[PATCH 0/2] Correctly handle plaform_get_irq()'s result in the libata drivers

[PATCH 0/2] Correctly handle plaform_get_irq()'s result in the libata drivers

[Sergey Shtylyov]

Here are 2 patches against the 'master' branch of the Jens Axboe's 'linux-block.git' repo.
The affected drivers call platform_get_irq() but mis-interprete its result -- they consider
IRQ0 as error and the real error codes as valid IRQs... :-/

[1/2] pata_arasan_cf: fix IRQ check
[2/2] pata_ipx4xx_cf: fix IRQ check

[PATCH 1/2] pata_arasan_cf: fix IRQ check

[Sergey Shtylyov]

The driver's probe() method is written as if platform_get_irq() returns 0
on error, while actually it returns a negative error code (with all the
other values considered valid IRQs). Rewrite the driver's IRQ checking code
to pass the positive IRQ #s to ata_host_activate(), propagate upstream
-EPROBE_DEFER, and set up the driver to polling mode on (negative) errors
and IRQ0 (libata treats IRQ #0 as a polling mode anyway)...

Fixes: a480167b23ef ("pata_arasan_cf: Adding support for arasan compact flash host controller")
Signed-off-by: Sergey Shtylyov <s.shtylyov@omprussia.ru>

---
This patch is against the 'master' branch of Jens Axboe's 'linux-block.git'
repo.

 drivers/ata/pata_arasan_cf.c |   15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

Index: linux-block/drivers/ata/pata_arasan_cf.c
===================================================================
--- linux-block.orig/drivers/ata/pata_arasan_cf.c
+++ linux-block/drivers/ata/pata_arasan_cf.c
@@ -818,12 +818,19 @@ static int arasan_cf_probe(struct platfo
 	else
 		quirk = CF_BROKEN_UDMA; /* as it is on spear1340 */
 
-	/* if irq is 0, support only PIO */
-	acdev->irq = platform_get_irq(pdev, 0);
-	if (acdev->irq)
+	/*
+	 * If there's an error getting IRQ (or we do get IRQ0),
+	 * support only PIO
+	 */
+	ret = platform_get_irq(pdev, 0);
+	if (ret > 0) {
+		acdev->irq = ret;
 		irq_handler = arasan_cf_interrupt;
-	else
+	} else	if (ret == -EPROBE_DEFER) {
+		return ret;
+	} else	{
 		quirk |= CF_BROKEN_MWDMA | CF_BROKEN_UDMA;
+	}
 
 	acdev->pbase = res->start;
 	acdev->vbase = devm_ioremap(&pdev->dev, res->start,

[PATCH 2/2] pata_ipx4xx_cf: fix IRQ check

[Sergey Shtylyov]

The driver's probe() method is written as if platform_get_irq() returns 0
on error, while actually it returns a negative error code (with all the
other values considered valid IRQs).  Rewrite the driver's IRQ checking
code to pass the positive IRQ #s to ata_host_activate(), propagate errors
upstream, and treat IRQ0 as error, returning -EINVAL, as the libata code
treats 0  as  an indication that polling should be used anyway...

Fixes: 0df0d0a0ea9f ("[libata] ARM: add ixp4xx PATA driver")
Signed-off-by: Sergey Shtylyov <s.shtylyov@omprussia.ru>

---
 drivers/ata/pata_ixp4xx_cf.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

Index: linux-block/drivers/ata/pata_ixp4xx_cf.c
===================================================================
--- linux-block.orig/drivers/ata/pata_ixp4xx_cf.c
+++ linux-block/drivers/ata/pata_ixp4xx_cf.c
@@ -165,8 +165,12 @@ static int ixp4xx_pata_probe(struct plat
 		return -ENOMEM;
 
 	irq = platform_get_irq(pdev, 0);
-	if (irq)
+	if (irq > 0)
 		irq_set_irq_type(irq, IRQ_TYPE_EDGE_RISING);
+	else if (irq < 0)
+		return irq;
+	else
+		return -EINVAL;
 
 	/* Setup expansion bus chip selects */
 	*data->cs0_cfg = data->cs0_bits;

Re: [PATCH 1/2] pata_arasan_cf: fix IRQ check

[Viresh Kumar]

On 25-03-21, 23:50, Sergey Shtylyov wrote:
> The driver's probe() method is written as if platform_get_irq() returns 0
> on error, while actually it returns a negative error code (with all the
> other values considered valid IRQs). Rewrite the driver's IRQ checking code
> to pass the positive IRQ #s to ata_host_activate(), propagate upstream
> -EPROBE_DEFER, and set up the driver to polling mode on (negative) errors
> and IRQ0 (libata treats IRQ #0 as a polling mode anyway)...
> 
> Fixes: a480167b23ef ("pata_arasan_cf: Adding support for arasan compact flash host controller")
> Signed-off-by: Sergey Shtylyov <s.shtylyov@omprussia.ru>
> 
> ---
> This patch is against the 'master' branch of Jens Axboe's 'linux-block.git'
> repo.
> 
>  drivers/ata/pata_arasan_cf.c |   15 +++++++++++----
>  1 file changed, 11 insertions(+), 4 deletions(-)
> 
> Index: linux-block/drivers/ata/pata_arasan_cf.c
> ===================================================================
> --- linux-block.orig/drivers/ata/pata_arasan_cf.c
> +++ linux-block/drivers/ata/pata_arasan_cf.c
> @@ -818,12 +818,19 @@ static int arasan_cf_probe(struct platfo
>  	else
>  		quirk = CF_BROKEN_UDMA; /* as it is on spear1340 */
>  
> -	/* if irq is 0, support only PIO */
> -	acdev->irq = platform_get_irq(pdev, 0);
> -	if (acdev->irq)
> +	/*
> +	 * If there's an error getting IRQ (or we do get IRQ0),
> +	 * support only PIO
> +	 */
> +	ret = platform_get_irq(pdev, 0);
> +	if (ret > 0) {
> +		acdev->irq = ret;
>  		irq_handler = arasan_cf_interrupt;
> -	else
> +	} else	if (ret == -EPROBE_DEFER) {
> +		return ret;
> +	} else	{
>  		quirk |= CF_BROKEN_MWDMA | CF_BROKEN_UDMA;
> +	}
>  
>  	acdev->pbase = res->start;
>  	acdev->vbase = devm_ioremap(&pdev->dev, res->start,

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

-- 
viresh

Re: [PATCH 0/2] Correctly handle plaform_get_irq()'s result in the libata drivers

[Jens Axboe]

On 3/25/21 2:46 PM, Sergey Shtylyov wrote:
> Here are 2 patches against the 'master' branch of the Jens Axboe's 'linux-block.git' repo.
> The affected drivers call platform_get_irq() but mis-interprete its result -- they consider
> IRQ0 as error and the real error codes as valid IRQs... :-/
> 
> [1/2] pata_arasan_cf: fix IRQ check
> [2/2] pata_ipx4xx_cf: fix IRQ check

Applied, thanks.

-- 
Jens Axboe