[PATCH] drivers/edac: NO_IRQ removal from powerpc-only drivers

We'd like to eventually remove NO_IRQ on powerpc, so remove usages of it
from powerpc-only drivers.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 drivers/edac/mpc85xx_edac.c | 6 +++---
 drivers/edac/mv64x60_edac.c | 8 ++++----
 drivers/edac/ppc4xx_edac.c  | 6 +++---
 3 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/edac/mpc85xx_edac.c b/drivers/edac/mpc85xx_edac.c
index ca63d0da8889..e12b8e166a53 100644
--- a/drivers/edac/mpc85xx_edac.c
+++ b/drivers/edac/mpc85xx_edac.c

@@ -267,7 +267,7 @@ static int mpc85xx_pci_err_probe(struct platform_device *op)
 
 	pdata = pci->pvt_info;
 	pdata->name = "mpc85xx_pci_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 
 	plat_data = op->dev.platform_data;
 	if (!plat_data) {

@@ -588,7 +588,7 @@ static int mpc85xx_l2_err_probe(struct platform_device *op)
 
 	pdata = edac_dev->pvt_info;
 	pdata->name = "mpc85xx_l2_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	edac_dev->dev = &op->dev;
 	dev_set_drvdata(edac_dev->dev, edac_dev);
 	edac_dev->ctl_name = pdata->name;

@@ -1058,7 +1058,7 @@ static int mpc85xx_mc_err_probe(struct platform_device *op)
 
 	pdata = mci->pvt_info;
 	pdata->name = "mpc85xx_mc_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	mci->pdev = &op->dev;
 	pdata->edac_idx = edac_mc_idx++;
 	dev_set_drvdata(mci->pdev, mci);

diff --git a/drivers/edac/mv64x60_edac.c b/drivers/edac/mv64x60_edac.c
index 6c54127e6eae..69c559189490 100644
--- a/drivers/edac/mv64x60_edac.c
+++ b/drivers/edac/mv64x60_edac.c

@@ -118,7 +118,7 @@ static int mv64x60_pci_err_probe(struct platform_device *pdev)
 
 	pdata->pci_hose = pdev->id;
 	pdata->name = "mpc85xx_pci_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	platform_set_drvdata(pdev, pci);
 	pci->dev = &pdev->dev;
 	pci->dev_name = dev_name(&pdev->dev);

@@ -291,7 +291,7 @@ static int mv64x60_sram_err_probe(struct platform_device *pdev)
 
 	pdata = edac_dev->pvt_info;
 	pdata->name = "mv64x60_sram_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	edac_dev->dev = &pdev->dev;
 	platform_set_drvdata(pdev, edac_dev);
 	edac_dev->dev_name = dev_name(&pdev->dev);

@@ -459,7 +459,7 @@ static int mv64x60_cpu_err_probe(struct platform_device *pdev)
 
 	pdata = edac_dev->pvt_info;
 	pdata->name = "mv64x60_cpu_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	edac_dev->dev = &pdev->dev;
 	platform_set_drvdata(pdev, edac_dev);
 	edac_dev->dev_name = dev_name(&pdev->dev);

@@ -727,7 +727,7 @@ static int mv64x60_mc_err_probe(struct platform_device *pdev)
 	mci->pdev = &pdev->dev;
 	platform_set_drvdata(pdev, mci);
 	pdata->name = "mv64x60_mc_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	mci->dev_name = dev_name(&pdev->dev);
 	pdata->edac_idx = edac_mc_idx++;
 

diff --git a/drivers/edac/ppc4xx_edac.c b/drivers/edac/ppc4xx_edac.c
index d3a64ba61fa3..ee84343f660f 100644
--- a/drivers/edac/ppc4xx_edac.c
+++ b/drivers/edac/ppc4xx_edac.c

@@ -1029,8 +1029,8 @@ static int ppc4xx_edac_mc_init(struct mem_ctl_info *mci,
 	pdata			= mci->pvt_info;
 
 	pdata->dcr_host		= *dcr_host;
-	pdata->irqs.sec		= NO_IRQ;
-	pdata->irqs.ded		= NO_IRQ;
+	pdata->irqs.sec		= 0;
+	pdata->irqs.ded		= 0;
 
 	/* Initialize controller capabilities and configuration */
 

@@ -1111,7 +1111,7 @@ static int ppc4xx_edac_register_irq(struct platform_device *op,
 	ded_irq = irq_of_parse_and_map(np, INTMAP_ECCDED_INDEX);
 	sec_irq = irq_of_parse_and_map(np, INTMAP_ECCSEC_INDEX);
 
-	if (ded_irq == NO_IRQ || sec_irq == NO_IRQ) {
+	if (!ded_irq || !sec_irq) {
 		ppc4xx_edac_mc_printk(KERN_ERR, mci,
 				      "Unable to map interrupts.\n");
 		status = -ENODEV;

-- 
2.7.4

On Sat, Sep 10, 2016 at 07:57:08PM +1000, Michael Ellerman wrote:

So all of those pdata structs come from kzalloc() so you don't really
have to assign to 0 - simply kill the line.

That part went into drivers/edac/fsl_ddr_edac.c which is only the
freescale memory controller being shared between ARM and PPC, see

http://git.kernel.org/cgit/linux/kernel/git/bp/bp.git/log/?h=for-next

But that shouldn't change the issue wrt ->irq as now it is implicitly 0.

IOW, you can drop this hunk.

Please send v2 against the above branch or linux-next.

Thanks.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.

Borislav Petkov [off-list ref] writes:

On Sat, Sep 10, 2016 at 07:57:08PM +1000, Michael Ellerman wrote:

quoted

We'd like to eventually remove NO_IRQ on powerpc, so remove usages of it
from powerpc-only drivers.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 drivers/edac/mpc85xx_edac.c | 6 +++---
 drivers/edac/mv64x60_edac.c | 8 ++++----
 drivers/edac/ppc4xx_edac.c  | 6 +++---
 3 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/edac/mpc85xx_edac.c b/drivers/edac/mpc85xx_edac.c
index ca63d0da8889..e12b8e166a53 100644
--- a/drivers/edac/mpc85xx_edac.c
+++ b/drivers/edac/mpc85xx_edac.c

@@ -267,7 +267,7 @@ static int mpc85xx_pci_err_probe(struct platform_device *op)
 
 	pdata = pci->pvt_info;
 	pdata->name = "mpc85xx_pci_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 
 	plat_data = op->dev.platform_data;
 	if (!plat_data) {

So all of those pdata structs come from kzalloc() so you don't really
have to assign to 0 - simply kill the line.

OK.

quoted

@@ -1058,7 +1058,7 @@ static int mpc85xx_mc_err_probe(struct platform_device *op)
 
 	pdata = mci->pvt_info;
 	pdata->name = "mpc85xx_mc_err";
-	pdata->irq = NO_IRQ;
+	pdata->irq = 0;
 	mci->pdev = &op->dev;
 	pdata->edac_idx = edac_mc_idx++;
 	dev_set_drvdata(mci->pdev, mci);

That part went into drivers/edac/fsl_ddr_edac.c which is only the
freescale memory controller being shared between ARM and PPC, see

http://git.kernel.org/cgit/linux/kernel/git/bp/bp.git/log/?h=for-next

But that shouldn't change the issue wrt ->irq as now it is implicitly 0.

Ah OK.

NO_IRQ is -1 on ARM (or at least it can be?), see arch/arm/include/asm/irq.h.

So I'll leave that one alone for now.

IOW, you can drop this hunk.

Please send v2 against the above branch or linux-next.

OK will do.

cheers

On Monday, September 12, 2016 2:47:37 PM CEST Michael Ellerman wrote:

quoted

quoted

@@ -1058,7 +1058,7 @@ static int mpc85xx_mc_err_probe(struct platform_device *op)
 
     pdata = mci->pvt_info;
     pdata->name = "mpc85xx_mc_err";
-    pdata->irq = NO_IRQ;
+    pdata->irq = 0;
     mci->pdev = &op->dev;
     pdata->edac_idx = edac_mc_idx++;
     dev_set_drvdata(mci->pdev, mci);

That part went into drivers/edac/fsl_ddr_edac.c which is only the
freescale memory controller being shared between ARM and PPC, see

http://git.kernel.org/cgit/linux/kernel/git/bp/bp.git/log/?h=for-next

But that shouldn't change the issue wrt ->irq as now it is implicitly 0.

Ah OK.

NO_IRQ is -1 on ARM (or at least it can be?), see arch/arm/include/asm/irq.h.

So I'll leave that one alone for now.

I don't think there is any way to build the mpc85xx on ARM, it depends
on the powerpc-specific "FSL_SOC" symbol and it uses the in_be32/out_be32
MMIO helpers that are not defined on ARM.

I've sent patches to remove all uses of NO_IRQ that happen on ARM.
Please remove this instance too.

	Arnd

Arnd Bergmann [off-list ref] writes:

On Monday, September 12, 2016 2:47:37 PM CEST Michael Ellerman wrote:

quoted

quoted

quoted

@@ -1058,7 +1058,7 @@ static int mpc85xx_mc_err_probe(struct platform_device *op)
 
     pdata = mci->pvt_info;
     pdata->name = "mpc85xx_mc_err";
-    pdata->irq = NO_IRQ;
+    pdata->irq = 0;
     mci->pdev = &op->dev;
     pdata->edac_idx = edac_mc_idx++;
     dev_set_drvdata(mci->pdev, mci);

That part went into drivers/edac/fsl_ddr_edac.c which is only the
freescale memory controller being shared between ARM and PPC, see

http://git.kernel.org/cgit/linux/kernel/git/bp/bp.git/log/?h=for-next

But that shouldn't change the issue wrt ->irq as now it is implicitly 0.

Ah OK.

NO_IRQ is -1 on ARM (or at least it can be?), see arch/arm/include/asm/irq.h.

So I'll leave that one alone for now.

I don't think there is any way to build the mpc85xx on ARM, it depends
on the powerpc-specific "FSL_SOC" symbol and it uses the in_be32/out_be32
MMIO helpers that are not defined on ARM.

Yeah you're right. The common code between powerpc and arm64 is in
fsl_ddr_edac.c, not mpc85xx_edac.c. There are no uses of NO_IRQ

Will send a v2 with all uses removed from drivers/edac.

cheers