Skip to contentSkip to conversation

Xilinx AMS fixes

PATCH Applied Robert Hancock 4 participants 15 messages Jan 20, 2022 – Jan 30, 2022

Subsystems (4)

v1 of 2 Applied Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover Xilinx AMS fixes

Various fixes for the Xilinx AMS driver, and addition of the required references in the ZynqMP device tree files to use it.

Robert Hancock (4):
arm64: dts: zynqmp: add AMS driver to device tree
iio: adc: xilinx-ams: Fixed missing PS channels
iio: adc: xilinx-ams: Fixed wrong sequencer register settings
iio: adc: xilinx-ams: Fix single channel switching sequence

 .../arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi |  4 +++
 arch/arm64/boot/dts/xilinx/zynqmp.dtsi        | 26 +++++++++++++++++++
 drivers/iio/adc/xilinx-ams.c                  | 15 ++++++++---
 3 files changed, 41 insertions(+), 4 deletions(-)
Signature · 1 line
2.31.1
#1 lore raw

PATCH 1/4 arm64: dts: zynqmp: add AMS driver to device tree

Add an entry to the ZynqMP device tree to support the AMS device which now has a driver in mainline.

---
 .../arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi |  4 +++
 arch/arm64/boot/dts/xilinx/zynqmp.dtsi        | 26 +++++++++++++++++++
 2 files changed, 30 insertions(+)
arch/arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi +4 -0
index 1e0b1bca7c94..108592104a1b 100644
@@ -239,6 +239,10 @@ &lpd_watchdog {
 	clocks = <&zynqmp_clk LPD_WDT>;
 };
 
+&xilinx_ams {
+	clocks = <&zynqmp_clk AMS_REF>;
+};
+
 &zynqmp_dpdma {
 	clocks = <&zynqmp_clk DPDMA_REF>;
 };
arch/arm64/boot/dts/xilinx/zynqmp.dtsi +26 -0
index 74e66443e4ce..d1fe1e5b46c1 100644
@@ -878,6 +878,32 @@ lpd_watchdog: watchdog@ff150000 {
 			timeout-sec = <10>;
 		};
 
+		xilinx_ams: ams@ffa50000 {
+			compatible = "xlnx,zynqmp-ams";
+			status = "disabled";
+			interrupt-parent = <&gic>;
+			interrupts = <0 56 4>;
+			reg = <0x0 0xffa50000 0x0 0x800>;
+			#address-cells = <1>;
+			#size-cells = <1>;
+			#io-channel-cells = <1>;
+			ranges = <0 0 0xffa50800 0x800>;
+
+			ams_ps: ams_ps@0 {
+				compatible = "xlnx,zynqmp-ams-ps";
+				status = "disabled";
+				reg = <0x0 0x400>;
+			};
+
+			ams_pl: ams_pl@400 {
+				compatible = "xlnx,zynqmp-ams-pl";
+				status = "disabled";
+				reg = <0x400 0x400>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+			};
+		};
+
 		zynqmp_dpdma: dma-controller@fd4c0000 {
 			compatible = "xlnx,zynqmp-dpdma";
 			status = "disabled";
Signature · 1 line
2.31.1
Reviewed-by on v1 1/4
#2 lore raw

Re: [PATCH 1/4] arm64: dts: zynqmp: add AMS driver to device tree

Quoted 4 lines · Robert Hancock

Add an entry to the ZynqMP device tree to support the AMS device which now has a driver in mainline.

Quoted 66 lines
---
 .../arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi |  4 +++
 arch/arm64/boot/dts/xilinx/zynqmp.dtsi        | 26 +++++++++++++++++++
 2 files changed, 30 insertions(+)
arch/arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi +4 -0
index 1e0b1bca7c94..108592104a1b 100644
@@ -239,6 +239,10 @@ &lpd_watchdog {
 	clocks = <&zynqmp_clk LPD_WDT>;
 };
 
+&xilinx_ams {
+	clocks = <&zynqmp_clk AMS_REF>;
+};
+
 &zynqmp_dpdma {
 	clocks = <&zynqmp_clk DPDMA_REF>;
 };
arch/arm64/boot/dts/xilinx/zynqmp.dtsi +26 -0
index 74e66443e4ce..d1fe1e5b46c1 100644
@@ -878,6 +878,32 @@ lpd_watchdog: watchdog@ff150000 {
 			timeout-sec = <10>;
 		};
 
+		xilinx_ams: ams@ffa50000 {
+			compatible = "xlnx,zynqmp-ams";
+			status = "disabled";
+			interrupt-parent = <&gic>;
+			interrupts = <0 56 4>;
+			reg = <0x0 0xffa50000 0x0 0x800>;
+			#address-cells = <1>;
+			#size-cells = <1>;
+			#io-channel-cells = <1>;
+			ranges = <0 0 0xffa50800 0x800>;
+
+			ams_ps: ams_ps@0 {
+				compatible = "xlnx,zynqmp-ams-ps";
+				status = "disabled";
+				reg = <0x0 0x400>;
+			};
+
+			ams_pl: ams_pl@400 {
+				compatible = "xlnx,zynqmp-ams-pl";
+				status = "disabled";
+				reg = <0x400 0x400>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+			};
+		};
+
 		zynqmp_dpdma: dma-controller@fd4c0000 {
 			compatible = "xlnx,zynqmp-dpdma";
 			status = "disabled";
Signature · 7 lines
2.31.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#3 lore raw

Re: [PATCH 1/4] arm64: dts: zynqmp: add AMS driver to device tree

In reply to Robert Hancock
Quoted 14 lines · Robert Hancock

Add an entry to the ZynqMP device tree to support the AMS device which now has a driver in mainline.

---
  .../arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi |  4 +++
  arch/arm64/boot/dts/xilinx/zynqmp.dtsi        | 26 +++++++++++++++++++
  2 files changed, 30 insertions(+)
arch/arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi +0 -0
index 1e0b1bca7c94..108592104a1b 100644

arch/arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi &lpd_watchdog { · line 244

Quoted 6 lines
arch/arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi +3 -0
@@ -239,6 +239,10 @@ &lpd_watchdog {
  	clocks = <&zynqmp_clk LPD_WDT>;
  };
 
+&xilinx_ams {
+	clocks = <&zynqmp_clk AMS_REF>;
+};

Please send this patch out of the series. It should go via soc tree not via iio tree. And unfortunately clock is not listed in DT binding document and needs to be added there. Can you please send that patch too? dtbs_check reports it as issue.

Quoted 31 lines
arch/arm64/boot/dts/xilinx/zynqmp-clk-ccf.dtsi +1 -0
+
  &zynqmp_dpdma {
  	clocks = <&zynqmp_clk DPDMA_REF>;
  };
arch/arm64/boot/dts/xilinx/zynqmp.dtsi +19 -0
index 74e66443e4ce..d1fe1e5b46c1 100644
@@ -878,6 +878,32 @@ lpd_watchdog: watchdog@ff150000 {
  			timeout-sec = <10>;
  		};
 
+		xilinx_ams: ams@ffa50000 {
+			compatible = "xlnx,zynqmp-ams";
+			status = "disabled";
+			interrupt-parent = <&gic>;
+			interrupts = <0 56 4>;
+			reg = <0x0 0xffa50000 0x0 0x800>;
+			#address-cells = <1>;
+			#size-cells = <1>;
+			#io-channel-cells = <1>;
+			ranges = <0 0 0xffa50800 0x800>;
+
+			ams_ps: ams_ps@0 {
+				compatible = "xlnx,zynqmp-ams-ps";
+				status = "disabled";
+				reg = <0x0 0x400>;
+			};
+
+			ams_pl: ams_pl@400 {
+				compatible = "xlnx,zynqmp-ams-pl";

arch/arm64/boot/dts/xilinx/zynqmp.dtsi lpd_watchdog: watchdog@ff150000 { · line 909

Quoted 10 lines
arch/arm64/boot/dts/xilinx/zynqmp.dtsi +7 -0
+				status = "disabled";
+				reg = <0x400 0x400>;
+				#address-cells = <1>;
+				#size-cells = <0>;
+			};
+		};
+
  		zynqmp_dpdma: dma-controller@fd4c0000 {
  			compatible = "xlnx,zynqmp-dpdma";
  			status = "disabled";

And this second piece is completely aligned with dt binding that's why it is fine.

Thanks,
Michal

#4 lore raw

PATCH 2/4 iio: adc: xilinx-ams: Fixed missing PS channels

In reply to Robert Hancock

The code forgot to increment num_channels for the PS channel inputs, resulting in them not being enabled as they should.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 1 +
 1 file changed, 1 insertion(+)
drivers/iio/adc/xilinx-ams.c +1 -0
index 8343c5f74121..b93864362dac 100644
@@ -1224,6 +1224,7 @@ static int ams_init_module(struct iio_dev *indio_dev,
 
 		/* add PS channels to iio device channels */
 		memcpy(channels, ams_ps_channels, sizeof(ams_ps_channels));
+		num_channels = ARRAY_SIZE(ams_ps_channels);
 	} else if (fwnode_property_match_string(fwnode, "compatible",
 						"xlnx,zynqmp-ams-pl") == 0) {
 		ams->pl_base = fwnode_iomap(fwnode, 0);
Signature · 1 line
2.31.1
#5 lore raw

Re: [PATCH 2/4] iio: adc: xilinx-ams: Fixed missing PS channels

Quoted 14 lines · Robert Hancock

The code forgot to increment num_channels for the PS channel inputs, resulting in them not being enabled as they should.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 1 +
 1 file changed, 1 insertion(+)
drivers/iio/adc/xilinx-ams.c +0 -0
index 8343c5f74121..b93864362dac 100644

drivers/iio/adc/xilinx-ams.c static int ams_init_module(struct iio_dev *indio_dev, · line 1230

Quoted 7 lines
drivers/iio/adc/xilinx-ams.c +1 -0
@@ -1224,6 +1224,7 @@ static int ams_init_module(struct iio_dev *indio_dev,
 
 		/* add PS channels to iio device channels */
 		memcpy(channels, ams_ps_channels, sizeof(ams_ps_channels));
+		num_channels = ARRAY_SIZE(ams_ps_channels);
 	} else if (fwnode_property_match_string(fwnode, "compatible",
 						"xlnx,zynqmp-ams-pl") == 0) {
 		ams->pl_base = fwnode_iomap(fwnode, 0);

Looks like this is the same change just submitted by Michael Tretter ("iio:
adc: xilinx-ams: Fix num_channels for PS channels").

Signature · 3 lines
Robert Hancock
Senior Hardware Designer, Calian Advanced Technologies
www.calian.com
Reviewed-by
#6 lore raw

Re: [PATCH 2/4] iio: adc: xilinx-ams: Fixed missing PS channels

Quoted 16 lines · Robert Hancock
Quoted 15 lines · Robert Hancock

The code forgot to increment num_channels for the PS channel inputs, resulting in them not being enabled as they should.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 1 +
 1 file changed, 1 insertion(+)
drivers/iio/adc/xilinx-ams.c +0 -0
index 8343c5f74121..b93864362dac 100644
@@ -1224,6 +1224,7 @@ static int ams_init_module(struct iio_dev *indio_dev,
 
Quoted 9 lines

drivers/iio/adc/xilinx-ams.c static int ams_init_module(struct iio_dev *indio_dev, · line 1230

Quoted 6 lines
drivers/iio/adc/xilinx-ams.c +1 -0
 		/* add PS channels to iio device channels */
 		memcpy(channels, ams_ps_channels, sizeof(ams_ps_channels));
+		num_channels = ARRAY_SIZE(ams_ps_channels);
 	} else if (fwnode_property_match_string(fwnode, "compatible",
 						"xlnx,zynqmp-ams-pl") == 0) {
 		ams->pl_base = fwnode_iomap(fwnode, 0);

Looks like this is the same change just submitted by Michael Tretter ("iio:
adc: xilinx-ams: Fix num_channels for PS channels").

Thanks for pointing me here.

Michael

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

maintainer
#7 lore raw

Re: [PATCH 2/4] iio: adc: xilinx-ams: Fixed missing PS channels

Quoted 27 lines · On Tue, 25 Jan 2022 09:06:34 +0100 Michael Tretter <[email protected]>…
Quoted 16 lines · Robert Hancock
Quoted 15 lines · Robert Hancock

The code forgot to increment num_channels for the PS channel inputs, resulting in them not being enabled as they should.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 1 +
 1 file changed, 1 insertion(+)
drivers/iio/adc/xilinx-ams.c +0 -0
index 8343c5f74121..b93864362dac 100644
@@ -1224,6 +1224,7 @@ static int ams_init_module(struct iio_dev *indio_dev,
 
Quoted 9 lines

drivers/iio/adc/xilinx-ams.c static int ams_init_module(struct iio_dev *indio_dev, · line 1230

Quoted 6 lines
drivers/iio/adc/xilinx-ams.c +1 -0
 		/* add PS channels to iio device channels */
 		memcpy(channels, ams_ps_channels, sizeof(ams_ps_channels));
+		num_channels = ARRAY_SIZE(ams_ps_channels);
 	} else if (fwnode_property_match_string(fwnode, "compatible",
 						"xlnx,zynqmp-ams-pl") == 0) {
 		ams->pl_base = fwnode_iomap(fwnode, 0);  

Looks like this is the same change just submitted by Michael Tretter ("iio:
adc: xilinx-ams: Fix num_channels for PS channels").

Quoted 3 lines

Thanks for pointing me here.

I picked up Michael's patch and will drop this one...
(based on random email reading order. I'd argue his was first in time, but then someone would point out next time I pick a patch up that is later than another one :)

Thanks,

Quoted 2 lines

Michael

#8 lore raw

PATCH 3/4 iio: adc: xilinx-ams: Fixed wrong sequencer register settings

In reply to Robert Hancock

Register settings used for the sequencer configuration register were incorrect, causing some inputs to not be read properly.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +2 -2
index b93864362dac..199027c93cdc 100644
@@ -91,8 +91,8 @@
 
 #define AMS_CONF1_SEQ_MASK		GENMASK(15, 12)
 #define AMS_CONF1_SEQ_DEFAULT		FIELD_PREP(AMS_CONF1_SEQ_MASK, 0)
-#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 1)
-#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK, 3)
 
 #define AMS_REG_SEQ0_MASK		GENMASK(15, 0)
 #define AMS_REG_SEQ2_MASK		GENMASK(21, 16)
Signature · 1 line
2.31.1
#9 lore raw

Re: [PATCH 3/4] iio: adc: xilinx-ams: Fixed wrong sequencer register settings

Quoted 14 lines · Robert Hancock

Register settings used for the sequencer configuration register were incorrect, causing some inputs to not be read properly.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +0 -0
index b93864362dac..199027c93cdc 100644

drivers/iio/adc/xilinx-ams.c line 95

Quoted 7 lines
drivers/iio/adc/xilinx-ams.c +2 -2
@@ -91,8 +91,8 @@
 
 #define AMS_CONF1_SEQ_MASK		GENMASK(15, 12)
 #define AMS_CONF1_SEQ_DEFAULT		FIELD_PREP(AMS_CONF1_SEQ_MASK, 0)
-#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 1)
-#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK, 3)

The TRM states that Continuous Loop Mode is 2, but Single Pass Sequence Mode is 1, not 3. Is there a reason, why you need to set both bits?

Michael

Quoted 12 lines
drivers/iio/adc/xilinx-ams.c +0 -0
 #define AMS_REG_SEQ0_MASK		GENMASK(15, 0)
 #define AMS_REG_SEQ2_MASK		GENMASK(21, 16)
Signature · 7 lines
2.31.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#10 lore raw

Re: [PATCH 3/4] iio: adc: xilinx-ams: Fixed wrong sequencer register settings

Quoted 16 lines · Michael Tretter
Quoted 15 lines · Robert Hancock

Register settings used for the sequencer configuration register were incorrect, causing some inputs to not be read properly.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +0 -0
index b93864362dac..199027c93cdc 100644
@@ -91,8 +91,8 @@
 
Quoted 12 lines

drivers/iio/adc/xilinx-ams.c line 98

Quoted 9 lines
drivers/iio/adc/xilinx-ams.c +2 -2
 #define AMS_CONF1_SEQ_MASK		GENMASK(15, 12)
 #define AMS_CONF1_SEQ_DEFAULT		FIELD_PREP(AMS_CONF1_SEQ_MASK,
 0)
-#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 1)
-#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK,
 2)
+#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK,
 3)

The TRM states that Continuous Loop Mode is 2, but Single Pass Sequence Mode is 1, not 3. Is there a reason, why you need to set both bits?

Single pass sequence mode (1) just runs the same sequence only once. To read these values it needs to switch to single channel mode (3).

The register bits are defined in Table 3-8 of
https://www.xilinx.com/support/documentation/user_guides/ug580-ultrascale-sysmon.pdf
.

Quoted 15 lines

Michael

Quoted 12 lines

#define AMS_REG_SEQ0_MASK GENMASK(15, 0)
#define AMS_REG_SEQ2_MASK GENMASK(21, 16)

Signature · 7 lines
2.31.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
https://urldefense.com/v3/__http://lists.infradead.org/mailman/listinfo/linux-arm-kernel__;!!IOGos0k!yGFEjSC1BL20lwurby914len0HCLXyzarwxKJP9Jx30qv_qrERSkRJUiVo_2MdusMVA$ 
Signature · 3 lines
Robert Hancock
Senior Hardware Designer, Calian Advanced Technologies
www.calian.com
Reviewed-by
#11 lore raw

Re: [PATCH 3/4] iio: adc: xilinx-ams: Fixed wrong sequencer register settings

Quoted 30 lines · Robert Hancock
Quoted 16 lines · Michael Tretter
Quoted 15 lines · Robert Hancock

Register settings used for the sequencer configuration register were incorrect, causing some inputs to not be read properly.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +0 -0
index b93864362dac..199027c93cdc 100644
@@ -91,8 +91,8 @@
 
Quoted 12 lines

drivers/iio/adc/xilinx-ams.c line 98

Quoted 9 lines
drivers/iio/adc/xilinx-ams.c +2 -2
 #define AMS_CONF1_SEQ_MASK		GENMASK(15, 12)
 #define AMS_CONF1_SEQ_DEFAULT		FIELD_PREP(AMS_CONF1_SEQ_MASK,
 0)
-#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 1)
-#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK,
 2)
+#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK,
 3)

The TRM states that Continuous Loop Mode is 2, but Single Pass Sequence Mode is 1, not 3. Is there a reason, why you need to set both bits?

Quoted 6 lines

Single pass sequence mode (1) just runs the same sequence only once. To read these values it needs to switch to single channel mode (3).

The register bits are defined in Table 3-8 of
https://www.xilinx.com/support/documentation/user_guides/ug580-ultrascale-sysmon.pdf
.

Thanks for the clarification.

Quoted 20 lines
Quoted 16 lines

Michael

Quoted 12 lines

#define AMS_REG_SEQ0_MASK GENMASK(15, 0)
#define AMS_REG_SEQ2_MASK GENMASK(21, 16)

Signature · 7 lines
2.31.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
https://urldefense.com/v3/__http://lists.infradead.org/mailman/listinfo/linux-arm-kernel__;!!IOGos0k!yGFEjSC1BL20lwurby914len0HCLXyzarwxKJP9Jx30qv_qrERSkRJUiVo_2MdusMVA$ 
Signature · 3 lines
Robert Hancock
Senior Hardware Designer, Calian Advanced Technologies
www.calian.com
maintainer Applied
#12 lore raw

Re: [PATCH 3/4] iio: adc: xilinx-ams: Fixed wrong sequencer register settings

Quoted 34 lines · On Wed, 26 Jan 2022 10:12:50 +0100 Michael Tretter <[email protected]>…
Quoted 33 lines · Robert Hancock
Quoted 16 lines · Michael Tretter
Quoted 15 lines · Robert Hancock

Register settings used for the sequencer configuration register were incorrect, causing some inputs to not be read properly.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +0 -0
index b93864362dac..199027c93cdc 100644
@@ -91,8 +91,8 @@
 
Quoted 12 lines

drivers/iio/adc/xilinx-ams.c line 98

Quoted 9 lines
drivers/iio/adc/xilinx-ams.c +2 -2
 #define AMS_CONF1_SEQ_MASK		GENMASK(15, 12)
 #define AMS_CONF1_SEQ_DEFAULT		FIELD_PREP(AMS_CONF1_SEQ_MASK,
 0)
-#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 1)
-#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK,
 2)
+#define AMS_CONF1_SEQ_CONTINUOUS	FIELD_PREP(AMS_CONF1_SEQ_MASK, 2)
+#define AMS_CONF1_SEQ_SINGLE_CHANNEL	FIELD_PREP(AMS_CONF1_SEQ_MASK,
 3)  

The TRM states that Continuous Loop Mode is 2, but Single Pass Sequence Mode is 1, not 3. Is there a reason, why you need to set both bits?

Single pass sequence mode (1) just runs the same sequence only once. To read these values it needs to switch to single channel mode (3).

Quoted 7 lines
Quoted 3 lines

The register bits are defined in Table 3-8 of
https://www.xilinx.com/support/documentation/user_guides/ug580-ultrascale-sysmon.pdf
.

Thanks for the clarification.

Applied to the fixes-togreg branch of iio.git

Thanks,

Jonathan

#13 lore raw

PATCH 4/4 iio: adc: xilinx-ams: Fix single channel switching sequence

In reply to Robert Hancock

Some of the AMS channels need to be read by switching into single-channel mode from the normal polling sequence. There was a logic issue in this switching code that could cause the first read of these channels to read back as zero.

It appears that the sequencer should be set back to default mode before changing the channel selection, and the channel should be set before switching the sequencer back into single-channel mode.

Also, write 1 to the EOC bit in the status register to clear it before waiting for it to become set, so that we actually wait for a new conversion to complete, and don't proceed based on a previous conversion completing.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")
---
 drivers/iio/adc/xilinx-ams.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +8 -2
index 199027c93cdc..7bf097fa10cb 100644
@@ -530,14 +530,18 @@ static int ams_enable_single_channel(struct ams *ams, unsigned int offset)
 		return -EINVAL;
 	}
 
-	/* set single channel, sequencer off mode */
+	/* put sysmon in a soft reset to change the sequence */
 	ams_ps_update_reg(ams, AMS_REG_CONFIG1, AMS_CONF1_SEQ_MASK,
-			  AMS_CONF1_SEQ_SINGLE_CHANNEL);
+			  AMS_CONF1_SEQ_DEFAULT);
 
 	/* write the channel number */
 	ams_ps_update_reg(ams, AMS_REG_CONFIG0, AMS_CONF0_CHANNEL_NUM_MASK,
 			  channel_num);
 
+	/* set single channel, sequencer off mode */
+	ams_ps_update_reg(ams, AMS_REG_CONFIG1, AMS_CONF1_SEQ_MASK,
+			  AMS_CONF1_SEQ_SINGLE_CHANNEL);
+
 	return 0;
 }
 
@@ -551,6 +555,8 @@ static int ams_read_vcc_reg(struct ams *ams, unsigned int offset, u32 *data)
 	if (ret)
 		return ret;
 
+	/* clear end-of-conversion flag, wait for next conversion to complete */
+	writel(expect, ams->base + AMS_ISR_1);
 	ret = readl_poll_timeout(ams->base + AMS_ISR_1, reg, (reg & expect),
 				 AMS_INIT_POLL_TIME_US, AMS_INIT_TIMEOUT_US);
 	if (ret)
Signature · 1 line
2.31.1
maintainer
#14 lore raw

Re: [PATCH 4/4] iio: adc: xilinx-ams: Fix single channel switching sequence

Quoted 9 lines · On Wed, 19 Jan 2022 19:02:46 -0600 Robert Hancock <[email protected]>…

Some of the AMS channels need to be read by switching into single-channel mode from the normal polling sequence. There was a logic issue in this switching code that could cause the first read of these channels to read back as zero.

It appears that the sequencer should be set back to default mode before changing the channel selection, and the channel should be set before switching the sequencer back into single-channel mode.

Quoted 7 lines

Also, write 1 to the EOC bit in the status register to clear it before waiting for it to become set, so that we actually wait for a new conversion to complete, and don't proceed based on a previous conversion completing.

Fixes d5c70627a794 ("iio: adc: Add Xilinx AMS driver")

Looking for an Ack from Anand or someone else familiar with this device.

Thanks,

Jonathan

Quoted 38 lines
---
 drivers/iio/adc/xilinx-ams.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/iio/adc/xilinx-ams.c +8 -2
index 199027c93cdc..7bf097fa10cb 100644
@@ -530,14 +530,18 @@ static int ams_enable_single_channel(struct ams *ams, unsigned int offset)
 		return -EINVAL;
 	}
 
-	/* set single channel, sequencer off mode */
+	/* put sysmon in a soft reset to change the sequence */
 	ams_ps_update_reg(ams, AMS_REG_CONFIG1, AMS_CONF1_SEQ_MASK,
-			  AMS_CONF1_SEQ_SINGLE_CHANNEL);
+			  AMS_CONF1_SEQ_DEFAULT);
 
 	/* write the channel number */
 	ams_ps_update_reg(ams, AMS_REG_CONFIG0, AMS_CONF0_CHANNEL_NUM_MASK,
 			  channel_num);
 
+	/* set single channel, sequencer off mode */
+	ams_ps_update_reg(ams, AMS_REG_CONFIG1, AMS_CONF1_SEQ_MASK,
+			  AMS_CONF1_SEQ_SINGLE_CHANNEL);
+
 	return 0;
 }
 
@@ -551,6 +555,8 @@ static int ams_read_vcc_reg(struct ams *ams, unsigned int offset, u32 *data)
 	if (ret)
 		return ret;
 
+	/* clear end-of-conversion flag, wait for next conversion to complete */
+	writel(expect, ams->base + AMS_ISR_1);
 	ret = readl_poll_timeout(ams->base + AMS_ISR_1, reg, (reg & expect),
 				 AMS_INIT_POLL_TIME_US, AMS_INIT_TIMEOUT_US);
 	if (ret)