Michael Knyszek
Michael Knyszek
I don't think this needs to be a proposal. The trace parsing API is still experimental.
> > We could just add this as a **new event type** that always lives in the same batch as EvFrequency... With respect to exposing this in the trace API,...
This work has landed, so I'm going to close off the issue. Thanks again @felixge!
@alexsaezm Could we possibly contact and involve a debugedit owner/maintainer? (Are you that person?)
CC @golang/command-line @griesemer
Oof, yeah, this is not a use-case supported by the `weak` package. The crypto libraries could ask the runtime if the slice is Go-managed, and if not, skip caching, maybe?...
FWIW, having something like #70224 (which is not going to be trivial) would make this possible, but there is still a hazard here of unmapping the memory prematurely (before the...
Looking back at your original post, it sounds like your main use-case for using a guarded page is to make sure the memory gets scrubbed and the private key doesn't...
Oh and, 'doh, a possible workaround here is if you're managing the memory yourself anyway, you can probably get away with just making a relatively large pointer-free allocation (>32 KiB)...