Skip to contentSkip to conversation

Implement nvmem support for mtd

PATCH In review Ansuel Smith 3 participants 14 messages Feb 16, 2021 – Mar 8, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v2 cover Implement nvmem support for mtd

The mtd support for the nvmem api has been stalled from 2018 with a patch half pushed hoping that a scheme is found for the mtd name later. This patchset try to address this.

The solution is simple.
New partitions scheme should always have the partitions {} structure and declare subnodes as partitions is deprecated and should not be used anymore. Fixed-partitions parser is changed to parse direct subnode as partitions only if the appropriate compatible is used. This change make possible the use of nvmem-partitions compatible and the entire partition node can be parsed by the nvmem of framework.
The current code register the partition to the nvmem framework every time but skip actually of_node parting. The new nvmem-partitions compatible is used to enable of_node parsing on the desired partitions.

Ansuel Smith (3):
mtd: partitions: ofpart: skip subnodes parse with compatible
mtd: core: add nvmem-partitions compatible to parse mtd as nvmem cells
dt-bindings: mtd: Document use of nvmem-partitions compatible

 .../mtd/partitions/nvmem-partitions.yaml      | 105 ++++++++++++++++++
 drivers/mtd/mtdcore.c                         |   3 +-
 drivers/mtd/parsers/ofpart.c                  |   5 +
 3 files changed, 112 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/mtd/partitions/nvmem-partitions.yaml
Signature · 1 line
2.30.0
#1 lore raw

PATCH v2 1/3 mtd: partitions: ofpart: skip subnodes parse with compatible

If a partitions structure is not used, parse direct subnodes as fixed-partitions only if a compatible is not found or is of type fixed-partition. A parser can be used directly on the subnode and subnodes should not be parsed as fixed-partitions by default.

---
 drivers/mtd/parsers/ofpart.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/mtd/parsers/ofpart.c +5 -0
index daf507c123e6..4b363dd0311c 100644
@@ -50,6 +50,11 @@ static int parse_fixed_partitions(struct mtd_info *master,
 			 master->name, mtd_node);
 		ofpart_node = mtd_node;
 		dedicated = false;
+
+		/* Skip parsing direct subnodes if a compatible is found and is not fixed-partitions */
+		if (node_has_compatible(ofpart_node) &&
+		    !of_device_is_compatible(ofpart_node, "fixed-partitions"))
+			return 0;
 	} else if (!of_device_is_compatible(ofpart_node, "fixed-partitions")) {
 		/* The 'partitions' subnode might be used by another parser */
 		return 0;
Signature · 6 lines
2.30.0


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/
#2 lore raw

Re: [PATCH v2 1/3] mtd: partitions: ofpart: skip subnodes parse with compatible

Quoted 16 lines · Ansuel Smith

If a partitions structure is not used, parse direct subnodes as fixed-partitions only if a compatible is not found or is of type fixed-partition. A parser can be used directly on the subnode and subnodes should not be parsed as fixed-partitions by default.

---
  drivers/mtd/parsers/ofpart.c | 5 +++++
  1 file changed, 5 insertions(+)
drivers/mtd/parsers/ofpart.c +0 -0
index daf507c123e6..4b363dd0311c 100644
@@ -50,6 +50,11 @@ static int parse_fixed_partitions(struct mtd_info *master,
  			 master->name, mtd_node);

drivers/mtd/parsers/ofpart.c static int parse_fixed_partitions(struct mtd_info *master, · line 60

Quoted 10 lines
drivers/mtd/parsers/ofpart.c +5 -0
  		ofpart_node = mtd_node;
  		dedicated = false;
+
+		/* Skip parsing direct subnodes if a compatible is found and is not fixed-partitions */
+		if (node_has_compatible(ofpart_node) &&
+		    !of_device_is_compatible(ofpart_node, "fixed-partitions"))
+			return 0;
  	} else if (!of_device_is_compatible(ofpart_node, "fixed-partitions")) {
  		/* The 'partitions' subnode might be used by another parser */
  		return 0;

I admit I'm not familiar with the old binding, so let me know if my understanding is incorrect.

It seems to me however that your change will break parsing in cases like:

spi-flash@0 {
	compatible = "jedec,spi-nor";
	reg = <0x0>;

	partition@0 {
		label = "bootloader";
		reg = <0x0 0x100000>;
	};
};
nandcs@0 {
	compatible = "brcm,nandcs";
	reg = <0>;

	partition@0 {
		label = "bootloader";
		reg = <0x0000000 0x10000>;
	};
};

Did we ever use "fixed-partitions" without partitions { } subnode?

#3 lore raw

Re: [PATCH v2 1/3] mtd: partitions: ofpart: skip subnodes parse with compatible

Quoted 54 lines · Rafał Miłecki
Quoted 16 lines · Ansuel Smith

If a partitions structure is not used, parse direct subnodes as fixed-partitions only if a compatible is not found or is of type fixed-partition. A parser can be used directly on the subnode and subnodes should not be parsed as fixed-partitions by default.

---
  drivers/mtd/parsers/ofpart.c | 5 +++++
  1 file changed, 5 insertions(+)
drivers/mtd/parsers/ofpart.c +0 -0
index daf507c123e6..4b363dd0311c 100644
@@ -50,6 +50,11 @@ static int parse_fixed_partitions(struct mtd_info *master,
  			 master->name, mtd_node);

drivers/mtd/parsers/ofpart.c static int parse_fixed_partitions(struct mtd_info *master, · line 60

Quoted 10 lines
drivers/mtd/parsers/ofpart.c +5 -0
  		ofpart_node = mtd_node;
  		dedicated = false;
+
+		/* Skip parsing direct subnodes if a compatible is found and is not fixed-partitions */
+		if (node_has_compatible(ofpart_node) &&
+		    !of_device_is_compatible(ofpart_node, "fixed-partitions"))
+			return 0;
  	} else if (!of_device_is_compatible(ofpart_node, "fixed-partitions")) {
  		/* The 'partitions' subnode might be used by another parser */
  		return 0;

I admit I'm not familiar with the old binding, so let me know if my understanding is incorrect.

It seems to me however that your change will break parsing in cases like:

spi-flash@0 {
	compatible = "jedec,spi-nor";
	reg = <0x0>;

	partition@0 {
		label = "bootloader";
		reg = <0x0 0x100000>;
	};
};
nandcs@0 {
	compatible = "brcm,nandcs";
	reg = <0>;

	partition@0 {
		label = "bootloader";
		reg = <0x0000000 0x10000>;
	};
};
Quoted 1 line

Did we ever use "fixed-partitions" without partitions { } subnode?

Hi, very good point. You are right and I didin't think about this case. I don't want to assume false statement, but since the ofpart parser and the partitions structure should have been pushed at the same time, there shouldn't be any use of "fixed-partitions" without partitions { } subnodes. With this assumtion, the current implementation looks to be the cleanest way to skip parsing. (if the parsing is dubious, don't parse at all... The idea was that)
The hacky and IMHO dirty way to solve this is add a bool to directly skip the subnode parsing and check for that. Something like "no-fixed-partition" that would disable the ofnode parser with no partitions { } subnode would accomplish the same result of this patch and keep compatibility with nodes scheme you pointed out.

#4 lore raw

PATCH v2 2/3 mtd: core: add nvmem-partitions compatible to parse mtd as nvmem cells

In reply to Ansuel Smith

Partitions that contains the nvmem-partitions compatible will register their direct subonodes as nvmem cells and the node will be treated as a nvmem provider.

---
 drivers/mtd/mtdcore.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mtd/mtdcore.c +2 -1
index 2d6423d89a17..29d257678a86 100644
@@ -531,6 +531,7 @@ static int mtd_nvmem_reg_read(void *priv, unsigned int offset,
 
 static int mtd_nvmem_add(struct mtd_info *mtd)
 {
+	struct device_node *node = mtd_get_of_node(mtd);
 	struct nvmem_config config = {};
 
 	config.id = -1;
@@ -543,7 +544,7 @@ static int mtd_nvmem_add(struct mtd_info *mtd)
 	config.stride = 1;
 	config.read_only = true;
 	config.root_only = true;
-	config.no_of_node = true;
+	config.no_of_node = !of_device_is_compatible(node, "nvmem-partitions");
 	config.priv = mtd;
 
 	mtd->nvmem = nvmem_register(&config);
Signature · 6 lines
2.30.0


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/
Tested-by on v2 2/3
#5 lore raw

Re: [PATCH v2 2/3] mtd: core: add nvmem-partitions compatible to parse mtd as nvmem cells

Quoted 5 lines · Ansuel Smith

Partitions that contains the nvmem-partitions compatible will register their direct subonodes as nvmem cells and the node will be treated as a nvmem provider.

I applied this patch on top of the:
[PATCH] mtd: parsers: ofpart: limit parsing of deprecated DT syntax

I succesfully used NVMEM cell defined in bootloader mtd partition for reading device MAC address.

partitions {
	compatible = "fixed-partitions";
	#address-cells = <1>;
	#size-cells = <1>;

	partition@0 {
		compatible = "nvmem-partitions";
		label = "bootloader";
		reg = <0x0 0x100000>;
		#address-cells = <1>;
		#size-cells = <1>;
		ranges = <0 0x0 0x100000>;

		base_mac_addr: mac@106a0 {
			reg = <0x106a0 0x6>;
		};
	};
}
#6 lore raw

PATCH v2 3/3 dt-bindings: mtd: Document use of nvmem-partitions compatible

In reply to Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

---
 .../mtd/partitions/nvmem-partitions.yaml      | 105 ++++++++++++++++++
 1 file changed, 105 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/mtd/partitions/nvmem-partitions.yaml
Documentation/devicetree/bindings/mtd/partitions/nvmem-partitions.yaml new file +105 -0
new file mode 100644
index 000000000000..1ff283febcaa
@@ -0,0 +1,105 @@
+# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/mtd/partitions/nvmem-partitions.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Nvmem partitions
+
+description: |
+  This binding can be used to treat the specific partition as a nvmem provider.
+  Each direct subnodes represents the nvmem cells and won't be parsed as fixed-partitions.
+  Fixed-partitions bindings described in fixed-partitions.yaml apply to the nvmem provider node.
+
+maintainers:
+  - Ansuel Smith <[email protected]>
+
+properties:
+  compatible:
+    const: nvmem-partitions
+
+  "#address-cells": true
+
+  "#size-cells": true
+
+  reg:
+    description: partition's offset and size within the flash
+    maxItems: 1
+
+required:
+  - compatible
+  - "#address-cells"
+  - "#size-cells"
+  - reg
+
+additionalProperties: true
+
+examples:
+  - |
+    partitions {
+      compatible = "fixed-partitions";
+      #address-cells = <1>;
+      #size-cells = <1>;
+
+      /* ... */
+
+      };
+      art: art@1200000 {
+        compatible = "nvmem-partitions";
+        reg = <0x1200000 0x0140000>;
+        label = "art";
+        read-only;
+        #address-cells = <1>;
+        #size-cells = <1>;
+
+        macaddr_gmac1: macaddr_gmac1@0 {
+          reg = <0x0 0x6>;
+        };
+
+        macaddr_gmac2: macaddr_gmac2@6 {
+          reg = <0x6 0x6>;
+        };
+
+        pre_cal_24g: pre_cal_24g@1000 {
+          reg = <0x1000 0x2f20>;
+        };
+
+        pre_cal_5g: pre_cal_5g@5000{
+          reg = <0x5000 0x2f20>;
+        };
+      };
+  - |
+    partitions {
+        compatible = "fixed-partitions";
+        #address-cells = <1>;
+        #size-cells = <1>;
+
+        partition@0 {
+            label = "bootloader";
+            reg = <0x000000 0x100000>;
+            read-only;
+        };
+
+        firmware@100000 {
+            compatible = "brcm,trx";
+            label = "firmware";
+            reg = <0x100000 0xe00000>;
+        };
+
+        calibration@f00000 {
+            compatible = "nvmem-partitions";
+            label = "calibration";
+            reg = <0xf00000 0x100000>;
+            ranges = <0 0xf00000 0x100000>;
+            #address-cells = <1>;
+            #size-cells = <1>;
+
+            wifi0@0 {
+                reg = <0x000000 0x080000>;
+            };
+
+            wifi1@80000 {
+                reg = <0x080000 0x080000>;
+            };
+        };
+    };
Signature · 6 lines
2.30.0


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/
#7 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

[Rob: please advise]

Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

Until now we were using "compatible" string in partition node only for parsers (looking for subpartitions). We need to think if this change can break anything from DT / Linux perspective.

Compatible strings should be unique, so there is no risk of conflict between NVMEM and parsers.

Now: can we ever need mtd partition to:
1. Contain subpartitions
2. Provide NVMEM
at the same time?

Let's say:

partition@0 {
	compatible = "vendor,dynamic-firmware-partitions", "nvmem-partitions";
	label = "firmware";
	reg = <0x0 0x100000>;
	#address-cells = <1>;
	#size-cells = <1>;
	ranges = <0 0x0 0x100000>;

	firmware-version@10 {
		reg = <0x10 0x4>;
	};

	firmware-date@10 {
		reg = <0x20 0x4>;
	};
};

Is that allowed to respect both "compatible" strings and have:
1. Linux parser parse "firmware" for subpartitions
2. Linux MTD register "firmware" as NVMEM device
?

If not, what other options do we have? Is that allowed to have a dangling MTD NVMEM node with phandle to MTD partition?

firmware: partition@0 {
	compatible = "vendor,dynamic-firmware-partitions";
	label = "firmware";
	reg = <0x0 0x100000>;
};

(...)

firmware-version@10 {
	compatible = "mtd-nvmem";
	reg = <0x10 0x4>;
	mtd = <&firmware>;
};
firmware-date@10 {
	compatible = "mtd-nvmem";
	reg = <0x20 0x4>;
	mtd = <&firmware>;
};

Rob: I'd really appreciate your input & help here.

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

maintainer
#8 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

Quoted 18 lines · Rafał Miłecki

[Rob: please advise]

Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

Until now we were using "compatible" string in partition node only for parsers (looking for subpartitions). We need to think if this change can break anything from DT / Linux perspective.

Compatible strings should be unique, so there is no risk of conflict between NVMEM and parsers.

Now: can we ever need mtd partition to:
1. Contain subpartitions
2. Provide NVMEM
at the same time?

Quoted 4 lines

Let's say:

partition@0 {
	compatible = "vendor,dynamic-firmware-partitions", "nvmem-partitions";

I think you'd want the "vendor,dynamic-firmware-partitions" parser/code to serve up any nvmem regions. Whether you have a fallback here depends if an OS could make use of the regions knowing nothing about "vendor,dynamic-firmware-partitions".

Quoted 37 lines
	label = "firmware";
	reg = <0x0 0x100000>;
	#address-cells = <1>;
	#size-cells = <1>;
	ranges = <0 0x0 0x100000>;

	firmware-version@10 {
		reg = <0x10 0x4>;
	};

	firmware-date@10 {
		reg = <0x20 0x4>;
	};
};

Is that allowed to respect both "compatible" strings and have:
1. Linux parser parse "firmware" for subpartitions
2. Linux MTD register "firmware" as NVMEM device
?

If not, what other options do we have? Is that allowed to have a dangling MTD NVMEM node with phandle to MTD partition?

firmware: partition@0 {
	compatible = "vendor,dynamic-firmware-partitions";
	label = "firmware";
	reg = <0x0 0x100000>;
};

(...)

firmware-version@10 {
	compatible = "mtd-nvmem";
	reg = <0x10 0x4>;
	mtd = <&firmware>;
};
Quoted 5 lines
firmware-date@10 {
	compatible = "mtd-nvmem";
	reg = <0x20 0x4>;
	mtd = <&firmware>;
};

This, I would not like to see.

Rob

#9 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

Quoted 21 lines · Rob Herring
Quoted 20 lines · Rafał Miłecki

[Rob: please advise]

Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

Until now we were using "compatible" string in partition node only for parsers (looking for subpartitions). We need to think if this change can break anything from DT / Linux perspective.

Compatible strings should be unique, so there is no risk of conflict between NVMEM and parsers.

Now: can we ever need mtd partition to:
1. Contain subpartitions
2. Provide NVMEM
at the same time?

Let's say:

Quoted 7 lines
Quoted 2 lines
partition@0 {
	compatible = "vendor,dynamic-firmware-partitions", "nvmem-partitions";

I think you'd want the "vendor,dynamic-firmware-partitions" parser/code to serve up any nvmem regions. Whether you have a fallback here depends if an OS could make use of the regions knowing nothing about "vendor,dynamic-firmware-partitions".

Perfect! I didn't think that driver handling "vendor,dynamic-firmware-partitions" may also take care of NVMEM.

Thank you.

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

#10 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

In reply to Ansuel Smith
Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

I'm just wondering if "nvmem-partitions" is accurate enough. Partitions bit sounds a bit ambiguous in the mtd context.

What do you think about "mtd-nvmem-cells" or just "nvmem-cells"?

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/

#11 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

Quoted 8 lines · Rafał Miłecki
Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

I'm just wondering if "nvmem-partitions" is accurate enough. Partitions bit sounds a bit ambiguous in the mtd context.

What do you think about "mtd-nvmem-cells" or just "nvmem-cells"?

I read somewhere that mtd is not so standard so "nvmem-cells" should be the right compatible.
To sum up, with v2 I should change the compatible name and just push the 2 and 3 patch. (waiting your fix to be accepted) Correct?

#12 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

Quoted 8 lines · Ansuel Smith
Quoted 7 lines · Rafał Miłecki
Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

I'm just wondering if "nvmem-partitions" is accurate enough. Partitions bit sounds a bit ambiguous in the mtd context.

Quoted 6 lines
Quoted 1 line

What do you think about "mtd-nvmem-cells" or just "nvmem-cells"?

I read somewhere that mtd is not so standard so "nvmem-cells" should be the right compatible.
To sum up, with v2 I should change the compatible name and just push the 2 and 3 patch. (waiting your fix to be accepted) Correct?

Yes, I believe so

#13 lore raw

Re: [PATCH v2 3/3] dt-bindings: mtd: Document use of nvmem-partitions compatible

In reply to Ansuel Smith
Quoted 8 lines · Ansuel Smith
Quoted 7 lines · Rafał Miłecki
Quoted 2 lines · Ansuel Smith

Document nvmem-partitions compatible used to treat mtd partitions as a nvmem provider.

I'm just wondering if "nvmem-partitions" is accurate enough. Partitions bit sounds a bit ambiguous in the mtd context.

Quoted 6 lines
Quoted 1 line

What do you think about "mtd-nvmem-cells" or just "nvmem-cells"?

I read somewhere that mtd is not so standard so "nvmem-cells" should be the right compatible.
To sum up, with v2 I should change the compatible name and just push the 2 and 3 patch. (waiting your fix to be accepted) Correct?

I'm also quite sure you're fine to send V2 now, if you just let maintainers know (e.g. in a comment below a --- tear line) that it depends on the:
[PATCH] mtd: parsers: ofpart: limit parsing of deprecated DT syntax

In other words: you don't need to wait for my patch to get accepted.

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/