On Wed, 12 Apr 2023, Ilpo Järvinen wrote:
Rewrite the CAT test to allocated a buffer sized to half of LLC. Then perform a sequence of tests with different LLC block sizes starting from 1-bit CBM to half of the CBM bits. Flush the buffer before each test and read the buffer twice. Observe the LLC misses on the second read through the buffer. As the allocated LLC block gets larger and larger, the LLC misses will become smaller and smaller giving a strong signal on CAT working properly.
I realized this description is now wrong way around as I at one point reversed the cache allocation mask to go from half of CBM bits to 1-bit but didn't remember to update the changelog (in case somebody ends up wondering why it doesn't match to what the code does). V2 will contain the correct way around description.