Test Results

396 tests  ±0   396 ✅ ±0   0s ⏱️ ±0s
  4 suites ±0     0 💤 ±0 
  1 files   ±0     0 ❌ ±0 

Results for commit 0f89f4e. ± Comparison against base commit b102d07.

♻️ This comment has been updated with latest results.

This doesn’t work for sail-cheri-riscv.

I think the enumeration should probably remain scattered, but it seems to me there are two cases:

• Standard extensions that are declared up-front, and that appear in misa. We could declare all those standard single letter extensions in one place

• Other extensions that don't appear in misa, e.g. Xcheri

Hmm good point about CHERI. Unfortunately it isn't just misa. There's also things like menvcfg/mseccfg which have circular dependencies.

I guess technically the right thing to do would be to make legalize_menvcfg etc. scattered functions, but that just seems like overkill.

How about we keep it scattered but still move them all into one file?

This doesn’t work for sail-cheri-riscv.

True, but IMO sail-cheri-riscv should change to just be a diff on top of the base model. That would make upstreaming a lot easier as well.

This doesn’t work for sail-cheri-riscv.

True, but IMO sail-cheri-riscv should change to just be a diff on top of the base model. That would make upstreaming a lot easier as well.

It should, but it won’t do so immediately, and the model has been designed to support the current approach, so we do need to uphold that.

Sorted according to the canonical ordering. Which actually doesn't make any sense for Zhinh since it's supposed to be related to Hypervisor. Eh whatever.

Hmm. Really should've been named Zfhinx or something. I'd say this is still an improvement though.