[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian and the desktop



Hi,

Josselin Mouette wrote:

[Permissions on device nodes]

Currently, there are two ways of handling this situation:
- The Debian way, where this is controlled by Unix groups, and where the
default user belongs to these groups. Your message seems to imply the
opposite, and I welcome you to install a sarge system and try plugging a
USB stick or playing sound.

This means that I have access to those devices even as a remote user, which is almost certainly not what anyone on a true multiuser system wants.

- The Redhat way, using pam_console. The user logging in gains rights on
some devices. The problem is that when the user logs out, there's no way
to force her to release the rights acquired. This is a limitation of the
Linux kernel, which cannot revoke privileges. AFAIK, that's why it isn't
used by default in Debian.

But would be a lot better than unconditionally granting access.

The "there is no way to revoke that" problem has an obvious solution: Go through a system service. The kernel is not the place to make or enforce policy, as it will not fit all use cases (and this is exactly what we're seeing here).

If you want things to move, you should provide a framework for the
kernel to handle a new revocation system call - far from an easy task.

I think it can and needs to be done in userspace.

[wireless configuration]

Some desktop tools doing that exist, but they seriously lack integration
in Debian.

I think they seriously lack integration period.

The overall trend on Linux in the last years has made it become an open-source Windows, with a lot of bloated applications fighting over the last 100 MiB of memory and a lot of "works for me" kind of solutions. Kernel capabilities have been misunderstood as a way to grant limited privileges to users, while in fact they are meant as a way to remove privileges from system services that do not need them.

   Simon

Attachment: signature.asc
Description: OpenPGP digital signature


Reply to: