Why Is There a Long Delay for CS Pull-High After SPI Data Exchange? (Need to Shorten It)

[HHJ]

We are not using the NRF5340DK board. According to tests, after initiating an SPI data exchange, the program is delayed in executing the next CS pull-high operation, resulting in excessively long execution time for the entire SPI command. However, the SPI clock line ends within a very short period. I would like to know what operations are being executed between the end of the SPI clock line and the chip's response to my CS pull-high operation, and whether I can shorten this time as much as possible.

[geoffF]

Ever get a resolution to this?  I am also trying to communicate with an RHS2116 and have found the Zephyr SPI driver calls very slow.  See my post here: https://devzone.nordicsemi.com/f/nordic-q-a/123440/nrf5340dk-high-frequency-spim-communication-failure-spi_transceive-delay-problem/556943?focus=true

Is there any way to use the Zephyr driver for SPIM4 and get anything approaching the speed of the actual SPI clock Sigurd ? Per the trace in my post above most of the time seems wasted in the spi_* calls.