[rv_dm,rtl] Fix acknowledgment of reset in dmstatus.allhavereset/dmstatus.anyhavereset

[andreaskurth]

pulp-platform/riscv-dbg#118:

Currently the debug module seems to assume any assertion of ndmreset_o will instantly cause a reset. The dmstatus.allhavereset/dmstatus.anyhavereset bits are set purely on ndmreset_o being asserted rather than any acknowledgement a reset has occurred.

This is incorrect behavior and should be fixed, but it is currently not urgent for us, because the rstmgr resets the system reset domain unconditionally whenever ndmreset_req is asserted (a few cycles after asserting this signal).

estimate 12

[andreaskurth]

Triaged for rv_dm. Assigning to M2.5 with Priority:P3 Priority: low because there are only a few clock cycles (between the assertion of ndmreset_o and the instant rstmgr resets the system) during which those bits are not correct -- so I think this can be easily worked around in software without noticeable impact, whereas the RTL changes involved in proper reset acknowledgment are out of scope for this release IMHO.

[msfschaffner]

Depending on how we intend to implement the delayed debug unlock feature, this may be something that we have to fix as well.

CC @rswarbrick