Bug#113036: not a bug
I just read this report, and I do not believe it is a valid bug. My
reasoning is that the 3ware card does not present an IDE interface to
the OS, so the OS should not be expected to emulate one.
I also agree that it is not an hdparm bug either.
hdparm's description says:
Get/set hard disk parameters for Linux IDE drives.
Primary use is for enabling irq-unmasking and IDE multiplemode.
Because you are going through a 3ware controller, your disks are no
longer Linux IDE drives - they are now 3ware logical drives, which
hdparm does not support.
If you want to tweak low-level settings of the disk, you should use a
controller that gives you this interface.
An analogy is asking for a USPS (US Postal Service) letter to be
delivered via bicycle instead of via truck. The USPS interface is a
mailbox at both ends, and the USPS gets to choose how its
implemented. If you insist on bicycle deliver, hire a bicycle courier
instead.
For a second opinion, I ran this by Adam Radford who maintains the
3ware driver:
Adam Radford said:
> The 3ware controllers purposefully don't allow someone to change the
> UDMA settings, etc, and as such, this functionality is not
> implemented through ioctls, and hdparm will not apply to these
> controllers in that regard. (N/A): Not applicable.
Because the controllers don't allow this - and this hasn't changed in
the 5 years this bug has been open, I think its time to close it.
--
dann frazier
Reply to: