SATA speed. Should be 150 or 133?

SATA speed. Should be 150 or 133?

[Krzysztof Oledzki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 2612 bytes --]

Hello,

Is it normal that each time kernel reports UDMA/133 instead of UDMA/150 
for sata? I noticed this on both ICH5 and ICH6 (AHCI).

* 2.6.12 with ICH5:
libata version 1.11 loaded.
ata_piix version 1.03
ACPI: PCI Interrupt 0000:00:1f.2[A] -> GSI 18 (level, low) -> IRQ 193
PCI: Setting latency timer of device 0000:00:1f.2 to 64
ata1: SATA max UDMA/133 cmd 0xCCB8 ctl 0xCCB2 bmdma 0xCC80 irq 193
ata2: SATA max UDMA/133 cmd 0xCCA0 ctl 0xCC9A bmdma 0xCC88 irq 193
ata1: dev 0 cfg 49:2f00 82:7c6b 83:7b09 84:4003 85:7c69 86:3801 87:4003 88:207f
ata1: dev 0 ATA, max UDMA/133, 156250000 sectors:
ata1: dev 0 configured for UDMA/133
scsi0 : ata_piix
ata2: dev 0 cfg 49:2f00 82:7c6b 83:7b09 84:4003 85:7c69 86:3801 87:4003 88:207f
ata2: dev 0 ATA, max UDMA/133, 156250000 sectors:
ata2: dev 0 configured for UDMA/133
scsi1 : ata_piix
   Vendor: ATA       Model: Maxtor 6Y080M0    Rev: YAR5
   Type:   Direct-Access                      ANSI SCSI revision: 05
   Vendor: ATA       Model: Maxtor 6Y080M0    Rev: YAR5
   Type:   Direct-Access                      ANSI SCSI revision: 05

* 2.6.12 with ICH6:
libata version 1.11 loaded.
ahci version 1.00
ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 217
PCI: Setting latency timer of device 0000:00:1f.2 to 64
ahci(0000:00:1f.2) AHCI 0001.0000 32 slots 4 ports 1.5 Gbps 0xf impl SATA mode
ahci(0000:00:1f.2) flags: 64bit ncq pm led slum part
ata1: SATA max UDMA/133 cmd 0xE0802100 ctl 0x0 bmdma 0x0 irq 217
ata2: SATA max UDMA/133 cmd 0xE0802180 ctl 0x0 bmdma 0x0 irq 217
ata3: SATA max UDMA/133 cmd 0xE0802200 ctl 0x0 bmdma 0x0 irq 217
ata4: SATA max UDMA/133 cmd 0xE0802280 ctl 0x0 bmdma 0x0 irq 217
ata1: dev 0 cfg 49:2f00 82:346b 83:7d01 84:4003 85:3469 86:3c01 87:4003 88:007f
ata1: dev 0 ATA, max UDMA/133, 156301488 sectors: lba48
ata1: dev 0 configured for UDMA/133
scsi0 : ahci
ata2: dev 0 cfg 49:2f00 82:346b 83:7d01 84:4003 85:3469 86:3c01 87:4003 88:007f
ata2: dev 0 ATA, max UDMA/133, 156301488 sectors: lba48
ata2: dev 0 configured for UDMA/133
scsi1 : ahci
ata3: no device found (phy stat 00000000)
scsi2 : ahci
ata4: no device found (phy stat 00000000)
scsi3 : ahci
   Vendor: ATA       Model: ST380817AS        Rev: 3.42
   Type:   Direct-Access                      ANSI SCSI revision: 05
   Vendor: ATA       Model: ST380817AS        Rev: 3.42
   Type:   Direct-Access                      ANSI SCSI revision: 05


Best regards,


 			Krzysztof Olędzki


PS: Thank you for updating the libata-dev/passthru repository and creating 
the 2.6.12-git4-passthru1 patch. :)

Re: SATA speed. Should be 150 or 133?

[Mark Lord]

True SATA drives ignore the "transfer speed",
as it really is meaningless and does not apply.

But most (all?) first-gen SATA drives are really
PATA drives with a SATA bridge built-in.
Some of those drives require that Linux set the
DMA transfer speed for them to work reliably.

Last I looked, the highest valid PATA transfer
speed was still "UDMA/133".  150 just plain
doesn't exist for PATA (and the whole concept
doesn't exist for SATA, so ..)

Cheers

Re: SATA speed. Should be 150 or 133?

[Krzysztof Oledzki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 732 bytes --]



On Thu, 23 Jun 2005, Mark Lord wrote:

> True SATA drives ignore the "transfer speed",
> as it really is meaningless and does not apply.
So, am I the the only person confused by this message? ;)
There is "SATA max UDMA/133" not "PATA max UDMA/133".

> But most (all?) first-gen SATA drives are really
> PATA drives with a SATA bridge built-in.
> Some of those drives require that Linux set the
> DMA transfer speed for them to work reliably.
Oh, so how to check true (current) speed?

> Last I looked, the highest valid PATA transfer
> speed was still "UDMA/133".  150 just plain
> doesn't exist for PATA (and the whole concept
> doesn't exist for SATA, so ..)
OK :)


Best regards,

 				Krzysztof Olędzki

Re: SATA speed. Should be 150 or 133?

[Mark Lord]

>> True SATA drives ignore the "transfer speed",
>> as it really is meaningless and does not apply.
> 
> So, am I the the only person confused by this message? ;)
> There is "SATA max UDMA/133" not "PATA max UDMA/133".

No, it really is as confusing as it sounds!
The drive is SATA, but the transfer speed gunk only
applies to the internal "PATA" portion of the drive,
which communicates to the built-in SATA bridge chip
of the same drive, which in turn presents a pure SATA
interface to the host computer.

> Oh, so how to check true (current) speed?

Same as always:  hdparm -I /dev/sd?
(requires the libata-dev "passthru" patch
recently reposted here by Jeff Garzik).

cheers

Re: SATA speed. Should be 150 or 133?

[Krzysztof Oledzki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 427 bytes --]



On Fri, 24 Jun 2005, Mark Lord wrote:

<CUT>
>> Oh, so how to check true (current) speed?
> Same as always:  hdparm -I /dev/sd?
> (requires the libata-dev "passthru" patch
> recently reposted here by Jeff Garzik).

Got it :)

# hdparm -I /dev/sda

/dev/sda:
(...)
         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
(...)

This one? udma6?

Best regards,


 			Krzysztof Olędzki

Re: SATA speed. Should be 150 or 133?

[Mark Lord]

Krzysztof Oledzki wrote:
>
> # hdparm -I /dev/sda
> 
> /dev/sda:
> (...)
>         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
> (...)
> 
> This one? udma6?

Yup.  udma6 == UDMA/133

Cheers

Re: SATA speed. Should be 150 or 133?

[Krzysztof Oledzki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 366 bytes --]



On Sat, 25 Jun 2005, Mark Lord wrote:

> Krzysztof Oledzki wrote:
>> 
>> # hdparm -I /dev/sda
>> 
>> /dev/sda:
>> (...)
>>         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
>> (...)
>> 
>> This one? udma6?
>
> Yup.  udma6 == UDMA/133

So, how this line would look like for SATA150?

Best regards,

 			Krzysztof Olędzki

Re: SATA speed. Should be 150 or 133?

[Jeff Garzik]

Krzysztof Oledzki wrote:
> 
> 
> On Sat, 25 Jun 2005, Mark Lord wrote:
> 
>> Krzysztof Oledzki wrote:
>>
>>>
>>> # hdparm -I /dev/sda
>>>
>>> /dev/sda:
>>> (...)
>>>         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 
>>> *udma6
>>> (...)
>>>
>>> This one? udma6?
>>
>>
>> Yup.  udma6 == UDMA/133
> 
> 
> So, how this line would look like for SATA150?

Unchanged.  SATA150 only goes to udma6.

	Jeff

Re: SATA speed. Should be 150 or 133?

[Krzysztof Oledzki]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 691 bytes --]



On Sun, 26 Jun 2005, Jeff Garzik wrote:

> Krzysztof Oledzki wrote:
>> 
>> 
>> On Sat, 25 Jun 2005, Mark Lord wrote:
>> 
>>> Krzysztof Oledzki wrote:
>>> 
>>>> 
>>>> # hdparm -I /dev/sda
>>>> 
>>>> /dev/sda:
>>>> (...)
>>>>         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
>>>> (...)
>>>> 
>>>> This one? udma6?
>>> 
>>> 
>>> Yup.  udma6 == UDMA/133
>> 
>> 
>> So, how this line would look like for SATA150?
>
> Unchanged.  SATA150 only goes to udma6.

So, excuse me for asking this question again: how to check true 
(current) speed? I still don't know if this disk uses UDMA/133 or SATA150.

Best regards,

 			Krzysztof Olędzki

Re: SATA speed. Should be 150 or 133?

[Jeff Garzik]

Krzysztof Oledzki wrote:
> So, excuse me for asking this question again: how to check true 
> (current) speed? I still don't know if this disk uses UDMA/133 or SATA150.

This is an irrelevant question, because the answer is "both."

The SATA bus speed is always 1.5Gbps, 3Gbps, etc. but the device must 
also set udma speed (udma0-6).

	Jeff

Re: SATA speed. Should be 150 or 133?

[Stefan Smietanowski]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi. This is directed toward Krzysztof and not you Jeff.

>> So, excuse me for asking this question again: how to check true
>> (current) speed? I still don't know if this disk uses UDMA/133 or
>> SATA150.
> 
> 
> This is an irrelevant question, because the answer is "both."
> 
> The SATA bus speed is always 1.5Gbps, 3Gbps, etc. but the device must
> also set udma speed (udma0-6).

Or simply speaking, if your device is SATA150 it has a 1.5Gbps link to
the IDE chipset always.
If the device is UDMA133 with a bridge (built-in or external) then
the link is 1.5Gbps between the IDE chipset and the bridge and UDMA0-6
between the drive and the bridge. Many older SATA drives were nothing
more than UDMA133 drives with a bridge and I believe most (all?) used
a UDMA133<->SATA150 bridge by Marvell.

So if you want a finite question : Look on the electronics board
of the harddrive and if you find a Marvell chip near the SATA connector
then your drive is UDMA133 with a bridge chip. You can naturally
check any resources on the internet if you so chose.

// Stefan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (MingW32)

iD8DBQFCvveoBrn2kJu9P78RArhJAJsEvbtbhvvUYjwYh46hRKVMmN6UiACgjGAr
AjR46vZNydzIR0tMDLg1CKc=
=t20w
-----END PGP SIGNATURE-----

Re: SATA speed. Should be 150 or 133?

[Mark Lord]

Krzysztof Oledzki wrote:
> 
> So, excuse me for asking this question again: how to check true 
> (current) speed? I still don't know if this disk uses UDMA/133 or SATA150.

Inside the drive itself, it *may* be using UDMA/133 to transfer data
between the PATA circuitry and the SATA bridge chip (inside the drive).
Not all drives are built that way (in which case the drive is merely
trying to keep people like you happy by reporting these obsolete fields)
but many SATA drives are.

Outside the drive, over the SATA cable, it is SATA150 just like
every other SATA1 drive.

-ml

Re: SATA speed. Should be 150 or 133?

[Mark Lord]

Krzysztof Oledzki wrote:
>
>>> # hdparm -I /dev/sda
>>>
>>> /dev/sda:
>>> (...)
>>>         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
..
> So, how this line would look like for SATA150?

It wouldn't exist.  No such concept.
But we display it anyway, because most SATA drives
are really PATA drives in disguise.

-ml