unsupported NVIDIA SATA controller

[Joseph Olatt]

On Tue, Sep 16, 2008 at 12:12:47PM +0100, Gavin Atkinson wrote:
> On Mon, 2008-09-15 at 19:25 -0700, Joseph Olatt wrote:
> > Hello,
> > 
> > I have the following SATA controller card on my system that appears to
> > be unsupported by FreeBSD 7-STABLE. Does anybody know if this card is
> > supported or will be supported in the near future?
> > 
> > /*** Begin change #2 ***/
> > [/usr/src/sys/dev/ata]
> > joji at snow> diff -u  ata-chipset.c ata-chipset.c.orig
> > --- ata-chipset.c       2008-09-15 20:47:55.000000000 -0500
> > +++ ata-chipset.c.orig  2008-09-15 20:40:43.000000000 -0500
> > @@ -3051,7 +3051,6 @@
> >       { ATA_NFORCE_MCP65,    0, AMDNVIDIA, NVIDIA,  ATA_UDMA6, "nForce MCP65" },
> >       { ATA_NFORCE_MCP67,    0, AMDNVIDIA, NVIDIA,  ATA_UDMA6, "nForce MCP67" },
> >       { ATA_NFORCE_MCP73,    0, AMDNVIDIA, NVIDIA,  ATA_UDMA6, "nForce MCP73" },
> > -     { ATA_NFORCE_MCP73_S1, 0, 0,         NV4|NVQ, ATA_SA300, "nForce MCP73" },
> >       { ATA_NFORCE_MCP77,    0, AMDNVIDIA, NVIDIA,  ATA_UDMA6, "nForce MCP77" },
> >       { 0, 0, 0, 0, 0, 0}} ;
> > /*** End change #2 ***/
> 
> Before you do anything, can you get a verbose DMESG and stick it online
> somewhere?  This may well help identify why the above isn't working.

http://www.eskimo.com/~joji/nvidia_sata/snow_dmesg-v.txt


> Secondly, you could try changing the line you've added to be:
> 
>         { ATA_NFORCE_MCP73_S1, 0, AMDNVIDIA, NVIDIA,  ATA_SA300, "nForce MCP73" },
> 
> although to be honest, I'm not expecting that to fix things for you.  If
> it doesn't, then it looks like this chip may need special support.

Made the change suggested above. Building a new kernel. Will provide
update after installation and boot-up of new kernel.


> Because the chip identifies itself as a RAID controller and not as a
> standard IDE controller, the generic code does not attach to it either.
> You could override this behaviour by removing your patch and using the
> attached patch.  If nothing else, that may well get you working at
> UDMA33, which is the most the "generic" ATA controller support can do.
> Again, if this fails, stick a verbose dmesg online somewhere.
> 
> The real solution, of proper support for the chip, may not be possible
> until either documentation is available for it, or another OS
> (Linux/*BSD/OpenSolaris) support it.  

I do have Ubuntu installed on a disk connected to the above mentioned
SATA controller. The dmesg from Ubuntu is at:

http://www.eskimo.com/~joji/nvidia_sata/dmesg_ubuntu.txt


The output of "lspci -vv" from Ubuntu is at:

http://www.eskimo.com/~joji/nvidia_sata/lspci-vv.txt


I am not clear if Ubuntu is recognizing the SATA hard disk as ATA or
SATA. I suspect it is being recognized as a ATA disk because of the
following lines from Ubuntu dmesg:

[   23.044251] ata1: SATA max UDMA/133 abar m8192 at 0xfea7c000 port
0xfea7c100 irq 508
[   23.044253] ata2: SATA max UDMA/133 abar m8192 at 0xfea7c000 port
0xfea7c180 irq 508
[   23.044256] ata3: SATA max UDMA/133 abar m8192 at 0xfea7c000 port
0xfea7c200 irq 508
[   23.044258] ata4: SATA max UDMA/133 abar m8192 at 0xfea7c000 port
0xfea7c280 irq 508
[   23.682254] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[   23.682854] ata1.00: ATA-7: ST3250410AS, 3.AAF, max UDMA/133


But the following line confusing:

[   23.682254] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)


> Out of interest, what motherboard is this on?
> 
> Gavin

Is there a way to find out the motherboard details without opening up
the box?


If nothing else works, I will back out all my changes and install your
patch and see how it goes. 

Thanks very much for helping.

regards,
joseph