[PATCHv2 0/4] memory: fsl_ifc: Switch to modern devm and platform IRQ APIs

[PATCHv2 0/4] memory: fsl_ifc: Switch to modern devm and platform IRQ APIs

[Rosen Penev]

Convert the Freescale IFC controller driver from legacy OF-specific
APIs to modern platform-device and devm-based equivalents. This
eliminates manual resource tracking in error paths and remove(),
simplifying the driver and reducing LoC by ~40%.

Patch 1 replaces irq_of_parse_and_map() with platform_get_irq() /
platform_get_irq_optional() and fixes unconditional free_irq() calls
on the optional NAND IRQ.
Patch 2 replaces of_iomap() with devm_platform_ioremap_resource(),
dropping the manual iounmap in remove/error paths and fixing a
missing free_irq on the main IRQ request_irq failure.
Patch 3 switches from request_irq to devm_request_irq, removing the
error-path labels entirely and reducing remove() to just
of_platform_depopulate.
Patch 4 removes of_platform_default_populate.

Rosen Penev (4):
  memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to
    platform IRQ API
  memory: fsl_ifc: Use devm_platform_ioremap_resource and fix error
    paths
  memory: fsl_ifc: Use devm_request_irq and simplify remove
  memory: fsl_ifc: drop simple-bus emulation

 drivers/memory/fsl_ifc.c | 101 +++++++++++++--------------------------
 1 file changed, 32 insertions(+), 69 deletions(-)

-- 
2.55.0

[PATCHv2 1/4] memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to platform IRQ API

[Rosen Penev]

Switch from irq_of_parse_and_map() to platform_get_irq() for the main
controller IRQ, and use platform_get_irq_optional() for the NAND IRQ
which is optional. Guard free_irq() calls on the NAND IRQ with a
nand_irq > 0 check to avoid passing invalid IRQ numbers on platforms
without a dedicated NAND interrupt line.

Get IRQs first as they can return a probe defferal error. Helps avoid
doing actual work in such a case.

Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/memory/fsl_ifc.c | 43 ++++++++++++++++++++--------------------
 1 file changed, 21 insertions(+), 22 deletions(-)

diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
index e89e0c6cc4bc..dce94d8b05b9 100644
--- a/drivers/memory/fsl_ifc.c
+++ b/drivers/memory/fsl_ifc.c
@@ -89,12 +89,10 @@ static void fsl_ifc_ctrl_remove(struct platform_device *dev)
 	struct fsl_ifc_ctrl *ctrl = dev_get_drvdata(&dev->dev);
 
 	of_platform_depopulate(&dev->dev);
-	free_irq(ctrl->nand_irq, ctrl);
+	if (ctrl->nand_irq > 0)
+		free_irq(ctrl->nand_irq, ctrl);
 	free_irq(ctrl->irq, ctrl);
 
-	irq_dispose_mapping(ctrl->nand_irq);
-	irq_dispose_mapping(ctrl->irq);
-
 	iounmap(ctrl->gregs);
 
 	dev_set_drvdata(&dev->dev, NULL);
@@ -204,9 +202,21 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 	int ret = 0;
 	int version, banks;
 	void __iomem *addr;
+	int nand_irq;
+	int irq;
 
 	dev_info(&dev->dev, "Freescale Integrated Flash Controller\n");
 
+	/* get the Controller level irq */
+	irq = platform_get_irq(dev, 0);
+	if (irq < 0)
+		return irq;
+
+	/* get the nand machine irq */
+	nand_irq = platform_get_irq_optional(dev, 1);
+	if (nand_irq == -EPROBE_DEFER)
+		return nand_irq;
+
 	fsl_ifc_ctrl_dev = devm_kzalloc(&dev->dev, sizeof(*fsl_ifc_ctrl_dev),
 					GFP_KERNEL);
 	if (!fsl_ifc_ctrl_dev)
@@ -246,23 +256,13 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 		addr += PGOFFSET_4K;
 	fsl_ifc_ctrl_dev->rregs = addr;
 
-	/* get the Controller level irq */
-	fsl_ifc_ctrl_dev->irq = irq_of_parse_and_map(dev->dev.of_node, 0);
-	if (fsl_ifc_ctrl_dev->irq == 0) {
-		dev_err(&dev->dev, "failed to get irq resource for IFC\n");
-		ret = -ENODEV;
-		goto err;
-	}
-
-	/* get the nand machine irq */
-	fsl_ifc_ctrl_dev->nand_irq =
-			irq_of_parse_and_map(dev->dev.of_node, 1);
+	fsl_ifc_ctrl_dev->irq = irq;
 
 	fsl_ifc_ctrl_dev->dev = &dev->dev;
 
 	ret = fsl_ifc_ctrl_init(fsl_ifc_ctrl_dev);
 	if (ret < 0)
-		goto err_unmap_nandirq;
+		goto err;
 
 	init_waitqueue_head(&fsl_ifc_ctrl_dev->nand_wait);
 
@@ -271,10 +271,11 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 	if (ret != 0) {
 		dev_err(&dev->dev, "failed to install irq (%d)\n",
 			fsl_ifc_ctrl_dev->irq);
-		goto err_unmap_nandirq;
+		goto err;
 	}
 
-	if (fsl_ifc_ctrl_dev->nand_irq) {
+	fsl_ifc_ctrl_dev->nand_irq = nand_irq;
+	if (fsl_ifc_ctrl_dev->nand_irq > 0) {
 		ret = request_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_nand_irq,
 				0, "fsl-ifc-nand", fsl_ifc_ctrl_dev);
 		if (ret != 0) {
@@ -292,12 +293,10 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 	return 0;
 
 err_free_nandirq:
-	free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
+	if (fsl_ifc_ctrl_dev->nand_irq > 0)
+		free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
 err_free_irq:
 	free_irq(fsl_ifc_ctrl_dev->irq, fsl_ifc_ctrl_dev);
-err_unmap_nandirq:
-	irq_dispose_mapping(fsl_ifc_ctrl_dev->nand_irq);
-	irq_dispose_mapping(fsl_ifc_ctrl_dev->irq);
 err:
 	iounmap(fsl_ifc_ctrl_dev->gregs);
 	return ret;
-- 
2.55.0

[PATCHv2 2/4] memory: fsl_ifc: Use devm_platform_ioremap_resource and fix error paths

[Rosen Penev]

Replace of_iomap() with devm_platform_ioremap_resource() for automatic
resource management, eliminating manual iounmap in remove and error
paths. devm_ioremap_resource() reserves the region and checks for
overlaps, which is safe here since the IFC controller register space
and child chip-select ranges are always at distinct addresses in all
DTS files.

As one of the errors is -EPROBE_DEFER, run it before anything else to
avoid doing actual work in such a case.

Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/memory/fsl_ifc.c | 23 +++++++++--------------
 1 file changed, 9 insertions(+), 14 deletions(-)

diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
index dce94d8b05b9..7e6800fc8480 100644
--- a/drivers/memory/fsl_ifc.c
+++ b/drivers/memory/fsl_ifc.c
@@ -92,10 +92,6 @@ static void fsl_ifc_ctrl_remove(struct platform_device *dev)
 	if (ctrl->nand_irq > 0)
 		free_irq(ctrl->nand_irq, ctrl);
 	free_irq(ctrl->irq, ctrl);
-
-	iounmap(ctrl->gregs);
-
-	dev_set_drvdata(&dev->dev, NULL);
 }
 
 /*
@@ -201,6 +197,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 {
 	int ret = 0;
 	int version, banks;
+	void __iomem *gregs;
 	void __iomem *addr;
 	int nand_irq;
 	int irq;
@@ -217,6 +214,11 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 	if (nand_irq == -EPROBE_DEFER)
 		return nand_irq;
 
+	/* IOMAP the entire IFC region */
+	gregs = devm_platform_ioremap_resource(dev, 0);
+	if (IS_ERR(gregs))
+		return PTR_ERR(gregs);
+
 	fsl_ifc_ctrl_dev = devm_kzalloc(&dev->dev, sizeof(*fsl_ifc_ctrl_dev),
 					GFP_KERNEL);
 	if (!fsl_ifc_ctrl_dev)
@@ -224,12 +226,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 
 	dev_set_drvdata(&dev->dev, fsl_ifc_ctrl_dev);
 
-	/* IOMAP the entire IFC region */
-	fsl_ifc_ctrl_dev->gregs = of_iomap(dev->dev.of_node, 0);
-	if (!fsl_ifc_ctrl_dev->gregs) {
-		dev_err(&dev->dev, "failed to get memory region\n");
-		return -ENODEV;
-	}
+	fsl_ifc_ctrl_dev->gregs = gregs;
 
 	if (of_property_read_bool(dev->dev.of_node, "little-endian")) {
 		fsl_ifc_ctrl_dev->little_endian = true;
@@ -262,7 +259,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 
 	ret = fsl_ifc_ctrl_init(fsl_ifc_ctrl_dev);
 	if (ret < 0)
-		goto err;
+		return ret;
 
 	init_waitqueue_head(&fsl_ifc_ctrl_dev->nand_wait);
 
@@ -271,7 +268,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 	if (ret != 0) {
 		dev_err(&dev->dev, "failed to install irq (%d)\n",
 			fsl_ifc_ctrl_dev->irq);
-		goto err;
+		goto err_free_irq;
 	}
 
 	fsl_ifc_ctrl_dev->nand_irq = nand_irq;
@@ -297,8 +294,6 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 		free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
 err_free_irq:
 	free_irq(fsl_ifc_ctrl_dev->irq, fsl_ifc_ctrl_dev);
-err:
-	iounmap(fsl_ifc_ctrl_dev->gregs);
 	return ret;
 }
 
-- 
2.55.0

[PATCHv2 3/4] memory: fsl_ifc: Use devm_request_irq and simplify remove

[Rosen Penev]

Switch from request_irq to devm_request_irq(), eliminating the error
path labels and free_irq() calls in remove(). Since both IRQs are now
devm-managed, the remove function only needs of_platform_depopulate
to tear down child devices before devm cleanup fires.

Remove dev_err calls on failure as devm_request_irq() already does this.

Don't bother using the irq variables in the struct. They were only used to
free the irqs in _remove.

Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/memory/fsl_ifc.c | 47 ++++++++++------------------------------
 1 file changed, 12 insertions(+), 35 deletions(-)

diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
index 7e6800fc8480..7fea6a356b55 100644
--- a/drivers/memory/fsl_ifc.c
+++ b/drivers/memory/fsl_ifc.c
@@ -86,12 +86,7 @@ static int fsl_ifc_ctrl_init(struct fsl_ifc_ctrl *ctrl)
 
 static void fsl_ifc_ctrl_remove(struct platform_device *dev)
 {
-	struct fsl_ifc_ctrl *ctrl = dev_get_drvdata(&dev->dev);
-
 	of_platform_depopulate(&dev->dev);
-	if (ctrl->nand_irq > 0)
-		free_irq(ctrl->nand_irq, ctrl);
-	free_irq(ctrl->irq, ctrl);
 }
 
 /*
@@ -253,8 +248,6 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 		addr += PGOFFSET_4K;
 	fsl_ifc_ctrl_dev->rregs = addr;
 
-	fsl_ifc_ctrl_dev->irq = irq;
-
 	fsl_ifc_ctrl_dev->dev = &dev->dev;
 
 	ret = fsl_ifc_ctrl_init(fsl_ifc_ctrl_dev);
@@ -263,38 +256,22 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 
 	init_waitqueue_head(&fsl_ifc_ctrl_dev->nand_wait);
 
-	ret = request_irq(fsl_ifc_ctrl_dev->irq, fsl_ifc_ctrl_irq, IRQF_SHARED,
-			  "fsl-ifc", fsl_ifc_ctrl_dev);
-	if (ret != 0) {
-		dev_err(&dev->dev, "failed to install irq (%d)\n",
-			fsl_ifc_ctrl_dev->irq);
-		goto err_free_irq;
-	}
+	ret = devm_request_irq(&dev->dev, irq,
+			       fsl_ifc_ctrl_irq, IRQF_SHARED,
+			       "fsl-ifc", fsl_ifc_ctrl_dev);
+	if (ret)
+		return ret;
 
-	fsl_ifc_ctrl_dev->nand_irq = nand_irq;
-	if (fsl_ifc_ctrl_dev->nand_irq > 0) {
-		ret = request_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_nand_irq,
-				0, "fsl-ifc-nand", fsl_ifc_ctrl_dev);
-		if (ret != 0) {
-			dev_err(&dev->dev, "failed to install irq (%d)\n",
-				fsl_ifc_ctrl_dev->nand_irq);
-			goto err_free_irq;
-		}
+	if (nand_irq > 0) {
+		ret = devm_request_irq(&dev->dev,
+				       nand_irq, fsl_ifc_nand_irq, 0,
+				       "fsl-ifc-nand", fsl_ifc_ctrl_dev);
+		if (ret)
+			return ret;
 	}
 
 	/* legacy dts may still use "simple-bus" compatible */
-	ret = of_platform_default_populate(dev->dev.of_node, NULL, &dev->dev);
-	if (ret)
-		goto err_free_nandirq;
-
-	return 0;
-
-err_free_nandirq:
-	if (fsl_ifc_ctrl_dev->nand_irq > 0)
-		free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
-err_free_irq:
-	free_irq(fsl_ifc_ctrl_dev->irq, fsl_ifc_ctrl_dev);
-	return ret;
+	return of_platform_default_populate(dev->dev.of_node, NULL, &dev->dev);
 }
 
 static const struct of_device_id fsl_ifc_match[] = {
-- 
2.55.0

[PATCHv2 4/4] memory: fsl_ifc: drop simple-bus emulation

[Rosen Penev]

of_platform_default_populate() in fsl_ifc_ctrl_probe() was a workaround
for legacy device trees that lacked "simple-bus" on the IFC node.
Commit 103478597591 ("powerpc/85xx: Add fsl,ifc to common device ids")
ensures the IFC node is populated as a platform device via
of_platform_bus_probe(), so the driver no longer needs to manually
populate its children -- they are handled by their own dedicated
drivers (NAND, NOR, etc.) matching on their own compatibles.

Drop the of_platform_default_populate() call and the corresponding
of_platform_depopulate() in fsl_ifc_ctrl_remove(), which is now dead
code and can be removed entirely.

Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/memory/fsl_ifc.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)

diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
index 7fea6a356b55..c591b89a1dd4 100644
--- a/drivers/memory/fsl_ifc.c
+++ b/drivers/memory/fsl_ifc.c
@@ -15,7 +15,6 @@
 #include <linux/slab.h>
 #include <linux/io.h>
 #include <linux/of.h>
-#include <linux/of_platform.h>
 #include <linux/platform_device.h>
 #include <linux/fsl_ifc.h>
 #include <linux/irqdomain.h>
@@ -84,11 +83,6 @@ static int fsl_ifc_ctrl_init(struct fsl_ifc_ctrl *ctrl)
 	return 0;
 }
 
-static void fsl_ifc_ctrl_remove(struct platform_device *dev)
-{
-	of_platform_depopulate(&dev->dev);
-}
-
 /*
  * NAND events are split between an operational interrupt which only
  * receives OPC, and an error interrupt that receives everything else,
@@ -270,8 +264,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
 			return ret;
 	}
 
-	/* legacy dts may still use "simple-bus" compatible */
-	return of_platform_default_populate(dev->dev.of_node, NULL, &dev->dev);
+	return 0;
 }
 
 static const struct of_device_id fsl_ifc_match[] = {
@@ -287,7 +280,6 @@ static struct platform_driver fsl_ifc_ctrl_driver = {
 		.of_match_table = fsl_ifc_match,
 	},
 	.probe       = fsl_ifc_ctrl_probe,
-	.remove  = fsl_ifc_ctrl_remove,
 };
 
 static int __init fsl_ifc_init(void)
-- 
2.55.0

Re: [PATCHv2 1/4] memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to platform IRQ API

[Krzysztof Kozlowski]

On 21/07/2026 22:24, Rosen Penev wrote:
> Switch from irq_of_parse_and_map() to platform_get_irq() for the main
> controller IRQ, and use platform_get_irq_optional() for the NAND IRQ
> which is optional. Guard free_irq() calls on the NAND IRQ with a
> nand_irq > 0 check to avoid passing invalid IRQ numbers on platforms
> without a dedicated NAND interrupt line.
> 
> Get IRQs first as they can return a probe defferal error. Helps avoid
> doing actual work in such a case.
> 
> Assisted-by: opencode:big-pickle
> Signed-off-by: Rosen Penev <rosenp@gmail.com>
> ---
>  drivers/memory/fsl_ifc.c | 43 ++++++++++++++++++++--------------------
>  1 file changed, 21 insertions(+), 22 deletions(-)
> 
> diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
> index e89e0c6cc4bc..dce94d8b05b9 100644
> --- a/drivers/memory/fsl_ifc.c
> +++ b/drivers/memory/fsl_ifc.c
> @@ -89,12 +89,10 @@ static void fsl_ifc_ctrl_remove(struct platform_device *dev)
>  	struct fsl_ifc_ctrl *ctrl = dev_get_drvdata(&dev->dev);
>  
>  	of_platform_depopulate(&dev->dev);
> -	free_irq(ctrl->nand_irq, ctrl);
> +	if (ctrl->nand_irq > 0)

0 is a valid IRQ for Linux, I think.

> +		free_irq(ctrl->nand_irq, ctrl);
>  	free_irq(ctrl->irq, ctrl);
>  
> -	irq_dispose_mapping(ctrl->nand_irq);
> -	irq_dispose_mapping(ctrl->irq);
> -
>  	iounmap(ctrl->gregs);
>  
>  	dev_set_drvdata(&dev->dev, NULL);
> @@ -204,9 +202,21 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  	int ret = 0;
>  	int version, banks;
>  	void __iomem *addr;
> +	int nand_irq;
> +	int irq;

Why do you need the intermediate variables?

>  
>  	dev_info(&dev->dev, "Freescale Integrated Flash Controller\n");
>  
> +	/* get the Controller level irq */
> +	irq = platform_get_irq(dev, 0);
> +	if (irq < 0)

... as also shown here.

> +		return irq;
> +
> +	/* get the nand machine irq */
> +	nand_irq = platform_get_irq_optional(dev, 1);
> +	if (nand_irq == -EPROBE_DEFER)
> +		return nand_irq;
> +
>  	fsl_ifc_ctrl_dev = devm_kzalloc(&dev->dev, sizeof(*fsl_ifc_ctrl_dev),
>  					GFP_KERNEL);
>  	if (!fsl_ifc_ctrl_dev)
> @@ -246,23 +256,13 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  		addr += PGOFFSET_4K;
>  	fsl_ifc_ctrl_dev->rregs = addr;
>  
> -	/* get the Controller level irq */
> -	fsl_ifc_ctrl_dev->irq = irq_of_parse_and_map(dev->dev.of_node, 0);
> -	if (fsl_ifc_ctrl_dev->irq == 0) {
> -		dev_err(&dev->dev, "failed to get irq resource for IFC\n");
> -		ret = -ENODEV;
> -		goto err;
> -	}
> -
> -	/* get the nand machine irq */
> -	fsl_ifc_ctrl_dev->nand_irq =
> -			irq_of_parse_and_map(dev->dev.of_node, 1);
> +	fsl_ifc_ctrl_dev->irq = irq;
>  
>  	fsl_ifc_ctrl_dev->dev = &dev->dev;
>  
>  	ret = fsl_ifc_ctrl_init(fsl_ifc_ctrl_dev);
>  	if (ret < 0)
> -		goto err_unmap_nandirq;
> +		goto err;
>  
>  	init_waitqueue_head(&fsl_ifc_ctrl_dev->nand_wait);
>  
> @@ -271,10 +271,11 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  	if (ret != 0) {
>  		dev_err(&dev->dev, "failed to install irq (%d)\n",
>  			fsl_ifc_ctrl_dev->irq);
> -		goto err_unmap_nandirq;
> +		goto err;
>  	}
>  
> -	if (fsl_ifc_ctrl_dev->nand_irq) {
> +	fsl_ifc_ctrl_dev->nand_irq = nand_irq;
> +	if (fsl_ifc_ctrl_dev->nand_irq > 0) {

...but not here.

>  		ret = request_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_nand_irq,
>  				0, "fsl-ifc-nand", fsl_ifc_ctrl_dev);
>  		if (ret != 0) {
> @@ -292,12 +293,10 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  	return 0;
>  
>  err_free_nandirq:
> -	free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
> +	if (fsl_ifc_ctrl_dev->nand_irq > 0)
> +		free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
>  err_free_irq:
>  	free_irq(fsl_ifc_ctrl_dev->irq, fsl_ifc_ctrl_dev);
> -err_unmap_nandirq:
> -	irq_dispose_mapping(fsl_ifc_ctrl_dev->nand_irq);
> -	irq_dispose_mapping(fsl_ifc_ctrl_dev->irq);
>  err:
>  	iounmap(fsl_ifc_ctrl_dev->gregs);
>  	return ret;


Best regards,
Krzysztof

Re: [PATCHv2 2/4] memory: fsl_ifc: Use devm_platform_ioremap_resource and fix error paths

[Krzysztof Kozlowski]

On 21/07/2026 22:24, Rosen Penev wrote:
> Replace of_iomap() with devm_platform_ioremap_resource() for automatic
> resource management, eliminating manual iounmap in remove and error
> paths. devm_ioremap_resource() reserves the region and checks for
> overlaps, which is safe here since the IFC controller register space
> and child chip-select ranges are always at distinct addresses in all
> DTS files.
> 
> As one of the errors is -EPROBE_DEFER, run it before anything else to
> avoid doing actual work in such a case.
> 
> Assisted-by: opencode:big-pickle
> Signed-off-by: Rosen Penev <rosenp@gmail.com>
> ---
>  drivers/memory/fsl_ifc.c | 23 +++++++++--------------
>  1 file changed, 9 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
> index dce94d8b05b9..7e6800fc8480 100644
> --- a/drivers/memory/fsl_ifc.c
> +++ b/drivers/memory/fsl_ifc.c
> @@ -92,10 +92,6 @@ static void fsl_ifc_ctrl_remove(struct platform_device *dev)
>  	if (ctrl->nand_irq > 0)
>  		free_irq(ctrl->nand_irq, ctrl);
>  	free_irq(ctrl->irq, ctrl);
> -
> -	iounmap(ctrl->gregs);
> -
> -	dev_set_drvdata(&dev->dev, NULL);
>  }
>  
>  /*
> @@ -201,6 +197,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  {
>  	int ret = 0;
>  	int version, banks;
> +	void __iomem *gregs;

Why another local variable?

>  	void __iomem *addr;
>  	int nand_irq;
>  	int irq;
> @@ -217,6 +214,11 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  	if (nand_irq == -EPROBE_DEFER)
>  		return nand_irq;
>  
> +	/* IOMAP the entire IFC region */
> +	gregs = devm_platform_ioremap_resource(dev, 0);
> +	if (IS_ERR(gregs))
> +		return PTR_ERR(gregs);
> +
>  	fsl_ifc_ctrl_dev = devm_kzalloc(&dev->dev, sizeof(*fsl_ifc_ctrl_dev),
>  					GFP_KERNEL);
>  	if (!fsl_ifc_ctrl_dev)
> @@ -224,12 +226,7 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>  
>  	dev_set_drvdata(&dev->dev, fsl_ifc_ctrl_dev);
>  
> -	/* IOMAP the entire IFC region */
> -	fsl_ifc_ctrl_dev->gregs = of_iomap(dev->dev.of_node, 0);

This is not equivalent and commit msg does not address that. Old code
allowed multiple maps of the same address space but new code bails with
EBUSY.

Again, I asked last time - how did you test it?

Please mark the patches as RFT.


Best regards,
Krzysztof

Re: [PATCHv2 1/4] memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to platform IRQ API

[Rosen Penev]

On Tue, Sep 1, 2026 at 8:29 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> On 21/07/2026 22:24, Rosen Penev wrote:
> > Switch from irq_of_parse_and_map() to platform_get_irq() for the main
> > controller IRQ, and use platform_get_irq_optional() for the NAND IRQ
> > which is optional. Guard free_irq() calls on the NAND IRQ with a
> > nand_irq > 0 check to avoid passing invalid IRQ numbers on platforms
> > without a dedicated NAND interrupt line.
> >
> > Get IRQs first as they can return a probe defferal error. Helps avoid
> > doing actual work in such a case.
> >
> > Assisted-by: opencode:big-pickle
> > Signed-off-by: Rosen Penev <rosenp@gmail.com>
> > ---
> >  drivers/memory/fsl_ifc.c | 43 ++++++++++++++++++++--------------------
> >  1 file changed, 21 insertions(+), 22 deletions(-)
> >
> > diff --git a/drivers/memory/fsl_ifc.c b/drivers/memory/fsl_ifc.c
> > index e89e0c6cc4bc..dce94d8b05b9 100644
> > --- a/drivers/memory/fsl_ifc.c
> > +++ b/drivers/memory/fsl_ifc.c
> > @@ -89,12 +89,10 @@ static void fsl_ifc_ctrl_remove(struct platform_device *dev)
> >       struct fsl_ifc_ctrl *ctrl = dev_get_drvdata(&dev->dev);
> >
> >       of_platform_depopulate(&dev->dev);
> > -     free_irq(ctrl->nand_irq, ctrl);
> > +     if (ctrl->nand_irq > 0)
>
> 0 is a valid IRQ for Linux, I think.
No. platform_get_irq has

WARN(ret == 0, "0 is an invalid IRQ number\n")

and 0 is an error with irq_of_parse_and_map.
>
> > +             free_irq(ctrl->nand_irq, ctrl);
> >       free_irq(ctrl->irq, ctrl);
> >
> > -     irq_dispose_mapping(ctrl->nand_irq);
> > -     irq_dispose_mapping(ctrl->irq);
> > -
> >       iounmap(ctrl->gregs);
> >
> >       dev_set_drvdata(&dev->dev, NULL);
> > @@ -204,9 +202,21 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
> >       int ret = 0;
> >       int version, banks;
> >       void __iomem *addr;
> > +     int nand_irq;
> > +     int irq;
>
> Why do you need the intermediate variables?
In order to call platform_get_irq before any allocation happens in probe.
>
> >
> >       dev_info(&dev->dev, "Freescale Integrated Flash Controller\n");
> >
> > +     /* get the Controller level irq */
> > +     irq = platform_get_irq(dev, 0);
> > +     if (irq < 0)
>
> ... as also shown here.
>
> > +             return irq;
> > +
> > +     /* get the nand machine irq */
> > +     nand_irq = platform_get_irq_optional(dev, 1);
> > +     if (nand_irq == -EPROBE_DEFER)
> > +             return nand_irq;
> > +
> >       fsl_ifc_ctrl_dev = devm_kzalloc(&dev->dev, sizeof(*fsl_ifc_ctrl_dev),
> >                                       GFP_KERNEL);
> >       if (!fsl_ifc_ctrl_dev)
> > @@ -246,23 +256,13 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
> >               addr += PGOFFSET_4K;
> >       fsl_ifc_ctrl_dev->rregs = addr;
> >
> > -     /* get the Controller level irq */
> > -     fsl_ifc_ctrl_dev->irq = irq_of_parse_and_map(dev->dev.of_node, 0);
> > -     if (fsl_ifc_ctrl_dev->irq == 0) {
> > -             dev_err(&dev->dev, "failed to get irq resource for IFC\n");
> > -             ret = -ENODEV;
> > -             goto err;
> > -     }
> > -
> > -     /* get the nand machine irq */
> > -     fsl_ifc_ctrl_dev->nand_irq =
> > -                     irq_of_parse_and_map(dev->dev.of_node, 1);
> > +     fsl_ifc_ctrl_dev->irq = irq;
> >
> >       fsl_ifc_ctrl_dev->dev = &dev->dev;
> >
> >       ret = fsl_ifc_ctrl_init(fsl_ifc_ctrl_dev);
> >       if (ret < 0)
> > -             goto err_unmap_nandirq;
> > +             goto err;
> >
> >       init_waitqueue_head(&fsl_ifc_ctrl_dev->nand_wait);
> >
> > @@ -271,10 +271,11 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
> >       if (ret != 0) {
> >               dev_err(&dev->dev, "failed to install irq (%d)\n",
> >                       fsl_ifc_ctrl_dev->irq);
> > -             goto err_unmap_nandirq;
> > +             goto err;
> >       }
> >
> > -     if (fsl_ifc_ctrl_dev->nand_irq) {
> > +     fsl_ifc_ctrl_dev->nand_irq = nand_irq;
> > +     if (fsl_ifc_ctrl_dev->nand_irq > 0) {
>
> ...but not here.
Will fix.
>
> >               ret = request_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_nand_irq,
> >                               0, "fsl-ifc-nand", fsl_ifc_ctrl_dev);
> >               if (ret != 0) {
> > @@ -292,12 +293,10 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
> >       return 0;
> >
> >  err_free_nandirq:
> > -     free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
> > +     if (fsl_ifc_ctrl_dev->nand_irq > 0)
> > +             free_irq(fsl_ifc_ctrl_dev->nand_irq, fsl_ifc_ctrl_dev);
> >  err_free_irq:
> >       free_irq(fsl_ifc_ctrl_dev->irq, fsl_ifc_ctrl_dev);
> > -err_unmap_nandirq:
> > -     irq_dispose_mapping(fsl_ifc_ctrl_dev->nand_irq);
> > -     irq_dispose_mapping(fsl_ifc_ctrl_dev->irq);
> >  err:
> >       iounmap(fsl_ifc_ctrl_dev->gregs);
> >       return ret;
>
>
> Best regards,
> Krzysztof

Re: [PATCHv2 1/4] memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to platform IRQ API

[Krzysztof Kozlowski]

On 06/09/2026 20:34, Rosen Penev wrote:
> On Tue, Sep 1, 2026 at 8:29 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>>>       dev_set_drvdata(&dev->dev, NULL);
>>> @@ -204,9 +202,21 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>>>       int ret = 0;
>>>       int version, banks;
>>>       void __iomem *addr;
>>> +     int nand_irq;
>>> +     int irq;
>>
>> Why do you need the intermediate variables?
> In order to call platform_get_irq before any allocation happens in probe.
>>

But allocation is the first thing the probe is doing.

Best regards,
Krzysztof

Re: [PATCHv2 1/4] memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to platform IRQ API

[Rosen Penev]

On Tue, Sep 8, 2026 at 11:00 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> On 06/09/2026 20:34, Rosen Penev wrote:
> > On Tue, Sep 1, 2026 at 8:29 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
> >>>       dev_set_drvdata(&dev->dev, NULL);
> >>> @@ -204,9 +202,21 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
> >>>       int ret = 0;
> >>>       int version, banks;
> >>>       void __iomem *addr;
> >>> +     int nand_irq;
> >>> +     int irq;
> >>
> >> Why do you need the intermediate variables?
> > In order to call platform_get_irq before any allocation happens in probe.
> >>
>
> But allocation is the first thing the probe is doing.
We're going in circles.

It's better to return -EPROBE_DEFER as early as possible to avoid
doing anything.
>
> Best regards,
> Krzysztof

Re: [PATCHv2 1/4] memory: fsl_ifc: Fix optional NAND IRQ handling and migrate to platform IRQ API

[Krzysztof Kozlowski]

On 10/09/2026 00:33, Rosen Penev wrote:
> On Tue, Sep 8, 2026 at 11:00 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>>
>> On 06/09/2026 20:34, Rosen Penev wrote:
>>> On Tue, Sep 1, 2026 at 8:29 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>>>>>       dev_set_drvdata(&dev->dev, NULL);
>>>>> @@ -204,9 +202,21 @@ static int fsl_ifc_ctrl_probe(struct platform_device *dev)
>>>>>       int ret = 0;
>>>>>       int version, banks;
>>>>>       void __iomem *addr;
>>>>> +     int nand_irq;
>>>>> +     int irq;
>>>>
>>>> Why do you need the intermediate variables?
>>> In order to call platform_get_irq before any allocation happens in probe.
>>>>
>>
>> But allocation is the first thing the probe is doing.
> We're going in circles.
> 
> It's better to return -EPROBE_DEFER as early as possible to avoid
> doing anything.

Then you would be getting all resources like that in every driver... and
we don't. Plus deferred probe on IRQ is highly unlikely.

Best regards,
Krzysztof