@rsnikhil
This set of changes tries to replaces, in the write-through cache, as many uses of word64 as possible with an abstracted cache-word instead. This is very helpful when updating caches with CHERI capability support (using 128-bit cache words, and storing capability tags at the cache-word granularity). The defaults are kept to be "just working" with word64.
Do you reckon this is an acceptable PR in its current state?
To what extent is a similar approach possible in the write-back caches?
@rsnikhil This set of changes tries to replaces, in the write-through cache, as many uses of word64 as possible with an abstracted cache-word instead. This is very helpful when updating caches with CHERI capability support (using 128-bit cache words, and storing capability tags at the cache-word granularity). The defaults are kept to be "just working" with word64.
Thanks.