Re: HighPoint RocketRAID 2320

[Mark Lord <liml@rtr.ca>]

Davíð Steinn Geirsson wrote:
> Hi Mark, thanks for the reply.
> 
> On Mon, Aug 03, 2009 at 06:53:52PM -0400, Mark Lord wrote:
>> Mark Lord wrote:
>>> Mark Lord wrote:
>>>> Davíð Steinn Geirsson wrote:
>>>>> Hi all,
>>>>>
>>>>> I'm throwing this out here in the hopes that someone smarter than me has
>>>>> a simple solution - never hurts to be optimistic. :)
>>>>>
>>>>> I have a HighPoint controller, RocketRAID 2320 (8-port PCIe SATA  
>>>>> fakeraid). It is only supported by an ugly binary blob deceptively  
>>>>> labeled as an "open source driver" from HighPoint (rr232x). Looking 
>>>>> at the wrapper around the blob, it seems this driver claims only 
>>>>> the 2320 and 2322 controllers:
>>>>>
>>>>> static const struct pci_device_id hpt_pci_tbl[] = {
>>>>>         {PCI_DEVICE(0x1103, 0x2320), 0, 0, 0},
>>>>>         {PCI_DEVICE(0x1103, 0x2322), 0, 0, 0},
>>>>>         {}
>>>>> };
>>>>>
>>>>> I've found that this controller contains a marvell 88SX6081 chip, which
>>>>> should be supported by the sata_mv driver. That driver claims device IDs
>>>>> 2300 and 2310:
>>>>>         { PCI_VDEVICE(TTI, 0x2300), chip_7042 },
>>>>>         { PCI_VDEVICE(TTI, 0x2310), chip_7042 },
>>>>>
>>>>> So, ever hopeful, I tried adding the 2320 into the table:
>>>>>         { PCI_VDEVICE(TTI, 0x2320), chip_608x },
>>>>>
>>>>> When I do this, the kernel successfully probes the attached disks and
>>>>> their capacity, but immediately errors out and starts resetting the
>>>>> ports repeatedly.
>>>> ..
>>>>
>>>> Send me a clear, in-focus detailed photograph of the board,
>>>> showing the chip markings very clearly.
>>> ..
>>>
>>> Never mind -- found one here:
>>> http://www.taipeitradeshows.com.tw/downloads/2007051104030512475/RR2320.jpg
>>>
>>> The SATA chip is clearly a 88SX6091-8CZ, which is a PCIX 8-port controller.
>>                             ^^^^^^^^
>> That was supposed to say 88SX6081 there.  :)
> 
> Ah, I see. I hadn't checked the datasheet, but it is indeed.
> 
>>> Since it is sitting on a PCIe card, one must assume there's a PCIe-to-PCIX
>>> bridge hidden under that huge heatsink.
>>>
>>> So if the card does not work in the chip_608x mode,
>>> there's probably some funny business in that bridge chip.
>>>
>>> Maybe it works *only* in the proprietary RAID mode (?)
> 
> I found some better chip pics (minus heatsink) here:
> http://www.pcdvd.com.tw/showthread.php?t=788031&page=3&pp=10
> 
> It shows that under the heatsink is an Intel QG41210 serial-to-parallel
> PCI bridge. I found it's datasheet here:
> http://www.intel.com/design/bridge/datashts/27887505.pdf
> 
> Hope this helps. :)
> 
> Is there any more debugging information I could send that could help
> diagnose this?
..

How about "lspci -vv", to go with the other info.

Then, perhaps Saeed Bishara from Marvell could have a peek ?
I'm on holiday for the next week and a bit, and Marvell is in arrears
for earlier work on this driver.

Cheers