I've been using the nRF52832 chip for over a year and it works great. However, for a new project I designed a board so that the nRF52 chip was the slave on a SPI bus, and I immediately ran into Errata 109. Fine, the workaround seemed simple, but it doesn't seem to work.
In the plot below, the orange trace is a GPIO pin that I pull down on the GPIOTE event callback. The cyan is SCLK from the master.
As you can see, the interrupt happens a full 18 usec after the SCLK is triggered. So, I don't think this is achieving the purpose of pre-starting the 64 MHz clock.
Am I missing something?