Re: [PATCH v2] xattr: Enable security.capability in user namespaces

From: Stefan Berger
Date: Fri Jul 14 2017 - 15:23:01 EST

On 07/14/2017 01:36 PM, Eric W. Biederman wrote:
Stefan Berger <stefanb@xxxxxxxxxxxxxxxxxx> writes:

On 07/14/2017 09:34 AM, Serge E. Hallyn wrote:
Quoting Stefan Berger (stefanb@xxxxxxxxxxxxxxxxxx):
On 07/13/2017 08:38 PM, Eric W. Biederman wrote:
Stefan Berger <stefanb@xxxxxxxxxxxxxxxxxx> writes:

On 07/13/2017 01:49 PM, Eric W. Biederman wrote:

My big question right now is can you implement Ted's suggested
restriction. Only one or attribute ?
We need to raw-list the xattrs and do the check before writing them. I am fairly sure this can be done.

So now you want to allow and one<> or just a single one[[:print:]]*)?

The latter.
That case would prevent a container user from overriding the xattr
on the host. Is that what we want? For limiting the number of xattrs
Not really. If the file is owned by a uid mapped into the container,
then the container root can chown the file which will clear the file
capability, after which he can set a new one. If the file is not
owned by a uid mapped into the container, then container root could
not set a filecap anyway.
Let's say I installed a container where all files are signed and thus have
security.ima. Now for some reason I want to re-sign some or all files inside
that container. How would I do that ? Would I need to get rid of security.ima
first, possibly by copying each file, deleting the original file, and renaming
the copied file to the original name, or should I just be able to write out a
new signature, thus creating security.ima@uid=1000 besides the security.ima ?
This gets us into some interesting territory, where the semantics of
these attributes matters.

The implementation of security.capable implements the security killpriv
hooks. Anyone merely by changing the file can cause the security
capability to go away. So it makes sense from the security.capable side
that anyone who has the capable_wrt_inode_uidgid(CAP_SETFCAP) will be
able to clear and set security.capable.

The integrity xattrs do not. Which results in very big semantic
difference between these two kinds of attributes. I am insufficiently
familiar with the rules for security.ima and security.evm to understand
what those rules should be.

That may be enough that we can not share code between these two cases.

On the host I can simply overwrite capabilities. I think the same model should apply to the virtualized world. The difference still is that removing an xattr, if written on the host, may only be possible by copy + file move to original filename.

On IMA, when appending a letter to an executable, the executable doesn't run anymore when appraisal is used, but the signature is still there and needs to be re-written. Though I think this aspect on how they disappear doesn't matter as much if they can simply be overwritten.

Some things could certainly be solved with flags indicating behaviors of xattrs for as long as these flags only affect the reading, listing, and re-writing of the virtualized xattrs, which is what the patch does. For example a flag for security.capability could say that only a single 'security.capability(@uid=<uid>)?' may exist while security.ima could have two.