SPIM4 IRQ Handler in SDK v2.7.0

[kelk]

Hello Nordic Team,

I am currently implementing low-level SPI communication using SDK v2.7.0 and nRF 5340dk . Could you please let me know the name of the IRQ handler for SPIM4?

In version 52 Board , I used the SPIM1_SPIS1_TWIM1_TWIS1_SPI1_TWI1_IRQHandler to handle SPI1 interrupts, but I couldn’t find the corresponding handler for SPIM4 in this version.

Looking forward to your guidance.

Best regards,
kelk.

[Vidar Berg]

Hello,

The SPIM4 instance differs from the other SPIM instances in that it is not shared with any other peripherals. Therefore it has its own IRQ number assigned to it. The handler is named  'SPIM4_IRQHandler'

https://github.com/zephyrproject-rtos/hal_nordic/blob/0f462f62cc0c347303e67f733fef369bb411c733/nrfx/mdk/nrf5340_application.h#L79 

Best regards,

Vidar

[kelk]

Hello Vidar,

Thank you for your prompt response.

Firstly, the "nrf5340_application.h" file I have differs from the one in the link you provided. For instance, it does not include the SPIM2_SPIS2_TWIM2_TWIS2_UARTE2_IRQn and SPIM1_SPIS1_TWIM1_TWIS1_UARTE1_IRQn . it just encludes the SPIM4_IRQn.

Secondly, I attempted to use SPIM4_IRQHandler, but encountered the following error:

[00:00:07.623,840] <err> os: Unhandled IRQn: 10
[00:00:07.623,870] <err> os: >>> ZEPHYR FATAL ERROR 1: Unhandled interrupt on CPU 0
[00:00:07.623,901] <err> os: Current thread: 0x20000708 (unknown)
[00:00:07.652,618] <err> os: Halting system

[Vidar Berg]

Hello,

SPIM1_SPIS1_TWIM1_TWIS1_UARTE1_IRQn is now named SERIAL_IRQn:

kelk said:

Secondly, I attempted to use SPIM4_IRQHandler, but encountered the following error:

Did you use the IRQ_CONNECT() macro to connect the IRQ to the interrupt handler? https://github.com/nrfconnect/sdk-zephyr/blob/ade49f081df6fab0db6288bad25be4f5eff7ef3a/samples/boards/nrf/nrfx/src/main.c#L50 

[kelk]

hello,

i used this:

#define SPIM_INST_IDX 4
 
    IRQ_CONNECT(NRFX_IRQ_NUMBER_GET(NRF_SPIM_INST_GET(SPIM_INST_IDX)), IRQ_PRIO_LOWEST,
                NRFX_SPIM_INST_HANDLER_GET(SPIM_INST_IDX), 0, 0);
    
  

   

and i received this Error:

ASSERTION FAIL [p_cb->handler] @ WEST_TOPDIR/modules/hal/nordic/nrfx/drivers/src/nrfx_spim.c:948
[00:00:00.425,567] <err> os: r0/a1:  0x00000004  r1/a2:  0x000003b4  r2/a3:  0x00000003
[00:00:00.425,598] <err> os: r3/a4:  0x00000004 r12/ip:  0x00000014 r14/lr:  0x0000612b
[00:00:00.425,598] <err> os:  xpsr:  0x2900001a
[00:00:00.425,598] <err> os: Faulting instruction address (r15/pc): 0x0000913e
[00:00:00.425,628] <err> os: >>> ZEPHYR FATAL ERROR 4: Kernel panic on CPU 0
[00:00:00.425,659] <err> os: Fault during interrupt handling

[00:00:00.425,659] <err> os: Current thread: 0x20000708 (unknown)
[00:00:00.493,743] <err> os: Halting system

[Vidar Berg]

This assert is raised if no callback is registered: https://github.com/NordicSemiconductor/nrfx/blob/4fb7ccb30a1db35c1cf48f1eac87310125ba1072/drivers/src/nrfx_spim.c#L948.  I don't know if you are using the nrfx_spim driver or writing your own. 

[kelk]

Hello,

It finally worked with the IRQ_CONNECT macro — thank you!

I have a question about SPI with Zephyr. There’s a parameter to set the duration  between the edge of the slave select (SS) and the clock edge.

I set it to NULL, but there’s still a delay. Could you show me the correct way to properly set it to NULL?

#define SS_DURATION 0
const spi_operation_t operation=SPI_WORD_SET(8)|SPI_OP_MODE_MASTER|SPI_LOCK_ON|SPI_CS_ACTIVE_HIGH;
const struct spi_dt_spec spi_dev=SPI_DT_SPEC_GET(DT_NODELABEL(mcp2518fdspi),operation,SS_DURATION);

[Vidar Berg]

Hello,

Glad to hear that it worked. Regarding the CS delay, even if you set it to 0 there will still be some delay because of overhead in the driver. How long is the delay? 

[kelk]

I have a delay of 6.568 µs before the transmission and 21.24 µs after.

[Vidar Berg]

That is about what I would have expected.