[ANNOUNCEMENT] Wiki for discussing P35/IHC9(R)/SATA issues set up

Jeremy Chadwick koitsu at FreeBSD.org
Tue Nov 6 02:35:26 PST 2007

On Tue, Nov 06, 2007 at 05:17:06AM -0500, Aryeh M. Friedman wrote:
> > Also, you do realise that having a SATA150 drive on your SATA bus does
> > not mean that the entire bus runs at 150MB/sec, correct?  It's not like
> > SCSI.  So there should be no performance hit having a single SATA150
> > drive on SATA controller also filled with SATA300 devices.
> My mobo uses seperate controllers for each SATA slot (I know you can
> chain them but I am using one per controler):
> Note ata2 is PATA all the rest are SATA
> ATA channel 2:
>     Master:  ad4 <Maxtor 6Y200P0/YAR41BW0> ATA/ATAPI revision 7
>     Slave:   ad5 <WDC WD2500JB-22REA0/20.00K20> ATA/ATAPI revision 7

Noted.  However...

> ATA channel 3:
>     Master:  ad6 <ST3500630AS/3.AAE> Serial ATA II
>     Slave:       no device present

This differs from what you have in the Wiki for the same device.
Note the "Serial ATA II" in your current one, versus "Serial ATA v1.0"
in the older revision:

ATA channel 4:
    Master:  ad8 <ST3500630AS/3.AAE> Serial ATA v1.0
    Slave:       no device present

What this means is, how you have things presently proves that the
device has SATA300 capability.  "Serial ATA II" == SATA300.

If dmesg on the machine shows "SATA150" for the device still, despite
what atacontrol shows, then there may be a driver bug.

Either way, this is a good improvement.

> > However: your PATA ports becoming unusable/disabled when you enable SATA
> > in the BIOS could be either a BIOS bug (or "feature") or a FreeBSD bug.
> > I would not put it past Gigabyte to have a BIOS bug (they are very
> > well-known for having such, but are also pretty good about fixing
> > such problems).  Have you tried a BIOS upgrade on your P35 since you
> > got it, or looked at the BIOS changelog?
> >   
> It appears to be a FreeBSD issue because:
> 1. After Xi Lin's patch they are seen
> 2. The boot manager and cmos boot order see and can boot from them
> > I do not have an ICH9 board to help confirm or deny -- I can purchase
> > one if needed, and/or send it to Xin Li free of cost.
> >   
> From what other people are saying I think it needs to be the p35/ihc9(r)
> combo specifically.

These two things are somewhat conflicting.

1) If the problem is fixed by Xi Lin's patch, then that seems to
   indicate it should be affecting everyone using the ICH9 controller,
   including those who use Asus, Abit, or MSI boards.
2) Boot manager as in the BIOS boot device menu, or as in FreeBSD BTX?
   If BIOS boot device menu, then that doesn't mean anything.  The
   BIOS could indeed be doing something wrong which FreeBSD expects
   to be right.  Problems like this should ideally be brought to the
   attention of the mainboard manufacturer.
3) The issue appears to be specific to the Gigabyte P35 board, which
   to me means someone needs to get Gigabyte involved in all of this.

| Jeremy Chadwick                                    jdc at parodius.com |
| Parodius Networking                           http://www.parodius.com/ |
| UNIX Systems Administrator                      Mountain View, CA, USA |
| Making life hard for others since 1977.                  PGP: 4BD6C0CB |

More information about the freebsd-current mailing list