[PATCH v2 00/14] iio: adc: stm32: add support for stm32mp23 and stm32mp25

[PATCH v2 00/14] iio: adc: stm32: add support for stm32mp23 and stm32mp25

[Fabrice Gasnier]

Add support for STM32MP23 and STM32MP25 Analog to Digital Converter.
It's a variant of other supported STM32MP ADCs. Add the DT bindings,
driver, SOC DTS and an example on stm32mp257f-ev1 board.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- dt-bindings: updates since discussion with Jonathan. Drop fallback
  compatible and "st,adc-trigger-sel".
- driver: address Jonathan and Andy comments and overall:
  - Split into several pieces to prepare the infrastructure without
    functional change
  - Add stm32mp23/25 support
  - Then add vddgpu feature.
- Rebased on top of fixes sent separately, since pre-existing issues have
  been reported by Sashiko on v1.
  https://patch.msgid.link/20260916-adc-fix-intchan-v1-v3-1-b5f05b86bc60@foss.st.com
  https://patch.msgid.link/20260916-adc-fix-div0-v3-1-bedf027a6e7e@foss.st.com
- Link to v1: https://patch.msgid.link/20260911-adc-stm32mp25-v1-v1-0-cef26eecdef2@foss.st.com

---
Cheick Traore (1):
      arm64: dts: st: add ADC nodes on stm32mp231

Fabrice Gasnier (10):
      dt-bindings: iio: adc: stm32-adc: add support for stm32mp23 and stm32mp25
      iio: adc: stm32-adc: core: manage clock prescaler diversity
      iio: adc: stm32-adc: core: configurable number of interrupts
      iio: adc: stm32-adc: rework internal channels data
      iio: adc: stm32-adc: add vreg enable option to manage diversity
      iio: adc: stm32-adc: add support for stm32mp25
      iio: adc: stm32-adc: add support for stm32mp23
      iio: adc: stm32: add support for vddgpu on stm32mp23 and stm32mp25
      arm64: dts: st: add ADC nodes on stm32mp251
      arm64: dts: st: add ADC support on stm32mp257f-ev1 board

Olivier Moysan (1):
      iio: adc: stm32-adc: manage characterization voltage diversity

Patrice Chotard (2):
      arm64: dts: st: add vrefint calibration on stm32mp23
      arm64: dts: st: add vrefint calibration on stm32mp25

 .../devicetree/bindings/iio/adc/st,stm32-adc.yaml  |  71 +++
 arch/arm64/boot/dts/st/stm32mp231.dtsi             | 116 +++++
 arch/arm64/boot/dts/st/stm32mp251.dtsi             | 116 +++++
 arch/arm64/boot/dts/st/stm32mp257f-ev1.dts         |  22 +
 drivers/iio/adc/stm32-adc-core.c                   |  95 +++-
 drivers/iio/adc/stm32-adc-core.h                   |  21 +
 drivers/iio/adc/stm32-adc.c                        | 570 ++++++++++++++++++++-
 7 files changed, 965 insertions(+), 46 deletions(-)
---
base-commit: 182c47735554925da9605dc9dd90a8c15a1bbae5
change-id: 20260911-adc-stm32mp25-v1-2ac7328559d9

Best regards,
--  
Fabrice Gasnier <fabrice.gasnier@foss.st.com>

[PATCH v2 01/14] dt-bindings: iio: adc: stm32-adc: add support for stm32mp23 and stm32mp25

[Fabrice Gasnier]

Add stm32mp23 and stm32mp25 compatibles. Three ADCs can be found on
these SoCs, split into two blocks:
- ADC1 and ADC2 are tightly coupled (common registers).
- ADC3 is managed independently.

Each ADC has:
- 12-bits successive approximation analog-to-digital converter
- up to 20 single ended or differential multiplexed channels
- different trigger list, or internal channels interconnects, hence
  add a dedicated compatible for each ADC1, ADC2 and ADC3.

The trigger list varies in between the two SoCs. So add a compatible for
each SoC/ADC child.

STM32 ADC may be in a power domain: allow a single 'power-domains'.

Restrict the last channel (19), as it is wired directly to ADC Vref+.
It's meaningless to have direct measurement. Internal Vref with
calibration data from bsec (nvmem) must be used instead e.g. vrefint.

Restrict deprecated properties (st,adc-channels, st,adc-diff-channels,
st,min-sample-time-nsecs).

Co-developed-by: Cheick Traore <cheick.traore-ext@st.com>
Signed-off-by: Cheick Traore <cheick.traore-ext@st.com>
Co-developed-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- Drop st,adc-trigger-sel: as discussed with Jonathan, better to
  describe each ADC child with a dedicated compatible. It is robust
  approach to distinguish variations in trigger list and internal
  channels interconnects.
- Drop unused fallback compatible.
---
 .../devicetree/bindings/iio/adc/st,stm32-adc.yaml  | 71 ++++++++++++++++++++++
 1 file changed, 71 insertions(+)

diff --git a/Documentation/devicetree/bindings/iio/adc/st,stm32-adc.yaml b/Documentation/devicetree/bindings/iio/adc/st,stm32-adc.yaml
index c4c4575d3fa9..e7e92fd4959a 100644
--- a/Documentation/devicetree/bindings/iio/adc/st,stm32-adc.yaml
+++ b/Documentation/devicetree/bindings/iio/adc/st,stm32-adc.yaml
@@ -28,6 +28,8 @@ properties:
       - st,stm32h7-adc-core
       - st,stm32mp1-adc-core
       - st,stm32mp13-adc-core
+      - st,stm32mp23-adc-core
+      - st,stm32mp25-adc-core
 
   reg:
     maxItems: 1
@@ -78,6 +80,9 @@ properties:
       Phandle to the vdd input voltage. It can be used to supply ADC analog
       input switches on stm32mp1.
 
+  power-domains:
+    maxItems: 1
+
   st,syscfg:
     description:
       Phandle to system configuration controller. It can be used to control the
@@ -215,6 +220,33 @@ allOf:
           maximum: 75000000
           default: 75000000
 
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - st,stm32mp23-adc-core
+              - st,stm32mp25-adc-core
+
+    then:
+      properties:
+        clocks:
+          maxItems: 1
+
+        clock-names:
+          const: adc
+
+        interrupts:
+          items:
+            - description: interrupt line for ADC1 or ADC3
+            - description: interrupt line for ADC2
+          minItems: 1
+
+        st,max-clk-rate-hz:
+          minimum: 700000
+          maximum: 70000000
+          default: 70000000
+
 additionalProperties: false
 
 required:
@@ -244,6 +276,12 @@ patternProperties:
           - st,stm32h7-adc
           - st,stm32mp1-adc
           - st,stm32mp13-adc
+          - st,stm32mp23-adc1
+          - st,stm32mp23-adc2
+          - st,stm32mp23-adc3
+          - st,stm32mp25-adc1
+          - st,stm32mp25-adc2
+          - st,stm32mp25-adc3
 
       reg:
         description: |
@@ -486,6 +524,39 @@ patternProperties:
               maxItems: 19
               items:
                 minimum: 40
+
+      - if:
+          properties:
+            compatible:
+              contains:
+                enum:
+                  - st,stm32mp23-adc1
+                  - st,stm32mp23-adc2
+                  - st,stm32mp23-adc3
+                  - st,stm32mp25-adc1
+                  - st,stm32mp25-adc2
+                  - st,stm32mp25-adc3
+
+        then:
+          properties:
+            reg:
+              enum:
+                - 0x0
+                - 0x100
+
+            interrupts:
+              minimum: 0
+              maximum: 1
+
+            assigned-resolution-bits:
+              enum: [6, 8, 10, 12]
+              default: 12
+
+            channel@19: false
+            st,adc-channels: false
+            st,adc-diff-channels: false
+            st,min-sample-time-nsecs: false
+
     additionalProperties: false
 
     required:

-- 
2.43.0

[PATCH v2 02/14] iio: adc: stm32-adc: core: manage clock prescaler diversity

[Fabrice Gasnier]

Make the clock prescaler selection configurable, to prepare stm32mp2
support.
Reuse existing stm32f4_adc_clk_sel(). Rename it so it's no longer
specific to stm32f4. Add configuration data to handle upcoming
differences on:
- registers and bits
- division ratio
No functional change intended.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- new precursor patch as advised by Andy and Jonathan
- unsigned prescaler array as advised by Andy
- cosmetic: spaces before and after braces
---
 drivers/iio/adc/stm32-adc-core.c | 40 ++++++++++++++++++++++++++--------------
 1 file changed, 26 insertions(+), 14 deletions(-)

diff --git a/drivers/iio/adc/stm32-adc-core.c b/drivers/iio/adc/stm32-adc-core.c
index a42d82d61cb8..14ed50bb3ab5 100644
--- a/drivers/iio/adc/stm32-adc-core.c
+++ b/drivers/iio/adc/stm32-adc-core.c
@@ -50,6 +50,8 @@
  * @ovr_msk:    array of ovr (overrun flag) masks in csr for adc1..n
  * @ier:	interrupt enable register offset for each adc
  * @eocie_msk:	end of conversion interrupt enable mask in @ier
+ * @presc_msk:	prescaler mask in ccr
+ * @presc_shift: prescaler bit shift in ccr
  */
 struct stm32_adc_common_regs {
 	u32 csr;
@@ -58,6 +60,8 @@ struct stm32_adc_common_regs {
 	u32 ovr_msk[STM32_ADC_MAX_ADCS];
 	u32 ier;
 	u32 eocie_msk;
+	u32 presc_msk;
+	u32 presc_shift;
 };
 
 struct stm32_adc_priv;
@@ -66,6 +70,8 @@ struct stm32_adc_priv;
  * struct stm32_adc_priv_cfg - stm32 core compatible configuration data
  * @regs:	common registers for all instances
  * @clk_sel:	clock selection routine
+ * @presc:	clock prescaler array
+ * @num_presc:	number of clock prescalers in presc array
  * @max_clk_rate_hz: maximum analog clock rate (Hz, from datasheet)
  * @ipid:	adc identification number
  * @has_syscfg: SYSCFG capability flags
@@ -75,6 +81,8 @@ struct stm32_adc_priv;
 struct stm32_adc_priv_cfg {
 	const struct stm32_adc_common_regs *regs;
 	int (*clk_sel)(struct platform_device *, struct stm32_adc_priv *);
+	unsigned int *presc;
+	unsigned int num_presc;
 	u32 max_clk_rate_hz;
 	u32 ipid;
 	unsigned int has_syscfg;
@@ -126,22 +134,22 @@ static struct stm32_adc_priv *to_stm32_adc_priv(struct stm32_adc_common *com)
 }
 
 /* STM32F4 ADC internal common clock prescaler division ratios */
-static int stm32f4_pclk_div[] = {2, 4, 6, 8};
+static unsigned int stm32f4_pclk_div[] = { 2, 4, 6, 8 };
 
 /**
- * stm32f4_adc_clk_sel() - Select stm32f4 ADC common clock prescaler
+ * stm32_adc_clk_sel() - Select stm32f4 ADC common clock prescaler
  * @pdev: platform device
  * @priv: stm32 ADC core private data
  * Select clock prescaler used for analog conversions, before using ADC.
  */
-static int stm32f4_adc_clk_sel(struct platform_device *pdev,
-			       struct stm32_adc_priv *priv)
+static int stm32_adc_clk_sel(struct platform_device *pdev,
+			     struct stm32_adc_priv *priv)
 {
 	unsigned long rate;
 	u32 val;
 	int i;
 
-	/* stm32f4 has one clk input for analog (mandatory), enforce it here */
+	/* This ADC has one clk input for analog (mandatory), enforce it here */
 	if (!priv->aclk) {
 		dev_err(&pdev->dev, "No 'adc' clock found\n");
 		return -ENOENT;
@@ -153,20 +161,20 @@ static int stm32f4_adc_clk_sel(struct platform_device *pdev,
 		return -EINVAL;
 	}
 
-	for (i = 0; i < ARRAY_SIZE(stm32f4_pclk_div); i++) {
-		if ((rate / stm32f4_pclk_div[i]) <= priv->max_clk_rate)
+	for (i = 0; i < priv->cfg->num_presc; i++) {
+		if ((rate / priv->cfg->presc[i]) <= priv->max_clk_rate)
 			break;
 	}
-	if (i >= ARRAY_SIZE(stm32f4_pclk_div)) {
+	if (i >= priv->cfg->num_presc) {
 		dev_err(&pdev->dev, "adc clk selection failed\n");
 		return -EINVAL;
 	}
 
-	priv->common.rate = rate / stm32f4_pclk_div[i];
-	val = readl_relaxed(priv->common.base + STM32F4_ADC_CCR);
-	val &= ~STM32F4_ADC_ADCPRE_MASK;
-	val |= i << STM32F4_ADC_ADCPRE_SHIFT;
-	writel_relaxed(val, priv->common.base + STM32F4_ADC_CCR);
+	priv->common.rate = rate / priv->cfg->presc[i];
+	val = readl_relaxed(priv->common.base + priv->cfg->regs->ccr);
+	val &= ~priv->cfg->regs->presc_msk;
+	val |= i << priv->cfg->regs->presc_shift;
+	writel_relaxed(val, priv->common.base + priv->cfg->regs->ccr);
 
 	dev_dbg(&pdev->dev, "Using analog clock source at %ld kHz\n",
 		priv->common.rate / 1000);
@@ -314,6 +322,8 @@ static const struct stm32_adc_common_regs stm32f4_adc_common_regs = {
 	.ovr_msk = { STM32F4_OVR1, STM32F4_OVR2, STM32F4_OVR3 },
 	.ier = STM32F4_ADC_CR1,
 	.eocie_msk = STM32F4_EOCIE,
+	.presc_msk = STM32F4_ADC_ADCPRE_MASK,
+	.presc_shift = STM32F4_ADC_ADCPRE_SHIFT,
 };
 
 /* STM32H7 common registers definitions */
@@ -850,7 +860,9 @@ static DEFINE_RUNTIME_DEV_PM_OPS(stm32_adc_core_pm_ops,
 
 static const struct stm32_adc_priv_cfg stm32f4_adc_priv_cfg = {
 	.regs = &stm32f4_adc_common_regs,
-	.clk_sel = stm32f4_adc_clk_sel,
+	.clk_sel = stm32_adc_clk_sel,
+	.presc = stm32f4_pclk_div,
+	.num_presc = ARRAY_SIZE(stm32f4_pclk_div),
 	.max_clk_rate_hz = 36000000,
 	.num_irqs = 1,
 	.num_adcs = 3,

-- 
2.43.0

Re: [PATCH v2 02/14] iio: adc: stm32-adc: core: manage clock prescaler diversity

[Jonathan Cameron]

On Wed, 23 Sep 2026 17:39:05 +0200
Fabrice Gasnier <fabrice.gasnier@foss.st.com> wrote:

> Make the clock prescaler selection configurable, to prepare stm32mp2
> support.
> Reuse existing stm32f4_adc_clk_sel(). Rename it so it's no longer
> specific to stm32f4. Add configuration data to handle upcoming
> differences on:
> - registers and bits
> - division ratio
> No functional change intended.
> 
> Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Hi Fabrice,

One small suggestion to use field_prep() rather than having
to carry the shift around.
>  
> -	priv->common.rate = rate / stm32f4_pclk_div[i];
> -	val = readl_relaxed(priv->common.base + STM32F4_ADC_CCR);
> -	val &= ~STM32F4_ADC_ADCPRE_MASK;
> -	val |= i << STM32F4_ADC_ADCPRE_SHIFT;
> -	writel_relaxed(val, priv->common.base + STM32F4_ADC_CCR);
> +	priv->common.rate = rate / priv->cfg->presc[i];
> +	val = readl_relaxed(priv->common.base + priv->cfg->regs->ccr);
> +	val &= ~priv->cfg->regs->presc_msk;
> +	val |= i << priv->cfg->regs->presc_shift;

I think this isn't a particularly hot path, so we can use
field_prep()  (the lower case one doesn't mind non const masks)

The we don't need to store the shift as well as the mask.

Maybe it is worth thinking about adding a field_modify() similar
to FIELD_MODIFY(). Perhaps that's not a thing to do in this patch set
though.

> +	writel_relaxed(val, priv->common.base + priv->cfg->regs->ccr);
>  
>  	dev_dbg(&pdev->dev, "Using analog clock source at %ld kHz\n",
>  		priv->common.rate / 1000);

> 

[PATCH v2 03/14] iio: adc: stm32-adc: core: configurable number of interrupts

[Fabrice Gasnier]

Make the number of interrupts configurable to prepare stm32mp2
support. Read from the SoC compatible data, and add it to
private data struct.
No functional change intended.

Co-developed-by: Cheick Traore <cheick.traore-ext@st.com>
Signed-off-by: Cheick Traore <cheick.traore-ext@st.com>
Co-developed-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- new precursor patch as advised by Andy and Jonathan to render
  num_irqs configurable
---
 drivers/iio/adc/stm32-adc-core.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/iio/adc/stm32-adc-core.c b/drivers/iio/adc/stm32-adc-core.c
index 14ed50bb3ab5..5845173e288c 100644
--- a/drivers/iio/adc/stm32-adc-core.c
+++ b/drivers/iio/adc/stm32-adc-core.c
@@ -94,6 +94,7 @@ struct stm32_adc_priv_cfg {
  * struct stm32_adc_priv - stm32 ADC core private data
  * @irq:		irq(s) for ADC block
  * @nb_adc_max:		actual maximum number of instance per ADC block
+ * @nb_irqs:		number of IRQs in the ADC block
  * @domain:		irq domain reference
  * @aclk:		clock reference for the analog circuitry
  * @bclk:		bus clock common for all ADCs, depends on part used
@@ -112,6 +113,7 @@ struct stm32_adc_priv_cfg {
 struct stm32_adc_priv {
 	int				irq[STM32_ADC_MAX_ADCS];
 	unsigned int			nb_adc_max;
+	unsigned int			nb_irqs;
 	struct irq_domain		*domain;
 	struct clk			*aclk;
 	struct clk			*bclk;
@@ -424,7 +426,7 @@ static int stm32_adc_irq_probe(struct platform_device *pdev,
 	 * - stm32f4/h7 shares a common interrupt line.
 	 * - stm32mp1, has one line per ADC
 	 */
-	for (i = 0; i < priv->cfg->num_irqs; i++) {
+	for (i = 0; i < priv->nb_irqs; i++) {
 		priv->irq[i] = platform_get_irq(pdev, i);
 		if (priv->irq[i] < 0)
 			return priv->irq[i];
@@ -439,7 +441,7 @@ static int stm32_adc_irq_probe(struct platform_device *pdev,
 		return -ENOMEM;
 	}
 
-	for (i = 0; i < priv->cfg->num_irqs; i++)
+	for (i = 0; i < priv->nb_irqs; i++)
 		irq_set_chained_handler_and_data(priv->irq[i],
 						 stm32_adc_irq_handler, priv);
 
@@ -456,7 +458,7 @@ static void stm32_adc_irq_remove(struct platform_device *pdev,
 		irq_dispose_mapping(irq_find_mapping(priv->domain, hwirq));
 	irq_domain_remove(priv->domain);
 
-	for (i = 0; i < priv->cfg->num_irqs; i++)
+	for (i = 0; i < priv->nb_irqs; i++)
 		irq_set_chained_handler(priv->irq[i], NULL);
 }
 
@@ -730,6 +732,7 @@ static int stm32_adc_probe(struct platform_device *pdev)
 
 	priv->cfg = device_get_match_data(dev);
 	priv->nb_adc_max = priv->cfg->num_adcs;
+	priv->nb_irqs = priv->cfg->num_irqs;
 	spin_lock_init(&priv->common.lock);
 
 	priv->common.base = devm_platform_get_and_ioremap_resource(pdev, 0, &res);

-- 
2.43.0

[PATCH v2 04/14] iio: adc: stm32-adc: manage characterization voltage diversity

[Fabrice Gasnier]

From: Olivier Moysan <olivier.moysan@foss.st.com>

The characterization voltage used to calibrate vrefint depends
on SoC family. Add vref_charac_mv field in stm32_adc_cfg structure
to store the characterization voltage for each SoC, and use it
in Vref computation formula.

Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- Andy's review comments: rename variable to vref_charac_mv (unit suffix).
  Drop the const attribute.
---
 drivers/iio/adc/stm32-adc.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index 90f0e257e6d1..ddf74c674da6 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -43,7 +43,6 @@
 #define STM32_ADC_TIMEOUT_US		100000
 #define STM32_ADC_TIMEOUT	(msecs_to_jiffies(STM32_ADC_TIMEOUT_US / 1000))
 #define STM32_ADC_HW_STOP_DELAY_MS	100
-#define STM32_ADC_VREFINT_VOLTAGE	3300
 
 #define STM32_DMA_BUFFER_SIZE		PAGE_SIZE
 
@@ -211,6 +210,7 @@ struct stm32_adc;
  * @set_ovs:		routine to set oversampling configuration
  * @smp_cycles:		programmable sampling time (ADC clock cycles)
  * @ts_int_ch:		pointer to array of internal channels minimum sampling time in ns
+ * @vref_charac_mv:	Vref characterization voltage used to measure vrefint calibration value
  */
 struct stm32_adc_cfg {
 	const struct stm32_adc_regspec	*regs;
@@ -230,6 +230,7 @@ struct stm32_adc_cfg {
 	void (*set_ovs)(struct iio_dev *indio_dev, u32 ovs_idx);
 	const unsigned int *smp_cycles;
 	const unsigned int *ts_int_ch;
+	int vref_charac_mv;
 };
 
 /**
@@ -1615,7 +1616,7 @@ static int stm32_adc_read_raw(struct iio_dev *indio_dev,
 		if (mask == IIO_CHAN_INFO_PROCESSED) {
 			if (*val == 0)
 				return -EINVAL;
-			*val = STM32_ADC_VREFINT_VOLTAGE * adc->vrefint.vrefint_cal / *val;
+			*val = adc->cfg->vref_charac_mv * adc->vrefint.vrefint_cal / *val;
 		}
 
 		return ret;
@@ -2711,6 +2712,7 @@ static const struct stm32_adc_cfg stm32f4_adc_cfg = {
 	.stop_conv = stm32f4_adc_stop_conv,
 	.smp_cycles = stm32f4_adc_smp_cycles,
 	.irq_clear = stm32f4_adc_irq_clear,
+	.vref_charac_mv = 3300,
 };
 
 static const unsigned int stm32_adc_min_ts_h7[] = { 0, 0, 0, 4300, 9000 };
@@ -2732,6 +2734,7 @@ static const struct stm32_adc_cfg stm32h7_adc_cfg = {
 	.irq_clear = stm32h7_adc_irq_clear,
 	.ts_int_ch = stm32_adc_min_ts_h7,
 	.set_ovs = stm32h7_adc_set_ovs,
+	.vref_charac_mv = 3300,
 };
 
 static const unsigned int stm32_adc_min_ts_mp1[] = { 100, 100, 100, 4300, 9800 };
@@ -2754,6 +2757,7 @@ static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
 	.irq_clear = stm32h7_adc_irq_clear,
 	.ts_int_ch = stm32_adc_min_ts_mp1,
 	.set_ovs = stm32h7_adc_set_ovs,
+	.vref_charac_mv = 3300,
 };
 
 static const unsigned int stm32_adc_min_ts_mp13[] = { 100, 0, 0, 4300, 9800 };
@@ -2772,6 +2776,7 @@ static const struct stm32_adc_cfg stm32mp13_adc_cfg = {
 	.irq_clear = stm32h7_adc_irq_clear,
 	.ts_int_ch = stm32_adc_min_ts_mp13,
 	.set_ovs = stm32mp13_adc_set_ovs,
+	.vref_charac_mv = 3300,
 };
 
 static const struct of_device_id stm32_adc_of_match[] = {

-- 
2.43.0

[PATCH v2 05/14] iio: adc: stm32-adc: rework internal channels data

[Fabrice Gasnier]

Diversity in internal channels data array lead to have zeroed data.
To facilitate addition of new internal channels, without touching
all array initializers, only focus on exiting channels on each
supported ADC variant (e.g. each SoC).

This also allows to get rid of static_assert() checks.

No functional change intended.
Cosmetic update: remove comma on STM32_ADC_INT_CH_NB as it should
always be the last element.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- new precursor patch as suggested by Jonathan
---
 drivers/iio/adc/stm32-adc.c | 51 +++++++++++++++++++++++----------------------
 1 file changed, 26 insertions(+), 25 deletions(-)

diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index ddf74c674da6..ac60c31de7a5 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -86,25 +86,16 @@ enum stm32_adc_int_ch {
 	STM32_ADC_INT_CH_VDDQ_DDR,
 	STM32_ADC_INT_CH_VREFINT,
 	STM32_ADC_INT_CH_VBAT,
-	STM32_ADC_INT_CH_NB,
+	STM32_ADC_INT_CH_NB
 };
 
-/**
- * struct stm32_adc_ic - ADC internal channels
- * @name:	name of the internal channel
- * @idx:	internal channel enum index
- */
-struct stm32_adc_ic {
-	const char *name;
-	u32 idx;
-};
-
-static const struct stm32_adc_ic stm32_adc_ic[STM32_ADC_INT_CH_NB] = {
-	{ "vddcore", STM32_ADC_INT_CH_VDDCORE },
-	{ "vddcpu", STM32_ADC_INT_CH_VDDCPU },
-	{ "vddq_ddr", STM32_ADC_INT_CH_VDDQ_DDR },
-	{ "vrefint", STM32_ADC_INT_CH_VREFINT },
-	{ "vbat", STM32_ADC_INT_CH_VBAT },
+/* STM32 ADC internal channel names */
+static const char stm32_adc_ic[STM32_ADC_INT_CH_NB][STM32_ADC_CH_SZ] = {
+	[STM32_ADC_INT_CH_VDDCORE] = "vddcore",
+	[STM32_ADC_INT_CH_VDDCPU] = "vddcpu",
+	[STM32_ADC_INT_CH_VDDQ_DDR] = "vddq_ddr",
+	[STM32_ADC_INT_CH_VREFINT] = "vrefint",
+	[STM32_ADC_INT_CH_VBAT] = "vbat",
 };
 
 /**
@@ -2268,7 +2259,7 @@ static int stm32_adc_populate_int_ch(struct iio_dev *indio_dev, const char *ch_n
 	int i, ret;
 
 	for (i = 0; i < STM32_ADC_INT_CH_NB; i++) {
-		if (!strncmp(stm32_adc_ic[i].name, ch_name, STM32_ADC_CH_SZ)) {
+		if (!strncmp(stm32_adc_ic[i], ch_name, STM32_ADC_CH_SZ)) {
 			bool na;
 
 			/* Check internal channel availability */
@@ -2302,7 +2293,7 @@ static int stm32_adc_populate_int_ch(struct iio_dev *indio_dev, const char *ch_n
 				return 0;
 			}
 
-			if (stm32_adc_ic[i].idx != STM32_ADC_INT_CH_VREFINT) {
+			if (i != STM32_ADC_INT_CH_VREFINT) {
 				adc->int_ch[i] = chan;
 				break;
 			}
@@ -2715,8 +2706,10 @@ static const struct stm32_adc_cfg stm32f4_adc_cfg = {
 	.vref_charac_mv = 3300,
 };
 
-static const unsigned int stm32_adc_min_ts_h7[] = { 0, 0, 0, 4300, 9000 };
-static_assert(ARRAY_SIZE(stm32_adc_min_ts_h7) == STM32_ADC_INT_CH_NB);
+static const unsigned int stm32_adc_min_ts_h7[STM32_ADC_INT_CH_NB] = {
+	[STM32_ADC_INT_CH_VREFINT] = 4300,
+	[STM32_ADC_INT_CH_VBAT] = 9000
+};
 
 static const struct stm32_adc_cfg stm32h7_adc_cfg = {
 	.regs = &stm32h7_adc_regspec,
@@ -2737,8 +2730,11 @@ static const struct stm32_adc_cfg stm32h7_adc_cfg = {
 	.vref_charac_mv = 3300,
 };
 
-static const unsigned int stm32_adc_min_ts_mp1[] = { 100, 100, 100, 4300, 9800 };
-static_assert(ARRAY_SIZE(stm32_adc_min_ts_mp1) == STM32_ADC_INT_CH_NB);
+static const unsigned int stm32_adc_min_ts_mp1[STM32_ADC_INT_CH_NB] = {
+	[STM32_ADC_INT_CH_VDDCORE] = 100,
+	[STM32_ADC_INT_CH_VREFINT] = 4300,
+	[STM32_ADC_INT_CH_VBAT] = 9800
+};
 
 static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
 	.regs = &stm32mp1_adc_regspec,
@@ -2760,8 +2756,13 @@ static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
 	.vref_charac_mv = 3300,
 };
 
-static const unsigned int stm32_adc_min_ts_mp13[] = { 100, 0, 0, 4300, 9800 };
-static_assert(ARRAY_SIZE(stm32_adc_min_ts_mp13) == STM32_ADC_INT_CH_NB);
+static const unsigned int stm32_adc_min_ts_mp13[STM32_ADC_INT_CH_NB] = {
+	[STM32_ADC_INT_CH_VDDCORE] = 1000,
+	[STM32_ADC_INT_CH_VDDCPU] = 1000,
+	[STM32_ADC_INT_CH_VDDQ_DDR] = 1000,
+	[STM32_ADC_INT_CH_VREFINT] = 4300,
+	[STM32_ADC_INT_CH_VBAT] = 9800
+};
 
 static const struct stm32_adc_cfg stm32mp13_adc_cfg = {
 	.regs = &stm32mp13_adc_regspec,

-- 
2.43.0

Re: [PATCH v2 05/14] iio: adc: stm32-adc: rework internal channels data

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] iio: adc: unintended functional change in stm32 internal channel delays
--

commit 1797bf77d1751b7e9502bd851d9e50359e42d644
Author: Fabrice Gasnier <fabrice.gasnier@foss.st.com>

iio: adc: stm32-adc: rework internal channels data

This commit reworks the internal channel data structures for STM32 ADCs by
using designated initializers mapped to explicit channel enumerations. This
removes zeroed data entries and the need for static assertions.

> diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
> index ddf74c674da6f..ac60c31de7a5b 100644
> --- a/drivers/iio/adc/stm32-adc.c
> +++ b/drivers/iio/adc/stm32-adc.c

[ ... ]

> @@ -2737,8 +2730,11 @@ static const struct stm32_adc_cfg stm32h7_adc_cfg = {
>  	.vref_charac_mv = 3300,
>  };
>  
> -static const unsigned int stm32_adc_min_ts_mp1[] = { 100, 100, 100, 4300, 9800 };
> -static_assert(ARRAY_SIZE(stm32_adc_min_ts_mp1) == STM32_ADC_INT_CH_NB);
> +static const unsigned int stm32_adc_min_ts_mp1[STM32_ADC_INT_CH_NB] = {
> +	[STM32_ADC_INT_CH_VDDCORE] = 100,
> +	[STM32_ADC_INT_CH_VREFINT] = 4300,
> +	[STM32_ADC_INT_CH_VBAT] = 9800
> +};

[Severity: Medium]
Does this unintentionally alter the minimum sampling times for the VDDCPU and
VDDQ_DDR channels? The commit message states "No functional change intended",
but the previous array initialized these values to 100, whereas omitting
them from the designated initializer will now implicitly set them to 0.

>  
>  static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
>  	.regs = &stm32mp1_adc_regspec,

[ ... ]

> @@ -2760,8 +2756,13 @@ static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
>  	.vref_charac_mv = 3300,
>  };
>  
> -static const unsigned int stm32_adc_min_ts_mp13[] = { 100, 0, 0, 4300, 9800 };
> -static_assert(ARRAY_SIZE(stm32_adc_min_ts_mp13) == STM32_ADC_INT_CH_NB);
> +static const unsigned int stm32_adc_min_ts_mp13[STM32_ADC_INT_CH_NB] = {
> +	[STM32_ADC_INT_CH_VDDCORE] = 1000,
> +	[STM32_ADC_INT_CH_VDDCPU] = 1000,
> +	[STM32_ADC_INT_CH_VDDQ_DDR] = 1000,
> +	[STM32_ADC_INT_CH_VREFINT] = 4300,
> +	[STM32_ADC_INT_CH_VBAT] = 9800
> +};

[Severity: Medium]
Is the minimum sampling time increase for the VDD channels intended? The
previous array initialized VDDCORE to 100, and VDDCPU and VDDQ_DDR to 0.
The new designated initializer explicitly sets all three to 1000. This
seems to conflict with the note in the commit message that no functional
change is intended.

>  
>  static const struct stm32_adc_cfg stm32mp13_adc_cfg = {
>  	.regs = &stm32mp13_adc_regspec,

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260923-adc-stm32mp25-v1-v2-0-46bc019537c6@foss.st.com?part=5

Re: [PATCH v2 05/14] iio: adc: stm32-adc: rework internal channels data

[Fabrice Gasnier]

On 9/23/26 17:39, Fabrice Gasnier wrote:
> Diversity in internal channels data array lead to have zeroed data.
> To facilitate addition of new internal channels, without touching
> all array initializers, only focus on exiting channels on each
> supported ADC variant (e.g. each SoC).
> 
> This also allows to get rid of static_assert() checks.
> 
> No functional change intended.

[ snip ]

> -static const unsigned int stm32_adc_min_ts_mp1[] = { 100, 100, 100, 4300, 9800 };
> -static_assert(ARRAY_SIZE(stm32_adc_min_ts_mp1) == STM32_ADC_INT_CH_NB);
> +static const unsigned int stm32_adc_min_ts_mp1[STM32_ADC_INT_CH_NB] = {
> +	[STM32_ADC_INT_CH_VDDCORE] = 100,
> +	[STM32_ADC_INT_CH_VREFINT] = 4300,
> +	[STM32_ADC_INT_CH_VBAT] = 9800
> +};

Hi all,

Oops, I just figured out I've mixed changes with a fix here. So, I just
have sent a separate fix for that [1].
It's been noticed by sashiko [2].

I'll wait a bit before sending a V3, to leave time to review other patches.

>  
>  static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
>  	.regs = &stm32mp1_adc_regspec,
> @@ -2760,8 +2756,13 @@ static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
>  	.vref_charac_mv = 3300,
>  };
>  
> -static const unsigned int stm32_adc_min_ts_mp13[] = { 100, 0, 0, 4300, 9800 };
> -static_assert(ARRAY_SIZE(stm32_adc_min_ts_mp13) == STM32_ADC_INT_CH_NB);
> +static const unsigned int stm32_adc_min_ts_mp13[STM32_ADC_INT_CH_NB] = {
> +	[STM32_ADC_INT_CH_VDDCORE] = 1000,
> +	[STM32_ADC_INT_CH_VDDCPU] = 1000,
> +	[STM32_ADC_INT_CH_VDDQ_DDR] = 1000,
> +	[STM32_ADC_INT_CH_VREFINT] = 4300,
> +	[STM32_ADC_INT_CH_VBAT] = 9800
> +};

Same here.

[1]
https://lore.kernel.org/linux-iio/20260924-adc-fix-sampling-v1-1-13814954d8d7@foss.st.com/

[2]
https://sashiko.dev/#/message/20260923155115.20E4E1F00893%40smtp.kernel.org

BR,
Fabrice

>  
>  static const struct stm32_adc_cfg stm32mp13_adc_cfg = {
>  	.regs = &stm32mp13_adc_regspec,
> 

Re: [PATCH v2 05/14] iio: adc: stm32-adc: rework internal channels data

[Andy Shevchenko]

On Wed, Sep 23, 2026 at 05:39:08PM +0200, Fabrice Gasnier wrote:
> Diversity in internal channels data array lead to have zeroed data.
> To facilitate addition of new internal channels, without touching
> all array initializers, only focus on exiting channels on each
> supported ADC variant (e.g. each SoC).
> 
> This also allows to get rid of static_assert() checks.
> 
> No functional change intended.
> Cosmetic update: remove comma on STM32_ADC_INT_CH_NB as it should
> always be the last element.

Yes, but...

> -static const unsigned int stm32_adc_min_ts_h7[] = { 0, 0, 0, 4300, 9000 };
> -static_assert(ARRAY_SIZE(stm32_adc_min_ts_h7) == STM32_ADC_INT_CH_NB);
> +static const unsigned int stm32_adc_min_ts_h7[STM32_ADC_INT_CH_NB] = {
> +	[STM32_ADC_INT_CH_VREFINT] = 4300,
> +	[STM32_ADC_INT_CH_VBAT] = 9000

...you should have it when it's not a terminator entry. Here is the case when
it's not a terminator, neither in other similar cases below.

> +};

-- 
With Best Regards,
Andy Shevchenko

[PATCH v2 06/14] iio: adc: stm32-adc: add vreg enable option to manage diversity

[Fabrice Gasnier]

Add a boolean (has_vregen) in adc configuration structure to prevent
setting the reserved bit (CR_ADVREGEN bit 28) on STM32MP2 SoCs.
This bit is present on earlier stm32h7, stm32mp13 and stm32mp15.
No functional change intended.

Co-developed-by: Cheick Traore <cheick.traore-ext@st.com>
Signed-off-by: Cheick Traore <cheick.traore-ext@st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- new precursor patch, e.g. split as suggested by Andy and Jonathan, to
  prepare the infrastructure for stm32mp23/25 support.
---
 drivers/iio/adc/stm32-adc.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index ac60c31de7a5..eecdc4cce1b4 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -193,6 +193,7 @@ struct stm32_adc;
  * @has_linearcal:	linear calibration support flag
  * @has_presel:		channel preselection support flag
  * @has_oversampling:	oversampling support flag
+ * @has_vregen:		voltage regulator enable/disable flag
  * @prepare:		optional prepare routine (power-up, enable)
  * @start_conv:		routine to start conversions
  * @stop_conv:		routine to stop conversions
@@ -213,6 +214,7 @@ struct stm32_adc_cfg {
 	bool has_linearcal;
 	bool has_presel;
 	bool has_oversampling;
+	bool has_vregen;
 	int (*prepare)(struct iio_dev *);
 	void (*start_conv)(struct iio_dev *, bool dma);
 	void (*stop_conv)(struct iio_dev *);
@@ -961,7 +963,8 @@ static int stm32h7_adc_exit_pwr_down(struct iio_dev *indio_dev)
 
 	/* Exit deep power down, then enable ADC voltage regulator */
 	stm32_adc_clr_bits(adc, STM32H7_ADC_CR, STM32H7_DEEPPWD);
-	stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADVREGEN);
+	if (adc->cfg->has_vregen)
+		stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADVREGEN);
 
 	if (adc->cfg->has_boostmode &&
 	    adc->common->rate > STM32H7_BOOST_CLKRATE)
@@ -2719,6 +2722,7 @@ static const struct stm32_adc_cfg stm32h7_adc_cfg = {
 	.has_linearcal = true,
 	.has_presel = true,
 	.has_oversampling = true,
+	.has_vregen = true,
 	.start_conv = stm32h7_adc_start_conv,
 	.stop_conv = stm32h7_adc_stop_conv,
 	.prepare = stm32h7_adc_prepare,
@@ -2745,6 +2749,7 @@ static const struct stm32_adc_cfg stm32mp1_adc_cfg = {
 	.has_linearcal = true,
 	.has_presel = true,
 	.has_oversampling = true,
+	.has_vregen = true,
 	.start_conv = stm32h7_adc_start_conv,
 	.stop_conv = stm32h7_adc_stop_conv,
 	.prepare = stm32h7_adc_prepare,
@@ -2769,6 +2774,7 @@ static const struct stm32_adc_cfg stm32mp13_adc_cfg = {
 	.adc_info = &stm32mp13_adc_info,
 	.trigs = stm32h7_adc_trigs,
 	.has_oversampling = true,
+	.has_vregen = true,
 	.start_conv = stm32mp13_adc_start_conv,
 	.stop_conv = stm32h7_adc_stop_conv,
 	.prepare = stm32h7_adc_prepare,

-- 
2.43.0

[PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[Fabrice Gasnier]

Add support for ADC on STM32MP25 SoC. It has 3 ADCs, split into two blocks:
- ADC1 & ADC2 are tightly coupled.
- ADC3 is managed independently.

Trigger list slightly changes between these blocks (ADC1 & ADC2). Other
differences are found on channels interconnects (similar between ADC2
and ADC3):
- ADC1 is connected to 18 external channels + 2 internal channels
- ADC2 is connected to 14 external channels + 6 internal channels
- ADC3 is connected to 14 external channels + 6 internal channels

Each ADC is a 12-bits successive approximation analog-to-digital converter,
with up to 20 multiplexed channels that can be configured as single ended
or differential. ADC resolution ranges from 6 to 12 bits.

It introduces diversity regarding IRQs, clocks, software calibration
procedure, internal voltage channels, sampling time (prescaler) and
trigger list. Most of the architecture, and the driver engine remains
similar. So, handle the differences w.r.t. other STM32 ADCs family with
a dedicated compatible and compatible data.

Co-developed-by: Cheick Traore <cheick.traore-ext@st.com>
Signed-off-by: Cheick Traore <cheick.traore-ext@st.com>
Co-developed-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- Split as advised by Jonathan and Andy. Added precursor patches to
  prepare all the infrastructure. Keep only the new hardware support.
- Adopt separate compatible for ADC1, ADC2 and ADC3 to better describe
  internal channels (split regs descriptions) and trigger list. Drop
  "st,adc-trigger-sel".
- Cosmetic updates whilst at it: remove trailing comma
- Refactor calibration routine, to avoid twisted goto chain, as advised
  by Andy
- One patch per new SoC
---
 drivers/iio/adc/stm32-adc-core.c |  43 ++++-
 drivers/iio/adc/stm32-adc-core.h |  20 +++
 drivers/iio/adc/stm32-adc.c      | 372 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 434 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/adc/stm32-adc-core.c b/drivers/iio/adc/stm32-adc-core.c
index 5845173e288c..7027d86f7457 100644
--- a/drivers/iio/adc/stm32-adc-core.c
+++ b/drivers/iio/adc/stm32-adc-core.c
@@ -138,6 +138,9 @@ static struct stm32_adc_priv *to_stm32_adc_priv(struct stm32_adc_common *com)
 /* STM32F4 ADC internal common clock prescaler division ratios */
 static unsigned int stm32f4_pclk_div[] = { 2, 4, 6, 8 };
 
+/* STM32MP25 ADC internal common clock prescaler division ratios */
+static unsigned int stm32mp25_presc_div[] = { 1, 2, 4, 6, 8, 10, 12, 16, 32, 64, 128, 256 };
+
 /**
  * stm32_adc_clk_sel() - Select stm32f4 ADC common clock prescaler
  * @pdev: platform device
@@ -348,6 +351,18 @@ static const struct stm32_adc_common_regs stm32mp13_adc_common_regs = {
 	.eocie_msk = STM32H7_EOCIE,
 };
 
+/* STM32MP25 common registers definitions */
+static const struct stm32_adc_common_regs stm32mp25_adc_common_regs = {
+	.csr = STM32H7_ADC_CSR,
+	.ccr = STM32H7_ADC_CCR,
+	.eoc_msk = { STM32H7_EOC_MST, STM32H7_EOC_SLV},
+	.ovr_msk = { STM32H7_OVR_MST, STM32H7_OVR_SLV},
+	.ier = STM32H7_ADC_IER,
+	.eocie_msk = STM32H7_EOCIE,
+	.presc_msk = STM32H7_PRESC_MASK,
+	.presc_shift = STM32H7_PRESC_SHIFT,
+};
+
 static const unsigned int stm32_adc_offset[STM32_ADC_MAX_ADCS] = {
 	0, STM32_ADC_OFFSET, STM32_ADC_OFFSET * 2,
 };
@@ -425,6 +440,8 @@ static int stm32_adc_irq_probe(struct platform_device *pdev,
 	 * Interrupt(s) must be provided, depending on the compatible:
 	 * - stm32f4/h7 shares a common interrupt line.
 	 * - stm32mp1, has one line per ADC
+	 * - stm32mp25, has a dual ADC1/ADC2 and standalone ADC3. Each ADC has an IRQ
+	 *   (two for ADC1/ADC2, one for ADC3)
 	 */
 	for (i = 0; i < priv->nb_irqs; i++) {
 		priv->irq[i] = platform_get_irq(pdev, i);
@@ -705,6 +722,17 @@ static int stm32_adc_probe_identification(struct platform_device *pdev,
 		return -EINVAL;
 	}
 
+	if (priv->cfg->ipid == STM32MP25_IPIDR_NUMBER) {
+		/*
+		 * ADC diversity on STM32MP25 is handled here. There may be:
+		 * - 2 IRQs for ADC1/ADC2 block
+		 * - 1 IRQ for ADC3 block.
+		 * There cannot be more IRQs than the actual number of ADCs.
+		 */
+		if (priv->nb_irqs > priv->nb_adc_max)
+			priv->nb_irqs = priv->nb_adc_max;
+	}
+
 	val = readl_relaxed(priv->common.base + STM32MP1_ADC_VERR);
 	dev_dbg(&pdev->dev, "ADC version: %lu.%lu\n",
 		FIELD_GET(STM32MP1_MAJREV_MASK, val),
@@ -897,6 +925,16 @@ static const struct stm32_adc_priv_cfg stm32mp13_adc_priv_cfg = {
 	.num_irqs = 1,
 };
 
+static const struct stm32_adc_priv_cfg stm32mp25_adc_priv_cfg = {
+	.regs = &stm32mp25_adc_common_regs,
+	.clk_sel = stm32_adc_clk_sel,
+	.presc = stm32mp25_presc_div,
+	.num_presc = ARRAY_SIZE(stm32mp25_presc_div),
+	.max_clk_rate_hz = 70000000,
+	.ipid = STM32MP25_IPIDR_NUMBER,
+	.num_irqs = 2, /* 2 IRQs for ADC1/ADC2, 1 irq for ADC3 */
+};
+
 static const struct of_device_id stm32_adc_of_match[] = {
 	{
 		.compatible = "st,stm32f4-adc-core",
@@ -911,7 +949,10 @@ static const struct of_device_id stm32_adc_of_match[] = {
 		.compatible = "st,stm32mp13-adc-core",
 		.data = (void *)&stm32mp13_adc_priv_cfg
 	}, {
-	},
+		.compatible = "st,stm32mp25-adc-core",
+		.data = (void *)&stm32mp25_adc_priv_cfg
+	}, {
+	}
 };
 MODULE_DEVICE_TABLE(of, stm32_adc_of_match);
 
diff --git a/drivers/iio/adc/stm32-adc-core.h b/drivers/iio/adc/stm32-adc-core.h
index db50a9f3b922..68b937e3819f 100644
--- a/drivers/iio/adc/stm32-adc-core.h
+++ b/drivers/iio/adc/stm32-adc-core.h
@@ -121,6 +121,9 @@
 #define STM32MP13_ADC_CALFACT		0xB4
 #define STM32MP13_ADC2_OR		0xC8
 
+/* STM32MP25 - ADC2 and ADC3 option register */
+#define STM32MP25_ADC23_OR		0xD0
+
 /* STM32H7 - common registers for all ADC instances */
 #define STM32H7_ADC_CSR			(STM32_ADCX_COMN_OFFSET + 0x00)
 #define STM32H7_ADC_CCR			(STM32_ADCX_COMN_OFFSET + 0x08)
@@ -257,8 +260,25 @@ enum stm32h7_adc_dmngt {
 #define STM32MP13_OP1			BIT(1)
 #define STM32MP13_OP0			BIT(0)
 
+/* STM32MP25 - Registers for each ADC instance */
+#define STM32MP25_ADC_CALFACT		0xC4
+
+/* STM32MP25_ADC_CALFACT - bit fields */
+#define STM32MP25_CALFACT_CALADDOS	BIT(31)
+#define STM32MP25_CALFACT_D_MASK	GENMASK(24, 16)
+#define STM32MP25_CALFACT_S_MASK	GENMASK(8, 0)
+
+/* STM32MP25_ADC_CFGR specific bit fields */
+#define STM32MP25_RES_SHIFT		2
+#define STM32MP25_RES_MASK		GENMASK(3, 2)
+
+/* STM32MP25_ADC23_OR - specific bit fields */
+#define STM32MP25_VDDCPUEN		BIT(3)
+#define STM32MP25_VDDCOREEN		BIT(2)
+
 #define STM32MP15_IPIDR_NUMBER		0x00110005
 #define STM32MP13_IPIDR_NUMBER		0x00110006
+#define STM32MP25_IPIDR_NUMBER		0x00110008
 
 /**
  * struct stm32_adc_common - stm32 ADC driver common data (for all instances)
diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index eecdc4cce1b4..a5bffc09898b 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -33,6 +33,9 @@
 /* Number of linear calibration shadow registers / LINCALRDYW control bits */
 #define STM32H7_LINCALFACT_NUM		6
 
+/* Number of loops in the calibration procedure to average data on STM32MP25 */
+#define STM32MP25_CALIB_LOOP		8
+
 /* BOOST bit must be set on STM32H7 when ADC clock is above 20MHz */
 #define STM32H7_BOOST_CLKRATE		20000000UL
 
@@ -77,6 +80,14 @@ enum stm32_adc_extsel {
 	STM32_EXT18,
 	STM32_EXT19,
 	STM32_EXT20,
+	STM32_EXT21,
+	STM32_EXT22,
+	STM32_EXT23,
+	STM32_EXT24,
+	STM32_EXT25,
+	STM32_EXT26,
+	STM32_EXT27,
+	STM32_EXT28
 };
 
 enum stm32_adc_int_ch {
@@ -348,6 +359,20 @@ static const struct stm32_adc_info stm32mp13_adc_info = {
 	.num_ovs = ARRAY_SIZE(stm32mp13_adc_oversampling_avail),
 };
 
+/*
+ * stm32mp25 can have up to 20 channels.
+ * Keep channel 19 (VREF+) hidden on STM32MP25: it is the ADC reference
+ * supply itself, so exposing it as a regular input is meaningless.
+ * Reference voltage must be derived from VREFINT with calibration data.
+ */
+static const struct stm32_adc_info stm32mp25_adc_info = {
+	.max_channels = STM32_ADC_CH_MAX - 1,
+	.resolutions = stm32f4_adc_resolutions,
+	.oversampling = stm32h7_adc_oversampling_avail,
+	.num_res = ARRAY_SIZE(stm32f4_adc_resolutions),
+	.num_ovs = ARRAY_SIZE(stm32h7_adc_oversampling_avail),
+};
+
 /*
  * stm32f4_sq - describe regular sequence registers
  * - L: sequence len (register & bit field)
@@ -538,6 +563,114 @@ static const struct stm32_adc_regspec stm32h7_adc_regspec = {
 	.smp_bits = stm32h7_smp_bits,
 };
 
+/* STM32MP25 external trigger sources for ADC1 and ADC2 */
+static struct stm32_adc_trig_info stm32mp25_adc1_adc2_trigs[] = {
+	{ TIM1_TRGO, STM32_EXT0 },
+	{ TIM1_TRGO2, STM32_EXT1 },
+	{ TIM8_TRGO, STM32_EXT2 },
+	{ TIM8_TRGO2, STM32_EXT3 },
+	{ TIM20_TRGO, STM32_EXT4 },
+	{ TIM20_TRGO2, STM32_EXT5 },
+	{ TIM2_TRGO, STM32_EXT6 },
+	{ TIM3_TRGO, STM32_EXT7 },
+	{ TIM4_TRGO, STM32_EXT8 },
+	{ TIM5_TRGO, STM32_EXT9 },
+	{ TIM6_TRGO, STM32_EXT10 },
+	{ TIM15_TRGO, STM32_EXT11 },
+	{ TIM1_CH1, STM32_EXT12 },
+	{ TIM1_CH2, STM32_EXT13 },
+	{ TIM1_CH3, STM32_EXT14 },
+	{ TIM20_OC1, STM32_EXT15 },
+	{ TIM20_OC2, STM32_EXT16 },
+	{ TIM20_OC3, STM32_EXT17 },
+	{ TIM2_CH2, STM32_EXT18 },
+	{ TIM3_CH4, STM32_EXT19 },
+	{ TIM4_CH4, STM32_EXT20 },
+	{ TIM5_CH1, STM32_EXT21 },
+	{ TIM12_CH1, STM32_EXT22 },
+	{ LPTIM1_CH1, STM32_EXT24 },
+	{ LPTIM2_CH1, STM32_EXT25 },
+	{ LPTIM3_CH1, STM32_EXT26 },
+	{ LPTIM4_CH1, STM32_EXT27 },
+	{ LPTIM5_OUT, STM32_EXT28 },
+	{ }
+};
+
+/* STM32MP25 external trigger sources for ADC3 */
+static struct stm32_adc_trig_info stm32mp25_adc3_trigs[] = {
+	{ TIM1_TRGO, STM32_EXT0 },
+	{ TIM1_TRGO2, STM32_EXT1 },
+	{ TIM8_TRGO, STM32_EXT2 },
+	{ TIM8_TRGO2, STM32_EXT3 },
+	{ TIM20_TRGO, STM32_EXT4 },
+	{ TIM20_TRGO2, STM32_EXT5 },
+	{ TIM2_TRGO, STM32_EXT6 },
+	{ TIM3_TRGO, STM32_EXT7 },
+	{ TIM4_TRGO, STM32_EXT8 },
+	{ TIM5_TRGO, STM32_EXT9 },
+	{ TIM6_TRGO, STM32_EXT10 },
+	{ TIM7_TRGO, STM32_EXT11 },
+	{ TIM15_TRGO, STM32_EXT12 },
+	{ TIM17_OC1, STM32_EXT13 },
+	{ TIM1_CH3, STM32_EXT14 },
+	{ TIM8_CH1, STM32_EXT15 },
+	{ TIM20_OC1, STM32_EXT16 },
+	{ TIM2_CH1, STM32_EXT17 },
+	{ TIM2_CH3, STM32_EXT18 },
+	{ TIM3_CH1, STM32_EXT19 },
+	{ TIM4_CH1, STM32_EXT20 },
+	{ TIM5_CH3, STM32_EXT21 },
+	{ TIM12_CH1, STM32_EXT22 },
+	{ LPTIM1_CH1, STM32_EXT24 },
+	{ LPTIM2_CH1, STM32_EXT25 },
+	{ LPTIM3_CH1, STM32_EXT26 },
+	{ LPTIM4_CH1, STM32_EXT27 },
+	{ LPTIM5_OUT, STM32_EXT28 },
+	{ }
+};
+
+/* STM32MP25 programmable sampling time (ADC clock cycles, rounded down) */
+static const unsigned int stm32mp25_adc_smp_cycles[STM32_ADC_MAX_SMP + 1] = {
+	2, 3, 7, 12, 24, 47, 247, 1501,
+};
+
+static const struct stm32_adc_regspec stm32mp25_adc1_regspec = {
+	.dr = STM32H7_ADC_DR,
+	.ier_eoc = { STM32H7_ADC_IER, STM32H7_EOCIE },
+	.ier_ovr = { STM32H7_ADC_IER, STM32H7_OVRIE },
+	.isr_eoc = { STM32H7_ADC_ISR, STM32H7_EOC },
+	.isr_ovr = { STM32H7_ADC_ISR, STM32H7_OVR },
+	.sqr = stm32h7_sq,
+	.exten = { STM32H7_ADC_CFGR, STM32H7_EXTEN_MASK, STM32H7_EXTEN_SHIFT },
+	.extsel = { STM32H7_ADC_CFGR, STM32H7_EXTSEL_MASK,
+		    STM32H7_EXTSEL_SHIFT },
+	.res = { STM32H7_ADC_CFGR, STM32MP25_RES_MASK, STM32MP25_RES_SHIFT },
+	.difsel = { STM32H7_ADC_DIFSEL, STM32H7_DIFSEL_MASK},
+	.smpr = { STM32H7_ADC_SMPR1, STM32H7_ADC_SMPR2 },
+	.smp_bits = stm32h7_smp_bits,
+	.ccr_vref = { STM32H7_ADC_CCR, STM32H7_VREFEN },
+};
+
+static const struct stm32_adc_regspec stm32mp25_adc2_adc3_regspec = {
+	.dr = STM32H7_ADC_DR,
+	.ier_eoc = { STM32H7_ADC_IER, STM32H7_EOCIE },
+	.ier_ovr = { STM32H7_ADC_IER, STM32H7_OVRIE },
+	.isr_eoc = { STM32H7_ADC_ISR, STM32H7_EOC },
+	.isr_ovr = { STM32H7_ADC_ISR, STM32H7_OVR },
+	.sqr = stm32h7_sq,
+	.exten = { STM32H7_ADC_CFGR, STM32H7_EXTEN_MASK, STM32H7_EXTEN_SHIFT },
+	.extsel = { STM32H7_ADC_CFGR, STM32H7_EXTSEL_MASK,
+		    STM32H7_EXTSEL_SHIFT },
+	.res = { STM32H7_ADC_CFGR, STM32MP25_RES_MASK, STM32MP25_RES_SHIFT },
+	.difsel = { STM32H7_ADC_DIFSEL, STM32H7_DIFSEL_MASK},
+	.smpr = { STM32H7_ADC_SMPR1, STM32H7_ADC_SMPR2 },
+	.smp_bits = stm32h7_smp_bits,
+	.or_vddcore = { STM32MP25_ADC23_OR, STM32MP25_VDDCOREEN },
+	.or_vddcpu = { STM32MP25_ADC23_OR, STM32MP25_VDDCPUEN },
+	.ccr_vbat = { STM32H7_ADC_CCR, STM32H7_VBATEN },
+	.ccr_vref = { STM32H7_ADC_CCR, STM32H7_VREFEN },
+};
+
 /* STM32MP13 programmable sampling time (ADC clock cycles, rounded down) */
 static const unsigned int stm32mp13_adc_smp_cycles[STM32_ADC_MAX_SMP + 1] = {
 	2, 6, 12, 24, 47, 92, 247, 640,
@@ -1304,6 +1437,184 @@ static void stm32h7_adc_unprepare(struct iio_dev *indio_dev)
 	stm32h7_adc_enter_pwr_down(adc);
 }
 
+/*
+ * STM32MP25 offset calibration software procedure. Basically the calibration routine is
+ * expected to average (for example) 8 samples in calibration mode, for single-ended  and
+ * differential channels, to calibrate the zero offset. In case offset is "negative", an
+ * additional offset can be added, to determine calibration factor. It must be kept later
+ * for all conversions.
+ */
+static int stm32mp25_adc_calib_get_average_data(struct iio_dev *indio_dev, u32 *average)
+{
+	struct stm32_adc *adc = iio_priv(indio_dev);
+	const struct stm32_adc_regspec *regs = adc->cfg->regs;
+	u32 val, avg = 0;
+	int ret;
+
+	/* Repeat several conversions in calibration mode, average the results */
+	for (unsigned int i = 0; i < STM32MP25_CALIB_LOOP; i++) {
+		stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADSTART);
+		ret = stm32_adc_readl_poll_timeout(STM32H7_ADC_CR, val,
+						   !(val & (STM32H7_ADSTART)),
+						   100, STM32_ADC_TIMEOUT_US);
+		if (ret) {
+			dev_err(&indio_dev->dev, "calibration average error %d\n", ret);
+			return ret;
+		}
+
+		val = stm32_adc_readl(adc, regs->dr);
+		dev_vdbg(&indio_dev->dev, "dr[%d]=0x%08x\n", i, val);
+		avg += val;
+	}
+
+	*average = DIV_ROUND_CLOSEST(avg, STM32MP25_CALIB_LOOP);
+	dev_vdbg(&indio_dev->dev, "average=0x%08x\n", *average);
+
+	return 0;
+}
+
+static int stm32mp25_adc_calib_single_ended_offset(struct iio_dev *indio_dev, bool *add_offset)
+{
+	struct stm32_adc *adc = iio_priv(indio_dev);
+	u32 average;
+	int ret;
+
+	/* Select single ended input calibration */
+	stm32_adc_clr_bits(adc, STM32H7_ADC_CR, STM32H7_ADCALDIF);
+
+retry:
+	/* Clears or set CALADDOS (also clear old calibration data if any) */
+	stm32_adc_writel(adc, STM32MP25_ADC_CALFACT,
+			 FIELD_PREP(STM32MP25_CALFACT_CALADDOS, *add_offset));
+
+	ret = stm32mp25_adc_calib_get_average_data(indio_dev, &average);
+	if (ret)
+		return ret;
+
+	/* Add offset and retry single-ended calibration if the averaged data is zero */
+	if (!average && !*add_offset) {
+		*add_offset = true;
+		goto retry;
+	}
+
+	if (!average) {
+		/* If average data is still zero with additional offset, just warn about it */
+		dev_warn(&indio_dev->dev, "Single-ended calibration average: 0\n");
+	} else {
+		u32 calfact = stm32_adc_readl(adc, STM32MP25_ADC_CALFACT);
+
+		calfact |= FIELD_PREP(STM32MP25_CALFACT_S_MASK, average);
+		stm32_adc_writel(adc, STM32MP25_ADC_CALFACT, calfact);
+	}
+
+	return 0;
+}
+
+static int stm32mp25_adc_calib(struct iio_dev *indio_dev)
+{
+	struct stm32_adc *adc = iio_priv(indio_dev);
+	bool add_offset = false;
+	bool diff_below_zero;
+	u32 average, calfact;
+	int ret;
+
+	stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADCAL);
+	/* Use default resolution (e.g. 12 bits) */
+	stm32_adc_clr_bits(adc, STM32H7_ADC_CFGR, STM32MP25_RES_MASK);
+
+retry:
+	/* Single ended input calibration */
+	ret = stm32mp25_adc_calib_single_ended_offset(indio_dev, &add_offset);
+	if (ret)
+		goto out;
+
+	/* Differential input calibration (keep previous CALADDOS value) */
+	stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADCALDIF);
+	ret = stm32mp25_adc_calib_get_average_data(indio_dev, &average);
+	if (ret)
+		goto out;
+
+	/* Averaged diff data is below 0x800 (half value in 12-bits mode) */
+	diff_below_zero = average < BIT(adc->cfg->adc_info->resolutions[0] - 1);
+
+	if (diff_below_zero && !add_offset) {
+		/* Retry the whole calibration with additional offset */
+		add_offset = true;
+		goto retry;
+	}
+
+	calfact = stm32_adc_readl(adc, STM32MP25_ADC_CALFACT);
+
+	if (diff_below_zero) {
+		/*
+		 * Averaged data is still below center value. It needs to be clamped to zero,
+		 * so don't use the result here, warn about it.
+		 */
+		dev_warn(&indio_dev->dev, "Differential calibration clamped(0): 0x%x\n", average);
+	} else {
+		calfact |= FIELD_PREP(STM32MP25_CALFACT_D_MASK, average);
+		stm32_adc_writel(adc, STM32MP25_ADC_CALFACT, calfact);
+	}
+
+	dev_dbg(&indio_dev->dev, "set calfact_s=0x%03lx, calfact_d=0x%03lx, calados=%ld\n",
+		FIELD_GET(STM32MP25_CALFACT_S_MASK, calfact),
+		FIELD_GET(STM32MP25_CALFACT_D_MASK, calfact),
+		FIELD_GET(STM32MP25_CALFACT_CALADDOS, calfact));
+out:
+	stm32_adc_clr_bits(adc, STM32H7_ADC_CR, STM32H7_ADCAL);
+	stm32_adc_set_res(adc);
+
+	return ret;
+}
+
+static int stm32mp25_adc_prepare(struct iio_dev *indio_dev)
+{
+	struct stm32_adc *adc = iio_priv(indio_dev);
+	int ret;
+
+	ret = stm32h7_adc_exit_pwr_down(indio_dev);
+	if (ret)
+		return ret;
+
+	/* Must enable the ADC before running software-assisted calibration */
+	ret = stm32h7_adc_enable(indio_dev);
+	if (ret)
+		goto pwr_dwn;
+
+	/* Always run offset calibration */
+	ret = stm32mp25_adc_calib(indio_dev);
+	if (ret)
+		goto adc_dis;
+
+	stm32_adc_int_ch_enable(indio_dev);
+
+	stm32_adc_writel(adc, adc->cfg->regs->difsel.reg, adc->difsel);
+
+	if (adc->cfg->has_presel)
+		stm32_adc_writel(adc, STM32H7_ADC_PCSEL, adc->pcsel);
+
+	return 0;
+
+adc_dis:
+	stm32h7_adc_disable(indio_dev);
+pwr_dwn:
+	stm32h7_adc_enter_pwr_down(adc);
+
+	return ret;
+}
+
+static void stm32mp25_adc_unprepare(struct iio_dev *indio_dev)
+{
+	struct stm32_adc *adc = iio_priv(indio_dev);
+
+	/* Undo things in the reverse order */
+	if (adc->cfg->has_presel)
+		stm32_adc_writel(adc, STM32H7_ADC_PCSEL, 0);
+	stm32_adc_int_ch_disable(adc);
+	stm32h7_adc_disable(indio_dev);
+	stm32h7_adc_enter_pwr_down(adc);
+}
+
 /**
  * stm32_adc_conf_scan_seq() - Build regular channels scan sequence
  * @indio_dev: IIO device
@@ -2786,11 +3097,72 @@ static const struct stm32_adc_cfg stm32mp13_adc_cfg = {
 	.vref_charac_mv = 3300,
 };
 
+static const unsigned int stm32_adc_min_ts_mp25[STM32_ADC_INT_CH_NB] = {
+	[STM32_ADC_INT_CH_VDDCORE] = 34,
+	[STM32_ADC_INT_CH_VDDCPU] = 34,
+	[STM32_ADC_INT_CH_VREFINT] = 34,
+	[STM32_ADC_INT_CH_VBAT] = 34,
+};
+
+static const struct stm32_adc_cfg stm32mp25_adc1_cfg = {
+	.regs = &stm32mp25_adc1_regspec,
+	.adc_info = &stm32mp25_adc_info,
+	.trigs = stm32mp25_adc1_adc2_trigs,
+	.has_oversampling = true,
+	.has_presel = true,
+	.start_conv = stm32h7_adc_start_conv,
+	.stop_conv = stm32h7_adc_stop_conv,
+	.prepare = stm32mp25_adc_prepare,
+	.unprepare = stm32mp25_adc_unprepare,
+	.smp_cycles = stm32mp25_adc_smp_cycles,
+	.irq_clear = stm32h7_adc_irq_clear,
+	.set_ovs = stm32h7_adc_set_ovs,
+	.ts_int_ch = stm32_adc_min_ts_mp25,
+	.vref_charac_mv = 1800,
+};
+
+static const struct stm32_adc_cfg stm32mp25_adc2_cfg = {
+	.regs = &stm32mp25_adc2_adc3_regspec,
+	.adc_info = &stm32mp25_adc_info,
+	.trigs = stm32mp25_adc1_adc2_trigs,
+	.has_oversampling = true,
+	.has_presel = true,
+	.start_conv = stm32h7_adc_start_conv,
+	.stop_conv = stm32h7_adc_stop_conv,
+	.prepare = stm32mp25_adc_prepare,
+	.unprepare = stm32mp25_adc_unprepare,
+	.smp_cycles = stm32mp25_adc_smp_cycles,
+	.irq_clear = stm32h7_adc_irq_clear,
+	.set_ovs = stm32h7_adc_set_ovs,
+	.ts_int_ch = stm32_adc_min_ts_mp25,
+	.vref_charac_mv = 1800,
+};
+
+static const struct stm32_adc_cfg stm32mp25_adc3_cfg = {
+	.regs = &stm32mp25_adc2_adc3_regspec,
+	.adc_info = &stm32mp25_adc_info,
+	.trigs = stm32mp25_adc3_trigs,
+	.has_oversampling = true,
+	.has_presel = true,
+	.start_conv = stm32h7_adc_start_conv,
+	.stop_conv = stm32h7_adc_stop_conv,
+	.prepare = stm32mp25_adc_prepare,
+	.unprepare = stm32mp25_adc_unprepare,
+	.smp_cycles = stm32mp25_adc_smp_cycles,
+	.irq_clear = stm32h7_adc_irq_clear,
+	.set_ovs = stm32h7_adc_set_ovs,
+	.ts_int_ch = stm32_adc_min_ts_mp25,
+	.vref_charac_mv = 1800,
+};
+
 static const struct of_device_id stm32_adc_of_match[] = {
 	{ .compatible = "st,stm32f4-adc", .data = (void *)&stm32f4_adc_cfg },
 	{ .compatible = "st,stm32h7-adc", .data = (void *)&stm32h7_adc_cfg },
 	{ .compatible = "st,stm32mp1-adc", .data = (void *)&stm32mp1_adc_cfg },
 	{ .compatible = "st,stm32mp13-adc", .data = (void *)&stm32mp13_adc_cfg },
+	{ .compatible = "st,stm32mp25-adc1", .data = (void *)&stm32mp25_adc1_cfg },
+	{ .compatible = "st,stm32mp25-adc2", .data = (void *)&stm32mp25_adc2_cfg },
+	{ .compatible = "st,stm32mp25-adc3", .data = (void *)&stm32mp25_adc3_cfg },
 	{ }
 };
 MODULE_DEVICE_TABLE(of, stm32_adc_of_match);

-- 
2.43.0

Re: [PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] iio: adc: stm32-adc: DIFSEL and CFGR writes fail silently when ADEN=1
--

commit 100a3f1abe7bc0887c43d25f1c79c051b323c2a8
Author: Fabrice Gasnier <fabrice.gasnier@foss.st.com>

iio: adc: stm32-adc: add support for stm32mp25

This commit adds support for the ADC on the STM32MP25 SoC. It introduces new
hardware configurations, trigger lists, and a dedicated software calibration
procedure for the new SoC.

> diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
> index eecdc4cce1b4b..a5bffc09898ba 100644
> --- a/drivers/iio/adc/stm32-adc.c
> +++ b/drivers/iio/adc/stm32-adc.c

[ ... ]

> @@ -1304,6 +1437,184 @@ static void stm32h7_adc_unprepare(struct iio_dev *indio_dev)
>  	stm32h7_adc_enter_pwr_down(adc);
>  }
>  
> +/*
> + * STM32MP25 offset calibration software procedure. Basically the calibration routine is

[ ... ]

> +static int stm32mp25_adc_calib(struct iio_dev *indio_dev)
> +{
> +	struct stm32_adc *adc = iio_priv(indio_dev);
> +	bool add_offset = false;
> +	bool diff_below_zero;
> +	u32 average, calfact;
> +	int ret;
> +
> +	stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADCAL);
> +	/* Use default resolution (e.g. 12 bits) */
> +	stm32_adc_clr_bits(adc, STM32H7_ADC_CFGR, STM32MP25_RES_MASK);

[Severity: High]
Does this clear the resolution bits correctly?

Since stm32h7_adc_enable() was already called in stm32mp25_adc_prepare(), ADEN
is set to 1. On STM32 ADCs, configuration registers like CFGR are locked by
hardware when the ADC is enabled, so this write might be silently ignored.

This could cause calibration to run at the wrong resolution if a non-default
resolution was configured by the user.

[ ... ]

> +static int stm32mp25_adc_prepare(struct iio_dev *indio_dev)
> +{
> +	struct stm32_adc *adc = iio_priv(indio_dev);
> +	int ret;
> +
> +	ret = stm32h7_adc_exit_pwr_down(indio_dev);
> +	if (ret)
> +		return ret;
> +
> +	/* Must enable the ADC before running software-assisted calibration */
> +	ret = stm32h7_adc_enable(indio_dev);
> +	if (ret)
> +		goto pwr_dwn;
> +
> +	/* Always run offset calibration */
> +	ret = stm32mp25_adc_calib(indio_dev);
> +	if (ret)
> +		goto adc_dis;
> +
> +	stm32_adc_int_ch_enable(indio_dev);
> +
> +	stm32_adc_writel(adc, adc->cfg->regs->difsel.reg, adc->difsel);

[Severity: High]
Will this configuration of differential channels be silently ignored?

The ADC is already enabled (ADEN=1) from the earlier call to
stm32h7_adc_enable(). Hardware locking prevents writes to DIFSEL while the
ADC is enabled, which would leave the channels configured as single-ended.

> +
> +	if (adc->cfg->has_presel)
> +		stm32_adc_writel(adc, STM32H7_ADC_PCSEL, adc->pcsel);
> +
> +	return 0;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260923-adc-stm32mp25-v1-v2-0-46bc019537c6@foss.st.com?part=7

Re: [PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[Andy Shevchenko]

On Wed, Sep 23, 2026 at 05:39:10PM +0200, Fabrice Gasnier wrote:
> Add support for ADC on STM32MP25 SoC. It has 3 ADCs, split into two blocks:
> - ADC1 & ADC2 are tightly coupled.
> - ADC3 is managed independently.
> 
> Trigger list slightly changes between these blocks (ADC1 & ADC2). Other
> differences are found on channels interconnects (similar between ADC2
> and ADC3):
> - ADC1 is connected to 18 external channels + 2 internal channels
> - ADC2 is connected to 14 external channels + 6 internal channels
> - ADC3 is connected to 14 external channels + 6 internal channels
> 
> Each ADC is a 12-bits successive approximation analog-to-digital converter,
> with up to 20 multiplexed channels that can be configured as single ended
> or differential. ADC resolution ranges from 6 to 12 bits.
> 
> It introduces diversity regarding IRQs, clocks, software calibration
> procedure, internal voltage channels, sampling time (prescaler) and
> trigger list. Most of the architecture, and the driver engine remains
> similar. So, handle the differences w.r.t. other STM32 ADCs family with
> a dedicated compatible and compatible data.

...

>  		.compatible = "st,stm32mp13-adc-core",
>  		.data = (void *)&stm32mp13_adc_priv_cfg
>  	}, {
> -	},
> +		.compatible = "st,stm32mp25-adc-core",
> +		.data = (void *)&stm32mp25_adc_priv_cfg
> +	}, {
> +	}

Same issue and now it's a regression from maintenance perspective: you added an
unnedeed churn that has to be handled from now on... TL;DR: do add trailing
commas to the non-terminator entries and remove trailing commas in the
terminator entries.

>  };

...

> +	STM32_EXT23,
> +	STM32_EXT24,
> +	STM32_EXT25,
> +	STM32_EXT26,
> +	STM32_EXT27,
> +	STM32_EXT28

Same issue and so on...

>  };

...

Are you doing patches with an assistance? LLMs might have a problem with the
style issues.

...

> +retry:
> +	/* Clears or set CALADDOS (also clear old calibration data if any) */
> +	stm32_adc_writel(adc, STM32MP25_ADC_CALFACT,
> +			 FIELD_PREP(STM32MP25_CALFACT_CALADDOS, *add_offset));
> +
> +	ret = stm32mp25_adc_calib_get_average_data(indio_dev, &average);
> +	if (ret)
> +		return ret;
> +
> +	/* Add offset and retry single-ended calibration if the averaged data is zero */
> +	if (!average && !*add_offset) {
> +		*add_offset = true;
> +		goto retry;
> +	}

Refactor to avoid a label. It's possible to achieve.

> +	if (!average) {

Why not positive conditional?

> +		/* If average data is still zero with additional offset, just warn about it */
> +		dev_warn(&indio_dev->dev, "Single-ended calibration average: 0\n");
> +	} else {
> +		u32 calfact = stm32_adc_readl(adc, STM32MP25_ADC_CALFACT);
> +
> +		calfact |= FIELD_PREP(STM32MP25_CALFACT_S_MASK, average);
> +		stm32_adc_writel(adc, STM32MP25_ADC_CALFACT, calfact);
> +	}

...

> +static int stm32mp25_adc_calib(struct iio_dev *indio_dev)

> +	struct stm32_adc *adc = iio_priv(indio_dev);
> +	bool add_offset = false;
> +	bool diff_below_zero;
> +	u32 average, calfact;
> +	int ret;
> +
> +	stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADCAL);
> +	/* Use default resolution (e.g. 12 bits) */
> +	stm32_adc_clr_bits(adc, STM32H7_ADC_CFGR, STM32MP25_RES_MASK);
> +
> +retry:
> +	/* Single ended input calibration */
> +	ret = stm32mp25_adc_calib_single_ended_offset(indio_dev, &add_offset);
> +	if (ret)
> +		goto out;
> +
> +	/* Differential input calibration (keep previous CALADDOS value) */
> +	stm32_adc_set_bits(adc, STM32H7_ADC_CR, STM32H7_ADCALDIF);
> +	ret = stm32mp25_adc_calib_get_average_data(indio_dev, &average);
> +	if (ret)
> +		goto out;
> +
> +	/* Averaged diff data is below 0x800 (half value in 12-bits mode) */
> +	diff_below_zero = average < BIT(adc->cfg->adc_info->resolutions[0] - 1);
> +
> +	if (diff_below_zero && !add_offset) {
> +		/* Retry the whole calibration with additional offset */
> +		add_offset = true;
> +		goto retry;
> +	}

Same comment, refactor to avoid label.

> +	calfact = stm32_adc_readl(adc, STM32MP25_ADC_CALFACT);
> +
> +	if (diff_below_zero) {
> +		/*
> +		 * Averaged data is still below center value. It needs to be clamped to zero,
> +		 * so don't use the result here, warn about it.
> +		 */
> +		dev_warn(&indio_dev->dev, "Differential calibration clamped(0): 0x%x\n", average);
> +	} else {
> +		calfact |= FIELD_PREP(STM32MP25_CALFACT_D_MASK, average);
> +		stm32_adc_writel(adc, STM32MP25_ADC_CALFACT, calfact);
> +	}
> +
> +	dev_dbg(&indio_dev->dev, "set calfact_s=0x%03lx, calfact_d=0x%03lx, calados=%ld\n",
> +		FIELD_GET(STM32MP25_CALFACT_S_MASK, calfact),
> +		FIELD_GET(STM32MP25_CALFACT_D_MASK, calfact),
> +		FIELD_GET(STM32MP25_CALFACT_CALADDOS, calfact));
> +out:

In any case if you ever have a label in the code, name it as an answer to the Q:
"What will be done if I goto *this* label?"


Here it is something like 'out_calibration_stop_and_reset'
(I haven't checked the real code and datasheet, just used below short context).

> +	stm32_adc_clr_bits(adc, STM32H7_ADC_CR, STM32H7_ADCAL);
> +	stm32_adc_set_res(adc);
> +
> +	return ret;

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[Fabrice Gasnier]

On 9/24/26 23:01, Andy Shevchenko wrote:
> On Wed, Sep 23, 2026 at 05:39:10PM +0200, Fabrice Gasnier wrote:
>> Add support for ADC on STM32MP25 SoC. It has 3 ADCs, split into two blocks:
>> - ADC1 & ADC2 are tightly coupled.
>> - ADC3 is managed independently.
>>
>> Trigger list slightly changes between these blocks (ADC1 & ADC2). Other
>> differences are found on channels interconnects (similar between ADC2
>> and ADC3):
>> - ADC1 is connected to 18 external channels + 2 internal channels
>> - ADC2 is connected to 14 external channels + 6 internal channels
>> - ADC3 is connected to 14 external channels + 6 internal channels
>>
>> Each ADC is a 12-bits successive approximation analog-to-digital converter,
>> with up to 20 multiplexed channels that can be configured as single ended
>> or differential. ADC resolution ranges from 6 to 12 bits.
>>
>> It introduces diversity regarding IRQs, clocks, software calibration
>> procedure, internal voltage channels, sampling time (prescaler) and
>> trigger list. Most of the architecture, and the driver engine remains
>> similar. So, handle the differences w.r.t. other STM32 ADCs family with
>> a dedicated compatible and compatible data.
> 
> ...
> 
>>  		.compatible = "st,stm32mp13-adc-core",
>>  		.data = (void *)&stm32mp13_adc_priv_cfg
>>  	}, {
>> -	},
>> +		.compatible = "st,stm32mp25-adc-core",
>> +		.data = (void *)&stm32mp25_adc_priv_cfg
>> +	}, {
>> +	}
> 
> Same issue and now it's a regression from maintenance perspective: you added an
> unnedeed churn that has to be handled from now on... TL;DR: do add trailing
> commas to the non-terminator entries and remove trailing commas in the
> terminator entries.

Hi Andy,

Do you wish I add a separate cleanup patch for existing match table ?
This can also fit into 100 chars line now:

static const struct of_device_id stm32_adc_of_match[] = {
	{ .compatible = "st,stm32f4-adc-core", .data = (void
*)&stm32f4_adc_priv_cfg, },
	{ .compatible = "st,stm32h7-adc-core", .data = (void
*)&stm32h7_adc_priv_cfg, },
	{ .compatible = "st,stm32mp1-adc-core",	.data = (void
*)&stm32mp1_adc_priv_cfg, },
	{ .compatible = "st,stm32mp13-adc-core", .data = (void
*)&stm32mp13_adc_priv_cfg, },
	{ }
};

> 
>>  };
> 
> ...
> 
>> +	STM32_EXT23,
>> +	STM32_EXT24,
>> +	STM32_EXT25,
>> +	STM32_EXT26,
>> +	STM32_EXT27,
>> +	STM32_EXT28
> 
> Same issue and so on...

Ack,

> 
>>  };
> 
> ...
> 
> Are you doing patches with an assistance? LLMs might have a problem with the
> style issues.

My apologies, that was just me between the keyboard and the chair.

> 
> ...
> 
>> +retry:
>> +	/* Clears or set CALADDOS (also clear old calibration data if any) */
>> +	stm32_adc_writel(adc, STM32MP25_ADC_CALFACT,
>> +			 FIELD_PREP(STM32MP25_CALFACT_CALADDOS, *add_offset));
>> +
>> +	ret = stm32mp25_adc_calib_get_average_data(indio_dev, &average);
>> +	if (ret)
>> +		return ret;
>> +
>> +	/* Add offset and retry single-ended calibration if the averaged data is zero */
>> +	if (!average && !*add_offset) {
>> +		*add_offset = true;
>> +		goto retry;
>> +	}
> 
> Refactor to avoid a label. It's possible to achieve.

Ok, I'll refactor in v3.

Thanks for reviewing,
BR,
Fabrice

Re: [PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[Andy Shevchenko]

On Tue, Sep 29, 2026 at 06:48:28PM +0200, Fabrice Gasnier wrote:
> On 9/24/26 23:01, Andy Shevchenko wrote:
> > On Wed, Sep 23, 2026 at 05:39:10PM +0200, Fabrice Gasnier wrote:

...

> >>  		.compatible = "st,stm32mp13-adc-core",
> >>  		.data = (void *)&stm32mp13_adc_priv_cfg
> >>  	}, {
> >> -	},
> >> +		.compatible = "st,stm32mp25-adc-core",
> >> +		.data = (void *)&stm32mp25_adc_priv_cfg
> >> +	}, {
> >> +	}
> > 
> > Same issue and now it's a regression from maintenance perspective: you added an
> > unnedeed churn that has to be handled from now on... TL;DR: do add trailing
> > commas to the non-terminator entries and remove trailing commas in the
> > terminator entries.

> Do you wish I add a separate cleanup patch for existing match table ?
> This can also fit into 100 chars line now:

This is, of course, is up to you, but my comment was solely against the
trailing comma removal.

> static const struct of_device_id stm32_adc_of_match[] = {
> 	{ .compatible = "st,stm32f4-adc-core", .data = (void
> *)&stm32f4_adc_priv_cfg, },

Here is actually the opposite, no inner trailing comma should be added as it's
a one-line entry.

> 	{ .compatible = "st,stm32h7-adc-core", .data = (void
> *)&stm32h7_adc_priv_cfg, },
> 	{ .compatible = "st,stm32mp1-adc-core",	.data = (void
> *)&stm32mp1_adc_priv_cfg, },
> 	{ .compatible = "st,stm32mp13-adc-core", .data = (void
> *)&stm32mp13_adc_priv_cfg, },
> 	{ }
> };

> >>  };

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[Fabrice Gasnier]

On 9/30/26 10:33, Andy Shevchenko wrote:
> On Tue, Sep 29, 2026 at 06:48:28PM +0200, Fabrice Gasnier wrote:
>> On 9/24/26 23:01, Andy Shevchenko wrote:
>>> On Wed, Sep 23, 2026 at 05:39:10PM +0200, Fabrice Gasnier wrote:
> 
> ...
> 
>>>>  		.compatible = "st,stm32mp13-adc-core",
>>>>  		.data = (void *)&stm32mp13_adc_priv_cfg
>>>>  	}, {
>>>> -	},
>>>> +		.compatible = "st,stm32mp25-adc-core",
>>>> +		.data = (void *)&stm32mp25_adc_priv_cfg
>>>> +	}, {
>>>> +	}
>>>
>>> Same issue and now it's a regression from maintenance perspective: you added an
>>> unnedeed churn that has to be handled from now on... TL;DR: do add trailing
>>> commas to the non-terminator entries and remove trailing commas in the
>>> terminator entries.
> 
>> Do you wish I add a separate cleanup patch for existing match table ?
>> This can also fit into 100 chars line now:
> 
> This is, of course, is up to you, but my comment was solely against the
> trailing comma removal.
> 
>> static const struct of_device_id stm32_adc_of_match[] = {
>> 	{ .compatible = "st,stm32f4-adc-core", .data = (void
>> *)&stm32f4_adc_priv_cfg, },
> 
> Here is actually the opposite, no inner trailing comma should be added as it's
> a one-line entry.

Hi Andy,

Sorry if I'm a bit confused :-), just to be sure, the preferred style should be as below ?

static const struct of_device_id stm32_adc_of_match[] = {
	{ .compatible = "st,stm32f4-adc-core", .data = (void *)&stm32f4_adc_priv_cfg },
	{ .compatible = "st,stm32h7-adc-core", .data = (void *)&stm32h7_adc_priv_cfg },
	{ .compatible = "st,stm32mp1-adc-core",	.data = (void *)&stm32mp1_adc_priv_cfg },
	{ .compatible = "st,stm32mp13-adc-core", .data = (void *)&stm32mp13_adc_priv_cfg },
	{ }
};

Thanks,
Best Regards,
Fabrice
> 
>> 	{ .compatible = "st,stm32h7-adc-core", .data = (void
>> *)&stm32h7_adc_priv_cfg, },
>> 	{ .compatible = "st,stm32mp1-adc-core",	.data = (void
>> *)&stm32mp1_adc_priv_cfg, },
>> 	{ .compatible = "st,stm32mp13-adc-core", .data = (void
>> *)&stm32mp13_adc_priv_cfg, },
>> 	{ }
>> };
> 
>>>>  };
> 

Re: [PATCH v2 07/14] iio: adc: stm32-adc: add support for stm32mp25

[Andy Shevchenko]

On Wed, Sep 30, 2026 at 10:44:12AM +0200, Fabrice Gasnier wrote:
> On 9/30/26 10:33, Andy Shevchenko wrote:

...

> Sorry if I'm a bit confused :-), just to be sure, the preferred style should be as below ?

Yes, in case if you want to update the whole ID table.

> static const struct of_device_id stm32_adc_of_match[] = {
> 	{ .compatible = "st,stm32f4-adc-core", .data = (void *)&stm32f4_adc_priv_cfg },
> 	{ .compatible = "st,stm32h7-adc-core", .data = (void *)&stm32h7_adc_priv_cfg },
> 	{ .compatible = "st,stm32mp1-adc-core",	.data = (void *)&stm32mp1_adc_priv_cfg },
> 	{ .compatible = "st,stm32mp13-adc-core", .data = (void *)&stm32mp13_adc_priv_cfg },
> 	{ }
> };

The rationale for the trailing comma is:
1) no need (doesn't make sense) to have it on a terminator entry;
2) do add it for non-terminator entry to avoid a churn in the future;
3) do not add it for any case when the whole array is located on a single line.

-- 
With Best Regards,
Andy Shevchenko

[PATCH v2 08/14] iio: adc: stm32-adc: add support for stm32mp23

[Fabrice Gasnier]

Add support for ADC on STM32MP23 SoC. It has 3 ADCs, split into two blocks:
- ADC1 & ADC2 are tightly coupled.
- ADC3 is managed independently.

It is quite similar to STM32MP25 ADC, except for the trigger list. TIM20
triggers have been removed. So, handle the differences with a dedicated
compatible and compatible data.

Co-developed-by: Cheick Traore <cheick.traore-ext@st.com>
Signed-off-by: Cheick Traore <cheick.traore-ext@st.com>
Co-developed-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- New patch, split as advised by Andy: one patch for this new SoC
---
 drivers/iio/adc/stm32-adc-core.c |   3 ++
 drivers/iio/adc/stm32-adc.c      | 112 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 115 insertions(+)

diff --git a/drivers/iio/adc/stm32-adc-core.c b/drivers/iio/adc/stm32-adc-core.c
index 7027d86f7457..418846f736ba 100644
--- a/drivers/iio/adc/stm32-adc-core.c
+++ b/drivers/iio/adc/stm32-adc-core.c
@@ -948,6 +948,9 @@ static const struct of_device_id stm32_adc_of_match[] = {
 	}, {
 		.compatible = "st,stm32mp13-adc-core",
 		.data = (void *)&stm32mp13_adc_priv_cfg
+	}, {
+		.compatible = "st,stm32mp23-adc-core",
+		.data = (void *)&stm32mp25_adc_priv_cfg
 	}, {
 		.compatible = "st,stm32mp25-adc-core",
 		.data = (void *)&stm32mp25_adc_priv_cfg
diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index a5bffc09898b..a14f858299ae 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -563,6 +563,64 @@ static const struct stm32_adc_regspec stm32h7_adc_regspec = {
 	.smp_bits = stm32h7_smp_bits,
 };
 
+/* STM32MP23 external trigger sources for ADC1 and ADC2 */
+static struct stm32_adc_trig_info stm32mp23_adc1_adc2_trigs[] = {
+	{ TIM1_TRGO, STM32_EXT0 },
+	{ TIM1_TRGO2, STM32_EXT1 },
+	{ TIM8_TRGO, STM32_EXT2 },
+	{ TIM8_TRGO2, STM32_EXT3 },
+	{ TIM2_TRGO, STM32_EXT6 },
+	{ TIM3_TRGO, STM32_EXT7 },
+	{ TIM4_TRGO, STM32_EXT8 },
+	{ TIM5_TRGO, STM32_EXT9 },
+	{ TIM6_TRGO, STM32_EXT10 },
+	{ TIM15_TRGO, STM32_EXT11 },
+	{ TIM1_CH1, STM32_EXT12 },
+	{ TIM1_CH2, STM32_EXT13 },
+	{ TIM1_CH3, STM32_EXT14 },
+	{ TIM2_CH2, STM32_EXT18 },
+	{ TIM3_CH4, STM32_EXT19 },
+	{ TIM4_CH4, STM32_EXT20 },
+	{ TIM5_CH1, STM32_EXT21 },
+	{ TIM12_CH1, STM32_EXT22 },
+	{ LPTIM1_CH1, STM32_EXT24 },
+	{ LPTIM2_CH1, STM32_EXT25 },
+	{ LPTIM3_CH1, STM32_EXT26 },
+	{ LPTIM4_CH1, STM32_EXT27 },
+	{ LPTIM5_OUT, STM32_EXT28 },
+	{ }
+};
+
+/* STM32MP23 external trigger sources for ADC3 */
+static struct stm32_adc_trig_info stm32mp23_adc3_trigs[] = {
+	{ TIM1_TRGO, STM32_EXT0 },
+	{ TIM1_TRGO2, STM32_EXT1 },
+	{ TIM8_TRGO, STM32_EXT2 },
+	{ TIM8_TRGO2, STM32_EXT3 },
+	{ TIM2_TRGO, STM32_EXT6 },
+	{ TIM3_TRGO, STM32_EXT7 },
+	{ TIM4_TRGO, STM32_EXT8 },
+	{ TIM5_TRGO, STM32_EXT9 },
+	{ TIM6_TRGO, STM32_EXT10 },
+	{ TIM7_TRGO, STM32_EXT11 },
+	{ TIM15_TRGO, STM32_EXT12 },
+	{ TIM17_OC1, STM32_EXT13 },
+	{ TIM1_CH3, STM32_EXT14 },
+	{ TIM8_CH1, STM32_EXT15 },
+	{ TIM2_CH1, STM32_EXT17 },
+	{ TIM2_CH3, STM32_EXT18 },
+	{ TIM3_CH1, STM32_EXT19 },
+	{ TIM4_CH1, STM32_EXT20 },
+	{ TIM5_CH3, STM32_EXT21 },
+	{ TIM12_CH1, STM32_EXT22 },
+	{ LPTIM1_CH1, STM32_EXT24 },
+	{ LPTIM2_CH1, STM32_EXT25 },
+	{ LPTIM3_CH1, STM32_EXT26 },
+	{ LPTIM4_CH1, STM32_EXT27 },
+	{ LPTIM5_OUT, STM32_EXT28 },
+	{ }
+};
+
 /* STM32MP25 external trigger sources for ADC1 and ADC2 */
 static struct stm32_adc_trig_info stm32mp25_adc1_adc2_trigs[] = {
 	{ TIM1_TRGO, STM32_EXT0 },
@@ -3104,6 +3162,57 @@ static const unsigned int stm32_adc_min_ts_mp25[STM32_ADC_INT_CH_NB] = {
 	[STM32_ADC_INT_CH_VBAT] = 34,
 };
 
+static const struct stm32_adc_cfg stm32mp23_adc1_cfg = {
+	.regs = &stm32mp25_adc1_regspec,
+	.adc_info = &stm32mp25_adc_info,
+	.trigs = stm32mp23_adc1_adc2_trigs,
+	.has_oversampling = true,
+	.has_presel = true,
+	.start_conv = stm32h7_adc_start_conv,
+	.stop_conv = stm32h7_adc_stop_conv,
+	.prepare = stm32mp25_adc_prepare,
+	.unprepare = stm32mp25_adc_unprepare,
+	.smp_cycles = stm32mp25_adc_smp_cycles,
+	.irq_clear = stm32h7_adc_irq_clear,
+	.set_ovs = stm32h7_adc_set_ovs,
+	.ts_int_ch = stm32_adc_min_ts_mp25,
+	.vref_charac_mv = 1800,
+};
+
+static const struct stm32_adc_cfg stm32mp23_adc2_cfg = {
+	.regs = &stm32mp25_adc2_adc3_regspec,
+	.adc_info = &stm32mp25_adc_info,
+	.trigs = stm32mp23_adc1_adc2_trigs,
+	.has_oversampling = true,
+	.has_presel = true,
+	.start_conv = stm32h7_adc_start_conv,
+	.stop_conv = stm32h7_adc_stop_conv,
+	.prepare = stm32mp25_adc_prepare,
+	.unprepare = stm32mp25_adc_unprepare,
+	.smp_cycles = stm32mp25_adc_smp_cycles,
+	.irq_clear = stm32h7_adc_irq_clear,
+	.set_ovs = stm32h7_adc_set_ovs,
+	.ts_int_ch = stm32_adc_min_ts_mp25,
+	.vref_charac_mv = 1800,
+};
+
+static const struct stm32_adc_cfg stm32mp23_adc3_cfg = {
+	.regs = &stm32mp25_adc2_adc3_regspec,
+	.adc_info = &stm32mp25_adc_info,
+	.trigs = stm32mp23_adc3_trigs,
+	.has_oversampling = true,
+	.has_presel = true,
+	.start_conv = stm32h7_adc_start_conv,
+	.stop_conv = stm32h7_adc_stop_conv,
+	.prepare = stm32mp25_adc_prepare,
+	.unprepare = stm32mp25_adc_unprepare,
+	.smp_cycles = stm32mp25_adc_smp_cycles,
+	.irq_clear = stm32h7_adc_irq_clear,
+	.set_ovs = stm32h7_adc_set_ovs,
+	.ts_int_ch = stm32_adc_min_ts_mp25,
+	.vref_charac_mv = 1800,
+};
+
 static const struct stm32_adc_cfg stm32mp25_adc1_cfg = {
 	.regs = &stm32mp25_adc1_regspec,
 	.adc_info = &stm32mp25_adc_info,
@@ -3160,6 +3269,9 @@ static const struct of_device_id stm32_adc_of_match[] = {
 	{ .compatible = "st,stm32h7-adc", .data = (void *)&stm32h7_adc_cfg },
 	{ .compatible = "st,stm32mp1-adc", .data = (void *)&stm32mp1_adc_cfg },
 	{ .compatible = "st,stm32mp13-adc", .data = (void *)&stm32mp13_adc_cfg },
+	{ .compatible = "st,stm32mp23-adc1", .data = (void *)&stm32mp23_adc1_cfg },
+	{ .compatible = "st,stm32mp23-adc2", .data = (void *)&stm32mp23_adc2_cfg },
+	{ .compatible = "st,stm32mp23-adc3", .data = (void *)&stm32mp23_adc3_cfg },
 	{ .compatible = "st,stm32mp25-adc1", .data = (void *)&stm32mp25_adc1_cfg },
 	{ .compatible = "st,stm32mp25-adc2", .data = (void *)&stm32mp25_adc2_cfg },
 	{ .compatible = "st,stm32mp25-adc3", .data = (void *)&stm32mp25_adc3_cfg },

-- 
2.43.0

[PATCH v2 09/14] iio: adc: stm32: add support for vddgpu on stm32mp23 and stm32mp25

[Fabrice Gasnier]

Add support for the vddgpu internal channel available on stm32mp23 and
stm32mp25. It is enabled by using ADC2 and ADC3 option register. Add
VDDGPU to the list of supported internal channels. Add sampling time
information from the datasheet.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- New patch split from V1, to avoid mixing new SoC / feature.
---
 drivers/iio/adc/stm32-adc-core.h |  1 +
 drivers/iio/adc/stm32-adc.c      | 18 ++++++++++++++++++
 2 files changed, 19 insertions(+)

diff --git a/drivers/iio/adc/stm32-adc-core.h b/drivers/iio/adc/stm32-adc-core.h
index 68b937e3819f..330c6c30e4f5 100644
--- a/drivers/iio/adc/stm32-adc-core.h
+++ b/drivers/iio/adc/stm32-adc-core.h
@@ -273,6 +273,7 @@ enum stm32h7_adc_dmngt {
 #define STM32MP25_RES_MASK		GENMASK(3, 2)
 
 /* STM32MP25_ADC23_OR - specific bit fields */
+#define STM32MP25_VDDGPUEN		BIT(4)
 #define STM32MP25_VDDCPUEN		BIT(3)
 #define STM32MP25_VDDCOREEN		BIT(2)
 
diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
index a14f858299ae..a2ed3abffa58 100644
--- a/drivers/iio/adc/stm32-adc.c
+++ b/drivers/iio/adc/stm32-adc.c
@@ -97,6 +97,7 @@ enum stm32_adc_int_ch {
 	STM32_ADC_INT_CH_VDDQ_DDR,
 	STM32_ADC_INT_CH_VREFINT,
 	STM32_ADC_INT_CH_VBAT,
+	STM32_ADC_INT_CH_VDDGPU,
 	STM32_ADC_INT_CH_NB
 };
 
@@ -107,6 +108,7 @@ static const char stm32_adc_ic[STM32_ADC_INT_CH_NB][STM32_ADC_CH_SZ] = {
 	[STM32_ADC_INT_CH_VDDQ_DDR] = "vddq_ddr",
 	[STM32_ADC_INT_CH_VREFINT] = "vrefint",
 	[STM32_ADC_INT_CH_VBAT] = "vbat",
+	[STM32_ADC_INT_CH_VDDGPU] = "vddgpu"
 };
 
 /**
@@ -167,6 +169,7 @@ struct stm32_adc_vrefint {
  * @smp_bits:		smpr1 & smpr2 index and bitfields
  * @or_vddcore:		option register & vddcore bitfield
  * @or_vddcpu:		option register & vddcpu bitfield
+ * @or_vddgpu:		option register & vddgpu bitfield
  * @or_vddq_ddr:	option register & vddq_ddr bitfield
  * @ccr_vbat:		common register & vbat bitfield
  * @ccr_vref:		common register & vrefint bitfield
@@ -186,6 +189,7 @@ struct stm32_adc_regspec {
 	const struct stm32_adc_regs *smp_bits;
 	const struct stm32_adc_regs or_vddcore;
 	const struct stm32_adc_regs or_vddcpu;
+	const struct stm32_adc_regs or_vddgpu;
 	const struct stm32_adc_regs or_vddq_ddr;
 	const struct stm32_adc_regs ccr_vbat;
 	const struct stm32_adc_regs ccr_vref;
@@ -725,6 +729,7 @@ static const struct stm32_adc_regspec stm32mp25_adc2_adc3_regspec = {
 	.smp_bits = stm32h7_smp_bits,
 	.or_vddcore = { STM32MP25_ADC23_OR, STM32MP25_VDDCOREEN },
 	.or_vddcpu = { STM32MP25_ADC23_OR, STM32MP25_VDDCPUEN },
+	.or_vddgpu = { STM32MP25_ADC23_OR, STM32MP25_VDDGPUEN },
 	.ccr_vbat = { STM32H7_ADC_CCR, STM32H7_VBATEN },
 	.ccr_vref = { STM32H7_ADC_CCR, STM32H7_VREFEN },
 };
@@ -953,6 +958,11 @@ static void stm32_adc_int_ch_enable(struct iio_dev *indio_dev)
 			stm32_adc_set_bits_common(adc, adc->cfg->regs->ccr_vbat.reg,
 						  adc->cfg->regs->ccr_vbat.mask);
 			break;
+		case STM32_ADC_INT_CH_VDDGPU:
+			dev_dbg(&indio_dev->dev, "Enable VDDGPU\n");
+			stm32_adc_set_bits(adc, adc->cfg->regs->or_vddgpu.reg,
+					   adc->cfg->regs->or_vddgpu.mask);
+			break;
 		}
 	}
 }
@@ -986,6 +996,10 @@ static void stm32_adc_int_ch_disable(struct stm32_adc *adc)
 			stm32_adc_clr_bits_common(adc, adc->cfg->regs->ccr_vbat.reg,
 						  adc->cfg->regs->ccr_vbat.mask);
 			break;
+		case STM32_ADC_INT_CH_VDDGPU:
+			stm32_adc_clr_bits(adc, adc->cfg->regs->or_vddgpu.reg,
+					   adc->cfg->regs->or_vddgpu.mask);
+			break;
 		}
 	}
 }
@@ -2651,6 +2665,9 @@ static int stm32_adc_populate_int_ch(struct iio_dev *indio_dev, const char *ch_n
 			case STM32_ADC_INT_CH_VBAT:
 				na = !adc->cfg->regs->ccr_vbat.reg;
 				break;
+			case STM32_ADC_INT_CH_VDDGPU:
+				na = !adc->cfg->regs->or_vddgpu.reg;
+				break;
 			default:
 				return -EINVAL;
 			}
@@ -3160,6 +3177,7 @@ static const unsigned int stm32_adc_min_ts_mp25[STM32_ADC_INT_CH_NB] = {
 	[STM32_ADC_INT_CH_VDDCPU] = 34,
 	[STM32_ADC_INT_CH_VREFINT] = 34,
 	[STM32_ADC_INT_CH_VBAT] = 34,
+	[STM32_ADC_INT_CH_VDDGPU] = 34
 };
 
 static const struct stm32_adc_cfg stm32mp23_adc1_cfg = {

-- 
2.43.0

Re: [PATCH v2 09/14] iio: adc: stm32: add support for vddgpu on stm32mp23 and stm32mp25

[Jonathan Cameron]

On Wed, 23 Sep 2026 17:39:12 +0200
Fabrice Gasnier <fabrice.gasnier@foss.st.com> wrote:

> Add support for the vddgpu internal channel available on stm32mp23 and
> stm32mp25. It is enabled by using ADC2 and ADC3 option register. Add
> VDDGPU to the list of supported internal channels. Add sampling time
> information from the datasheet.
> 
> Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
I looked at everything before this point and only had comments
to add to Andy's existing review for patch 2.

You will need to rebase on the fix for v3.  That will probably
confuse Sashiko, so if you end up including that patch as
the first one in this series, add a clear note under the ---
that it is already applied.

Thanks,

Jonathan

> ---
> Changes in v2:
> - New patch split from V1, to avoid mixing new SoC / feature.
> ---
>  drivers/iio/adc/stm32-adc-core.h |  1 +
>  drivers/iio/adc/stm32-adc.c      | 18 ++++++++++++++++++
>  2 files changed, 19 insertions(+)
> 
> diff --git a/drivers/iio/adc/stm32-adc-core.h b/drivers/iio/adc/stm32-adc-core.h
> index 68b937e3819f..330c6c30e4f5 100644
> --- a/drivers/iio/adc/stm32-adc-core.h
> +++ b/drivers/iio/adc/stm32-adc-core.h
> @@ -273,6 +273,7 @@ enum stm32h7_adc_dmngt {
>  #define STM32MP25_RES_MASK		GENMASK(3, 2)
>  
>  /* STM32MP25_ADC23_OR - specific bit fields */
> +#define STM32MP25_VDDGPUEN		BIT(4)
>  #define STM32MP25_VDDCPUEN		BIT(3)
>  #define STM32MP25_VDDCOREEN		BIT(2)
>  
> diff --git a/drivers/iio/adc/stm32-adc.c b/drivers/iio/adc/stm32-adc.c
> index a14f858299ae..a2ed3abffa58 100644
> --- a/drivers/iio/adc/stm32-adc.c
> +++ b/drivers/iio/adc/stm32-adc.c
> @@ -97,6 +97,7 @@ enum stm32_adc_int_ch {
>  	STM32_ADC_INT_CH_VDDQ_DDR,
>  	STM32_ADC_INT_CH_VREFINT,
>  	STM32_ADC_INT_CH_VBAT,
> +	STM32_ADC_INT_CH_VDDGPU,
>  	STM32_ADC_INT_CH_NB
>  };
>  
> @@ -107,6 +108,7 @@ static const char stm32_adc_ic[STM32_ADC_INT_CH_NB][STM32_ADC_CH_SZ] = {
>  	[STM32_ADC_INT_CH_VDDQ_DDR] = "vddq_ddr",
>  	[STM32_ADC_INT_CH_VREFINT] = "vrefint",
>  	[STM32_ADC_INT_CH_VBAT] = "vbat",
> +	[STM32_ADC_INT_CH_VDDGPU] = "vddgpu"
>  };
>  
>  /**
> @@ -167,6 +169,7 @@ struct stm32_adc_vrefint {
>   * @smp_bits:		smpr1 & smpr2 index and bitfields
>   * @or_vddcore:		option register & vddcore bitfield
>   * @or_vddcpu:		option register & vddcpu bitfield
> + * @or_vddgpu:		option register & vddgpu bitfield
>   * @or_vddq_ddr:	option register & vddq_ddr bitfield
>   * @ccr_vbat:		common register & vbat bitfield
>   * @ccr_vref:		common register & vrefint bitfield
> @@ -186,6 +189,7 @@ struct stm32_adc_regspec {
>  	const struct stm32_adc_regs *smp_bits;
>  	const struct stm32_adc_regs or_vddcore;
>  	const struct stm32_adc_regs or_vddcpu;
> +	const struct stm32_adc_regs or_vddgpu;
>  	const struct stm32_adc_regs or_vddq_ddr;
>  	const struct stm32_adc_regs ccr_vbat;
>  	const struct stm32_adc_regs ccr_vref;
> @@ -725,6 +729,7 @@ static const struct stm32_adc_regspec stm32mp25_adc2_adc3_regspec = {
>  	.smp_bits = stm32h7_smp_bits,
>  	.or_vddcore = { STM32MP25_ADC23_OR, STM32MP25_VDDCOREEN },
>  	.or_vddcpu = { STM32MP25_ADC23_OR, STM32MP25_VDDCPUEN },
> +	.or_vddgpu = { STM32MP25_ADC23_OR, STM32MP25_VDDGPUEN },
>  	.ccr_vbat = { STM32H7_ADC_CCR, STM32H7_VBATEN },
>  	.ccr_vref = { STM32H7_ADC_CCR, STM32H7_VREFEN },
>  };
> @@ -953,6 +958,11 @@ static void stm32_adc_int_ch_enable(struct iio_dev *indio_dev)
>  			stm32_adc_set_bits_common(adc, adc->cfg->regs->ccr_vbat.reg,
>  						  adc->cfg->regs->ccr_vbat.mask);
>  			break;
> +		case STM32_ADC_INT_CH_VDDGPU:
> +			dev_dbg(&indio_dev->dev, "Enable VDDGPU\n");
> +			stm32_adc_set_bits(adc, adc->cfg->regs->or_vddgpu.reg,
> +					   adc->cfg->regs->or_vddgpu.mask);
> +			break;
>  		}
>  	}
>  }
> @@ -986,6 +996,10 @@ static void stm32_adc_int_ch_disable(struct stm32_adc *adc)
>  			stm32_adc_clr_bits_common(adc, adc->cfg->regs->ccr_vbat.reg,
>  						  adc->cfg->regs->ccr_vbat.mask);
>  			break;
> +		case STM32_ADC_INT_CH_VDDGPU:
> +			stm32_adc_clr_bits(adc, adc->cfg->regs->or_vddgpu.reg,
> +					   adc->cfg->regs->or_vddgpu.mask);
> +			break;
>  		}
>  	}
>  }
> @@ -2651,6 +2665,9 @@ static int stm32_adc_populate_int_ch(struct iio_dev *indio_dev, const char *ch_n
>  			case STM32_ADC_INT_CH_VBAT:
>  				na = !adc->cfg->regs->ccr_vbat.reg;
>  				break;
> +			case STM32_ADC_INT_CH_VDDGPU:
> +				na = !adc->cfg->regs->or_vddgpu.reg;
> +				break;
>  			default:
>  				return -EINVAL;
>  			}
> @@ -3160,6 +3177,7 @@ static const unsigned int stm32_adc_min_ts_mp25[STM32_ADC_INT_CH_NB] = {
>  	[STM32_ADC_INT_CH_VDDCPU] = 34,
>  	[STM32_ADC_INT_CH_VREFINT] = 34,
>  	[STM32_ADC_INT_CH_VBAT] = 34,
> +	[STM32_ADC_INT_CH_VDDGPU] = 34
>  };
>  
>  static const struct stm32_adc_cfg stm32mp23_adc1_cfg = {
> 

[PATCH v2 10/14] arm64: dts: st: add vrefint calibration on stm32mp23

[Fabrice Gasnier]

From: Patrice Chotard <patrice.chotard@foss.st.com>

Describe vrefint calibration cell to be retrieved by ADCs through bsec,
on STM32MP23.

Co-developed-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
 arch/arm64/boot/dts/st/stm32mp231.dtsi | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm64/boot/dts/st/stm32mp231.dtsi b/arch/arm64/boot/dts/st/stm32mp231.dtsi
index 9e1d240888ff..938d6f0098ac 100644
--- a/arch/arm64/boot/dts/st/stm32mp231.dtsi
+++ b/arch/arm64/boot/dts/st/stm32mp231.dtsi
@@ -790,6 +790,10 @@ part-number-otp@24 {
 				reg = <0x24 0x4>;
 			};
 
+			vrefint: vrefin-cal@1b8 {
+				reg = <0x1b8 0x2>;
+			};
+
 			package-otp@1e8 {
 				reg = <0x1e8 0x1>;
 				bits = <0 3>;

-- 
2.43.0

[PATCH v2 11/14] arm64: dts: st: add vrefint calibration on stm32mp25

[Fabrice Gasnier]

From: Patrice Chotard <patrice.chotard@foss.st.com>

Describe vrefint calibration cell to be retrieved by ADCs through bsec,
on STM32MP25.

Co-developed-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Olivier Moysan <olivier.moysan@foss.st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
 arch/arm64/boot/dts/st/stm32mp251.dtsi | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm64/boot/dts/st/stm32mp251.dtsi b/arch/arm64/boot/dts/st/stm32mp251.dtsi
index 9c63fdb5a885..82a8a6f1f881 100644
--- a/arch/arm64/boot/dts/st/stm32mp251.dtsi
+++ b/arch/arm64/boot/dts/st/stm32mp251.dtsi
@@ -1777,6 +1777,10 @@ part-number-otp@24 {
 				reg = <0x24 0x4>;
 			};
 
+			vrefint: vrefin-cal@1b8 {
+				reg = <0x1b8 0x2>;
+			};
+
 			package-otp@1e8 {
 				reg = <0x1e8 0x1>;
 				bits = <0 3>;

-- 
2.43.0

[PATCH v2 12/14] arm64: dts: st: add ADC nodes on stm32mp231

[Fabrice Gasnier]

From: Cheick Traore <cheick.traore@foss.st.com>

Add analog-to-digital converter support on STM32MP23 SoC. It has 3 ADCs,
split into two blocks:
- ADC1 & ADC2 are tightly coupled.
- ADC3 is managed independently.
Define internal voltage channels in the SOC dtsi file, exception made of
vbat/4 channel (to avoid undesired load on it, in case no conversion is
needed on this channel). Choice is let to board implementation (via DT)
to define it.

Signed-off-by: Cheick Traore <cheick.traore@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- Adopt dedicated compatible for each child ADC, drop "st,adc-trigger-sel"
- Drop fallback generic match (serve no purpose)
---
 arch/arm64/boot/dts/st/stm32mp231.dtsi | 112 +++++++++++++++++++++++++++++++++
 1 file changed, 112 insertions(+)

diff --git a/arch/arm64/boot/dts/st/stm32mp231.dtsi b/arch/arm64/boot/dts/st/stm32mp231.dtsi
index 938d6f0098ac..578ab9d4cffc 100644
--- a/arch/arm64/boot/dts/st/stm32mp231.dtsi
+++ b/arch/arm64/boot/dts/st/stm32mp231.dtsi
@@ -646,6 +646,118 @@ uart7: serial@40370000 {
 				status = "disabled";
 			};
 
+			adc_12: adc@404e0000 {
+				compatible = "st,stm32mp23-adc-core";
+				reg = <0x404e0000 0x400>;
+				interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>,
+					<GIC_SPI 88 IRQ_TYPE_LEVEL_HIGH>;
+				clocks = <&rcc CK_KER_ADC12>;
+				clock-names = "adc";
+				interrupt-controller;
+				#interrupt-cells = <1>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+				access-controllers = <&rifsc 58>;
+				power-domains = <&cluster_pd>;
+				status = "disabled";
+
+				adc1: adc@0 {
+					compatible = "st,stm32mp23-adc1";
+					reg = <0x0>;
+					interrupt-parent = <&adc_12>;
+					interrupts = <0>;
+					dmas = <&hpdma 81 0x20 0x12>;
+					dma-names = "rx";
+					#io-channel-cells = <1>;
+					#address-cells = <1>;
+					#size-cells = <0>;
+					nvmem-cells = <&vrefint>;
+					nvmem-cell-names = "vrefint";
+					status = "disabled";
+					channel@14 {
+						reg = <14>;
+						label = "vrefint";
+					};
+				};
+
+				adc2: adc@100 {
+					compatible = "st,stm32mp23-adc2";
+					reg = <0x100>;
+					interrupt-parent = <&adc_12>;
+					interrupts = <1>;
+					dmas = <&hpdma 82 0x20 0x12>;
+					dma-names = "rx";
+					#io-channel-cells = <1>;
+					#address-cells = <1>;
+					#size-cells = <0>;
+					nvmem-cells = <&vrefint>;
+					nvmem-cell-names = "vrefint";
+					status = "disabled";
+					channel@14 {
+						reg = <14>;
+						label = "vrefint";
+					};
+					channel@15 {
+						reg = <15>;
+						label = "vddcore";
+					};
+					channel@17 {
+						reg = <17>;
+						label = "vddcpu";
+					};
+					channel@18 {
+						reg = <18>;
+						label = "vddgpu";
+					};
+				};
+			};
+
+			adc_3: adc@404f0000 {
+				compatible = "st,stm32mp23-adc-core";
+				reg = <0x404f0000 0x400>;
+				interrupts = <GIC_SPI 89 IRQ_TYPE_LEVEL_HIGH>;
+				clocks = <&rcc CK_KER_ADC3>;
+				clock-names = "adc";
+				interrupt-controller;
+				#interrupt-cells = <1>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+				access-controllers = <&rifsc 59>;
+				power-domains = <&cluster_pd>;
+				status = "disabled";
+
+				adc3: adc@0 {
+					compatible = "st,stm32mp23-adc3";
+					reg = <0x0>;
+					interrupt-parent = <&adc_3>;
+					interrupts = <0>;
+					dmas = <&hpdma 83 0x20 0x12>;
+					dma-names = "rx";
+					#io-channel-cells = <1>;
+					#address-cells = <1>;
+					#size-cells = <0>;
+					nvmem-cells = <&vrefint>;
+					nvmem-cell-names = "vrefint";
+					status = "disabled";
+					channel@14 {
+						reg = <14>;
+						label = "vrefint";
+					};
+					channel@15 {
+						reg = <15>;
+						label = "vddcore";
+					};
+					channel@17 {
+						reg = <17>;
+						label = "vddcpu";
+					};
+					channel@18 {
+						reg = <18>;
+						label = "vddgpu";
+					};
+				};
+			};
+
 			rng: rng@42020000 {
 				compatible = "st,stm32mp25-rng";
 				reg = <0x42020000 0x400>;

-- 
2.43.0

[PATCH v2 13/14] arm64: dts: st: add ADC nodes on stm32mp251

[Fabrice Gasnier]

Add analog-to-digital converter support on STM32MP25 SoC. It has 3 ADCs,
split into two blocks:
- ADC1 and ADC2 are tightly coupled.
- ADC3 is managed independently.
Define internal voltage channels in the SOC dtsi file, exception made of
vbat/4 channel (to avoid undesired load on it, in case no conversion is
needed on this channel). Choice is let to board implementation (via DT)
to define it.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
Changes in v2:
- Adopt dedicated compatible for each child ADC, drop "st,adc-trigger-sel"
- Drop fallback generic match (serve no purpose)
---
 arch/arm64/boot/dts/st/stm32mp251.dtsi | 112 +++++++++++++++++++++++++++++++++
 1 file changed, 112 insertions(+)

diff --git a/arch/arm64/boot/dts/st/stm32mp251.dtsi b/arch/arm64/boot/dts/st/stm32mp251.dtsi
index 82a8a6f1f881..5834755194a1 100644
--- a/arch/arm64/boot/dts/st/stm32mp251.dtsi
+++ b/arch/arm64/boot/dts/st/stm32mp251.dtsi
@@ -1450,6 +1450,118 @@ uart8: serial@40380000 {
 				status = "disabled";
 			};
 
+			adc_12: adc@404e0000 {
+				compatible = "st,stm32mp25-adc-core";
+				reg = <0x404e0000 0x400>;
+				interrupts = <GIC_SPI 87 IRQ_TYPE_LEVEL_HIGH>,
+					<GIC_SPI 88 IRQ_TYPE_LEVEL_HIGH>;
+				clocks = <&rcc CK_KER_ADC12>;
+				clock-names = "adc";
+				interrupt-controller;
+				#interrupt-cells = <1>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+				power-domains = <&CLUSTER_PD>;
+				access-controllers = <&rifsc 58>;
+				status = "disabled";
+
+				adc1: adc@0 {
+					compatible = "st,stm32mp25-adc1";
+					reg = <0x0>;
+					interrupt-parent = <&adc_12>;
+					interrupts = <0>;
+					dmas = <&hpdma 81 0x20 0x12>;
+					dma-names = "rx";
+					#io-channel-cells = <1>;
+					#address-cells = <1>;
+					#size-cells = <0>;
+					nvmem-cells = <&vrefint>;
+					nvmem-cell-names = "vrefint";
+					status = "disabled";
+					channel@14 {
+						reg = <14>;
+						label = "vrefint";
+					};
+				};
+
+				adc2: adc@100 {
+					compatible = "st,stm32mp25-adc2";
+					reg = <0x100>;
+					interrupt-parent = <&adc_12>;
+					interrupts = <1>;
+					dmas = <&hpdma 82 0x20 0x12>;
+					dma-names = "rx";
+					#io-channel-cells = <1>;
+					#address-cells = <1>;
+					#size-cells = <0>;
+					nvmem-cells = <&vrefint>;
+					nvmem-cell-names = "vrefint";
+					status = "disabled";
+					channel@14 {
+						reg = <14>;
+						label = "vrefint";
+					};
+					channel@15 {
+						reg = <15>;
+						label = "vddcore";
+					};
+					channel@17 {
+						reg = <17>;
+						label = "vddcpu";
+					};
+					channel@18 {
+						reg = <18>;
+						label = "vddgpu";
+					};
+				};
+			};
+
+			adc_3: adc@404f0000 {
+				compatible = "st,stm32mp25-adc-core";
+				reg = <0x404f0000 0x400>;
+				interrupts = <GIC_SPI 89 IRQ_TYPE_LEVEL_HIGH>;
+				clocks = <&rcc CK_KER_ADC3>;
+				clock-names = "adc";
+				interrupt-controller;
+				#interrupt-cells = <1>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+				access-controllers = <&rifsc 59>;
+				power-domains = <&CLUSTER_PD>;
+				status = "disabled";
+
+				adc3: adc@0 {
+					compatible = "st,stm32mp25-adc3";
+					reg = <0x0>;
+					interrupt-parent = <&adc_3>;
+					interrupts = <0>;
+					dmas = <&hpdma 83 0x20 0x12>;
+					dma-names = "rx";
+					#io-channel-cells = <1>;
+					#address-cells = <1>;
+					#size-cells = <0>;
+					nvmem-cells = <&vrefint>;
+					nvmem-cell-names = "vrefint";
+					status = "disabled";
+					channel@14 {
+						reg = <14>;
+						label = "vrefint";
+					};
+					channel@15 {
+						reg = <15>;
+						label = "vddcore";
+					};
+					channel@17 {
+						reg = <17>;
+						label = "vddcpu";
+					};
+					channel@18 {
+						reg = <18>;
+						label = "vddgpu";
+					};
+				};
+			};
+
 			rng: rng@42020000 {
 				compatible = "st,stm32mp25-rng";
 				reg = <0x42020000 0x400>;

-- 
2.43.0

[PATCH v2 14/14] arm64: dts: st: add ADC support on stm32mp257f-ev1 board

[Fabrice Gasnier]

Add ADC support on stm32mp257f-ev1 board. ANA0 dedicated pin is wired on
mikroBUS connector, through a resistor divided and operational amplifier.
Configure adc1 channel0, and voltage divider in the DT. Keep it disabled
by default, to allow ADC to be used freely for other purpose.

--> | 470K | -+-- > | OpAmp (gain = 1) | --> ANA0 (to ADC1, ADC2 or ADC3)
              |
              v
           | 560K |
              |
              v
             ---
             GND

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
---
 arch/arm64/boot/dts/st/stm32mp257f-ev1.dts | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/arch/arm64/boot/dts/st/stm32mp257f-ev1.dts b/arch/arm64/boot/dts/st/stm32mp257f-ev1.dts
index b5db62a0a550..ed61ff70c0de 100644
--- a/arch/arm64/boot/dts/st/stm32mp257f-ev1.dts
+++ b/arch/arm64/boot/dts/st/stm32mp257f-ev1.dts
@@ -160,6 +160,28 @@ mm_ospi1: mm-ospi@60000000 {
 			no-map;
 		};
 	};
+
+	vdiv_ana0: voltage-divider-ana0 {
+		compatible = "voltage-divider";
+		io-channels = <&adc1 0>;
+		output-ohms = <560000>;
+		full-ohms = <1030000>;
+		status = "disabled";
+	};
+};
+
+&adc_12 {
+	/* Don't need a pinctrl for ANA0 dedicated pin e.g. Channel@0 */
+	vdda-supply = <&scmi_vdda18adc>;
+	vref-supply = <&scmi_v1v8>;
+	status = "disabled";
+	adc1: adc@0 {
+		status = "okay";
+		channel@0 {
+			reg = <0>;
+			st,min-sample-time-ns = <400>;
+		};
+	};
 };
 
 &arm_wdt {

-- 
2.43.0