m_pkthdr.label now moved to m_tag

[Robert Watson]

Per a discussion here last year some time, I've gone ahead and moved the
MAC label from the mbuf packet header structure to optional meta-data in
an m_tag.  This should improve the cleanliness of the mbuf header, which
is possible now we have m_tags (thanks Sam!), which of course weren't
available when we originally did the MAC adaptation of the network stack. 
As we now use the common meta-data infrastructure, it makes the failure
modes for IPsec, MAC, etc, the same when it comes to poor handling of
meta-data on mbufs, so hopefully that will improve the time to correct
meta-data handling problems in the stack.  Because the MAC m_tags are
"deep", they required tweaks in the m_tag handling code to properly copy
and free the tags--Sam and I have talked a bit about possible
generalizations and slippery slopes on generality.  I'd like to let this
stuff settle a bit before we take on generalizing :-). 

I do have some initial performance results from a bit of experimentation; 
there was a small but measurable performance improvement for GENERIC that
I believe corresponds to reducing the bzero() overhead on the mbuf header.
If I unconditionally allocated m_tags with MAC enabled, I saw a slight
additional performance overhead against m_pkthdr.label.  However, I took
this opportunity to move to conditionally allocating MAC m_tags, and in
general observed a substantial performance improvement for un-labeled MAC
policies, eliminating that overhead (and in fact improving performance
substantially) in most cases.  Labeled network policies actually see a
per-byte throughput improvement with large packet sizes; they do see a
per-packet overhead increase, however.

Two graphs, for the bored, are available at: 

    http://www.watson.org/~robert/freebsd/label-to-tag/

They explore the impact of three labeling approaches (m_pkthdr.label,
m_tag always alloc, m_tag conditional alloc) across five kernel
configurations: GENERIC, GENERIC+MAC, GENERIC+MAC with three policies:
mac_none (stub), mac_bsdextended (file system firewall), and mac_biba
(comprehensive system integrity policy).  Of these, mac_biba is the only
labeled network policy; mac_none does implement per-packet stub entry
points, however, so does pay a per-packet overhead compared to no policy
or mac_bsdextended. 

All measurements were over the loopback interface on an 800mhz box (SMP
turned off, I should re-do them sometime with SMP).

Incidentally, this should also fix a number of WITNESS warnings when
running with labeled MAC policies because it cleans up some bugs in the
mapping of mbuf allocator flags to malloc flags, and also more cleanly
avoids blocking allocations during certain classes of mbuf head copies.

Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
robert at fledge.watson.org      Network Associates Laboratories

---------- Forwarded message ----------
Date: Mon, 14 Apr 2003 13:39:06 -0700 (PDT)
From: Robert Watson <rwatson at FreeBSD.org>
To: src-committers at FreeBSD.org, cvs-src at FreeBSD.org, cvs-all at FreeBSD.org
Subject: cvs commit: src/sys/kern kern_mac.c subr_mbuf.c uipc_mbuf.c         uipc_mbuf2.c src/sys/net if_loop.c src/sys/sys mac.h mac_policy.h         mbuf.h

rwatson     2003/04/14 13:39:06 PDT

  FreeBSD src repository

  Modified files:
    sys/kern             kern_mac.c subr_mbuf.c uipc_mbuf.c 
                         uipc_mbuf2.c 
    sys/net              if_loop.c 
    sys/sys              mac.h mac_policy.h mbuf.h 
  Log:
  Move MAC label storage for mbufs into m_tags from the m_pkthdr structure,
  returning some additional room in the first mbuf in a chain, and
  avoiding feature-specific contents in the mbuf header.  To do this:
  
  - Modify mbuf_to_label() to extract the tag, returning NULL if not
    found.
  
  - Introduce mac_init_mbuf_tag() which does most of the work
    mac_init_mbuf() used to do, except on an m_tag rather than an
    mbuf.
  
  - Scale back mac_init_mbuf() to perform m_tag allocation and invoke
    mac_init_mbuf_tag().
  
  - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since
    m_tag's are now GC'd deep in the m_tag/mbuf code rather than
    at a higher level when mbufs are directly free()'d.
  
  - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related
    notions.
  
  - Generally change all references to mbuf labels so that they use
    mbuf_to_label() rather than &mbuf->m_pkthdr.label.  This
    required no changes in the MAC policies (yay!).
  
  - Tweak mbuf release routines to not call mac_destroy_mbuf(),
    tag destruction takes care of it for us now.
  
  - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() --
    the existing m_tag support does all this for us.  Note that
    we can no longer just zero the m_tag list on the target mbuf,
    rather, we have to delete the chain because m_tag's will
    already be hung off freshly allocated mbuf's.
  
  - Tweak m_tag copying routines so that if we're copying a MAC
    m_tag, we don't do a binary copy, rather, we initialize the
    new storage and do a deep copy of the label.
  
  - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places
    having to do with mbuf header copies previously.
  
  - When an mbuf is copied in ip_input(), we no longer need to
    explicitly copy the label because it will get handled by the
    m_tag code now.
  
  - No longer any weird handling of MAC labels in if_loop.c during
    header copies.
  
  - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test.
    In mac_test, handle the label==NULL case, since it can be
    dynamically loaded.
  
  In order to improve performance with this change, introduce the notion
  of "lazy MAC label allocation" -- only allocate m_tag storage for MAC
  labels if we're running with a policy that uses MAC labels on mbufs.
  Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS
  flag in their load-time flags field during declaration.  Note: this
  opens up the possibility of post-boot policy modules getting back NULL
  slot entries even though they have policy invariants of non-NULL slot
  entries, as the policy might have been loaded after the mbuf was
  allocated, leaving the mbuf without label storage.  Policies that cannot
  handle this case must be declared as NOTLATE, or must be modified.
  
  - mac_labelmbufs holds the current cumulative status as to whether
    any policies require mbuf labeling or not.  This is updated whenever
    the active policy set changes by the function mac_policy_updateflags().
    The function iterates the list and checks whether any have the
    flag set.  Write access to this variable is protected by the policy
    list; read access is currently not protected for performance reasons.
    This might change if it causes problems.
  
  - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update
    function to assert appropriate locks.
  
  - This makes allocation in mac_init_mbuf() conditional on the flag.
  
  Reviewed by:    sam
  Obtained from:  TrustedBSD Project
  Sponsored by:   DARPA, Network Associates Laboratories
  
  Revision  Changes    Path
  1.85      +118 -17   src/sys/kern/kern_mac.c
  1.45      +9 -17     src/sys/kern/subr_mbuf.c
  1.115     +8 -15     src/sys/kern/uipc_mbuf.c
  1.19      +22 -1     src/sys/kern/uipc_mbuf2.c
  1.82      +0 -8      src/sys/net/if_loop.c
  1.39      +5 -2      src/sys/sys/mac.h
  1.38      +3 -0      src/sys/sys/mac_policy.h
  1.120     +1 -2      src/sys/sys/mbuf.h