harddisk type not recognized with hdparm

Bug #21585 reported by Frank Niedermann
6
Affects Status Importance Assigned to Milestone
linux-source-2.6.15 (Ubuntu)
Invalid
Low
Ben Collins

Bug Description

I'm running Breezy on my system (updated today) and I've noticed that the
harddisk LED is always on / in use, also if there is no activity on the harddisk
at all.

System is a Shuttle XPC SB95P with Intel i925X and ICH6R/ICH6RW SATA Controller
with a Maxtor SATA disk.

I can use the disk (root partition is on the disk) but hdparm gives me an error:
fbn@paul:~$ mount
/dev/sda5 on / type ext3 (rw,errors=remount-ro)
fbn@paul:~$ hdparm -i /dev/sda
/dev/sda:
HDIO_GET_IDENTITY failed: Inappropriate ioctl for device

It's the only harddisk available for this system so I can not test another one.

Revision history for this message
Matt Zimmerman (mdz) wrote :

Please attach dmesg output

Revision history for this message
Frank Niedermann (fbn) wrote :
Revision history for this message
Matt Zimmerman (mdz) wrote :

Please attach lspci output as well.
The disk is working fine otherwise, right?

Revision history for this message
Frank Niedermann (fbn) wrote :

(In reply to comment #3)
> Please attach lspci output as well.
> The disk is working fine otherwise, right?

I think the disk is working fine as I can use it like I did with Hoary or WinXP
but I don't know if the settings for the disk are okay (DMA for example) because
hdparm doesn't work.

I'm a bit concerned about the situation, don't want to damage hardware ...

Revision history for this message
Frank Niedermann (fbn) wrote :

Created an attachment (id=3807)
lspci -v

Revision history for this message
Frank Niedermann (fbn) wrote :

(In reply to comment #3)
> Please attach lspci output as well.
> The disk is working fine otherwise, right?

Matt, although the harddisk is working fine I'm not sure if the behavior could
be damaging the hardware.

As I know the harddisk LED is hardware controlled, the harddisk may be in
writing state all the time Hoary is running ...

Is there a way to identify the issue?

Revision history for this message
Matt Zimmerman (mdz) wrote :

You should be able to hear the disk seeking if it is actually writing. You
would surely notice the system feeling sluggish as well. You can add an I/O
meter to the panel also, as an additional check

Revision history for this message
Matt Zimmerman (mdz) wrote :

Please use bug #21852 to discuss the LED problem, and this one only for the
hdparm issue.

I assume that on a different kernel, hdparm -i was working for you? Which one?

Revision history for this message
Frank Niedermann (fbn) wrote :

The default kernel with Ubuntu Hoary worked but I don't know which version it
was as I switched completely to Breezy

Revision history for this message
Matt Zimmerman (mdz) wrote :

(In reply to comment #9)
> The default kernel with Ubuntu Hoary worked but I don't know which version it
> was as I switched completely to Breezy

Unless you explicitly purged the old kernel, it is still installed and available
from the grub menu for testing.

Revision history for this message
Frank Niedermann (fbn) wrote :

(In reply to comment #10)
> Unless you explicitly purged the old kernel, it is still installed and available
> from the grub menu for testing.

Unfortunately not because I did a brand new installation at the time Breezy
testing CDs came out. But I still have Hoary installed on my notebook, updated
yesterday and kernel is 2.6.10-5-686 package
ii linux-image-686 2.6.10-7 Linux kernel image on PPro/Celeron/PII/PIII/

Revision history for this message
Frank Niedermann (fbn) wrote :

hdparm -i does not work:
/dev/sda:
 HDIO_GET_IDENTITY failed: Inappropriate ioctl for device

but hdparm -I does:
/dev/sda:

ATA device, with non-removable media
        Model Number: Maxtor 6Y160M0
        Serial Number: Y457PW3E
        Firmware Revision: YAR51KW0
...

on Breezy final :)

Revision history for this message
Ben Collins (ben-collins) wrote :

The hdparm manpage says about -i, that it obtains info "obtained from the drive
at boot time, if available". The "if available" says to me that it doesn't have
to return the info, and that it isn't always expected to. Also, given that -I
works, which is info obtained from the driver directly, I think things are
working as expected, and that there is no bug.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.