 2.6.18. No idea if it's correct or not;
haven't even compiled it yet.

Fails to compile.
RQ_SCSI_DONE and RQ_SCSI_BUSY are no longer defined. We can easily redefine
them, but there are other issues with struct changes.

I'll monitor the hdaps lists for new patches, as I don't have the time for this
right now.

BTW, thanks for your effort :)

http://sourceforge.net/mailarchive/forum.php?thread_id=30578933&forum_id=45521

Heh, way beyond my scope. We probably won't see a 2.6.18 patch for some time :(

Created an attachment (id=97981) [details]
hdaps_protect-2.6.18.patch

Ugh. Here's a patch with the scsi stuff disabled, so it will only work on ide
disks.

Ed, you so are going to hate me ..... I cannot test this as my laptop disks
have to go through the SCSI interface to work correctly with DVD's and burning
CD's :(

Hopefully some other thinkpad user with IDE disks can test and comment here.

removing block on 148429: hdapsd is not in stable tree

Hi!
It really is a pain in the ass that the hard park patch is not in the official
kernel. Why? Isn't it to be considered as "generally a good idea"?

Anyway, there is also a newer hdapsd out there (look at
http://www.thinkwiki.org/wiki/HDAPS).
You can find it at: ftp://rohrmoser-engineering.de/pub/hdapsd-20061007.c

So, I applied your patch to the stable gentoo kernel: 2.6.18-gentoo-r2 and
compiled it. No errors so far, I will boot it and find out if the HDD is
actually parking on axis movement.

My system is a IBM ThinkPad R51 1830-DG4
(http://www.thinkwiki.org/wiki/1830-DG4) which is an IDE-only system.

Will reboot into 2.6.18-gentoo-r2-hdaps_patch soon...

I'm now running 2.6.18-gentoo-r2-hdaps_patch and it works fine so far. tp_smapi
compiled with USE="hdaps", hdaps modules loaded, hdapsd started, everything
looks good.

When I flip the entire notebook backwards (in direction to the display) I can
hear the HDD parking. It does that IMHO a little less often than it used to
with the 2.6.17 kernel.

Is there any other way to check this?

I hope that helped. BTW, keep up the great work! And thanks, Andreas.

(In reply to comment #8)
> When I flip the entire notebook backwards (in direction to the display) I can
> hear the HDD parking. It does that IMHO a little less often than it used to
> with the 2.6.17 kernel.
> 
> Is there any other way to check this?

If you're in KDE or GNOME, use khdapsmonitor, khdapsmon or gnome-hdaps-applet.
(An ebuild for the latter is in my overlay; see
http://sources.catmur.co.uk/viewcvs/svn/gentoo/tree/app-laptop/gnome-hdaps-applet/
.) Alternatively, just watch cat /sys/block/hda/queue/protect.

Oh, I almost forgot about that... I used khdapsmon. I just didn't think about
it anymore because I normally don't want the taskbar area filled up with icons. 

The HDD heads get parked when they should, so it is working.

Any idea if or when this patch will be available through "ebuild hdapsd
--config"?

(In reply to comment #10)
> Any idea if or when this patch will be available through "ebuild hdapsd
> --config"?

I won't put it in as I can only test with SCSI, which doesn't work anymore in
2.6.18. Only IDE works.

I don't have the time or the knowledge to make SCSI work as they seem to be
changing the SCSI subsystem a fair bit atm.

Ok, so now there is an _experimental_ patch for 2.6.18.3 that supports
SCSI/SATA as well as IDE.
My source of information: http://www.thinkwiki.org/wiki/HDAPS

The patch: http://article.gmane.org/gmane.linux.drivers.hdaps.devel/943

Hopefully that works for you all! (I'm very happy with the IDE-only version
right now...)

*** Bug 156171 has been marked as a duplicate of this bug. ***

-r1 added with patches for 2.6.18 (untested by me) and 2.6.19 (confirmed
working)