libata update posted (was Re: VIA Serial ATA chipset)

libata update posted (was Re: VIA Serial ATA chipset)

[Jeff Garzik]

[-- Attachment #1: Type: text/plain, Size: 2971 bytes --]

Daniela Engert wrote:
> On Tue, 12 Aug 2003 23:49:23 -0700, Misha Nasledov wrote:
> 
> 
>>Thanks a lot, this patch works for me. I have the following in my dmesg:
>>
>>VIA8237SATA: IDE controller at PCI slot 0000:00:0f.0
>>PCI: Found IRQ 3 for device 0000:00:0f.0
>>IRQ routing conflict for 0000:00:0f.0, have irq 11, want irq 3
>>PCI: Sharing IRQ 3 with 0000:00:0f.1
>>IRQ routing conflict for 0000:00:10.0, have irq 11, want irq 3
>>IRQ routing conflict for 0000:00:10.1, have irq 11, want irq 3
>>VIA8237SATA: chipset revision 128
>>VIA8237SATA: 100% native mode on irq 11
>>   ide4: BM-DMA at 0x8400-0x8407, BIOS settings: hdi:pio, hdj:pio
>>   ide5: BM-DMA at 0x8408-0x840f, BIOS settings: hdk:pio, hdl:pio
>>
>>The IRQ routing conflict lines have always existed, and there are more
>>similar lines in other parts of the dmesg. I'm not sure if they are
>>pertinent at all. Unfortunately, the drive does not seem to perform as
>>well as a SATA drive should; my UDMA/133 drive currently outperforms
>>this. However, it seems to significantly faster than the UDMA/66 drive
>>I run Linux on, so I am glad that it works as fast it does because now
>>I can perhaps use this as my new system drive.
> 
> 
> Well, the patch was meant to be a quick shot from the hips only, just
> to get your system going and to check if my assumptions about the
> structure of the VIA SATA host controller implementation wer correct.
> It looks like they are, this controller seems to follow the T13 ATA
> host controller model.

Correct.


>>Are there any plans on the horizon for a more complete driver for this
>>chipset?
> 
> 
> Jeff Garzik is working on a SATA driver. My conclusions from the
> results of the feasability test above are that adding the VIA SATA
> controller to the already existing libata source should boil down to
> pretty simple programming excercise.

It was.  The sata_via.c driver is absolutely generic except for its PCI 
ID.  I only chose to create a separate driver because I now have docs 
(thanks, VIA), so I can begin programming the SATA-specific registers.

So, here are the 2.4 and 2.6 versions of the VIA SATA support for 
libata, contained in the latest libata update.

2.4 BK: bk://kernel.bkbits.net/jgarzik/atascsi-2.4
2.4 Patch: 
ftp://ftp.kernel.org/pub/linux/kernel/people/jgarzik/patchkits/2.4/2.4.23-pre1-libata3.patch.bz2

2.6 BK: bk://kernel.bkbits.net/jgarzik/atascsi-2.6
2.6 Patch: 
ftp://ftp.kernel.org/pub/linux/kernel/people/jgarzik/patchkits/2.6/2.6.0-test4-bk2-libata3.patch.bz2


Changes:
* add VIA SATA driver
* fixes to software reset.
* other fixes
* split scsi-related code into separate file, libata-scsi.c.
* continue work on phy layer
* continue work towards fully async taskfile API:  you call submit_tf(), 
and later on, your callback is called when the taskfile completes or 
times out.   async taskfile API is required for ATAPI and supporting 
more advanced host controllers like Promise or AHCI (SATA2).
* some cleanups


[-- Attachment #2: patch.srst --]
[-- Type: text/plain, Size: 1029 bytes --]

===== drivers/scsi/ata_piix.c 1.22 vs edited =====
--- 1.22/drivers/scsi/ata_piix.c	Thu Aug 28 22:07:27 2003
+++ edited/drivers/scsi/ata_piix.c	Fri Aug 29 00:51:44 2003
@@ -137,7 +137,7 @@
 	/* ich5_pata */
 	{
 		.sht		= &piix_sht,
-		.host_flags	= ATA_FLAG_SLAVE_POSS,
+		.host_flags	= ATA_FLAG_SLAVE_POSS | ATA_FLAG_SRST,
 		.pio_mask	= 0x03,	/* pio3-4 */
 		.udma_mask	= ATA_UDMA_MASK_40C, /* FIXME: cbl det */
 		.host_info	= &piix_pata_ops,
@@ -146,7 +146,8 @@
 	/* ich5_sata */
 	{
 		.sht		= &piix_sht,
-		.host_flags	= ATA_FLAG_SATA | PIIX_FLAG_COMBINED,
+		.host_flags	= ATA_FLAG_SATA | PIIX_FLAG_COMBINED
+				  | ATA_FLAG_SRST,
 		.pio_mask	= 0x03,	/* pio3-4 */
 		.udma_mask	= 0x7f,	/* udma0-6 ; FIXME */
 		.host_info	= &piix_sata_ops,
@@ -155,7 +156,7 @@
 	/* piix4_pata */
 	{
 		.sht		= &piix_sht,
-		.host_flags	= ATA_FLAG_SLAVE_POSS,
+		.host_flags	= ATA_FLAG_SLAVE_POSS | ATA_FLAG_SRST,
 		.pio_mask	= 0x03, /* pio3-4 */
 		.udma_mask	= ATA_UDMA_MASK_40C, /* FIXME: cbl det */
 		.host_info	= &piix_pata_ops,

Re: libata update posted (was Re: VIA Serial ATA chipset)

[Samuel Flory]

Jeff Garzik wrote:

>
> So, here are the 2.4 and 2.6 versions of the VIA SATA support for 
> libata, contained in the latest libata update.
>
> 2.4 BK: bk://kernel.bkbits.net/jgarzik/atascsi-2.4
> 2.4 Patch: 
> ftp://ftp.kernel.org/pub/linux/kernel/people/jgarzik/patchkits/2.4/2.4.23-pre1-libata3.patch.bz2 
>
>
> 2.6 BK: bk://kernel.bkbits.net/jgarzik/atascsi-2.6
> 2.6 Patch: 
> ftp://ftp.kernel.org/pub/linux/kernel/people/jgarzik/patchkits/2.6/2.6.0-test4-bk2-libata3.patch.bz2 
>
>
>
> Changes:
> * add VIA SATA driver
> * fixes to software reset.
> * other fixes
> * split scsi-related code into separate file, libata-scsi.c.
> * continue work on phy layer
> * continue work towards fully async taskfile API:  you call 
> submit_tf(), and later on, your callback is called when the taskfile 
> completes or times out.   async taskfile API is required for ATAPI and 
> supporting more advanced host controllers like Promise or AHCI (SATA2).
> * some cleanups
>
>  
>

   I'm guessing there is no support for Promise yet? 

  I'm just asking because I've got this dual amd64 with 4 Promise sata 
ports.  Which for some reason there aren't drivers any 64 bit OS;-(


PS-  The driver works great on the silcon image chipset.  (Once I 
realized that my Seagate drive needed newer firmware.)

-- 
Once you have their hardware. Never give it back.
(The First Rule of Hardware Acquisition)
Sam Flory  <sflory@rackable.com>

Re: libata update posted (was Re: VIA Serial ATA chipset)

[Jeff Garzik]

Samuel Flory wrote:
> Jeff Garzik wrote:
>> Changes:
>> * continue work towards fully async taskfile API:  you call 
>> submit_tf(), and later on, your callback is called when the taskfile 
>> completes or times out.   async taskfile API is required for ATAPI and 
>> supporting more advanced host controllers like Promise or AHCI (SATA2).
>> * some cleanups

>   I'm guessing there is no support for Promise yet?

Not yet.  Once I finish the item mentioned above, "async taskfile API", 
Promise support will appear quite rapidly.


> PS-  The driver works great on the silcon image chipset.  (Once I 
> realized that my Seagate drive needed newer firmware.)

Um... libata doesn't support Silicon Image yet?

	Jeff

Re: libata update posted (was Re: VIA Serial ATA chipset)

[Samuel Flory]

Jeff Garzik wrote:

> Samuel Flory wrote:
>
>> Jeff Garzik wrote:
>>
>>> Changes:
>>> * continue work towards fully async taskfile API:  you call 
>>> submit_tf(), and later on, your callback is called when the taskfile 
>>> completes or times out.   async taskfile API is required for ATAPI 
>>> and supporting more advanced host controllers like Promise or AHCI 
>>> (SATA2).
>>> * some cleanups
>>
>
>>   I'm guessing there is no support for Promise yet?
>
>
> Not yet.  Once I finish the item mentioned above, "async taskfile 
> API", Promise support will appear quite rapidly.
>
>
>> PS-  The driver works great on the silcon image chipset.  (Once I 
>> realized that my Seagate drive needed newer firmware.)
>
>
> Um... libata doesn't support Silicon Image yet?
>


  The Intel PIIX/ICH support in the ac kernel seems to work on all my 
onboard  Silicon Image controllers.  I think that's only your ata-scsi 
work, however.

-- 
Once you have their hardware. Never give it back.
(The First Rule of Hardware Acquisition)
Sam Flory  <sflory@rackable.com>