On Thu, 2020-01-23 at 10:41 +0200, Janne Karhunen wrote:

On Wed, Jan 22, 2020 at 5:56 PM Mimi Zohar [off-list ref] wrote:

quoted

quoted

While it can now be argued that since this is an admin-driven event,
kernel does not need to write the file. However, the intention is to
bring out a second patch a bit later that adds a variable to define
the max number of entries to be kept in the kernel memory and
workqueue based automatic flushing. In those cases the kernel has to
be able to write the file without any help from the admin..

I don't think it is common, and probably not acceptable, for the
kernel to open a file for writing.

Ok. It just means that the kernel cannot do its own memory management
and will depend on the user flushing the memory often enough to
prevent something bad from happening. Is this more common in the
kernel than writing out a file?

Ok, there are examples of both passing a file descriptor and passing a
pathname from userspace, but even in the case of passing a pathname,
userspace normally creates the file.

There's been discussion in the past of defining an integrity
capability.  Are we at that point where we really do need to define an
integrity capability or is everyone comfortable with relying on
CAP_SYS_ADMIN?

When implementing this feature of exporting and truncating the
measurement list, please keep in mind how this would work in the
context of IMA namespaces.

thanks,

Mimi