[BUG] PMPCFG : Dead code in pmp module

[AyoubJalali]

Is there an existing CVA6 bug for this?

• I have searched the existing bug issues

Bug Description

Hello,

I noticed that the write of the value (2'b10 == NA4) in the pmpcfg.A is ignore for all the CVA6 configuration, That's means we don't support NA4, that's why we don't NA4 code in the pmp module.

We had some code coverage holes :

@OlivierBetschi what do you think ?

[OlivierBetschi]

According to https://docs.openhwgroup.org/projects/cva6-user-manual/01_cva6_user/PMP.html the NA4 mode is not selectable because of the granularity of the PMP being 1. This is confirmed in the RISC-V privileged ISA (3.7.1 Physical memory protection CSRs) : When G ≥ 1, the NA4 mode is not selectable.

So NA4 not being supported makes sense. Then the question if why the granularity is set to 1, as I don't see anything in the CVA6 requirements ( https://docs.openhwgroup.org/projects/cva6-user-manual/02_cva6_requirements/cva6_requirements_specification.html ) related to the granularity of the PMP. I don't know who could answer that ?

[AyoubJalali]

@zchamski we had some discussion on this before