Hi everyone. I am modifying the o3 cpu model and trying to collect all the squashed instructions in a buffer. Is there a single point from which all squashed instructions are passing? From what I understand squashes due to mispeculation are detected either in decode or iew with the latter case initiating the squash from the commit, whether squashes due to memory order violation or traps are also initiated by the commit stage. If I'm missing something, could someone give me a high-level overview on how squashes are handled in o3? Also, any word of advice on how to enforce a redirect of fetching to a different pc than the next one? Would resetting the pc after/in fetch's lookupAndUpdate function be enough? Thanks in advance! Constantinos