[RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[Adrian Bunk]

Users should use the libata based drivers for SATA drives.

Signed-off-by: Adrian Bunk <bunk@stusta.de>

---
--- linux-2.6.22-rc4-mm2/Documentation/feature-removal-schedule.txt.old	2007-06-21 23:41:03.000000000 +0200
+++ linux-2.6.22-rc4-mm2/Documentation/feature-removal-schedule.txt	2007-06-21 23:42:03.000000000 +0200
@@ -347,3 +347,10 @@
 Who:	Jean Delvare <khali@linux-fr.org>
 
 ---------------------------
+
+What:  CONFIG_BLK_DEV_IDE_SATA
+When:  December 2007
+Why:   users should use the libata based drivers for SATA
+Who:   Adrian Bunk <bunk@stusta.de>
+
+---------------------------

Re: [RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[David Woodhouse]

On Fri, 2007-06-22 at 01:52 +0200, Adrian Bunk wrote:
> Users should use the libata based drivers for SATA drives.

NAK. Not all IDE drivers are converted yet. Not even all the relatively
common ones.

-- 
dwmw2

Re: [RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[David Woodhouse]

On Fri, 2007-06-22 at 11:39 +0800, David Woodhouse wrote:
> On Fri, 2007-06-22 at 01:52 +0200, Adrian Bunk wrote:
> > Users should use the libata based drivers for SATA drives.
> 
> NAK. Not all IDE drivers are converted yet. Not even all the relatively
> common ones.

Ignore me. I thought you were trying to deprecate BLK_DEV_IDE. Sorry.

/me wanders off in search of more caffeine....

-- 
dwmw2

Re: [RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[Alan Cox]

On Fri, 22 Jun 2007 11:39:45 +0800
David Woodhouse <dwmw2@infradead.org> wrote:

> On Fri, 2007-06-22 at 01:52 +0200, Adrian Bunk wrote:
> > Users should use the libata based drivers for SATA drives.
> 
> NAK. Not all IDE drivers are converted yet. Not even all the relatively
> common ones.

All the common SATA ones are. Just the PowerPC people are falling behind.
Really we should obsolete powerpc instead ;)

Alan

Re: [RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[Paul Mackerras]

Alan Cox writes:

> On Fri, 22 Jun 2007 11:39:45 +0800
> David Woodhouse <dwmw2@infradead.org> wrote:
> 
> > On Fri, 2007-06-22 at 01:52 +0200, Adrian Bunk wrote:
> > > Users should use the libata based drivers for SATA drives.
> > 
> > NAK. Not all IDE drivers are converted yet. Not even all the relatively
> > common ones.
> 
> All the common SATA ones are. Just the PowerPC people are falling behind.
> Really we should obsolete powerpc instead ;)

What SATA drive uses a non-libata driver on PowerPC?  In fact the
sata_svw driver used for SATA drives on G5s was one of the early
libata drivers.  Or are you talking about PATA?

Paul.

Re: [RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[Jeff Garzik]

Paul Mackerras wrote:
> Alan Cox writes:
> 
>> On Fri, 22 Jun 2007 11:39:45 +0800
>> David Woodhouse <dwmw2@infradead.org> wrote:
>>
>>> On Fri, 2007-06-22 at 01:52 +0200, Adrian Bunk wrote:
>>>> Users should use the libata based drivers for SATA drives.
>>> NAK. Not all IDE drivers are converted yet. Not even all the relatively
>>> common ones.
>> All the common SATA ones are. Just the PowerPC people are falling behind.
>> Really we should obsolete powerpc instead ;)
> 
> What SATA drive uses a non-libata driver on PowerPC?  In fact the
> sata_svw driver used for SATA drives on G5s was one of the early
> libata drivers.  Or are you talking about PATA?

This entire thread is confused :)

Going back to $subject, it is relevant only to the scant few old-IDE 
drivers that do indeed successfully probe and talk to SATA drives, from 
SATA controllers.

The vast majority have always used libata for SATA, but another option 
did indeed exist before libata appeared, for a few controllers.

Alan was just having a bit of fun at dwmw2's expense, since we do indeed 
want to see libata drivers for PowerPC PATA as well :)

	Jeff

Re: [RFC: 2.6 patch] schedule BLK_DEV_IDE_SATA for removal

[Alan Cox]

On Fri, 22 Jun 2007 01:52:11 +0200
Adrian Bunk <bunk@stusta.de> wrote:

> Users should use the libata based drivers for SATA drives.
> 
> Signed-off-by: Adrian Bunk <bunk@stusta.de>
> 
> ---
> --- linux-2.6.22-rc4-mm2/Documentation/feature-removal-schedule.txt.old	2007-06-21 23:41:03.000000000 +0200
> +++ linux-2.6.22-rc4-mm2/Documentation/feature-removal-schedule.txt	2007-06-21 23:42:03.000000000 +0200
> @@ -347,3 +347,10 @@
>  Who:	Jean Delvare <khali@linux-fr.org>
>  
>  ---------------------------
> +
> +What:  CONFIG_BLK_DEV_IDE_SATA
> +When:  December 2007
> +Why:   users should use the libata based drivers for SATA
> +Who:   Adrian Bunk <bunk@stusta.de>
> +
> +---------------------------

NAK this for now

For the standard SATA controllers it is probably true now and will
definitely be true by December 2007 but there is lots of stuff out there
which uses SATA/PATA bridges on older PATA controllers such as the
VIA and HPT controllers. While support for them varies between rock solid
(VIA) and passable (HPT), I don't think its yet time to bump those people
towards libata forcibly at all.

About the only thing this would meaningfully obsolete would be
ide/pci/siimage.c for SI3112 and some PCI identifiers from ide/pci/piix.c.