sata_mv 7042 support and port multipliers

sata_mv 7042 support and port multipliers

[Robin H. Johnson]

[-- Attachment #1: Type: text/plain, Size: 1180 bytes --]

Hi,

I recently picked up a new eSATA controller
(http://www.sonnettech.com/product/tempo_sata_e4p.html), along with a
port-multiplier enclosure (Fusion 500P from the same company) and a
bunch of disks, since I was running out of space.

I added this to the table mv_pci_tbl[]:
{ PCI_VDEVICE(MARVELL, 0x7042), chip_7042 },

However, I see it doesn't quite work. 7042 support doesn't seem to have
been touched or even mentioned since Jeff posted the 6042 changes in
January.

The card reports as revision 2, which is not in the code either.

I've got the hardware, so I'm willing to do heavy lifting of code (NCQ,
Hotplug, PMP) if it hasn't been done yet, if somebody point me to some
docs on the matter.  I've seen the mvSata-3.6.1 driver, and it appears
to have most of the stuff on revision 1, but nothing on hardware
revision 2 either.

Jeff: you mentioned back in September about Marvell being a bit more
generous with docs? Were 7042 docs amongst them or do I have to slog
through the messy mvSata driver to find everything?

-- 
Robin Hugh Johnson
E-Mail     : robbat2@gentoo.org
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: sata_mv 7042 support and port multipliers

[Jeff Garzik]

Robin H. Johnson wrote:
> Hi,
> 
> I recently picked up a new eSATA controller
> (http://www.sonnettech.com/product/tempo_sata_e4p.html), along with a
> port-multiplier enclosure (Fusion 500P from the same company) and a
> bunch of disks, since I was running out of space.
> 
> I added this to the table mv_pci_tbl[]:
> { PCI_VDEVICE(MARVELL, 0x7042), chip_7042 },
> 
> However, I see it doesn't quite work. 7042 support doesn't seem to have
> been touched or even mentioned since Jeff posted the 6042 changes in
> January.
> 
> The card reports as revision 2, which is not in the code either.
> 
> I've got the hardware, so I'm willing to do heavy lifting of code (NCQ,
> Hotplug, PMP) if it hasn't been done yet, if somebody point me to some
> docs on the matter.  I've seen the mvSata-3.6.1 driver, and it appears
> to have most of the stuff on revision 1, but nothing on hardware
> revision 2 either.
> 
> Jeff: you mentioned back in September about Marvell being a bit more
> generous with docs? Were 7042 docs amongst them or do I have to slog
> through the messy mvSata driver to find everything?


That's what I had to do.  I have the docs, which cover the programming 
interface, but it doesn't cover a lot of the associated errata.  I had 
to dig that out of the mvSata driver myself.

I had thought that Mark Lord was working on improving sata_mv, but 
outside of a locking bugfixes, I haven't seen anything since mid-May.

So I guess sata_mv is in limbo again...

	Jeff

Re: sata_mv 7042 support and port multipliers

[Mark Lord]

Jeff Garzik wrote:
> Robin H. Johnson wrote:
..
>> The card reports as revision 2, which is not in the code either.
>>
>> I've got the hardware, so I'm willing to do heavy lifting of code (NCQ,
>> Hotplug, PMP) if it hasn't been done yet, if somebody point me to some
>> docs on the matter.  I've seen the mvSata-3.6.1 driver, and it appears
>> to have most of the stuff on revision 1, but nothing on hardware
>> revision 2 either.
>>
>> Jeff: you mentioned back in September about Marvell being a bit more
>> generous with docs? Were 7042 docs amongst them or do I have to slog
>> through the messy mvSata driver to find everything?
> 
> That's what I had to do.  I have the docs, which cover the programming 
> interface, but it doesn't cover a lot of the associated errata.  I had 
> to dig that out of the mvSata driver myself.
> 
> I had thought that Mark Lord was working on improving sata_mv, but 
> outside of a locking bugfixes, I haven't seen anything since mid-May.

I no longer use a system with that chipset (6081), so if someone new (Robin?)
wants to take up the cause, then go for it!

Cheers

Re: sata_mv 7042 support and port multipliers

[Robin H. Johnson]

[-- Attachment #1: Type: text/plain, Size: 1033 bytes --]

On Mon, Nov 06, 2006 at 01:22:24AM -0500, Jeff Garzik wrote:
> That's what I had to do.  I have the docs, which cover the programming 
> interface, but it doesn't cover a lot of the associated errata.  I had 
> to dig that out of the mvSata driver myself.
> 
> I had thought that Mark Lord was working on improving sata_mv, but 
> outside of a locking bugfixes, I haven't seen anything since mid-May.
> 
> So I guess sata_mv is in limbo again...
Ok, since Mark says he doesn't have the 6081 anymore I guess if I want
this working at all, I get to write and maintain it myself :-).

Are those Marvell docs under some form of protection (NDA etc)? Can I
get a copy of them? There a few questions (around NCQ/PMP) that the
mvSata driver raises for me, and I think the docs would help there.

-- 
Robin Hugh Johnson
E-Mail     : robbat2@orbis-terrarum.net
Home Page  : http://www.orbis-terrarum.net/?l=people.robbat2
ICQ#       : 30269588 or 41961639
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: sata_mv 7042 support and port multipliers

[Jeff Garzik]

Robin H. Johnson wrote:
> On Mon, Nov 06, 2006 at 01:22:24AM -0500, Jeff Garzik wrote:
>> That's what I had to do.  I have the docs, which cover the programming 
>> interface, but it doesn't cover a lot of the associated errata.  I had 
>> to dig that out of the mvSata driver myself.
>>
>> I had thought that Mark Lord was working on improving sata_mv, but 
>> outside of a locking bugfixes, I haven't seen anything since mid-May.
>>
>> So I guess sata_mv is in limbo again...
> Ok, since Mark says he doesn't have the 6081 anymore I guess if I want
> this working at all, I get to write and maintain it myself :-).
> 
> Are those Marvell docs under some form of protection (NDA etc)? Can I
> get a copy of them? There a few questions (around NCQ/PMP) that the
> mvSata driver raises for me, and I think the docs would help there.

Yes, they're under NDA.  If you have specific programming questions that 
deal directly with the functioning of the driver (i.e. imparting facts 
necessary for the operation of the open source driver), I can help out 
though.

	Jeff