v1.8.9

Feature

• Add clone3(CLONE_INTO_CGROUP) syscall support on kernel >= 5.7 and cgroup v2 enabled
  • ~20% faster in latency than previous implementation, which requires pause of child process, waiting to be written into cgroup.procs
• Add procPeak measurement from cgroup v2 on kernel >= 6.1

Changelog

• 8b9b6f3 linuxc: add cached cgroup implementation
• 9ae8487 linuxc: fix cgroup fd & add clean up to containers when exit
• ffdf3e3 linuxc: try support faster new clone3(CLONE_INTO_CGROUP) syscall
• 60a1859 monitor: add stat for worker queue