On Mon, Sep 2, 2019 at 11:02 AM Tony Chuang <yhchuang@xxxxxxxxxxx> wrote: > Indeed it is a little strange but it works like that, if I don't disable HIMR, > interrupt could be lost and never gets into interrupt handler again. > If so, the interrupt will be stopped, and TX/RX path will be freezed. OK, thanks for checking & explaining. It's a bit hard to put in words as we have seen, but it would be good if you could try to summarise this unusual design in the commit message or in a code comment. Thanks! Daniel