Andres Amaya Garcia
Andres Amaya Garcia
This seems to be a problem if the user requires writing a characteristic with a value that is longer than 21 bytes. For instance, the device name characteristic value can...
@LawrenceEsswood: > The distinction is what should happen when loading via a capability that does have "ElevateWrite" but DOES have "Write". Did you mean this instead? > via a capability...
@arichardson: Any thoughts on @LawrenceEsswood's proposal?
@LawrenceEsswood : Is the new proposal aligned with what you had in mind?
@nwf: Thanks for the explanation and your proposal. Its very informative! First of all, I am new to temporal safety using CHERI! Could you elaborate a bit on how could...
@nwf: Thanks for the info! My understanding so far is that we there are 3 key goals for the PTEs: * Limit capability propagation -- TN state (1) * Revocation...
@jonwoodruff has updated the proposal as discussed last week! Thanks for taking the time to do this @jonwoodruff! https://github.com/andresag01/riscv-cheri/pull/2
> 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. The key...
> > 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. >...
> (v)sstatus would make the most sense for it. There's precedent from SUM and MXR (and TVM in the full mstatus). @jrtc27 I agree that (v)sstatus would be a good...