[PATCH v4 0/2] Amazon's Annapurna Labs POS Driver

[PATCH v4 0/2] Amazon's Annapurna Labs POS Driver

[Talel Shenhar]

The Amazon's Annapurna Labs SoCs includes Point Of Serialization error
logging unit that reports an error in case of write error (e.g. attempt to
write to a read only register).

This patch series introduces the support for this unit.

Changes since v3:
=================
- ported to be edac device
- converted dt-bindings to new scheme
- added unit address to dt example

Changes since v2:
=================
- squashed left shifting fix to the driver

Changes since v1:
=================
- move MODULE_ to the end of the file
- simplified resource remapping devm_platform_ioremap_resource()
- use platform_get_irq() instead of irq_of_parse_and_map()
- removed the use of _relaxed accessor in favor to the regular ones
- removed driver selected based on arch
- added casting to u64 before left shifting (reported by kbuild test robot)


Talel Shenhar (2):
  dt-bindings: soc: al-pos: Amazon's Annapurna Labs POS
  soc: amazon: al-pos-edac: Introduce Amazon's Annapurna Labs POS EDAC
    driver

 .../bindings/edac/amazon,al-pos-edac.yaml          |  39 +++++
 MAINTAINERS                                        |   7 +
 drivers/edac/Kconfig                               |   6 +
 drivers/edac/Makefile                              |   1 +
 drivers/edac/al_pos_edac.c                         | 173 +++++++++++++++++++++
 5 files changed, 226 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
 create mode 100644 drivers/edac/al_pos_edac.c

-- 
2.7.4

[PATCH v4 1/2] dt-bindings: soc: al-pos: Amazon's Annapurna Labs POS

[Talel Shenhar]

Document Amazon's Annapurna Labs POS SoC binding.

Signed-off-by: Talel Shenhar <talel@amazon.com>
---
 .../bindings/edac/amazon,al-pos-edac.yaml          | 39 ++++++++++++++++++++++
 1 file changed, 39 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml

diff --git a/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml b/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
new file mode 100644
index 00000000..7936da0
--- /dev/null
+++ b/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
@@ -0,0 +1,39 @@
+# SPDX-License-Identifier: GPL-2.0
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/edac/amazon,al-pos-edac.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Amazon's Annapurna Labs POS
+
+maintainers:
+  - Talel Shenhar <talel@amazon.com>
+  - Talel Shenhar <talelshenhar@gmail.com>
+
+description: |
+  POS node is defined to describe the Point Of Serialization (POS) error
+  detection capability.
+
+properties:
+
+  compatible:
+    - const: "amazon,al-pos-edac"
+
+  reg:
+    maxItems: 1
+
+  interrupts:
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+
+examples:
+  - |
+    al_pos_edac@f0070084 {
+      compatible = "amazon,al-pos-edac";
+      reg = <0x0 0xf0070084 0x0 0x00000008>;
+      interrupt-parent = <&amazon_system_fabric>;
+      interrupts = <24 IRQ_TYPE_LEVEL_HIGH>;
+    };
-- 
2.7.4

[PATCH v4 2/2] soc: amazon: al-pos-edac: Introduce Amazon's Annapurna Labs POS EDAC driver

[Talel Shenhar]

The Amazon's Annapurna Labs SoCs includes Point Of Serialization error
logging unit that reports an error in case write error (e.g . Attempt to
write to a read only register).
This error shall be reported to EDAC subsystem as uncorrectable-error.

Signed-off-by: Talel Shenhar <talel@amazon.com>
---
 MAINTAINERS                |   7 ++
 drivers/edac/Kconfig       |   6 ++
 drivers/edac/Makefile      |   1 +
 drivers/edac/al_pos_edac.c | 173 +++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 187 insertions(+)
 create mode 100644 drivers/edac/al_pos_edac.c

diff --git a/MAINTAINERS b/MAINTAINERS
index e7a47b5..f5ce446 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -751,6 +751,13 @@ F:	drivers/tty/serial/altera_jtaguart.c
 F:	include/linux/altera_uart.h
 F:	include/linux/altera_jtaguart.h
 
+AMAZON ANNAPURNA LABS POS EDAC DRIVER
+M:	Talel Shenhar <talel@amazon.com>
+M:	Talel Shenhar <talelshenhar@gmail.com>
+S:	Maintained
+F:	Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
+F:	drivers/edac/al-pos-edac.c
+
 AMAZON ANNAPURNA LABS THERMAL MMIO DRIVER
 M:	Talel Shenhar <talel@amazon.com>
 S:	Maintained
diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig
index 200c04c..bb5805f 100644
--- a/drivers/edac/Kconfig
+++ b/drivers/edac/Kconfig
@@ -100,6 +100,12 @@ config EDAC_AMD64_ERROR_INJECTION
 	  In addition, there are two control files, inject_read and inject_write,
 	  which trigger the DRAM ECC Read and Write respectively.
 
+config EDAC_AL_POS
+	tristate "Amazon's Annapurna Labs POS EDAC driver"
+	depends on (ARCH_ALPINE || COMPILE_TEST)
+	help
+	  Include support for the SoC POS EDAC error capability.
+
 config EDAC_AMD76X
 	tristate "AMD 76x (760, 762, 768)"
 	depends on PCI && X86_32
diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
index 165ca65e..3571936 100644
--- a/drivers/edac/Makefile
+++ b/drivers/edac/Makefile
@@ -22,6 +22,7 @@ obj-$(CONFIG_EDAC_GHES)			+= ghes_edac.o
 edac_mce_amd-y				:= mce_amd.o
 obj-$(CONFIG_EDAC_DECODE_MCE)		+= edac_mce_amd.o
 
+obj-$(CONFIG_EDAC_AL_POS)		+= al_pos_edac.o
 obj-$(CONFIG_EDAC_AMD76X)		+= amd76x_edac.o
 obj-$(CONFIG_EDAC_CPC925)		+= cpc925_edac.o
 obj-$(CONFIG_EDAC_I5000)		+= i5000_edac.o
diff --git a/drivers/edac/al_pos_edac.c b/drivers/edac/al_pos_edac.c
new file mode 100644
index 00000000..bd6cd87
--- /dev/null
+++ b/drivers/edac/al_pos_edac.c
@@ -0,0 +1,173 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
+ */
+#include <linux/bitfield.h>
+#include <linux/edac.h>
+#include <linux/of_irq.h>
+#include "edac_module.h"
+
+#define DRV_NAME "al_pos_edac"
+#define AL_POS_EDAC_MSG_MAX 256
+
+/* Registers Offset */
+#define AL_POS_ERROR_LOG_1	0x0
+#define AL_POS_ERROR_LOG_0	0x4
+
+/* Registers Fields */
+#define AL_POS_ERROR_LOG_1_VALID	BIT(31)
+#define AL_POS_ERROR_LOG_1_BRESP	GENMASK(18, 17)
+#define AL_POS_ERROR_LOG_1_REQUEST_ID	GENMASK(16, 8)
+#define AL_POS_ERROR_LOG_1_ADDR_HIGH	GENMASK(7, 0)
+
+#define AL_POS_ERROR_LOG_0_ADDR_LOW	GENMASK(31, 0)
+
+struct al_pos_edac {
+	struct edac_device_ctl_info *edac_dev;
+	void __iomem *mmio_base;
+	int irq;
+};
+
+static int al_pos_handle(struct al_pos_edac *al_pos)
+{
+	u32 log0, log1;
+	u64 addr;
+	u16 request_id;
+	u8 bresp;
+	char msg[AL_POS_EDAC_MSG_MAX];
+
+	log1 = readl(al_pos->mmio_base + AL_POS_ERROR_LOG_1);
+	if (!FIELD_GET(AL_POS_ERROR_LOG_1_VALID, log1))
+		return 0;
+
+	log0 = readl(al_pos->mmio_base + AL_POS_ERROR_LOG_0);
+	writel(0, al_pos->mmio_base + AL_POS_ERROR_LOG_1);
+
+	addr = FIELD_GET(AL_POS_ERROR_LOG_0_ADDR_LOW, log0);
+	addr |= (((u64)FIELD_GET(AL_POS_ERROR_LOG_1_ADDR_HIGH, log1)) << 32);
+	request_id = FIELD_GET(AL_POS_ERROR_LOG_1_REQUEST_ID, log1);
+	bresp = FIELD_GET(AL_POS_ERROR_LOG_1_BRESP, log1);
+
+	snprintf(msg, sizeof(msg),
+		 "addr=0x%llx request_id=0x%x bresp=0x%x\n",
+		 addr, request_id, bresp);
+
+	edac_device_handle_ue(al_pos->edac_dev, 0, 0, msg);
+
+	return 1;
+}
+
+static void al_pos_edac_check(struct edac_device_ctl_info *edac_dev)
+{
+	struct al_pos_edac *al_pos = edac_dev->pvt_info;
+
+	al_pos_handle(al_pos);
+}
+
+static irqreturn_t al_pos_irq_handler(int irq, void *info)
+{
+	struct platform_device *pdev = info;
+	struct al_pos_edac *al_pos = platform_get_drvdata(pdev);
+
+	if (al_pos_handle(al_pos))
+		return IRQ_HANDLED;
+	return IRQ_NONE;
+}
+
+static int al_pos_probe(struct platform_device *pdev)
+{
+	struct edac_device_ctl_info *edac_dev;
+	struct al_pos_edac *al_pos;
+	int ret;
+
+	edac_dev = edac_device_alloc_ctl_info(sizeof(*al_pos), DRV_NAME, 1,
+					      DRV_NAME, 1, 0, NULL, 0,
+					      edac_device_alloc_index());
+	if (!edac_dev)
+		return -ENOMEM;
+
+	al_pos = edac_dev->pvt_info;
+	al_pos->edac_dev = edac_dev;
+	platform_set_drvdata(pdev, al_pos);
+
+	al_pos->mmio_base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(al_pos->mmio_base)) {
+		dev_err(&pdev->dev, "failed to ioremap memory (%ld)\n",
+			PTR_ERR(al_pos->mmio_base));
+		return PTR_ERR(al_pos->mmio_base);
+	}
+
+	al_pos->irq = platform_get_irq(pdev, 0);
+	if (al_pos->irq <= 0)
+		edac_dev->edac_check = al_pos_edac_check;
+
+	edac_dev->dev = &pdev->dev;
+	edac_dev->mod_name = DRV_NAME;
+	edac_dev->dev_name = dev_name(&pdev->dev);
+	edac_dev->ctl_name = "POS";
+
+	ret = edac_device_add_device(edac_dev);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to add edac device\n");
+		goto err_free_edac;
+	}
+
+	if (al_pos->irq > 0) {
+		ret = devm_request_irq(&pdev->dev,
+				       al_pos->irq,
+				       al_pos_irq_handler,
+				       0,
+				       pdev->name,
+				       pdev);
+		if (ret != 0) {
+			dev_err(&pdev->dev,
+				"failed to register to irq %d (%d)\n",
+				al_pos->irq, ret);
+			goto err_remove_edac;
+		}
+	}
+
+	return 0;
+
+err_remove_edac:
+	edac_device_del_device(edac_dev->dev);
+err_free_edac:
+	edac_device_free_ctl_info(edac_dev);
+
+	return ret;
+}
+
+static int al_pos_remove(struct platform_device *pdev)
+{
+	struct al_pos_edac *al_pos = platform_get_drvdata(pdev);
+
+	if (al_pos->irq > 0)
+		devm_free_irq(&pdev->dev, al_pos->irq, pdev);
+
+	edac_device_del_device(al_pos->edac_dev->dev);
+	edac_device_free_ctl_info(al_pos->edac_dev);
+
+	return 0;
+}
+
+static const struct of_device_id al_pos_of_match[] = {
+	{ .compatible = "amazon,al-pos-edac", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, al_pos_of_match);
+
+static struct platform_driver al_pos_driver = {
+	.probe = al_pos_probe,
+	.remove = al_pos_remove,
+	.driver = {
+		.name = DRV_NAME,
+		.of_match_table = al_pos_of_match,
+	},
+};
+
+module_platform_driver(al_pos_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Talel Shenhar");
+MODULE_DESCRIPTION("Amazon's Annapurna Labs POS driver");
-- 
2.7.4

Re: [PATCH v4 1/2] dt-bindings: soc: al-pos: Amazon's Annapurna Labs POS

[Rob Herring]

On Thu, Oct 3, 2019 at 6:33 AM Talel Shenhar <talel@amazon.com> wrote:
>
> Document Amazon's Annapurna Labs POS SoC binding.
>
> Signed-off-by: Talel Shenhar <talel@amazon.com>
> ---
>  .../bindings/edac/amazon,al-pos-edac.yaml          | 39 ++++++++++++++++++++++
>  1 file changed, 39 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
>
> diff --git a/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml b/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
> new file mode 100644
> index 00000000..7936da0
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
> @@ -0,0 +1,39 @@
> +# SPDX-License-Identifier: GPL-2.0
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/edac/amazon,al-pos-edac.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: Amazon's Annapurna Labs POS
> +
> +maintainers:
> +  - Talel Shenhar <talel@amazon.com>
> +  - Talel Shenhar <talelshenhar@gmail.com>
> +
> +description: |
> +  POS node is defined to describe the Point Of Serialization (POS) error
> +  detection capability.
> +
> +properties:
> +
> +  compatible:
> +    - const: "amazon,al-pos-edac"

Fails 'make dt_binding_check' (drop the '-'):

warning: no schema found in file:
Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
/builds/robherring/linux-dt-review/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml:
ignoring, error in schema 'compatible'
/builds/robherring/linux-dt-review/Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml:
properties:compatible: [{'const': 'amazon,al-pos-edac'}] is not of
type 'object', 'boolean'
Documentation/devicetree/bindings/Makefile:12: recipe for target
'Documentation/devicetree/bindings/edac/amazon,al-pos-edac.example.dts'
failed
make[1]: *** [Documentation/devicetree/bindings/edac/amazon,al-pos-edac.example.dts]
Error 1
Makefile:1282: recipe for target 'dt_binding_check' failed
make: *** [dt_binding_check] Error 2

> +
> +  reg:
> +    maxItems: 1
> +
> +  interrupts:
> +    maxItems: 1
> +
> +required:
> +  - compatible
> +  - reg
> +
> +examples:
> +  - |
> +    al_pos_edac@f0070084 {

edac@...

> +      compatible = "amazon,al-pos-edac";
> +      reg = <0x0 0xf0070084 0x0 0x00000008>;
> +      interrupt-parent = <&amazon_system_fabric>;
> +      interrupts = <24 IRQ_TYPE_LEVEL_HIGH>;

Not documented.

> +    };
> --
> 2.7.4
>

Re: [PATCH v4 1/2] dt-bindings: soc: al-pos: Amazon's Annapurna Labs POS

[Shenhar, Talel]

thanks for the review.

On 10/4/2019 6:05 PM, Rob Herring wrote:
> On Thu, Oct 3, 2019 at 6:33 AM Talel Shenhar <talel@amazon.com> wrote:
>
> Fails 'make dt_binding_check' (drop the '-'):
ack, shall be part of v5
>
>> +
>> +examples:
>> +  - |
>> +    al_pos_edac@f0070084 {
> edac@...
ack, shall be part of v5
>
>> +      compatible = "amazon,al-pos-edac";
>> +      reg = <0x0 0xf0070084 0x0 0x00000008>;
>> +      interrupt-parent = <&amazon_system_fabric>;
>> +      interrupts = <24 IRQ_TYPE_LEVEL_HIGH>;
> Not documented.

looks good?

    interrupts:

     description: Interrupt for the error event.
     maxItems: 1

Re: [PATCH v4 2/2] soc: amazon: al-pos-edac: Introduce Amazon's Annapurna Labs POS EDAC driver

[Marc Zyngier]

On Thu, 03 Oct 2019 12:32:41 +0100,
Talel Shenhar <talel@amazon.com> wrote:
> 
> The Amazon's Annapurna Labs SoCs includes Point Of Serialization error
> logging unit that reports an error in case write error (e.g . Attempt to
> write to a read only register).
> This error shall be reported to EDAC subsystem as uncorrectable-error.
> 
> Signed-off-by: Talel Shenhar <talel@amazon.com>
> ---
>  MAINTAINERS                |   7 ++
>  drivers/edac/Kconfig       |   6 ++
>  drivers/edac/Makefile      |   1 +
>  drivers/edac/al_pos_edac.c | 173 +++++++++++++++++++++++++++++++++++++++++++++
>  4 files changed, 187 insertions(+)
>  create mode 100644 drivers/edac/al_pos_edac.c
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index e7a47b5..f5ce446 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -751,6 +751,13 @@ F:	drivers/tty/serial/altera_jtaguart.c
>  F:	include/linux/altera_uart.h
>  F:	include/linux/altera_jtaguart.h
>  
> +AMAZON ANNAPURNA LABS POS EDAC DRIVER
> +M:	Talel Shenhar <talel@amazon.com>
> +M:	Talel Shenhar <talelshenhar@gmail.com>
> +S:	Maintained
> +F:	Documentation/devicetree/bindings/edac/amazon,al-pos-edac.yaml
> +F:	drivers/edac/al-pos-edac.c
> +
>  AMAZON ANNAPURNA LABS THERMAL MMIO DRIVER
>  M:	Talel Shenhar <talel@amazon.com>
>  S:	Maintained
> diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig
> index 200c04c..bb5805f 100644
> --- a/drivers/edac/Kconfig
> +++ b/drivers/edac/Kconfig
> @@ -100,6 +100,12 @@ config EDAC_AMD64_ERROR_INJECTION
>  	  In addition, there are two control files, inject_read and inject_write,
>  	  which trigger the DRAM ECC Read and Write respectively.
>  
> +config EDAC_AL_POS
> +	tristate "Amazon's Annapurna Labs POS EDAC driver"
> +	depends on (ARCH_ALPINE || COMPILE_TEST)
> +	help
> +	  Include support for the SoC POS EDAC error capability.
> +
>  config EDAC_AMD76X
>  	tristate "AMD 76x (760, 762, 768)"
>  	depends on PCI && X86_32
> diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
> index 165ca65e..3571936 100644
> --- a/drivers/edac/Makefile
> +++ b/drivers/edac/Makefile
> @@ -22,6 +22,7 @@ obj-$(CONFIG_EDAC_GHES)			+= ghes_edac.o
>  edac_mce_amd-y				:= mce_amd.o
>  obj-$(CONFIG_EDAC_DECODE_MCE)		+= edac_mce_amd.o
>  
> +obj-$(CONFIG_EDAC_AL_POS)		+= al_pos_edac.o
>  obj-$(CONFIG_EDAC_AMD76X)		+= amd76x_edac.o
>  obj-$(CONFIG_EDAC_CPC925)		+= cpc925_edac.o
>  obj-$(CONFIG_EDAC_I5000)		+= i5000_edac.o
> diff --git a/drivers/edac/al_pos_edac.c b/drivers/edac/al_pos_edac.c
> new file mode 100644
> index 00000000..bd6cd87
> --- /dev/null
> +++ b/drivers/edac/al_pos_edac.c
> @@ -0,0 +1,173 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
> + */
> +#include <linux/bitfield.h>
> +#include <linux/edac.h>
> +#include <linux/of_irq.h>
> +#include "edac_module.h"
> +
> +#define DRV_NAME "al_pos_edac"
> +#define AL_POS_EDAC_MSG_MAX 256
> +
> +/* Registers Offset */
> +#define AL_POS_ERROR_LOG_1	0x0
> +#define AL_POS_ERROR_LOG_0	0x4
> +
> +/* Registers Fields */
> +#define AL_POS_ERROR_LOG_1_VALID	BIT(31)
> +#define AL_POS_ERROR_LOG_1_BRESP	GENMASK(18, 17)
> +#define AL_POS_ERROR_LOG_1_REQUEST_ID	GENMASK(16, 8)
> +#define AL_POS_ERROR_LOG_1_ADDR_HIGH	GENMASK(7, 0)
> +
> +#define AL_POS_ERROR_LOG_0_ADDR_LOW	GENMASK(31, 0)
> +
> +struct al_pos_edac {
> +	struct edac_device_ctl_info *edac_dev;
> +	void __iomem *mmio_base;
> +	int irq;
> +};
> +
> +static int al_pos_handle(struct al_pos_edac *al_pos)
> +{
> +	u32 log0, log1;
> +	u64 addr;
> +	u16 request_id;
> +	u8 bresp;
> +	char msg[AL_POS_EDAC_MSG_MAX];
> +
> +	log1 = readl(al_pos->mmio_base + AL_POS_ERROR_LOG_1);

I already commented on the misuse of strict accesses. Unless you can
explain and document *why* you need the extra ordering, please use
relaxed accesses.

> +	if (!FIELD_GET(AL_POS_ERROR_LOG_1_VALID, log1))
> +		return 0;
> +
> +	log0 = readl(al_pos->mmio_base + AL_POS_ERROR_LOG_0);
> +	writel(0, al_pos->mmio_base + AL_POS_ERROR_LOG_1);
> +
> +	addr = FIELD_GET(AL_POS_ERROR_LOG_0_ADDR_LOW, log0);
> +	addr |= (((u64)FIELD_GET(AL_POS_ERROR_LOG_1_ADDR_HIGH, log1)) << 32);
> +	request_id = FIELD_GET(AL_POS_ERROR_LOG_1_REQUEST_ID, log1);
> +	bresp = FIELD_GET(AL_POS_ERROR_LOG_1_BRESP, log1);
> +
> +	snprintf(msg, sizeof(msg),
> +		 "addr=0x%llx request_id=0x%x bresp=0x%x\n",
> +		 addr, request_id, bresp);
> +
> +	edac_device_handle_ue(al_pos->edac_dev, 0, 0, msg);
> +
> +	return 1;
> +}
> +
> +static void al_pos_edac_check(struct edac_device_ctl_info *edac_dev)
> +{
> +	struct al_pos_edac *al_pos = edac_dev->pvt_info;
> +
> +	al_pos_handle(al_pos);
> +}
> +
> +static irqreturn_t al_pos_irq_handler(int irq, void *info)
> +{
> +	struct platform_device *pdev = info;
> +	struct al_pos_edac *al_pos = platform_get_drvdata(pdev);
> +
> +	if (al_pos_handle(al_pos))
> +		return IRQ_HANDLED;
> +	return IRQ_NONE;
> +}
> +
> +static int al_pos_probe(struct platform_device *pdev)
> +{
> +	struct edac_device_ctl_info *edac_dev;
> +	struct al_pos_edac *al_pos;
> +	int ret;
> +
> +	edac_dev = edac_device_alloc_ctl_info(sizeof(*al_pos), DRV_NAME, 1,
> +					      DRV_NAME, 1, 0, NULL, 0,
> +					      edac_device_alloc_index());
> +	if (!edac_dev)
> +		return -ENOMEM;
> +
> +	al_pos = edac_dev->pvt_info;
> +	al_pos->edac_dev = edac_dev;
> +	platform_set_drvdata(pdev, al_pos);
> +
> +	al_pos->mmio_base = devm_platform_ioremap_resource(pdev, 0);
> +	if (IS_ERR(al_pos->mmio_base)) {
> +		dev_err(&pdev->dev, "failed to ioremap memory (%ld)\n",
> +			PTR_ERR(al_pos->mmio_base));
> +		return PTR_ERR(al_pos->mmio_base);
> +	}
> +
> +	al_pos->irq = platform_get_irq(pdev, 0);
> +	if (al_pos->irq <= 0)
> +		edac_dev->edac_check = al_pos_edac_check;
> +
> +	edac_dev->dev = &pdev->dev;
> +	edac_dev->mod_name = DRV_NAME;
> +	edac_dev->dev_name = dev_name(&pdev->dev);
> +	edac_dev->ctl_name = "POS";
> +
> +	ret = edac_device_add_device(edac_dev);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed to add edac device\n");
> +		goto err_free_edac;
> +	}
> +
> +	if (al_pos->irq > 0) {
> +		ret = devm_request_irq(&pdev->dev,
> +				       al_pos->irq,
> +				       al_pos_irq_handler,
> +				       0,
> +				       pdev->name,
> +				       pdev);
> +		if (ret != 0) {
> +			dev_err(&pdev->dev,
> +				"failed to register to irq %d (%d)\n",
> +				al_pos->irq, ret);
> +			goto err_remove_edac;

Would it be worth continuing without interrupts? After all, the
interrupt seems to be an optional part of the device...

Thanks,

	M.

-- 
Jazz is not dead, it just smells funny.

Re: [UNVERIFIED SENDER] Re: [PATCH v4 2/2] soc: amazon: al-pos-edac: Introduce Amazon's Annapurna Labs POS EDAC driver

[Shenhar, Talel]

thanks for the review

On 10/7/2019 2:26 PM, Marc Zyngier wrote:
> On Thu, 03 Oct 2019 12:32:41 +0100,
> Talel Shenhar <talel@amazon.com> wrote:
>> +	log1 = readl(al_pos->mmio_base + AL_POS_ERROR_LOG_1);
> I already commented on the misuse of strict accesses. Unless you can
> explain and document *why* you need the extra ordering, please use
> relaxed accesses.
agreeing on relaxed, shall be part of v5
>
>> +
>> +	if (al_pos->irq > 0) {
>> +		ret = devm_request_irq(&pdev->dev,
>> +				       al_pos->irq,
>> +				       al_pos_irq_handler,
>> +				       0,
>> +				       pdev->name,
>> +				       pdev);
>> +		if (ret != 0) {
>> +			dev_err(&pdev->dev,
>> +				"failed to register to irq %d (%d)\n",
>> +				al_pos->irq, ret);
>> +			goto err_remove_edac;
> Would it be worth continuing without interrupts? After all, the
> interrupt seems to be an optional part of the device...

indeed interrupts are optional, however, this is optional for some of 
the systems.

in some cases (and some systems), this error event is critical and 
should cause fast handling. for those, we define the interrupts.

so bottom line, i would like to keep this error in case of error in 
interrupt.

>
> Thanks,
>
> 	M.
>