Am I using atacontrol attach properly? (sata hotplug issue).

[Kevin Oberman]

> From: George Hartzell <hartzell at alerce.com>
> Date: Sat, 7 Mar 2009 15:39:36 -0800
> Sender: owner-freebsd-current at freebsd.org
> 
> 
> I have a Gigabyte GA-6KIEH-RH motherboard.  It has 5 SATA ports, 4
> attached to a Sil 3114 controller and 1 attach to an ICH8m.  The
> system is running -CURRENT from yesterday.
> 
> I have two disks attaching as ad4 and ad6 to ata2 and ata3 resp.
> 
> I can set ACHI mode for ICH8m in the BIOS.  When I do not enable it,
> the 5th SATA conector shows up as slave on ata6.  When I do enable it
> the 5th connector shows up as master on ata8.
> 
> If I boot without anything connected to the fifth port, then connect a
> drive I thought that the following should work to get the drive
> recognized:
> 
>   sudo atacontrol detach ata8
>   sudo atacontrol attach ata8 
> 
> (or a similar dance with ata6 if AHCI isn't enabled).
> 
> When I set boot_verbose in loader.conf I see the following when I
> attach (two separate attempts with different SATA drives):
> 
> ata8: AHCI reset...
> ata8: SATA connect status=00000004
> ata8: AHCI reset done: phy reset found no device
> ata8: [MPSAFE]
> ata8: [ITHREAD]
> ata8: Identifying devices: 00000000
> ata8: New devices: 00000000
> ata8: AHCI reset...
> ata8: SATA connect status=00000004
> ata8: AHCI reset done: phy reset found no device
> ata8: [MPSAFE]
> ata8: [ITHREAD]
> ata8: Identifying devices: 00000000
> ata8: New devices: 00000000
> 
> If the drive's attached at boot time then it's recognized.
> 
> Am I doing something wrong, or is my hardware balky?

I don't think so.  (Either of the above.)

I have a hot-swappable PATA drive that I used to be able to insert after
booting. At some time, about 2 years ago, that stopped working. If I
have a drive in the slot at boot time, I can detach the controller,
replace the drive, and attach the replacement drive with no problem,
but, if the slot is empty at boot time, I can't detach, add a drive and
attach it.

I can't be sure, but I suspect that this is the same problem.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634
Key fingerprint:059B 2DDF 031C 9BA3 14A4  EADA 927D EBB3 987B 3751