Re: ptyfs and chroot environments

[gary%duzan.org@localhost]

=> In article <20100401015815.0E21AE4035%pluto.mumble.net@localhost>,
=> Taylor R Campbell  <campbell+netbsd%mumble.net@localhost> wrote:
=>>(I am not subscribed to this list, so please cc me in replies.)
=>>
=>>ptyfs interacts badly with chroot environments.  Since there can be
=>>only one ptyfs mount at any given time, a chroot environment cannot
=>>use ptm(4) to allocate ptys, because it will return /dev/pts/N to the
=>>userland, which does not (and cannot) exist in the chroot environment.
=>>
=>>Consequently, for example, when I try to run xterm inside a chroot
=>>environment on a system with a ptyfs mounted, it fails.  Testing
=>>programs in chroot environments is invaluable to me -- it lets me make
=>>sure that, say, a pkgsrc upgrade or a new X server works before I
=>>clobber my main file system.  I didn't notice this in NetBSD 4 because
=>>sysinst didn't start mounting ptyfs by default until NetBSD 5.
=>>
=>>I can work around this by unmounting ptyfs, and that's what I'll do
=>>for now.  Are there any other ways around this?  Is there a danger to
=>>having multiple ptyfs mounts, or could the ptyfs code be adapted to
=>>support this?
=>
=> It needs to be adapted, but it is a non trivial amount of work.

   Just a random idea: would a null mount of /dev/pts into the chroot
work? (I don't have it set up at the moment to test it.)

                                            Gary Duzan