Re: [RFC v7 10/41] richacl: Permission check algorithm

On Sat, Sep 12, 2015 at 12:12:16AM +0200, Andreas Grünbacher wrote:

2015-09-11 23:16 GMT+02:00 J. Bruce Fields [off-list ref]:

quoted

On Sat, Sep 05, 2015 at 12:27:05PM +0200, Andreas Gruenbacher wrote:

quoted

+             /*
+              * Apply the group file mask to entries other than owner@ and
+              * everyone@ or user entries matching the owner.  This ensures
+              * that we grant the same permissions as the acl computed by
+              * richacl_apply_masks().
+              *
+              * Without this restriction, the following richacl would grant
+              * rw access to processes which are both the owner and in the
+              * owning group, but not to other users in the owning group,
+              * which could not be represented without masks:
+              *
+              *  owner:rw::mask
+              *  group@:rw::allow
+              */
+             if ((acl->a_flags & RICHACL_MASKED) && richace_is_allow(ace))
+                     ace_mask &= acl->a_group_mask;

I'm having trouble understanding this.  I think the problem is that I
don't really understand the notation in your example.  Is a_group_mask
zero in that example?  I think it must be, in which case, OK I think I
get it.

Yes. I'm not sure if the example becomes easier to understand when the
empty group mask and perhaps also the other mask is included.

I think it would have been for me.

In general I find it confusing to present the mask bits as additional
ACEs--they're really pretty different.

quoted

(Though I still have to think about it a little more to convince myself
that richacl_apply_masks() always gets the same result.)

I have tried to break the algorithm into digestible pieces. Do you see
another way to make things easier to understand?

I just haven't reread those carefully enough yet, working on it....

--b.