Sil3512 (sii, SATA) driver on Sun Ultra60/blade 1000

[R J]

> Have you tried newer SiI3124/3132 cards family with new siis(4) or old
> ata(4) driver?


I have tried the Sil3124, but not with the new siis driver.  I tried it 
with the ata driver;  I had the card with 4 ports, and when I tried it 
with ata, it was exhibiting the same behaviour as the Sil3512 card in the 
ultra 60.

But the Sil3124 card died, and not sure how that happened.  I still have 
it, but the system doesn't see it atall.

How can I help?  I am willing to do testing utilizing the Sil3512 that I 
have, and if I can get a hold of a working Sil3124, I can test with that 
too.  I could probably provide remote access to a sparc64 box for testing, 
but I will have to check on that.

Let me know how I can help.

Thanks,

Robert


On Thu, 24 Sep 2009, Alexander Motin wrote:

> Hi.
>
> R J wrote:
>> I have been referred to you by Marius regarding the sii (Silicon Image)
>> ata driver on sparc64.  The card I have is a two-port SATA card with the
>> Sil3512 chipset.
>>
>> You have perhaps seen some of my communication to the freebsd-sparc list
>> earlier, but briefly, here's the problem I have had with it.
>>
>> UNDER REL_ENG 7:
>>
>> I have had a silicon image Sil 3512 in the ultra 60, and attached two
>> disk drives.  The drives are seen by the ata driver.  I was able to lay
>> a ufs file system on them, and even copied files onto the drives, but
>> the system was not stable. I.e, I would copy files to the sata hard
>> drives for a few minutes before it would hang and remain that way till I
>> did a hard reset (power-off).  Usually there would be some message about
>> the device detaching from the system, and the /dev/ad* entries disappear.
>>
>> I then take the Sil 3512 out of the ultra60 and put it in the Sun blade
>> 1000, and attach the disk drives; upon booting, the drives are not seen
>> by the ata driver, although the controller is seen.  Remember, the
>> drives are seen with the same card in the ultra60.
>>
>> REBUILD TO 8.0 BETA on the Sun Blade 1000:
>>
>> After I upgraded the Sun blade 1000 to 8.0 BETA, the ata driver could
>> now see the disk attached to the Sil3512 controller, but the behaviour
>> remains pretty much like on the Ultra60 under 7 stable if I put any i/o
>> on the disks.
>>
>> I have worked with Marius to get the ata-marvell driver sane for
>> big-endian sparc64, and that driver now works on the Sun blade 1000
>> under 8.0 BETA.  Before Marius fixed that driver, it was exhibiting the
>> same behaviour as the Sil3512 under 8.0 Beta; i.e, the controller is
>> seen, but no drives are seen, and the system wouldn't even finish
>> booting if disks were attached to the controller.
>>
>> Is it possible to for  you to take a look at the Sil ata driver again
>> in-light of this information?
>
> I have only tested SiI3114 chip from this family and only on i386/amd64.
> It worked fine for me. Unfortunately, I have neither Sil3512 controller
> nor sparc64 hardware for testing. Driver mentions about some errata for
> this chip, but I am not aware about it and have no card to test it.
>
> Have you tried newer SiI3124/3132 cards family with new siis(4) or old
> ata(4) driver?
>
> -- 
> Alexander Motin
>
>