And here is a some more extreme cleanup:

It seems bigger as it is at first glance, however if you start to read 
it at ide.h, the rest should
be, well,  obivous...

Anyway please let me explain some bits: the end_request() function familiy
(not the global one, but the IDE specific ones), did bear a permuted 
parameter ordering.
After fixing this it turned out that at all places the huk parameter 
wasn't the
hwgroup, but just the drive in question itself. I have changed this to
be more sane, which allowed to remove many unneccessary code duplication,
or rather obfuscation, in between the __ide_end_request() and 
ide_end_request() functions.
This simplification is actually the "spreading" part of the game.

In a next step rather as much as possible should be moved from beeing 
hooked on
the disk to be hooked on the request itself - which seems more natural 
and would make
the overall code treamlined with other similar driver mid-layers (SCSI 
comes to mind).

In a third step one should take care to remove the excessive usage of 
single linked
lists inside the ide driver, where possible and make it be handled the 
same way like
in SCSI and elsewere in the kernel... The overall perspecive is to make 
as much as possible
common between all block device handlers no matter whatever SCSI/IDE/I2O 
or FW or iSCSI.

I have taken as much care as possible to not break anything. In esp. it 
all has ben tested
in life on my home system with L120 (aka ide-floppy), an CD-RW, and two 
disks.
The only blind fly is ide-tape... but the patch can be easly verifyed 
for correctness by reading
through it.

PS. I have killed deliberately some one-shoot functions as well, since 
those where only
obscuring the overall code structure even more....