As some processors designed for embedded applications won’t necessarily run software with a user/supervisor-mode split of operation, it’s not necessary for them to implement user-mode features (eg. protect from unprivileged access to SPRs and the like.) Obviously this lack of support for user mode should be indicated via a bit in the CPUCFGR. So I propose a bit is added so that implementations without user-mode support can be detected by software which requires user-mode.


Wallento 11:51, 3 March 2015 (CET)

I think this is maybe too much of generalization. Except there is a specific plan to also implement a core without it, I don’t think we should over-complicate the spec.