On 6/19/20 6:44 AM, Peter Maydell wrote: >> +/* No-fault version of mte_check1, to be used by SVE for MemSingleNF. */ >> +bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr, uintptr_t ra) > > If this is a no-fault version, why do we need to pass in the ra ?
Excellent question. At first blush this doesn't actually implement no-fault at all, since it looks as if probe_access_flags would in fact fault. Except that within sve, we've already probed the page and this is just to handle the mte check. I think I'll remove the argument, pass 0 down to allocation_tag_mem, assert that the page is valid, and document the process. r~