Does setting mcontrol6 trigger to NAPOT limit the size of tdata2 due to maskmax restriction?

[ThimayaPR]

With regards to NAPOT match trigger, if an implementation chooses to support maskmax6 an N.
We would expect tdata2[N-1] to be set to zero to comply with 5.7.12.

e.g:
In a system with XLEN=64.
We setup maxmask6 to be 63 (tdata2[62] will be set to zero) and configure tdata1 to setup a NAPOT trigger.

What is the expected behaviour of tdata2 when we attempt to write tdata2[61:0] with all ones?

Are we expected to read back 0x3fff...ffff [the same value we attempt to write]
or
are we expected to read back 0xbfff...ffff since we are anyway limiting the NAPOT range to 2^63?

[en-sc]

AFAIU, from [5.7.12. Match Control Type 6 (mcontrol6, at 0x7a1), match field description ]:

1 (napot): Matches when the top M bits of any compare
value match the top M bits of tdata2. M is XLEN-1 minus the
index of the least-significant bit containing 0 in tdata2.
tdata2 is WARL and if bits maskmax6-1:0 are written with
all ones then bit maskmax6-1 will be set to 0 while the
values of bits maskmax6-2:0 are UNSPECIFIED. Legal
values for tdata2 require M + maskmax6 ≥ XLEN and M > 0.
See above for how to determine maskmax6.

In your example:

1. M == 63 - 62 == 1
2. Bit maskmax6 - 1 == 62 is written with zero.

So my understanding is tdata2 will be read back as 0x3fff...ffff since it's a valid value.