On Sat Jan 24, 2026 at 3:28 AM JST, Gary Guo wrote: >> - let wptr = self.cpu_write_ptr().wrapping_add(elem_count) & >> MSGQ_NUM_PAGES; >> + let wptr = self.cpu_write_ptr().wrapping_add(elem_count) % >> MSGQ_NUM_PAGES; > > So the actual number of pages is indeed 0x3F, not 0x40? That's interesting...
Yes, that's right. I thought it was weird as well, but the reason is that the first page stores the rx/tx pointers (see `Msgq`) and the remaining 63 are used for the actual ring buffer.
