Promise 20319 chipset specs opened

Promise 20319 chipset specs opened

[Jeff Garzik]

Promise has given me permission to post hardware programming info for 
one of their chips (Linux driver: sata_promise), PDC20319.  This also 
marks the first open chipset for Promise (AFAIK), so let's give them a 
round of applause.

The PDC20319 is a reasonably representative example of the hardware 
programming interface covered by sata_promise.  Kernel developers 
wishing to work on sata_promise can study this doc, and deduce how the 
two-port PDC2037x chips work.

In addition to describing the "packet" format for ATA and ATAPI 
commands, this doc also describes how to use the chip's XOR RAID-assist 
features.  I think it would be cool if someone so motivated found a way 
to use this efficiently under Linux.

The doc:
	http://gkernel.sourceforge.net/specs/promise/pdc20319.pdf.bz2

I have also updated the list open chipsets and developer resources on 
linux-ata.org:
	http://linux-ata.org/driver-status.html#open_chipsets
	http://linux-ata.org/devel.html

Have fun!

	Jeff

More Promise chipset specs opened

[Jeff Garzik]

Jeff Garzik wrote:
> Promise has given me permission to post hardware programming info for 
> one of their chips (Linux driver: sata_promise), PDC20319.  This also 
> marks the first open chipset for Promise (AFAIK), so let's give them a 
> round of applause.

I missed an email from Promise which supplied even more chipset docs to 
open.  Whoops.  The entire line supported by sata_promise.c should now 
be open:

2037x:
http://gkernel.sourceforge.net/specs/promise/pdc2037x%20series%20development%20guide.pdf.bz2

20319 (updated link, updated doc):
http://gkernel.sourceforge.net/specs/promise/pdc20319%20development%20guide.pdf.bz2

205xx (the newer NCQ chips):
http://gkernel.sourceforge.net/specs/promise/pdc205xx%20development%20guide%201.0.pdf.bz2

Cheers,

	Jeff

More Promise chipset specs opened

[otto Meier]

> Jeff Garzik wrote:
>
>     Promise has given me permission to post hardware programming info
>     for one of their chips (Linux driver: sata_promise), PDC20319.
>     This also marks the first open chipset for Promise (AFAIK), so
>     let's give them a round of applause. 
>
>
> I missed an email from Promise which supplied even more chipset docs
> to open. Whoops. The entire line supported by sata_promise.c should
> now be open:

I use the following Promise Card : (lspci output):

02:06.0 Mass storage controller: Promise Technology, Inc. PDC20718 (SATA
300 TX4) (rev 02)
        Subsystem: Promise Technology, Inc. PDC20718 (SATA 300 TX4)
        Flags: bus master, 66MHz, medium devsel, latency 72, IRQ 185
        I/O ports at bc00 [size=128]
        I/O ports at b800 [size=256]
        Memory at fdeff000 (32-bit, non-prefetchable) [size=4K]
        Memory at fdec0000 (32-bit, non-prefetchable) [size=128K]
        [virtual] Expansion ROM at fdd80000 [disabled] [size=32K]
        Capabilities: [60] Power Management version 2

lspci -vvn:

02:06.0 Class 0180: 105a:3d17 (rev 02)
        Subsystem: 105a:3d17
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
        Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
        Latency: 72 (1000ns min, 4500ns max), Cache Line Size 01
        Interrupt: pin A routed to IRQ 185
        Region 0: I/O ports at bc00 [size=128]
        Region 2: I/O ports at b800 [size=256]
        Region 3: Memory at fdeff000 (32-bit, non-prefetchable) [size=4K]
        Region 4: Memory at fdec0000 (32-bit, non-prefetchable) [size=128K]
        [virtual] Expansion ROM at fdd80000 [disabled] [size=32K]
        Capabilities: [60] Power Management version 2
                Flags: PMEClk- DSI+ D1+ D2- AuxCurrent=0mA
PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-





Dmesg:
<7>sata_promise 0000:02:06.0: version 1.04
<4>ACPI: PCI Interrupt Link [APC1] enabled at IRQ 16
<6>GSI 17 sharing vector 0xB9 and IRQ 17
<6>ACPI: PCI Interrupt 0000:02:06.0[A] -> Link [APC1] -> GSI 16 (level,
low) -> IRQ 185
<6>ata5: SATA max UDMA/133 cmd 0xFFFFC20000022200 ctl 0xFFFFC20000022238
bmdma 0x0 irq 185
<6>ata6: SATA max UDMA/133 cmd 0xFFFFC20000022280 ctl 0xFFFFC200000222B8
bmdma 0x0 irq 185
<6>ata7: SATA max UDMA/133 cmd 0xFFFFC20000022300 ctl 0xFFFFC20000022338
bmdma 0x0 irq 185
<6>ata8: SATA max UDMA/133 cmd 0xFFFFC20000022380 ctl 0xFFFFC200000223B8
bmdma 0x0 irq 185
<6>scsi4 : sata_promise
<6>ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
<6>ata5.00: ATA-7, max UDMA7, 488397168 sectors: LBA48 NCQ (depth 0/32)
<6>ata5.00: configured for UDMA/133

In which controller classes does it belong?
Promise claims it should  have NCQ  how to enable ?

Best regards

Re: More Promise chipset specs opened

[Jeff Garzik]

otto Meier wrote:
> In which controller classes does it belong?
> Promise claims it should  have NCQ  how to enable ?

If it claims NCQ and 3.0Gbps support, I would guess 205xx

	Jeff

Re: More Promise chipset specs opened

[otto Meier]

Jeff Garzik schrieb:
> otto Meier wrote:
>> In which controller classes does it belong?
>> Promise claims it should  have NCQ  how to enable ?
>
> If it claims NCQ and 3.0Gbps support, I would guess 205xx
>
>     Jeff
>
Does this mean, changing :

 { PCI_VDEVICE(PROMISE, 0x3d17), board_20319 },
to
 { PCI_VDEVICE(PROMISE, 0x3d17), board_2057X },

in sata_promise.c would do the trick to activate NCQ of this board?

Does this hurt?
Do I need to do something else to get NCQ working?

Otto

Re: More Promise chipset specs opened

[Jeff Garzik]

otto Meier wrote:
> Jeff Garzik schrieb:
>> otto Meier wrote:
>>> In which controller classes does it belong?
>>> Promise claims it should  have NCQ  how to enable ?
>> If it claims NCQ and 3.0Gbps support, I would guess 205xx
>>
>>     Jeff
>>
> Does this mean, changing :
> 
>  { PCI_VDEVICE(PROMISE, 0x3d17), board_20319 },
> to
>  { PCI_VDEVICE(PROMISE, 0x3d17), board_2057X },
> 
> in sata_promise.c would do the trick to activate NCQ of this board?
> 
> Does this hurt?
> Do I need to do something else to get NCQ working?

Heh, no, a lot more than that...  NCQ is not supported at all in the 
driver.  If someone is looking for a project, the docs are now available 
for them to work on NCQ.

	Jeff

Re: More Promise chipset specs opened

[otto Meier]

Jeff Garzik schrieb:
> otto Meier wrote:
>> Jeff Garzik schrieb:
>>> otto Meier wrote:
>>>> In which controller classes does it belong?
>>>> Promise claims it should  have NCQ  how to enable ?
>>> If it claims NCQ and 3.0Gbps support, I would guess 205xx
>>>
>>>     Jeff
>>>
>> Does this mean, changing :
>>
>>  { PCI_VDEVICE(PROMISE, 0x3d17), board_20319 },
>> to
>>  { PCI_VDEVICE(PROMISE, 0x3d17), board_2057X },
>>
>> in sata_promise.c would do the trick to activate NCQ of this board?
>>
>> Does this hurt?
>> Do I need to do something else to get NCQ working?
>
> Heh, no, a lot more than that...  NCQ is not supported at all in the
> driver.  If someone is looking for a project, the docs are now
> available for them to work on NCQ.
>
I would take it but I am not able to do the programming, but I would
volunteer in testing.

Otto