drivers/iio/adc/rockchip_saradc.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)
The Successive Approximation ADC (SARADC) in RV1106 uses the v2
controller and supports:
- 10-bit resolution
- Up to 1MS/s sampling rate
- 2 single-ended input channels
"dt-bindings: iio: adc: rockchip-saradc: Add RV1106 compatible"
documents "rockchip,rv1106-saradc" as falling back to
"rockchip,rk3588-saradc" (see Link below). The two IP blocks are not
equivalent, though: per the RV1106 TRM v0.3, chapter 20 (SAR-ADC),
RV1106 only has 2 channels at 10-bit resolution, versus 8 channels
at 12-bit on RK3588. Without a dedicated match entry, the driver
binds via the fallback compatible and uses the RK3588 channel table,
which reports 8 channels instead of 2, and an in_voltage_scale that
is off by exactly 4x (4096 vs 1024 full-scale).
Add a dedicated compatible/data pair for RV1106 with the correct
2-channel, 10-bit layout.
Tested on a custom RV1106-based board, with:
&saradc {
vref-supply = <&vcc_saradc>; /* 1.8V fixed regulator */
status = "okay";
};
Before this patch:
# ls /sys/bus/iio/devices/iio:device0/in_voltage*_raw | wc -l
8
# cat /sys/bus/iio/devices/iio:device0/in_voltage_scale
0.439453125
After this patch:
# ls /sys/bus/iio/devices/iio:device0/in_voltage*_raw | wc -l
2
# cat /sys/bus/iio/devices/iio:device0/in_voltage_scale
1.757812500
Link: https://lore.kernel.org/all/20260714131631.v2.1.0b846080833e4b836793c43fc26d1c6d817cf9b4@changeid/
Signed-off-by: Vladislav Leonov <vlad@zlab.su>
---
drivers/iio/adc/rockchip_saradc.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/drivers/iio/adc/rockchip_saradc.c b/drivers/iio/adc/rockchip_saradc.c
index 0f0bf2906af0..5c45fae11890 100644
--- a/drivers/iio/adc/rockchip_saradc.c
+++ b/drivers/iio/adc/rockchip_saradc.c
@@ -348,6 +348,19 @@ static const struct rockchip_saradc_data rk3588_saradc_data = {
.read = rockchip_saradc_read_v2,
};
+static const struct iio_chan_spec rockchip_rv1106_saradc_iio_channels[] = {
+ SARADC_CHANNEL(0, "adc0", 10),
+ SARADC_CHANNEL(1, "adc1", 10),
+};
+
+static const struct rockchip_saradc_data rv1106_saradc_data = {
+ .channels = rockchip_rv1106_saradc_iio_channels,
+ .num_channels = ARRAY_SIZE(rockchip_rv1106_saradc_iio_channels),
+ .clk_rate = 1000000,
+ .start = rockchip_saradc_start_v2,
+ .read = rockchip_saradc_read_v2,
+};
+
static const struct of_device_id rockchip_saradc_match[] = {
{
.compatible = "rockchip,saradc",
@@ -370,6 +383,9 @@ static const struct of_device_id rockchip_saradc_match[] = {
}, {
.compatible = "rockchip,rk3588-saradc",
.data = &rk3588_saradc_data,
+ }, {
+ .compatible = "rockchip,rv1106-saradc",
+ .data = &rv1106_saradc_data,
},
{ }
};
--
2.47.3
On Mon, 10 Aug 2026 12:08:26 +0300
Vladislav Leonov <vlad@zlab.su> wrote:
> The Successive Approximation ADC (SARADC) in RV1106 uses the v2
> controller and supports:
> - 10-bit resolution
> - Up to 1MS/s sampling rate
> - 2 single-ended input channels
>
> "dt-bindings: iio: adc: rockchip-saradc: Add RV1106 compatible"
> documents "rockchip,rv1106-saradc" as falling back to
> "rockchip,rk3588-saradc" (see Link below). The two IP blocks are not
> equivalent, though: per the RV1106 TRM v0.3, chapter 20 (SAR-ADC),
> RV1106 only has 2 channels at 10-bit resolution, versus 8 channels
> at 12-bit on RK3588. Without a dedicated match entry, the driver
> binds via the fallback compatible and uses the RK3588 channel table,
> which reports 8 channels instead of 2, and an in_voltage_scale that
> is off by exactly 4x (4096 vs 1024 full-scale).
>
> Add a dedicated compatible/data pair for RV1106 with the correct
> 2-channel, 10-bit layout.
>
> Tested on a custom RV1106-based board, with:
>
> &saradc {
> vref-supply = <&vcc_saradc>; /* 1.8V fixed regulator */
> status = "okay";
> };
>
> Before this patch:
> # ls /sys/bus/iio/devices/iio:device0/in_voltage*_raw | wc -l
> 8
> # cat /sys/bus/iio/devices/iio:device0/in_voltage_scale
> 0.439453125
>
> After this patch:
> # ls /sys/bus/iio/devices/iio:device0/in_voltage*_raw | wc -l
> 2
> # cat /sys/bus/iio/devices/iio:device0/in_voltage_scale
> 1.757812500
>
> Link: https://lore.kernel.org/all/20260714131631.v2.1.0b846080833e4b836793c43fc26d1c6d817cf9b4@changeid/
> Signed-off-by: Vladislav Leonov <vlad@zlab.su>
Hi Vladislav,
Could you also send a fix for the dt-binding as it clearly
shouldn't claim the fallback compatible!
Whilst I can see why you did here, in general don't reply
to an existing series when sending a new one. Your link tag
is sufficient to associate the two.
I'll give this a little longer on list to let other comment, but
looks good to me.
Thanks,
Jonathan
> ---
> drivers/iio/adc/rockchip_saradc.c | 16 ++++++++++++++++
> 1 file changed, 16 insertions(+)
>
> diff --git a/drivers/iio/adc/rockchip_saradc.c b/drivers/iio/adc/rockchip_saradc.c
> index 0f0bf2906af0..5c45fae11890 100644
> --- a/drivers/iio/adc/rockchip_saradc.c
> +++ b/drivers/iio/adc/rockchip_saradc.c
> @@ -348,6 +348,19 @@ static const struct rockchip_saradc_data rk3588_saradc_data = {
> .read = rockchip_saradc_read_v2,
> };
>
> +static const struct iio_chan_spec rockchip_rv1106_saradc_iio_channels[] = {
> + SARADC_CHANNEL(0, "adc0", 10),
> + SARADC_CHANNEL(1, "adc1", 10),
> +};
> +
> +static const struct rockchip_saradc_data rv1106_saradc_data = {
> + .channels = rockchip_rv1106_saradc_iio_channels,
> + .num_channels = ARRAY_SIZE(rockchip_rv1106_saradc_iio_channels),
> + .clk_rate = 1000000,
> + .start = rockchip_saradc_start_v2,
> + .read = rockchip_saradc_read_v2,
> +};
> +
> static const struct of_device_id rockchip_saradc_match[] = {
> {
> .compatible = "rockchip,saradc",
> @@ -370,6 +383,9 @@ static const struct of_device_id rockchip_saradc_match[] = {
> }, {
> .compatible = "rockchip,rk3588-saradc",
> .data = &rk3588_saradc_data,
> + }, {
> + .compatible = "rockchip,rv1106-saradc",
> + .data = &rv1106_saradc_data,
> },
> { }
> };
Hi Jonathan, Sent a separate patch for the dt-binding fix: https://lore.kernel.org/all/20260816113636.33635-1-vlad@zlab.su Regards, Vladislav
On Sun, 16 Aug 2026 14:45:44 +0300 Vladislav Leonov <vlad@zlab.su> wrote: > Hi Jonathan, > > Sent a separate patch for the dt-binding fix: > > https://lore.kernel.org/all/20260816113636.33635-1-vlad@zlab.su > Just for completeness (as I replied to that just now) please send them both as a series so all the info is easily available to anyone looking at either part. Thanks Jonathan > Regards, > Vladislav
© 2016 - 2026 Red Hat, Inc.