Re: Capability tests in netlink and oom_kill

From: Stephen Smalley (sdsat_private)
Date: Wed Aug 22 2001 - 05:37:51 PDT

  • Next message: Stephen Smalley: "Re: syscall convention"

    On Wed, 22 Aug 2001, James Morris wrote:
    
    > Yep, looks like a good solution.  I'm just wondering (not sure if it
    > really matters) why we wouldn't create a new netlink_security_ops
    > structure for these hooks though.
    
    I could put these hooks into a new substructure, but I don't know if 
    it is worthwhile to do so.  In most cases, the substructures correspond
    to different kernel object types, so I was thinking that the netlink
    hooks should stay in the top-level structure.  But we also have 
    substructures for grouping related operations even when there is no
    particular kernel object, like the module_ops and ip_ops.  And as
    both Chris Wright and I have previously observed, some of the current
    top-level hooks could be moved into substructures, possibly with
    some alterations (e.g. ptrace, capget, capset_check, and capset_set
    could all go into task_security_ops; all of the mount-related hooks could
    go into super_block_security_ops, especially if they were changed to
    use struct super_block instead of vfsmount - but that might break
    some modules).  
    
    It doesn't really matter to me.  Does anyone else care?
    
    --
    Stephen D. Smalley, NAI Labs
    ssmalleyat_private
    
    
    
    
    
    _______________________________________________
    linux-security-module mailing list
    linux-security-moduleat_private
    http://mail.wirex.com/mailman/listinfo/linux-security-module
    



    This archive was generated by hypermail 2b30 : Wed Aug 22 2001 - 05:39:58 PDT