PeterRugg
PeterRugg
Oh, it appears that pexpect also fails to launch it when I remove the assert...
Indeed, we do use completely separate processes. It's possible that makes this a somewhat unusual use-case. Agreed that it would be nice for it to be cross-platform. I suppose that...
> I think that menvcfg should also get the CW, CRG bits, otherwise this scheme is different to how the other bit-fields work where menvcfg is a superset. Looking through...
> It seems that we can do better and only clear the tags on a cache line without touching the data, when running a revocation sweep. Hmm, this might be...
@Timmmm I think this was already fixed? The current spec pages for the prefetch instructions say "the instruction does not perform a prefetch if...." rather than listing exception conditions. Is...
If you haven't made a write to it, then you shouldn't care whether you invalidate or flush. It seems like software that _requires_ an invalidate rather than a flush is...
For that example though, flush would do exactly the same as invalidate.
So cheri needs read and write for inval, because of something like the following: ``` k = secret key; // do some work k = 0; // clear the key...
Yeah, it's one thing I noticed when I first ran through the spec: the conditions are a random mix of saying "exception when..." and "exception unless...". Could be good to...
@andresag01 > ASR is because invalidate needs to read a CSR to work out how it behaves. Hmm, that wasn't quite my understanding: it was more because it's so dangerous....