[PATCH v3 0/3] rockchip,otp: Add support for rk3562 and rk3568

[PATCH v3 0/3] rockchip,otp: Add support for rk3562 and rk3568

[Kever Yang]

Rockchip rk3562 and rk3568 has the same otp controller, which is
different with those already support in the driver, add support for it.

This patch set depends on Jonas's change[1].

[1] https://lore.kernel.org/r/20250316191900.1858944-1-jonas@kwiboo.se

Changes in v3:
- update the clock name from "usr" to "otp" and re-order to sync with
 other soc.
- rebase on rk3576 and rk3528, changes suggest by Jonas
- Use rk3568 data structure for rk3562

Changes in v2:
- Update the commit message and add maxItems in schema.

Finley Xiao (2):
  nvmem: rockchip-otp: Add support for rk3568-otp
  nvmem: rockchip-otp: Add support for rk3562

Kever Yang (1):
  dt-bindings: nvmem: rockchip,otp: Add support for rk3562 and rk3568

 .../bindings/nvmem/rockchip,otp.yaml          | 53 ++++++++++++--
 drivers/nvmem/rockchip-otp.c                  | 73 +++++++++++++++++++
 2 files changed, 119 insertions(+), 7 deletions(-)

-- 
2.25.1

[PATCH v3 1/3] dt-bindings: nvmem: rockchip,otp: Add support for rk3562 and rk3568

[Kever Yang]

Add compatible entry for the otp controller in rk3562 and rk3568, add schema
for different clock names for new entry.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
---

Changes in v3:
- update the clock name from "usr" to "otp" and re-order to sync with
 other soc.

Changes in v2:
- Update the commit message and add maxItems in schema.

 .../bindings/nvmem/rockchip,otp.yaml          | 53 ++++++++++++++++---
 1 file changed, 46 insertions(+), 7 deletions(-)

diff --git a/Documentation/devicetree/bindings/nvmem/rockchip,otp.yaml b/Documentation/devicetree/bindings/nvmem/rockchip,otp.yaml
index a44d44b32809..8b257662f282 100644
--- a/Documentation/devicetree/bindings/nvmem/rockchip,otp.yaml
+++ b/Documentation/devicetree/bindings/nvmem/rockchip,otp.yaml
@@ -14,6 +14,8 @@ properties:
     enum:
       - rockchip,px30-otp
       - rockchip,rk3308-otp
+      - rockchip,rk3562-otp
+      - rockchip,rk3568-otp
       - rockchip,rk3588-otp
 
   reg:
@@ -25,19 +27,15 @@ properties:
 
   clock-names:
     minItems: 3
-    items:
-      - const: otp
-      - const: apb_pclk
-      - const: phy
-      - const: arb
+    maxItems: 4
 
   resets:
     minItems: 1
-    maxItems: 3
+    maxItems: 4
 
   reset-names:
     minItems: 1
-    maxItems: 3
+    maxItems: 4
 
 required:
   - compatible
@@ -62,12 +60,45 @@ allOf:
       properties:
         clocks:
           maxItems: 3
+        clock-names:
+          items:
+            - const: otp
+            - const: apb_pclk
+            - const: phy
         resets:
           maxItems: 1
         reset-names:
           items:
             - const: phy
 
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - rockchip,rk3562-otp
+              - rockchip,rk3568-otp
+    then:
+      properties:
+        clocks:
+          minItems: 4
+          maxItems: 4
+        clock-names:
+          items:
+            - const: otp
+            - const: apb_pclk
+            - const: phy
+            - const: sbpi
+        resets:
+          minItems: 4
+          maxItems: 4
+        reset-names:
+          items:
+            - const: otp
+            - const: apb
+            - const: phy
+            - const: sbpi
+
   - if:
       properties:
         compatible:
@@ -78,8 +109,16 @@ allOf:
       properties:
         clocks:
           minItems: 4
+          maxItems: 4
+        clock-names:
+          items:
+            - const: otp
+            - const: apb_pclk
+            - const: phy
+            - const: arb
         resets:
           minItems: 3
+          maxItems: 3
         reset-names:
           items:
             - const: otp
-- 
2.25.1

[PATCH v3 2/3] nvmem: rockchip-otp: Add support for rk3568-otp

[Kever Yang]

From: Finley Xiao <finley.xiao@rock-chips.com>

This adds the necessary data for handling otp the rk3568.

Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
---

Changes in v3:
- rebase on rk3576 and rk3528, changes suggest by Jonas

Changes in v2: None

 drivers/nvmem/rockchip-otp.c | 69 ++++++++++++++++++++++++++++++++++++
 1 file changed, 69 insertions(+)

diff --git a/drivers/nvmem/rockchip-otp.c b/drivers/nvmem/rockchip-otp.c
index 45bbb6147fb7..cfb69bc58869 100644
--- a/drivers/nvmem/rockchip-otp.c
+++ b/drivers/nvmem/rockchip-otp.c
@@ -27,6 +27,7 @@
 #define OTPC_USER_CTRL			0x0100
 #define OTPC_USER_ADDR			0x0104
 #define OTPC_USER_ENABLE		0x0108
+#define OTPC_USER_QP			0x0120
 #define OTPC_USER_Q			0x0124
 #define OTPC_INT_STATUS			0x0304
 #define OTPC_SBPI_CMD0_OFFSET		0x1000
@@ -184,6 +185,58 @@ static int px30_otp_read(void *context, unsigned int offset,
 	return ret;
 }
 
+static int rk3568_otp_read(void *context, unsigned int offset, void *val,
+			   size_t count)
+{
+	struct rockchip_otp *otp = context;
+	u16 *buf = val;
+	u32 otp_qp;
+	int ret;
+
+	ret = rockchip_otp_reset(otp);
+	if (ret) {
+		dev_err(otp->dev, "failed to reset otp phy\n");
+		return ret;
+	}
+
+	ret = rockchip_otp_ecc_enable(otp, true);
+	if (ret) {
+		dev_err(otp->dev, "rockchip_otp_ecc_enable err\n");
+		return ret;
+	}
+
+	writel(OTPC_USE_USER | OTPC_USE_USER_MASK, otp->base + OTPC_USER_CTRL);
+	udelay(5);
+
+	while (count--) {
+		writel(offset++ | OTPC_USER_ADDR_MASK,
+		       otp->base + OTPC_USER_ADDR);
+		writel(OTPC_USER_FSM_ENABLE | OTPC_USER_FSM_ENABLE_MASK,
+		       otp->base + OTPC_USER_ENABLE);
+
+		ret = rockchip_otp_wait_status(otp, OTPC_INT_STATUS,
+					       OTPC_USER_DONE);
+		if (ret) {
+			dev_err(otp->dev, "timeout during read setup\n");
+			goto read_end;
+		}
+
+		otp_qp = readl(otp->base + OTPC_USER_QP);
+		if (((otp_qp & 0xc0) == 0xc0) || (otp_qp & 0x20)) {
+			ret = -EIO;
+			dev_err(otp->dev, "ecc check error during read setup\n");
+			goto read_end;
+		}
+
+		*buf++ = readl(otp->base + OTPC_USER_Q);
+	}
+
+read_end:
+	writel(0x0 | OTPC_USE_USER_MASK, otp->base + OTPC_USER_CTRL);
+
+	return ret;
+}
+
 static int rk3588_otp_read(void *context, unsigned int offset,
 			   void *val, size_t count)
 {
@@ -280,6 +333,18 @@ static const struct rockchip_data px30_data = {
 	.reg_read = px30_otp_read,
 };
 
+static const char * const rk3568_otp_clocks[] = {
+	"otp", "apb_pclk", "phy", "sbpi",
+};
+
+static const struct rockchip_data rk3568_data = {
+	.size = 0x80,
+	.word_size = sizeof(u16),
+	.clks = rk3568_otp_clocks,
+	.num_clks = ARRAY_SIZE(rk3568_otp_clocks),
+	.reg_read = rk3568_otp_read,
+};
+
 static const struct rockchip_data rk3576_data = {
 	.size = 0x100,
 	.read_offset = 0x700,
@@ -311,6 +376,10 @@ static const struct of_device_id rockchip_otp_match[] = {
 		.compatible = "rockchip,rk3308-otp",
 		.data = &px30_data,
 	},
+	{
+		.compatible = "rockchip,rk3568-otp",
+		.data = &rk3568_data,
+	},
 	{
 		.compatible = "rockchip,rk3576-otp",
 		.data = &rk3576_data,
-- 
2.25.1

[PATCH v3 3/3] nvmem: rockchip-otp: Add support for rk3562

[Kever Yang]

From: Finley Xiao <finley.xiao@rock-chips.com>

This adds the necessary data for handling otp on the rk3562.

Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
---

Changes in v3:
- Use rk3568 data structure for rk3562

Changes in v2: None

 drivers/nvmem/rockchip-otp.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/nvmem/rockchip-otp.c b/drivers/nvmem/rockchip-otp.c
index cfb69bc58869..62ce22d72586 100644
--- a/drivers/nvmem/rockchip-otp.c
+++ b/drivers/nvmem/rockchip-otp.c
@@ -376,6 +376,10 @@ static const struct of_device_id rockchip_otp_match[] = {
 		.compatible = "rockchip,rk3308-otp",
 		.data = &px30_data,
 	},
+	{
+		.compatible = "rockchip,rk3562-otp",
+		.data = &rk3568_data,
+	},
 	{
 		.compatible = "rockchip,rk3568-otp",
 		.data = &rk3568_data,
-- 
2.25.1

Re: [PATCH v3 1/3] dt-bindings: nvmem: rockchip,otp: Add support for rk3562 and rk3568

[Rob Herring (Arm)]

On Tue, 15 Apr 2025 18:32:01 +0800, Kever Yang wrote:
> Add compatible entry for the otp controller in rk3562 and rk3568, add schema
> for different clock names for new entry.
> 
> Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
> ---
> 
> Changes in v3:
> - update the clock name from "usr" to "otp" and re-order to sync with
>  other soc.
> 
> Changes in v2:
> - Update the commit message and add maxItems in schema.
> 
>  .../bindings/nvmem/rockchip,otp.yaml          | 53 ++++++++++++++++---
>  1 file changed, 46 insertions(+), 7 deletions(-)
> 

Reviewed-by: Rob Herring (Arm) <robh@kernel.org>

Re: [PATCH v3 1/3] dt-bindings: nvmem: rockchip,otp: Add support for rk3562 and rk3568

[Heiko Stübner]

Am Dienstag, 15. April 2025, 12:32:01 Mitteleuropäische Sommerzeit schrieb Kever Yang:
> Add compatible entry for the otp controller in rk3562 and rk3568, add schema
> for different clock names for new entry.
> 
> Signed-off-by: Kever Yang <kever.yang@rock-chips.com>

Reviewed-by: Heiko Stuebner <heiko@sntech.de>

Re: [PATCH v3 2/3] nvmem: rockchip-otp: Add support for rk3568-otp

[Heiko Stübner]

Am Dienstag, 15. April 2025, 12:32:02 Mitteleuropäische Sommerzeit schrieb Kever Yang:
> From: Finley Xiao <finley.xiao@rock-chips.com>
> 
> This adds the necessary data for handling otp the rk3568.
> 
> Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
> Signed-off-by: Kever Yang <kever.yang@rock-chips.com>


Reviewed-by: Heiko Stuebner <heiko@sntech.de>

On a Quartz64b
Tested-by: Heiko Stuebner <heiko@sntech.de>


> ---
> 
> Changes in v3:
> - rebase on rk3576 and rk3528, changes suggest by Jonas
> 
> Changes in v2: None
> 
>  drivers/nvmem/rockchip-otp.c | 69 ++++++++++++++++++++++++++++++++++++
>  1 file changed, 69 insertions(+)
> 
> diff --git a/drivers/nvmem/rockchip-otp.c b/drivers/nvmem/rockchip-otp.c
> index 45bbb6147fb7..cfb69bc58869 100644
> --- a/drivers/nvmem/rockchip-otp.c
> +++ b/drivers/nvmem/rockchip-otp.c
> @@ -27,6 +27,7 @@
>  #define OTPC_USER_CTRL			0x0100
>  #define OTPC_USER_ADDR			0x0104
>  #define OTPC_USER_ENABLE		0x0108
> +#define OTPC_USER_QP			0x0120
>  #define OTPC_USER_Q			0x0124
>  #define OTPC_INT_STATUS			0x0304
>  #define OTPC_SBPI_CMD0_OFFSET		0x1000
> @@ -184,6 +185,58 @@ static int px30_otp_read(void *context, unsigned int offset,
>  	return ret;
>  }
>  
> +static int rk3568_otp_read(void *context, unsigned int offset, void *val,
> +			   size_t count)
> +{
> +	struct rockchip_otp *otp = context;
> +	u16 *buf = val;
> +	u32 otp_qp;
> +	int ret;
> +
> +	ret = rockchip_otp_reset(otp);
> +	if (ret) {
> +		dev_err(otp->dev, "failed to reset otp phy\n");
> +		return ret;
> +	}
> +
> +	ret = rockchip_otp_ecc_enable(otp, true);
> +	if (ret) {
> +		dev_err(otp->dev, "rockchip_otp_ecc_enable err\n");
> +		return ret;
> +	}
> +
> +	writel(OTPC_USE_USER | OTPC_USE_USER_MASK, otp->base + OTPC_USER_CTRL);
> +	udelay(5);
> +
> +	while (count--) {
> +		writel(offset++ | OTPC_USER_ADDR_MASK,
> +		       otp->base + OTPC_USER_ADDR);
> +		writel(OTPC_USER_FSM_ENABLE | OTPC_USER_FSM_ENABLE_MASK,
> +		       otp->base + OTPC_USER_ENABLE);
> +
> +		ret = rockchip_otp_wait_status(otp, OTPC_INT_STATUS,
> +					       OTPC_USER_DONE);
> +		if (ret) {
> +			dev_err(otp->dev, "timeout during read setup\n");
> +			goto read_end;
> +		}
> +
> +		otp_qp = readl(otp->base + OTPC_USER_QP);
> +		if (((otp_qp & 0xc0) == 0xc0) || (otp_qp & 0x20)) {
> +			ret = -EIO;
> +			dev_err(otp->dev, "ecc check error during read setup\n");
> +			goto read_end;
> +		}
> +
> +		*buf++ = readl(otp->base + OTPC_USER_Q);
> +	}
> +
> +read_end:
> +	writel(0x0 | OTPC_USE_USER_MASK, otp->base + OTPC_USER_CTRL);
> +
> +	return ret;
> +}
> +
>  static int rk3588_otp_read(void *context, unsigned int offset,
>  			   void *val, size_t count)
>  {
> @@ -280,6 +333,18 @@ static const struct rockchip_data px30_data = {
>  	.reg_read = px30_otp_read,
>  };
>  
> +static const char * const rk3568_otp_clocks[] = {
> +	"otp", "apb_pclk", "phy", "sbpi",
> +};
> +
> +static const struct rockchip_data rk3568_data = {
> +	.size = 0x80,
> +	.word_size = sizeof(u16),
> +	.clks = rk3568_otp_clocks,
> +	.num_clks = ARRAY_SIZE(rk3568_otp_clocks),
> +	.reg_read = rk3568_otp_read,
> +};
> +
>  static const struct rockchip_data rk3576_data = {
>  	.size = 0x100,
>  	.read_offset = 0x700,
> @@ -311,6 +376,10 @@ static const struct of_device_id rockchip_otp_match[] = {
>  		.compatible = "rockchip,rk3308-otp",
>  		.data = &px30_data,
>  	},
> +	{
> +		.compatible = "rockchip,rk3568-otp",
> +		.data = &rk3568_data,
> +	},
>  	{
>  		.compatible = "rockchip,rk3576-otp",
>  		.data = &rk3576_data,
> 

Re: [PATCH v3 3/3] nvmem: rockchip-otp: Add support for rk3562

[Heiko Stübner]

Am Dienstag, 15. April 2025, 12:32:03 Mitteleuropäische Sommerzeit schrieb Kever Yang:
> From: Finley Xiao <finley.xiao@rock-chips.com>
> 
> This adds the necessary data for handling otp on the rk3562.
> 
> Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
> Signed-off-by: Kever Yang <kever.yang@rock-chips.com>

Reviewed-by: Heiko Stuebner <heiko@sntech.de>

Re: [PATCH v3 3/3] nvmem: rockchip-otp: Add support for rk3562

[Willy Tarreau]

On Tue, Apr 15, 2025 at 06:32:03PM +0800, Kever Yang wrote:
> From: Finley Xiao <finley.xiao@rock-chips.com>
> 
> This adds the necessary data for handling otp on the rk3562.
> 
> Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
> Signed-off-by: Kever Yang <kever.yang@rock-chips.com>

Successfully tested with Jonas' patches on top of it on my
Radxa E20C (RK3528):

Tested-by: Willy Tarreau <w@1wt.eu>

Willy

Re: [PATCH v3 3/3] nvmem: rockchip-otp: Add support for rk3562

[Heiko Stübner]

Hi Willy,

Am Montag, 28. Juli 2025, 21:01:54 Mitteleuropäische Normalzeit schrieb Willy Tarreau:
> On Tue, Apr 15, 2025 at 06:32:03PM +0800, Kever Yang wrote:
> > From: Finley Xiao <finley.xiao@rock-chips.com>
> > 
> > This adds the necessary data for handling otp on the rk3562.
> > 
> > Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
> > Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
> 
> Successfully tested with Jonas' patches on top of it on my
> Radxa E20C (RK3528):
> 
> Tested-by: Willy Tarreau <w@1wt.eu>

as you replied to the rk3562 variant, does that mean the rk3528
uses the same variant data?

Thanks
Heiko

Re: [PATCH v3 3/3] nvmem: rockchip-otp: Add support for rk3562

[Willy Tarreau]

Hi Heiko,

On Thu, Feb 05, 2026 at 01:50:34PM +0100, Heiko Stübner wrote:
> Hi Willy,
> 
> Am Montag, 28. Juli 2025, 21:01:54 Mitteleuropäische Normalzeit schrieb Willy Tarreau:
> > On Tue, Apr 15, 2025 at 06:32:03PM +0800, Kever Yang wrote:
> > > From: Finley Xiao <finley.xiao@rock-chips.com>
> > > 
> > > This adds the necessary data for handling otp on the rk3562.
> > > 
> > > Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
> > > Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
> > 
> > Successfully tested with Jonas' patches on top of it on my
> > Radxa E20C (RK3528):
> > 
> > Tested-by: Willy Tarreau <w@1wt.eu>
> 
> as you replied to the rk3562 variant, does that mean the rk3528
> uses the same variant data?

Now you make me doubt, I'll have to recheck the original message
in its series and the patches that I had applied. I seem to
rememeber there were certain common patches between the two but
I don't remember which ones, and it's also possibly that I replied
by accident to the wrong message in the series or that I mistook
one patch for another when comparing them :-/

Ah, I seem to have found it, there's indeed the 3562 in my local
patches and the 3528 as well, both of which rely on the 3568 one:

  commit a68433c468c8a9d4789641eb6a40ae571fb7f01f
  Author: Finley Xiao <finley.xiao@rock-chips.com>
  Date:   Tue Apr 15 18:32:03 2025 +0800

    nvmem: rockchip-otp: Add support for RK3562
    
    This adds the necessary data for handling otp on the rk3562.
    
    Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
    Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
    Reviewed-by: Heiko Stuebner <heiko@sntech.de>
    Signed-off-by: Jonas Karlman <jonas@kwiboo.se>

and:

  commit 249b07e24d3d1d47b7ec23d5f09a56837b66d7f5
  Author: Jonas Karlman <jonas@kwiboo.se>
  Date:   Sun Mar 16 00:05:45 2025 +0000

    nvmem: rockchip-otp: Add support for RK3528
    
    Add support for the OTP controller in RK3528. The OTPC is similar to the
    OTPC in RK3562 and RK3568, exept for a missing phy clock and reset.
    
    Signed-off-by: Jonas Karlman <jonas@kwiboo.se>

It's very possible that I visually mistook one for the other when
responding to the series, and if so, I'm really sorry about that!

The former relies on rk3568_data and the latter on rk3528_data, which
is defined like this:

+static const char * const rk3528_otp_clocks[] = {
+       "otp", "apb_pclk", "sbpi",
+};
+
+static const struct rockchip_data rk3528_data = {
+       .size = 0x80,
+       .word_size = sizeof(u16),
+       .clks = rk3528_otp_clocks,
+       .num_clks = ARRAY_SIZE(rk3528_otp_clocks),
+       .reg_read = rk3568_otp_read,
+};

Looks like the patches originally come from this branch that I
manually cherry-picked:

  https://github.com/Kwiboo/linux-rockchip/commits/next-20250620-rk3528

I have one extra E20C that's not yet in prod if you're interested in
me running some specific tests. It will not happen quickly but just let
me know if I can help, as it's a nice little machine.

Willy

Re: [PATCH v3 3/3] nvmem: rockchip-otp: Add support for rk3562

[Heiko Stübner]

Hi Willy,

Am Donnerstag, 5. Februar 2026, 16:36:33 Mitteleuropäische Normalzeit schrieb Willy Tarreau:
> On Thu, Feb 05, 2026 at 01:50:34PM +0100, Heiko Stübner wrote:
> > Am Montag, 28. Juli 2025, 21:01:54 Mitteleuropäische Normalzeit schrieb Willy Tarreau:
> > > On Tue, Apr 15, 2025 at 06:32:03PM +0800, Kever Yang wrote:
> > > > From: Finley Xiao <finley.xiao@rock-chips.com>
> > > > 
> > > > This adds the necessary data for handling otp on the rk3562.
> > > > 
> > > > Signed-off-by: Finley Xiao <finley.xiao@rock-chips.com>
> > > > Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
> > > 
> > > Successfully tested with Jonas' patches on top of it on my
> > > Radxa E20C (RK3528):
> > > 
> > > Tested-by: Willy Tarreau <w@1wt.eu>
> > 
> > as you replied to the rk3562 variant, does that mean the rk3528
> > uses the same variant data?

[...]

> It's very possible that I visually mistook one for the other when
> responding to the series, and if so, I'm really sorry about that!

Really, no worries ... it just confused me a bit, while I was
resurrecting the whole thing :-)

[...]

> I have one extra E20C that's not yet in prod if you're interested in
> me running some specific tests. It will not happen quickly but just let
> me know if I can help, as it's a nice little machine.

With Jonas' blessing I added his patches for the RK3528 and send
the driver thing rebased onto a recent kernel just now [0] with you
in Cc.

I've also send in the corresponding DTS patches for all 3 [1], again
with you in Cc :-) .

So if you can give them a spin on your RK3528 board, that would be great.
Because I need to find my RK3528 board first :-) .


Thanks a lot
Heiko


[0] https://lore.kernel.org/linux-rockchip/20260205211901.490181-1-heiko@sntech.de/T/#t
[1] https://lore.kernel.org/linux-rockchip/20260205212347.490636-1-heiko@sntech.de/T/#t