[PATCH] pata_sis: Remove bogus cable match

[PATCH] pata_sis: Remove bogus cable match

[Alan Cox]

From: Alan Cox <alan@redhat.com>

Some systems report SIS 5513 as both vendor/id and subvendor/id string. In
that case we can't distinguish the system by the id svid/sdid and in fact
the entry here breaks some boxes. At some point we need to find another way
to detect the Targa Visionary 1000, until then this a hang for some users with
lower performance for others.

Closes: #12092

Signed-off-by: Alan Cox <alan@redhat.com>
---

 drivers/ata/pata_sis.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)


diff --git a/drivers/ata/pata_sis.c b/drivers/ata/pata_sis.c
index d342366..e4be55e 100644
--- a/drivers/ata/pata_sis.c
+++ b/drivers/ata/pata_sis.c
@@ -56,7 +56,6 @@ static const struct sis_laptop sis_laptop[] = {
 	{ 0x5513, 0x1043, 0x1107 },	/* ASUS A6K */
 	{ 0x5513, 0x1734, 0x105F },	/* FSC Amilo A1630 */
 	{ 0x5513, 0x1071, 0x8640 },     /* EasyNote K5305 */
-	{ 0x5513, 0x1039, 0x5513 },	/* Targa Visionary 1000 */
 	/* end marker */
 	{ 0, }
 };

Re: [PATCH] pata_sis: Remove bogus cable match

[Jeff Garzik]

Alan Cox wrote:
> From: Alan Cox <alan@redhat.com>
> 
> Some systems report SIS 5513 as both vendor/id and subvendor/id string. In
> that case we can't distinguish the system by the id svid/sdid and in fact
> the entry here breaks some boxes. At some point we need to find another way
> to detect the Targa Visionary 1000, until then this a hang for some users with
> lower performance for others.
> 
> Closes: #12092
> 
> Signed-off-by: Alan Cox <alan@redhat.com>

Should that last sentence be "this SOLVES a hang"?

As it is currently worded, it is unclear what we are trading off with 
this patch...

Thanks,

	Jeff

Re: [PATCH] pata_sis: Remove bogus cable match

[Alan Cox]

On Fri, 05 Dec 2008 15:31:40 -0500
Jeff Garzik <jeff@garzik.org> wrote:

> Alan Cox wrote:
> > From: Alan Cox <alan@redhat.com>
> > 
> > Some systems report SIS 5513 as both vendor/id and subvendor/id string. In
> > that case we can't distinguish the system by the id svid/sdid and in fact
> > the entry here breaks some boxes. At some point we need to find another way
> > to detect the Targa Visionary 1000, until then this a hang for some users with
> > lower performance for others.
> > 
> > Closes: #12092
> > 
> > Signed-off-by: Alan Cox <alan@redhat.com>
> 
> Should that last sentence be "this SOLVES a hang"?
> 
> As it is currently worded, it is unclear what we are trading off with 
> this patch...

I missed out the word "trades" ...

Re: [PATCH] pata_sis: Remove bogus cable match

[Kai Krakow]

Am Samstag, 06. Dezember 2008 schrieb Alan Cox:

> On Fri, 05 Dec 2008 15:31:40 -0500
>
> Jeff Garzik <jeff@garzik.org> wrote:
> > Alan Cox wrote:
> > > From: Alan Cox <alan@redhat.com>
> > >
> > > Some systems report SIS 5513 as both vendor/id and subvendor/id string.
> > > In that case we can't distinguish the system by the id svid/sdid and in
> > > fact the entry here breaks some boxes. At some point we need to find
> > > another way to detect the Targa Visionary 1000, until then this a hang
> > > for some users with lower performance for others.
> > >
> > > Closes: #12092
> > >
> > > Signed-off-by: Alan Cox <alan@redhat.com>
> >
> > Should that last sentence be "this SOLVES a hang"?
> >
> > As it is currently worded, it is unclear what we are trading off with
> > this patch...
>
> I missed out the word "trades" ...

Is there away to bypass this bogus-detection and force the driver into using 
high-speed mode on the short 40-wire cable? Maybe by introducing a 
kernel/module parameter? Or is it there? I found nothing in the docs.

Can I provide any information you need to properly detect the Targa Visionary 
1000?

Thanks,
Kai

Re: [PATCH] pata_sis: Remove bogus cable match

[Alan Cox]

> Is there away to bypass this bogus-detection and force the driver into using 
> high-speed mode on the short 40-wire cable? Maybe by introducing a 
> kernel/module parameter? Or is it there? I found nothing in the docs.
> 
> Can I provide any information you need to properly detect the Targa Visionary 
> 1000?

Thats why I cc'd you on the changes. I am sure we can find something to
identify this machine.

Can you send the output of dmidecode ?

Alan

Re: [PATCH] pata_sis: Remove bogus cable match

[Jeff Garzik]

Alan Cox wrote:
> From: Alan Cox <alan@redhat.com>
> 
> Some systems report SIS 5513 as both vendor/id and subvendor/id string. In
> that case we can't distinguish the system by the id svid/sdid and in fact
> the entry here breaks some boxes. At some point we need to find another way
> to detect the Targa Visionary 1000, until then this a hang for some users with
> lower performance for others.
> 
> Closes: #12092
> 
> Signed-off-by: Alan Cox <alan@redhat.com>
> ---
> 
>  drivers/ata/pata_sis.c |    1 -
>  1 files changed, 0 insertions(+), 1 deletions(-)
> 
> 
> diff --git a/drivers/ata/pata_sis.c b/drivers/ata/pata_sis.c
> index d342366..e4be55e 100644
> --- a/drivers/ata/pata_sis.c
> +++ b/drivers/ata/pata_sis.c
> @@ -56,7 +56,6 @@ static const struct sis_laptop sis_laptop[] = {
>  	{ 0x5513, 0x1043, 0x1107 },	/* ASUS A6K */
>  	{ 0x5513, 0x1734, 0x105F },	/* FSC Amilo A1630 */
>  	{ 0x5513, 0x1071, 0x8640 },     /* EasyNote K5305 */
> -	{ 0x5513, 0x1039, 0x5513 },	/* Targa Visionary 1000 */
>  	/* end marker */

applied #upstream-fixes