VIA VT6330 PCI-E Combo Controller not fully supported

VIA VT6330 PCI-E Combo Controller not fully supported

[Chris M]

Hi everyone!  I'll quickly introduce myself before I get to the crux of 
my reason for posting to the mailing list.  My name is Chris Merrett and 
although I'm not a subscriber to the list (my mail server would likely 
cry), I check a well-known archive quite regularly just to try and keep 
on top of developments in the community.

The reason I'm mailing is due to a partially-supported VIA chip present 
on a small handful of motherboards currently on sale.  I mailed the 
maintainer (Jeff Garzik) to bring this to his attention quite some time 
ago, but I'd imagine he's not found the time to reply to me yet which is 
understandable.  I figured that it'd be a good idea to make a post here 
with the info I've gathered, so that people affected by this same issue 
are aware that this particular chip isn't yet fully supported in the kernel.

The VT6330 is a combo controller chip that handles both Firewire and 
PATA, and is present on my Asrock N7AD-SLI.  I'm also aware of it being 
present on other Asrock boards, and presumably other boards from 
different manufacturers.  The Firewire functionality is supported 
perfectly and without issue, but unfortunately the PATA support is 
lacking at the moment.  VIA have some information on this at the 
following URL: http://www.via.com.tw/en/products/peripherals/1394/vt6330/

 From the info on the URL above, it appears that the PATA functionality 
is identical to that of the VT6415 PATA controller, support for which 
was added in 2.6.28.  The mailing list post made regarding the patch 
that was submitted can be found here: 
http://lkml.indiana.edu/hypermail/linux/kernel/0902.2/01428.html

Finally, some relevant output from lshw:

           *-firewire
                description: FireWire (IEEE 1394)
                product: VIA Technologies, Inc.
                vendor: VIA Technologies, Inc.
                physical id: 0
                bus info: pci@0000:01:00.0
                version: 00
                width: 64 bits
                clock: 33MHz
                capabilities: pm msi pciexpress bus_master cap_list
                configuration: driver=ohci1394 latency=0 module=ohci1394
           *-ide UNCLAIMED
                description: IDE interface
                product: PATA IDE Host Controller
                vendor: VIA Technologies, Inc.
                physical id: 0.1
                bus info: pci@0000:01:00.1
                version: a0
                width: 32 bits
                clock: 33MHz
                capabilities: ide pm msi pciexpress cap_list
                configuration: latency=0

I'm going to get my hands dirty and have a try at fixing it myself one 
evening this week, so if I have any luck (which I'll no doubt need) I'll 
be sure to reply soon.

Warmest regards to everyone,

Chris Merrett

****

Re: VIA VT6330 PCI-E Combo Controller not fully supported

[Robert Hancock]

On 09/29/2009 01:30 PM, Chris M wrote:
> Hi everyone!  I'll quickly introduce myself before I get to the crux of
> my reason for posting to the mailing list. My name is Chris Merrett and
> although I'm not a subscriber to the list (my mail server would likely
> cry), I check a well-known archive quite regularly just to try and keep
> on top of developments in the community.
>
> The reason I'm mailing is due to a partially-supported VIA chip present
> on a small handful of motherboards currently on sale. I mailed the
> maintainer (Jeff Garzik) to bring this to his attention quite some time
> ago, but I'd imagine he's not found the time to reply to me yet which is
> understandable. I figured that it'd be a good idea to make a post here
> with the info I've gathered, so that people affected by this same issue
> are aware that this particular chip isn't yet fully supported in the
> kernel.
>
> The VT6330 is a combo controller chip that handles both Firewire and
> PATA, and is present on my Asrock N7AD-SLI. I'm also aware of it being
> present on other Asrock boards, and presumably other boards from
> different manufacturers. The Firewire functionality is supported
> perfectly and without issue, but unfortunately the PATA support is
> lacking at the moment. VIA have some information on this at the
> following URL: http://www.via.com.tw/en/products/peripherals/1394/vt6330/
>
>  From the info on the URL above, it appears that the PATA functionality
> is identical to that of the VT6415 PATA controller, support for which
> was added in 2.6.28. The mailing list post made regarding the patch that
> was submitted can be found here:
> http://lkml.indiana.edu/hypermail/linux/kernel/0902.2/01428.html

If it is similar enough to the existing controllers then you may be able 
to just add the device ID into pata_via in a similar fashion..

Re: VIA VT6330 PCI-E Combo Controller not fully supported

[chrisfu]

> On 09/29/2009 01:30 PM, Chris M wrote:
>> Hi everyone!  I'll quickly introduce myself before I get to the crux of
>> my reason for posting to the mailing list. My name is Chris Merrett and
>> although I'm not a subscriber to the list (my mail server would likely
>> cry), I check a well-known archive quite regularly just to try and keep
>> on top of developments in the community.
>>
>> The reason I'm mailing is due to a partially-supported VIA chip present
>> on a small handful of motherboards currently on sale. I mailed the
>> maintainer (Jeff Garzik) to bring this to his attention quite some time
>> ago, but I'd imagine he's not found the time to reply to me yet which is
>> understandable. I figured that it'd be a good idea to make a post here
>> with the info I've gathered, so that people affected by this same issue
>> are aware that this particular chip isn't yet fully supported in the
>> kernel.
>>
>> The VT6330 is a combo controller chip that handles both Firewire and
>> PATA, and is present on my Asrock N7AD-SLI. I'm also aware of it being
>> present on other Asrock boards, and presumably other boards from
>> different manufacturers. The Firewire functionality is supported
>> perfectly and without issue, but unfortunately the PATA support is
>> lacking at the moment. VIA have some information on this at the
>> following URL:
>> http://www.via.com.tw/en/products/peripherals/1394/vt6330/
>>
>>  From the info on the URL above, it appears that the PATA functionality
>> is identical to that of the VT6415 PATA controller, support for which
>> was added in 2.6.28. The mailing list post made regarding the patch that
>> was submitted can be found here:
>> http://lkml.indiana.edu/hypermail/linux/kernel/0902.2/01428.html
>
> If it is similar enough to the existing controllers then you may be able
> to just add the device ID into pata_via in a similar fashion..
>

Thanks for your feedback Robert.  I went to do exactly that yesterday
evening,
and found some interesting stuff.  The PATA functionality of the VT6330
identifies itself as the VT6415 and uses the exact same pci_id, meaning that
support was already added in the mailing list post I quoted in my original
mail.

The perplexing thing is that the device is unclaimed even though that by all
accounts, it should be picked up.

A snippet from dmesg:

[    4.970261] pata_via 0000:01:00.1: version 0.3.3
[    4.970637] ACPI: PCI Interrupt Link [LN6A] enabled at IRQ 19
[    4.970643] pata_via 0000:01:00.1: PCI INT A -> Link[LN6A] -> GSI 19
(level,
low) -> IRQ 19
[    4.970709] pata_via 0000:01:00.1: PCI INT A disabled
[    4.970774] pata_acpi 0000:01:00.1: PCI INT A -> Link[LN6A] -> GSI 19
(level,
low) -> IRQ 19
[    4.970799] pata_acpi 0000:01:00.1: setting latency timer to 64
[    4.970809] pata_acpi 0000:01:00.1: PCI INT A disabled

I guess I'll just keep probing and see what I can come up with.  The next
step
is trying to find out if the changes to support the VT6415 worked for
*anyone*. :)

Warm regards,

Chris M

Re: VIA VT6330 PCI-E Combo Controller not fully supported

[Robert Hancock]

On Wed, Sep 30, 2009 at 6:23 AM,  <chrisfu@mastodon.hardc0re.org.uk> wrote:
>>>  From the info on the URL above, it appears that the PATA functionality
>>> is identical to that of the VT6415 PATA controller, support for which
>>> was added in 2.6.28. The mailing list post made regarding the patch that
>>> was submitted can be found here:
>>> http://lkml.indiana.edu/hypermail/linux/kernel/0902.2/01428.html
>>
>> If it is similar enough to the existing controllers then you may be able
>> to just add the device ID into pata_via in a similar fashion..
>>
>
> Thanks for your feedback Robert.  I went to do exactly that yesterday
> evening,
> and found some interesting stuff.  The PATA functionality of the VT6330
> identifies itself as the VT6415 and uses the exact same pci_id, meaning that
> support was already added in the mailing list post I quoted in my original
> mail.
>
> The perplexing thing is that the device is unclaimed even though that by all
> accounts, it should be picked up.
>
> A snippet from dmesg:
>
> [    4.970261] pata_via 0000:01:00.1: version 0.3.3
> [    4.970637] ACPI: PCI Interrupt Link [LN6A] enabled at IRQ 19
> [    4.970643] pata_via 0000:01:00.1: PCI INT A -> Link[LN6A] -> GSI 19
> (level,
> low) -> IRQ 19
> [    4.970709] pata_via 0000:01:00.1: PCI INT A disabled
> [    4.970774] pata_acpi 0000:01:00.1: PCI INT A -> Link[LN6A] -> GSI 19
> (level,
> low) -> IRQ 19
> [    4.970799] pata_acpi 0000:01:00.1: setting latency timer to 64
> [    4.970809] pata_acpi 0000:01:00.1: PCI INT A disabled
>
> I guess I'll just keep probing and see what I can come up with.  The next
> step
> is trying to find out if the changes to support the VT6415 worked for
> *anyone*. :)

I think the PCI ID portion is OK, you wouldn't even get the lines
about enabling the interrupt if it wasn't detecting a matching device.
I'm suspecting that ata_pci_sff_init_one is failing out for some
reason, you might need to scatter some printks around in that code to
figure out why..