[patch 25/30] pata_hpt3x3: clean up by using cable method

[patch 25/30] pata_hpt3x3: clean up by using cable method

[akpm]

From: Alan Cox <alan@lxorguk.ukuu.org.uk>

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 drivers/ata/pata_hpt3x3.c |   22 ++--------------------
 1 file changed, 2 insertions(+), 20 deletions(-)

diff -puN drivers/ata/pata_hpt3x3.c~pata_hpt3x3-clean-up-by-using-cable-method drivers/ata/pata_hpt3x3.c
--- a/drivers/ata/pata_hpt3x3.c~pata_hpt3x3-clean-up-by-using-cable-method
+++ a/drivers/ata/pata_hpt3x3.c
@@ -25,25 +25,6 @@
 #define DRV_NAME	"pata_hpt3x3"
 #define DRV_VERSION	"0.4.2"
 
-static int hpt3x3_probe_init(struct ata_port *ap)
-{
-	ap->cbl = ATA_CBL_PATA40;
-	return ata_std_prereset(ap);
-}
-
-/**
- *	hpt3x3_probe_reset	-	reset the hpt3x3 bus
- *	@ap: ATA port to reset
- *
- *	Perform the housekeeping when doing an ATA bus reeset. We just
- *	need to force the cable type.
- */
-
-static void hpt3x3_error_handler(struct ata_port *ap)
-{
-	return ata_bmdma_drive_eh(ap, hpt3x3_probe_init, ata_std_softreset, NULL, ata_std_postreset);
-}
-
 /**
  *	hpt3x3_set_piomode		-	PIO setup
  *	@ap: ATA interface
@@ -139,8 +120,9 @@ static struct ata_port_operations hpt3x3
 
 	.freeze		= ata_bmdma_freeze,
 	.thaw		= ata_bmdma_thaw,
-	.error_handler	= hpt3x3_error_handler,
+	.error_handler	= ata_bmdma_error_handler,
 	.post_internal_cmd = ata_bmdma_post_internal_cmd,
+	.cable_detect	= ata_cable_40wire,
 
 	.bmdma_setup 	= ata_bmdma_setup,
 	.bmdma_start 	= ata_bmdma_start,
_

Re: [patch 25/30] pata_hpt3x3: clean up by using cable method

[Jeff Garzik]

akpm@linux-foundation.org wrote:
> From: Alan Cox <alan@lxorguk.ukuu.org.uk>
> 
> Signed-off-by: Alan Cox <alan@redhat.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
> 
>  drivers/ata/pata_hpt3x3.c |   22 ++--------------------
>  1 file changed, 2 insertions(+), 20 deletions(-)

dropped patches 9-25.

they should all be rolled up into a single "update ->cable_detect hook" 
patch, as discussed.

Re: [patch 25/30] pata_hpt3x3: clean up by using cable method

[Jeff Garzik]

Alan Cox wrote:
>> dropped patches 9-25.
>>
>> they should all be rolled up into a single "update ->cable_detect hook" 
>> patch, as discussed.
> 
> Strongly disagree. Each cable detect for a driver is testable alone which
> means each one is a git bisect point and you can find which driver has
> broken that way.

For most testers, you can generally find out which driver is broken 
through intuition and simple observation, without having to resort to 
git bisect.

	Jeff

Re: [patch 25/30] pata_hpt3x3: clean up by using cable method

[Alan Cox]

> dropped patches 9-25.
> 
> they should all be rolled up into a single "update ->cable_detect hook" 
> patch, as discussed.

Strongly disagree. Each cable detect for a driver is testable alone which
means each one is a git bisect point and you can find which driver has
broken that way.

Alan

Re: [patch 25/30] pata_hpt3x3: clean up by using cable method

[Alan Cox]

On Tue, 06 Mar 2007 09:25:34 -0500
Jeff Garzik <jeff@garzik.org> wrote:

> through intuition and simple observation, without having to resort to 
> git bisect.

Well if you want one patch you can have one patch. It'll be a while as
I'm still converting drivers over and I'll feed them one by one for real
world testing/verifying to Andrew first. Some like the Jmicron and
pata_sis aren't so trivial although are well worth doing.

Alan