[PATCH] sata_sil24: Identify which card suffered IRQ status error

[PATCH] sata_sil24: Identify which card suffered IRQ status error

[Tim Small]

In machines with multiple Silicon Image 3124 and/or 3132 cards, there is no
way to tell which card is the culprit when the sata_sil24 interrupt handler
gets a bad status.

Tested-by: tim@seoss.co.uk
Signed-off-by: tim@seoss.co.uk

---

I got the following fault on a machine with five SiI3124 / 3132 cards in total
so I patched the driver to report the PCI slot number too.  The resulting code
seems to identify the card as desired...

old output:

sata_sil24: IRQ status == 0xffffffff, PCI fault or device removal?


new output:

sata_sil24 0000:08:04.0: IRQ status == 0xffffffff, PCI fault or device removal?


Regards,

Tim.



diff --git a/drivers/ata/sata_sil24.c b/drivers/ata/sata_sil24.c
index 0534890..d81b20d 100644
--- a/drivers/ata/sata_sil24.c
+++ b/drivers/ata/sata_sil24.c
@@ -1154,8 +1154,8 @@ static irqreturn_t sil24_interrupt(int irq, void *dev_instance)
 	status = readl(host_base + HOST_IRQ_STAT);
 
 	if (status == 0xffffffff) {
-		printk(KERN_ERR DRV_NAME ": IRQ status == 0xffffffff, "
-		       "PCI fault or device removal?\n");
+		dev_err(host->dev, "IRQ status == 0xffffffff, "
+			"PCI fault or device removal?\n");
 		goto out;
 	}
 

Re: [PATCH] sata_sil24: Identify which card suffered IRQ status error

[Tejun Heo]

On Tue, Jul 22, 2014 at 02:28:00PM +0100, Tim Small wrote:
> In machines with multiple Silicon Image 3124 and/or 3132 cards, there is no
> way to tell which card is the culprit when the sata_sil24 interrupt handler
> gets a bad status.
> 
> Tested-by: tim@seoss.co.uk
> Signed-off-by: tim@seoss.co.uk

Applied to libata/for-3.17.  Please use "Full Name <EMAIL>" for tags.

Thanks.

-- 
tejun

Re: [PATCH] sata_sil24: Identify which card suffered IRQ status error

[Sergei Shtylyov]

Hello.

On 07/22/2014 05:28 PM, Tim Small wrote:

> In machines with multiple Silicon Image 3124 and/or 3132 cards, there is no
> way to tell which card is the culprit when the sata_sil24 interrupt handler
> gets a bad status.

> Tested-by: tim@seoss.co.uk
> Signed-off-by: tim@seoss.co.uk

    Full name is required before email, and email should be enclosed in <>. 
Also, I guess you want the same email as the commit author and in the signoff? 
Use the From: line at the start of mail to override the one in the header. See 
Documentation/SubmittingPatches for more details...

[...]

> diff --git a/drivers/ata/sata_sil24.c b/drivers/ata/sata_sil24.c
> index 0534890..d81b20d 100644
> --- a/drivers/ata/sata_sil24.c
> +++ b/drivers/ata/sata_sil24.c
> @@ -1154,8 +1154,8 @@ static irqreturn_t sil24_interrupt(int irq, void *dev_instance)
>   	status = readl(host_base + HOST_IRQ_STAT);
>
>   	if (status == 0xffffffff) {
> -		printk(KERN_ERR DRV_NAME ": IRQ status == 0xffffffff, "
> -		       "PCI fault or device removal?\n");
> +		dev_err(host->dev, "IRQ status == 0xffffffff, "
> +			"PCI fault or device removal?\n");

    Don't break the string anymore please. I guess you haven't run the patch 
thru scripts/checkpatch.pl?

>   		goto out;
>   	}

WBR, Sergei