I finally wrote an article about the last #MirageOS retreat (#OCaml). It's available here: https://blog.osau.re/articles/last_mirageos_retreat.html Enjoy reading!
Dinosaure's website — The 15th MirageOS retreat

An overview about the last MirageOS retreat

@dinosaure damn, I just casually clicked here because I'm somewhat interested in ocaml but ended up reading everything.

There's so much amazing work done here and I'm super astonished 🤯.

Wasm on a unikernel and sqlite working... It didn't even cross my mind, but really amazing stuff.

All I have to say is great work everyone. Super interesting seeing or at least reading about so much stuff being reimplemented from the basics 😌

@dinosaure Thanks for the writeup, love reading these reports from the MirageOS retreat.

It's such a great space to focus and do some interesting work while being with other great people to collaborate, discuss, or just motivate each other!

@dinosaure interesting, thanks for sharing. Do you have more details about what the final `perf kvm` command looked like? Also were stack traces (flamegraphs) working? Using an ocaml compiler that has frame pointers enabled might help (although you'd have to compile Solo5 with frame pointers too on the C side).
@edwintorok Thanks for your interest! So cfcs wrote something into the filesystem of the retreat' APE :D I need to search it but the idea was to produce a kallsyms file and run perf-kvm with such a file and be sure that we have a certain layout of directories like the Linux kernel (I think the unikernel must be renamed as a vmlinuz...). So a weird thing. If I found the document, I will obviously try to share it!