On Fri, May 22, 2026 at 04:00:05PM -0700, Richard Henderson wrote:
> I did this myself 9 months ago, but never posted it for some reason.
>
> https://gitlab.com/rth7680/qemu/-/commit/2598d4896e19376fbae3926728ac737e7bbea190
>
> static CPAccessResult access_rndr(CPUARMState *env, const ARMCPRegInfo *ri,
> bool isread)
> {
> if (env->cp15.scr_el3 & SCR_TRNDR) {
> return CP_ACCESS_TRAP_EL3;
> }
> /*
> * Note that FEAT_RNDR_TRAP may be implemented without FEAT_RNDR.
> * In that case, if the trap is not enabled, the read undefs.
> */
> if (!cpu_isar_feature(aa64_rndr, env_archcpu(env))) {
> return CP_ACCESS_UNDEFINED;
> }
Should I pull this into my diff and resubmit? It's definitely the
right idea (I didn't think about this possibility).
>
> if (cpu_isar_feature(aa64_rndr, cpu) ||
> cpu_isar_feature(aa64_rndr_trap, cpu)) {
> define_arm_cp_regs(cpu, rndr_reginfo);
> }
>
Same with this? (adding the || rndr_trap to what is already there).
--Jason Wright
