Some tweaks to kern_chroot(2)
ClosedPublic
Actions

Authored by kib on Sat, Aug 2, 6:03 PM.

Details

Reviewers

emaste
markj
kevans

Commits

rGcf0f693f0361: vfs_syscall.c: EXTERROR-ise EPERM from chroot(2)
rGa523d882a467: vfs_syscall.c: do not take process lock around the read of P2_NO_NEW_PRIVS

Summary

vfs_syscall.c: do not take process lock around the read of P2_NO_NEW_PRIVS

The flag is write-only, and if we raced with reading the action of
setting the flag, we can as well get it under the lock.


vfs_syscall.c: EXTERROR-ise EPERM from chroot(2)

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Skipped

Unit

Tests Skipped

Event Timeline

kib created this revision.Sat, Aug 2, 6:03 PM

Herald added subscribers: olce, imp. · View Herald TranscriptSat, Aug 2, 6:03 PM

kib requested review of this revision.Sat, Aug 2, 6:03 PM

Maybe be explicit that unprivileged_chroot is a sysctl/tunable, e.g. security.bsd.unprivileged_chroot sysctl not enabled

This revision is now accepted and ready to land.Sat, Aug 2, 6:10 PM

EXTERROR msg for disabled sysctl

This revision now requires review to proceed.Sat, Aug 2, 6:13 PM

kevans accepted this revision.Sat, Aug 2, 6:16 PM

kevans added inline comments.

sys/kern/vfs_syscalls.c
993	I don't feel that strongly here, but I think if I were to read this as someone unfamiliar, my question would be: "How does P2_NO_NEW_PRIVS get set?" -> a string that appears in a manpage to answer that would probably curtail that.

This revision is now accepted and ready to land.Sat, Aug 2, 6:16 PM

kevans added inline comments.Sat, Aug 2, 6:16 PM