Should sata_mv work with Highpoint RocketRAID 2300?

Should sata_mv work with Highpoint RocketRAID 2300?

[Karim 'Kasi Mir' Senoucci]

Hello all,
sorry if this it too dumb a question, but since I haven't gotten any 
results with the sata_mv driver (kernel 2.6.22.2), I was wondering if 
the Highpoint RocketRaid 2300 controller - which features a Marvell chip 
- is even supposed to work with sata_mv.

lspci shows in regard to the controller card:

06:00.0 SCSI storage controller [0100]: Triones Technologies, Inc. 
Unknown device [1103:2300] (rev 02)
        Subsystem: Marvell Technology Group Ltd. Unknown device [11ab:11ab]
        Flags: bus master, fast devsel, latency 0, IRQ 11
        Memory at fd500000 (64-bit, non-prefetchable) [size=1M]
        I/O ports at 4800 [size=256]
        Capabilities: [40] Power Management version 2
        Capabilities: [50] Message Signalled Interrupts: Mask- 64bit+ 
Queue=0/0 Enable-
        Capabilities: [60] Express Legacy Endpoint IRQ 0

Is this one of the cards the sata_mv driver should handle, or do I need 
to use another kernel module, or is that chip without linux support so far?

Thanks for any insight into the matter.

Greetings, Karim.

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Alan Cox]

On Tue, 14 Aug 2007 18:47:55 +0200
"Karim 'Kasi Mir' Senoucci" <ks@double-you.de> wrote:

> Hello all,
> sorry if this it too dumb a question, but since I haven't gotten any 
> results with the sata_mv driver (kernel 2.6.22.2), I was wondering if 
> the Highpoint RocketRaid 2300 controller - which features a Marvell chip 
> - is even supposed to work with sata_mv.

It should but the vendor issued the chip with different PCI identifiers
to a normal Marvell chip. Add it to the sata_mv driver and it should work.

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Karim 'Kasi Mir' Senoucci]

Alan Cox schrieb:
> On Tue, 14 Aug 2007 18:47:55 +0200
> "Karim 'Kasi Mir' Senoucci" <ks@double-you.de> wrote:
>
>   
>> Hello all,
>> sorry if this it too dumb a question, but since I haven't gotten any 
>> results with the sata_mv driver (kernel 2.6.22.2), I was wondering if 
>> the Highpoint RocketRaid 2300 controller - which features a Marvell chip 
>> - is even supposed to work with sata_mv.
>>     
>
> It should but the vendor issued the chip with different PCI identifiers
> to a normal Marvell chip. Add it to the sata_mv driver and it should work.
>
>   
I'm not entirely sure what to add there - I'm gathering you mean those 
lines:

       { PCI_VDEVICE(MARVELL, 0x6040), chip_604x },
       { PCI_VDEVICE(MARVELL, 0x6041), chip_604x },
       { PCI_VDEVICE(MARVELL, 0x6042), chip_6042 },
       { PCI_VDEVICE(MARVELL, 0x6080), chip_608x },
       { PCI_VDEVICE(MARVELL, 0x6081), chip_608x },

and so on, and thus would add something like:

      { PCI_VDEVICE(MARVELL, 0x11ab), chip_614x },

(for the 88SE614x apparently used on the RocketRAID 2300) except that 
there is no  "chip_614x" in the module.
Should I write "chip_604x" instead or do I need to do something else 
entirely.

Along the same lines, theres is an entry:

        { PCI_VDEVICE(TTI, 0x2310), chip_7042 },

in the sata_mv source. Should I add something like:

        { PCI_VDEVICE(TTI, 0x2300), chip_604x },

as well?

Last but not least: I'm assuming that this driver - when working 
correctly - will let me access the RAID arrays I setup via BIOS or 
Windows driver, but there will be no way to create or modify the arrays 
from within linux itself, right?

Thank for any help, past, present or future. :-)

Greetings, Karim.

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Alan Cox]

> in the sata_mv source. Should I add something like:
> 
>         { PCI_VDEVICE(TTI, 0x2300), chip_604x },
> 
> as well?

Yes


> Last but not least: I'm assuming that this driver - when working 
> correctly - will let me access the RAID arrays I setup via BIOS or 
> Windows driver, but there will be no way to create or modify the arrays 
> from within linux itself, right?

The RAID is entirely software so its up to software what is handled. I
don't know if dmraid will handle whatever format HPT have used this time
but you may be lucky.

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Karim 'Kasi Mir' Senoucci]

Alan Cox schrieb:
>> in the sata_mv source. Should I add something like:
>>
>>         { PCI_VDEVICE(TTI, 0x2300), chip_604x },
>>
>> Yes
>>     
Just to not leave this thread hanging in the mail archives: this didn't 
work for me. With the above patch, the controller is detected, but the 
system hangs upon inclusion of the sata_mv driver. dmess shows:

[   32.240000] sata_mv 0000:06:00.0: version 0.81
[   32.272000] ACPI: PCI Interrupt 0000:06:00.0[A] -> Done.Link [APC6] 
->GSI 16 (level, low) -> IRQ 17
[   32.336000] sata_mv 0000:06:00.0: Applying B2 workarounds to unknown rev
[   32.388000] sata_mv 0000:06:00.0: 32 slots 4 ports SCSI mode IRQ via INTx
[   32.432000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
[   32.464000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
[   32.496000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
[   32.528000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
[   32.560000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
[   32.592000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000

and many many more PCI ERRORS until I reboot the system. So, I guess 
that particular Highpoint controller does not work yet with sata_mv.

Thanks very much for your efforts, anyway.

Greetings, Karim.

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Jeff Garzik]

Karim 'Kasi Mir' Senoucci wrote:
> Alan Cox schrieb:
>>> in the sata_mv source. Should I add something like:
>>>
>>>         { PCI_VDEVICE(TTI, 0x2300), chip_604x },
>>>
>>> Yes
>>>     
> Just to not leave this thread hanging in the mail archives: this didn't 
> work for me. With the above patch, the controller is detected, but the 
> system hangs upon inclusion of the sata_mv driver. dmess shows:
> 
> [   32.240000] sata_mv 0000:06:00.0: version 0.81
> [   32.272000] ACPI: PCI Interrupt 0000:06:00.0[A] -> Done.Link [APC6] 
> ->GSI 16 (level, low) -> IRQ 17
> [   32.336000] sata_mv 0000:06:00.0: Applying B2 workarounds to unknown rev
> [   32.388000] sata_mv 0000:06:00.0: 32 slots 4 ports SCSI mode IRQ via 
> INTx
> [   32.432000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
> [   32.464000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
> [   32.496000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
> [   32.528000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
> [   32.560000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
> [   32.592000] sata_mv: PCI ERROR; PCI IRQ cause=0x00000000
> 
> and many many more PCI ERRORS until I reboot the system. So, I guess 
> that particular Highpoint controller does not work yet with sata_mv.
> 
> Thanks very much for your efforts, anyway.

You need to use the sata_mv from kernel 2.6.23-rcX.

Patching into older drivers isn't guaranteed to work.

	Jeff

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Alan Cox]

On Fri, 24 Aug 2007 05:55:39 -0400
Jeff Garzik <jeff@garzik.org> wrote:

> Karim 'Kasi Mir' Senoucci wrote:
> > Alan Cox schrieb:
> >>> in the sata_mv source. Should I add something like:
> >>>
> >>>         { PCI_VDEVICE(TTI, 0x2300), chip_604x }

> You need to use the sata_mv from kernel 2.6.23-rcX.
> 
> Patching into older drivers isn't guaranteed to work.

May also be the actual chip on the 2300 is different to the older ones
and isn't a 604x.

Re: Should sata_mv work with Highpoint RocketRAID 2300?

[Krzysztof Oledzki]

On 2007-08-24 15:48, Alan Cox wrote:
> On Fri, 24 Aug 2007 05:55:39 -0400
> Jeff Garzik <jeff@garzik.org> wrote:
> 
>> Karim 'Kasi Mir' Senoucci wrote:
>>> Alan Cox schrieb:
>>>>> in the sata_mv source. Should I add something like:
>>>>>
>>>>>         { PCI_VDEVICE(TTI, 0x2300), chip_604x }
> 
>> You need to use the sata_mv from kernel 2.6.23-rcX.
>>
>> Patching into older drivers isn't guaranteed to work.
> 
> May also be the actual chip on the 2300 is different to the older ones
> and isn't a 604x.

Indeed, it seems to be a 7042:

http://www.highpoint-tech.com/image/products/SATAII/RR2300_Card_big.jpg

Best regards,

				Krzysztof Olędzki