Result of spi_transceive() isn't correct

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 04:37:34 GMT

devzone.nordicsemi.com/.../8233.custom_5F00_nrf52832.zip

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 04:36:27 GMT

devzone.nordicsemi.com/.../0763.minimal_5F00_log_5F00_gsensor.zip

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 04:36:10 GMT

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 04:33:16 GMT

Thanks for your great support! I can read WHO_AM_I register successfully after changed _lis2dh12RegRead() as follows.

static int _lis2dh12RegRead(const uint8_t regAddr, uint8_t* regVal) {
    int ret;
    uint8_t txBuf[1] = { regAddr | LIS2DH12_REG_READ_MSK };

    const struct spi_buf txSpiBuf = {
        .buf = txBuf,
        .len = sizeof(txBuf)
    };

    const struct spi_buf_set txSpiBufSet = {
        .buffers = &txSpiBuf,
        .count   = 1
    };

    uint8_t rxBuf[2];
    struct spi_buf rxSpiBuf = {
        .buf = rxBuf,
        .len = sizeof(rxBuf)
    };
    const struct spi_buf_set rxSpiBufSet = { .buffers = &rxSpiBuf, .count = 1 };
    ret = spi_transceive(_gSpiDev, &_gSpiCfg, &txSpiBufSet, &rxSpiBufSet);
    if (ret < 0) {
        printk("Reading LIS2DH12 register failed, regAddr=0x%x, err=%d\n", regAddr, ret);
        return ERROR_GSENSOR_READ_FAIL;
    }

    *regVal = rxBuf[1];

    return 0;
}

00> *** Booting Zephyr OS build v3.3.99-ncs1 ***
00> Starting gsensor test
00> Starting button Test example
00> device_is_ready() ok
00> gpio_pin_configure_dt() ok
00> gpio_pin_interrupt_configure_dt() ok
00> Set up button at gpio@50000000 pin 27
00> button state=1
00> start a timer
00> now-1-11-30 0:0:0
00> 0
00> lis2dh12 device is ready!
00> LIS2DH12 ID check ok, id=0x33
00> lis2dh12 ID check pass!

RE: Result of spi_transceive() isn't correct

hmolesworth — Fri, 25 Aug 2023 03:30:39 GMT

I didn't bother reading anything but this snippet (feeling lazy) but to read the LIS12DH12 whoami response 2 Rx bytes are required not just 1 byte: Change

    uint8_t rxBuf[1];

    uint8_t rxBuf[2];

and make sure the expected receive length is set to 2 not 1. Be aware the "I am LIS2DH12" response is in rxBuf[1] not rxBuf[0]. rxBuff[0] is a dummy byte clocked out while the command byte is being sent; until the LIS2DH12 receives that command byte it can't respond with Iam as it doesn't know what the command is until the command has been received.

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 01:58:02 GMT

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 01:42:57 GMT

devzone.nordicsemi.com/.../gsensor.c

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 01:22:10 GMT

devzone.nordicsemi.com/.../custom_5F00_nrf52832.overlay

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 01:10:03 GMT

devzone.nordicsemi.com/.../1172.custom_5F00_nrf52832.zip

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 01:09:48 GMT

devzone.nordicsemi.com/.../3312.minimal_5F00_log_5F00_gsensor.zip

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 00:57:52 GMT

devzone.nordicsemi.com/.../6201.zephyr.dts

RE: Result of spi_transceive() isn't correct

snowuyl — Fri, 25 Aug 2023 00:57:40 GMT

Thanks for your reply! I have uploaded zephyr.dts file. And I have three questions.

1. I found cs-gpios = < &gpio0 0xf 0x1 >; in gsspi device node and interrupts = < 0xf 0x1 >; in ccm device node of zephyr.dts file. Does it mean pin 0xf conflict?

2. I have specified cs-gpios = <&gpio0 15 GPIO_ACTIVE_LOW>; in gsspi device node. Why cs-gpios = < &gpio0 0xf 0x1 >; in gsspi device node of zephyr.dts file? Does 0x1 means active low? How can I change GPIO_ACTIVE_LOW mapping to 0x0? From p9 of lis2dh12.pdf, it should be 0x0 for SPI communication mode.

3. I can't find pinctrl of custom_nrf52832.overlay in zephyr.dts file. Which file can I check for pinctrl of custom_nrf52832.overlay?

4. spim_cs of gsensor.c specified GPIO pin used by SPI chip select. How can I make sure it get correct GPIO pin for SPI chip select?

struct spi_cs_control spim_cs = {
    .gpio = SPI_CS_GPIOS_DT_SPEC_GET(DT_NODELABEL(reg_gsspi)),
    .delay = 0,
};

static const struct spi_config _gSpiCfg = {
    .operation = SPI_WORD_SET(8)  |
                 SPI_TRANSFER_MSB |
                 SPI_MODE_CPOL    |
                 SPI_MODE_CPHA,
    .frequency = 8000000,
    .slave = 0,
    .cs = &spim_cs,
};

        spi1: gsspi: spi@40004000 {
			compatible = "nordic,nrf-spi";
			#address-cells = < 0x1 >;
			#size-cells = < 0x0 >;
			reg = < 0x40004000 0x1000 >;
			interrupts = < 0x4 0x1 >;
			max-frequency = < 0x7a1200 >;
			easydma-maxcnt-bits = < 0x8 >;
			status = "okay";
			pinctrl-0 = < &spi1_default >;
			pinctrl-1 = < &spi1_sleep >;
			pinctrl-names = "default", "sleep";
			cs-gpios = < &gpio0 0xf 0x1 >;
			reg_gsspi: spi-dev-a@0 {
				reg = < 0x0 >;
			};
		};
		ccm: ccm@4000f000 {
			compatible = "nordic,nrf-ccm";
			reg = < 0x4000f000 0x1000 >;
			interrupts = < 0xf 0x1 >;
			length-field-length-8-bits;
			status = "okay";
		};

gsspi: &spi1 {
    compatible = "nordic,nrf-spi";
    status = "okay";
    pinctrl-0 = <&spi1_default>;
    pinctrl-1 = <&spi1_sleep>;
    pinctrl-names = "default", "sleep";
    cs-gpios = <&gpio0 15 GPIO_ACTIVE_LOW>;
    reg_gsspi: spi-dev-a@0 {
        reg = <0>;
    };
};

RE: Result of spi_transceive() isn't correct

Jørgen Holmefjord — Thu, 24 Aug 2023 11:38:36 GMT

Hi,

Might be some overlapping GPIO assignment to different peripherals.

Can you upload the compiled DTS file (found in build\zephyr\zephyr.dts)?

Best regards,
Jørgen

RE: Result of spi_transceive() isn't correct

snowuyl — Thu, 24 Aug 2023 10:24:12 GMT

devzone.nordicsemi.com/.../4431.custom_5F00_nrf52832.zip

RE: Result of spi_transceive() isn't correct

snowuyl — Thu, 24 Aug 2023 10:23:41 GMT

devzone.nordicsemi.com/.../minimal_5F00_log_5F00_gsensor.zip

RE: Result of spi_transceive() isn't correct

snowuyl — Thu, 24 Aug 2023 10:22:53 GMT

devzone.nordicsemi.com/.../lis2dh12.pdf

RE: Result of spi_transceive() isn't correct

snowuyl — Thu, 24 Aug 2023 10:22:22 GMT