Remove Asus EEE UDMA/33 limitation.

Remove Asus EEE UDMA/33 limitation.

[Dave Jones]

The Asus Eee PC prints this during bootup..

ata2.00: limited to UDMA/33 due to 40-wire cable
Which is unfortunate, as it has..
ata2.00: ATA-4: SILICONMOTION SM223AC, , max UDMA/66

I believe there actually is no cable, and the flash is directly
connected to the controller.

[root@eee ~]# hdparm -t /dev/sda

/dev/sda:
 Timing buffered disk reads:   64 MB in  3.08 seconds =  20.76 MB/sec

With the patch below it returns..

/dev/sda:
 Timing buffered disk reads:   76 MB in  3.04 seconds =  25.04 MB/sec


Signed-off-by: Dave Jones <davej@redhat.com>

--- linux-2.6.24.noarch/drivers/ata/ata_piix.c~	2008-03-31 23:43:05.000000000 -0400
+++ linux-2.6.24.noarch/drivers/ata/ata_piix.c	2008-03-31 23:44:34.000000000 -0400
@@ -716,6 +716,7 @@ struct ich_laptop {
 
 static const struct ich_laptop ich_laptop[] = {
 	/* devid, subvendor, subdev */
+	{ 0x2653, 0x1043, 0x82d8 },	/* ICH6M on Asus Eee PC */
 	{ 0x27DF, 0x0005, 0x0280 },	/* ICH7 on Acer 5602WLMi */
 	{ 0x27DF, 0x1025, 0x0102 },	/* ICH7 on Acer 5602aWLMi */
 	{ 0x27DF, 0x1025, 0x0110 },	/* ICH7 on Acer 3682WLMi */

-- 
http://www.codemonkey.org.uk

Re: Remove Asus EEE UDMA/33 limitation.

[Dave Jones]

On Fri, May 02, 2008 at 11:44:58AM -0400, Dave Jones wrote:
 > The Asus Eee PC prints this during bootup..

never mind, someone just pointed out 01ce2601e4ba354fe1e25bb940817570d0c8ed4f
which does the same thing.

	Dave

-- 
http://www.codemonkey.org.uk

Re: Remove Asus EEE UDMA/33 limitation.

[Matthew Garrett]

On Fri, May 02, 2008 at 11:44:58AM -0400, Dave Jones wrote:

>  static const struct ich_laptop ich_laptop[] = {
>  	/* devid, subvendor, subdev */
> +	{ 0x2653, 0x1043, 0x82d8 },	/* ICH6M on Asus Eee PC */
>  	{ 0x27DF, 0x0005, 0x0280 },	/* ICH7 on Acer 5602WLMi */
>  	{ 0x27DF, 0x1025, 0x0102 },	/* ICH7 on Acer 5602aWLMi */
>  	{ 0x27DF, 0x1025, 0x0110 },	/* ICH7 on Acer 3682WLMi */

Don't the ACPI timing methods let us work this out without requiring 
machine specific knowledge? Or is the concern that some firmware will 
let us program high modes even if the cabling isn't up to it?

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: Remove Asus EEE UDMA/33 limitation.

[Alan Cox]

On Sat, 3 May 2008 10:58:17 +0100
Matthew Garrett <mjg59@srcf.ucam.org> wrote:

> On Fri, May 02, 2008 at 11:44:58AM -0400, Dave Jones wrote:
> 
> >  static const struct ich_laptop ich_laptop[] = {
> >  	/* devid, subvendor, subdev */
> > +	{ 0x2653, 0x1043, 0x82d8 },	/* ICH6M on Asus Eee PC */
> >  	{ 0x27DF, 0x0005, 0x0280 },	/* ICH7 on Acer 5602WLMi */
> >  	{ 0x27DF, 0x1025, 0x0102 },	/* ICH7 on Acer 5602aWLMi */
> >  	{ 0x27DF, 0x1025, 0x0110 },	/* ICH7 on Acer 3682WLMi */
> 
> Don't the ACPI timing methods let us work this out without requiring 
> machine specific knowledge? 

We don't currently sniff the ACPI timing data on Intel boxes - we can do
so and there is a library routine but I'd be worried about the usual
standard of BIOS code. Right now we only do it for chipsets which are so
braindamaged we have no other choice or where the vendor specifically
implemented and intended the ACPI data to be used for this (eg Nvidia)

Re: Remove Asus EEE UDMA/33 limitation.

[Robert Hancock]

Alan Cox wrote:
> On Sat, 3 May 2008 10:58:17 +0100
> Matthew Garrett <mjg59@srcf.ucam.org> wrote:
> 
>> On Fri, May 02, 2008 at 11:44:58AM -0400, Dave Jones wrote:
>>
>>>  static const struct ich_laptop ich_laptop[] = {
>>>  	/* devid, subvendor, subdev */
>>> +	{ 0x2653, 0x1043, 0x82d8 },	/* ICH6M on Asus Eee PC */
>>>  	{ 0x27DF, 0x0005, 0x0280 },	/* ICH7 on Acer 5602WLMi */
>>>  	{ 0x27DF, 0x1025, 0x0102 },	/* ICH7 on Acer 5602aWLMi */
>>>  	{ 0x27DF, 0x1025, 0x0110 },	/* ICH7 on Acer 3682WLMi */
>> Don't the ACPI timing methods let us work this out without requiring 
>> machine specific knowledge? 
> 
> We don't currently sniff the ACPI timing data on Intel boxes - we can do
> so and there is a library routine but I'd be worried about the usual
> standard of BIOS code. Right now we only do it for chipsets which are so
> braindamaged we have no other choice or where the vendor specifically
> implemented and intended the ACPI data to be used for this (eg Nvidia)

Given that the bootup ACPI _GTM results will be used by Windows to 
determine the DMA mode setting with the standard Microsoft atapi.sys 
driver which PATA controllers would typically use, and Windows is what 
the majority of this hardware was designed to run, I think it should be 
quite trustworthy..