tree: e3630a21e5f72b87d9e8580ba14cf94b25af8eb2 [path history] [tgz]
  1. atom-d510.json
  2. BUILD.gn
  3. converter.py
  4. core2-6300.json
  5. haswell-xeon-e5-2690v3.json
  6. ivy-bridge-i3-3240.json
  7. nehalem-xeon-e5520.json
  8. README.md
  9. ryzen-2700x.json
  10. ryzen-3950x-virtualbox-hyperv.json
  11. ryzen-3950x-virtualbox-kvm.json
  12. ryzen-3950x-vmware.json
  13. ryzen-3950x-wsl2.json
  14. ryzen-3950x.json
  15. sandy-bridge-i7-2600k.json
  16. skylake-i3-6100.json
  17. threadripper-1950x.json
zircon/kernel/lib/arch/test/data/cpuid/README.md

CPUID value corpus

This directory holds a corpus of CPUID values from a variety of processors. The format is a simple, self-documenting JSON schema.

For the convenience of extending the corpus without writing the JSON by hand, one can use converter.py to convert the raw output of the cpuid tool to our JSON format. (The tool is only available on Linux.) The python script expects the raw cpuid output on stdin and writes the JSON to stdout.

Example usage:

cpuid -r1 | converter.py > cpuid.json

(cpuid -r may also be piped into the script.)

An extensive suite of expectation tests against the corpus is written in cpuid-corpus-tests.cc. Any addition to the corpus is expected to be referenced in that file - with a similar set of tests - as well as in the source list in BUILD.gn.