could use better documentation on how to properly debug acceleration
Does mean add a detailed description akin to:
- Check to see if your system supports Accel
- Check to see if the modules are loaded on linux
- Check to see if you are part of the
kvmgroup on linux - Try running
hello worldexample with accel enabled
Is that the kind of debugging we are talking about? To help users get accel working? Or are we talking abut something more low level?
Note that on MacOS accel doesn't really need much documentation. If you are using a version of MacOS that has hv_support and you are using a version of qemu that has hv_support then you are good if you added the accel flag to your config. Most modern Macs fall into this category.
A command ops accel-check could be implemented rather easily. The command would provide some debug info:
Your system does not support hardware acceleration
Your system support hardware acceleration but no hypervisors found. Supported hypervisors are ...
Your user <username> is not part of the kvm group. Please add your user to the kvm group ...
nah - I was just thinking of a help section for docs listing some of the checks you mentioned - eg: are you in a vm already, do you have rights to kvm, etc.. but if you want to roll an accel-check I'm not opposed to that either - I'm already thinking of ways to automate help/debugging
https://github.com/nanovms/ops-documentation/pull/42