qyliss changed the topic of #spectrum to: A compartmentalized operating system | https://spectrum-os.org/ | Logs: https://logs.spectrum-os.org/spectrum/
cole-h has quit [Ping timeout: 250 seconds]
dgreid_ has quit [*.net *.split]
dgreid_ has joined #spectrum
nicooo has quit [Remote host closed the connection]
nicooo has joined #spectrum
DrWhax_ has quit [Ping timeout: 264 seconds]
DrWhax has joined #spectrum
cole-h has joined #spectrum
<qyliss> grr… update to crosvm has stopped my Nix derivation from it starting
<MichaelRaskin> Unfortunate indeed
<qyliss> I got help in #crosvm :)
<qyliss> The seccomp policy files need to be compiled to bpf ahead of time, because the compiler built into crosvm can't handle some new feature
<qyliss> But that should also boost startup time anyway, so I should have been doing that all along
<MichaelRaskin> Oh, better startup time
<qyliss> It doesn't seem to be that slow anyway, so I wouldn't get your hopes up
<MichaelRaskin> Hm true, it is before the Linux starts booting inside, which is indeed a small part of the boot in any case
<qyliss> Optimising the kernel build is what's going to bring down the startup time most I suspect
<qyliss> Currently I'm using a full kernel meant for booting laptops
<MichaelRaskin> And my tests were even worse as I was using rust-9p
<qyliss> Anyone know about LLVM?
<Shell> that is never a good question
<Shell> what's broken?
<qyliss> minijail wanted some LLVM IR but I think I've found a better way
<tilpner> For the bpf compilation?
<qyliss> Yeah
<qyliss> What Google actually does in their packaging is different, and just involves compiling and running a C program.
<qyliss> The only bad thing about that is that it doesn't work when cross-compiling
<qyliss> Or, well, this is for a dependency of bpf compilation; a JSON syscall table
<tilpner> Huh, how do you get a syscall table from LLVM IR?
<qyliss> Like this! pkgsCross.aarch64-multiplatform.minijail
<qyliss> whoops bad paste
<tilpner> ph no
<tilpner> *oh no