[PATCH #upstream-fixes] sata_nv: use ata_pci_sff_activate_host() instead of ata_host_activate()

[PATCH #upstream-fixes] sata_nv: use ata_pci_sff_activate_host() instead of ata_host_activate()

[Tejun Heo]

sata_nv was incorrectly using ata_host_activate() instead of
ata_pci_sff_activate_host() leading to IRQ assignment failure in
legacy mode.  Fix it.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Robert Hancock <hancockr@shaw.ca>
Cc: stable@kernel.org
---
 drivers/ata/sata_nv.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

Index: ata/drivers/ata/sata_nv.c
===================================================================
--- ata.orig/drivers/ata/sata_nv.c
+++ ata/drivers/ata/sata_nv.c
@@ -2479,8 +2479,7 @@ static int nv_init_one(struct pci_dev *p
 	}

 	pci_set_master(pdev);
-	return ata_host_activate(host, pdev->irq, ipriv->irq_handler,
-				 IRQF_SHARED, ipriv->sht);
+	return ata_pci_sff_activate_host(host, ipriv->irq_handler, ipriv->sht);
 }

 #ifdef CONFIG_PM

Re: [PATCH #upstream-fixes] sata_nv: use ata_pci_sff_activate_host() instead of ata_host_activate()

[Jeff Garzik]

On 05/14/2010 05:48 AM, Tejun Heo wrote:
> sata_nv was incorrectly using ata_host_activate() instead of
> ata_pci_sff_activate_host() leading to IRQ assignment failure in
> legacy mode.  Fix it.
>
> Signed-off-by: Tejun Heo<tj@kernel.org>
> Cc: Robert Hancock<hancockr@shaw.ca>
> Cc: stable@kernel.org
> ---
>   drivers/ata/sata_nv.c |    3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
>
> Index: ata/drivers/ata/sata_nv.c
> ===================================================================
> --- ata.orig/drivers/ata/sata_nv.c
> +++ ata/drivers/ata/sata_nv.c
> @@ -2479,8 +2479,7 @@ static int nv_init_one(struct pci_dev *p
>   	}
>
>   	pci_set_master(pdev);
> -	return ata_host_activate(host, pdev->irq, ipriv->irq_handler,
> -				 IRQF_SHARED, ipriv->sht);
> +	return ata_pci_sff_activate_host(host, ipriv->irq_handler, ipriv->sht);

hmmmm.  I agree 100% with this patch, but is it a recent regression?

At -rc7, we are trying hard to only include fixes for regressions.

Around -rc1, I would stuff this into #upstream-fixes without hesitation. 
  But now, I'm thinking #upstream.

	Jeff

Re: [PATCH #upstream-fixes] sata_nv: use ata_pci_sff_activate_host() instead of ata_host_activate()

[Tejun Heo]

Hello,

On 05/14/2010 11:12 PM, Jeff Garzik wrote:
> hmmmm.  I agree 100% with this patch, but is it a recent regression?

It's more than a year old at least.  It's a bit surprising that no one
reported this before.

> At -rc7, we are trying hard to only include fixes for regressions.
> 
> Around -rc1, I would stuff this into #upstream-fixes without hesitation.
> But now, I'm thinking #upstream.

Yeap, sure.  We can wait some time after it makes -rc1 and then push
it through -stable.

Thanks.

-- 
tejun

Re: [PATCH #upstream-fixes] sata_nv: use ata_pci_sff_activate_host() instead of ata_host_activate()

[Jeff Garzik]

On 05/14/2010 05:48 AM, Tejun Heo wrote:
> sata_nv was incorrectly using ata_host_activate() instead of
> ata_pci_sff_activate_host() leading to IRQ assignment failure in
> legacy mode.  Fix it.
>
> Signed-off-by: Tejun Heo<tj@kernel.org>
> Cc: Robert Hancock<hancockr@shaw.ca>
> Cc: stable@kernel.org
> ---
>   drivers/ata/sata_nv.c |    3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)

applied #upstream