The function nrf_qspi_cinstrdata_get has logic bug that didn't combine data from both register
e.g if the length = 6, it will go into the first switch, get the first byte from CINSTRDAT1. The next switch that got data from CINSTRDAT0 will never executed since the length is 6 (out of all cases).
Thank you for notifying us about this. It is highly appreciated! Yes, it looks like a bug, and I have registered it in our internal tracker.
If you have a look at the corresponding ..._set() function you will see that there everything is in the same switch block, with the change of the "reg" variable halfway through the cases. I think most likely that approach would be correct for the ..._get() as well, and if not there should at least be a comment explaining why.
Ah thanks, hope you guyes fixes it soon