VPP-221 CLI auto-documentation infrastructure

[Note: This is an amalgamation of several patches cherry-picked from
master all related to introducing auto-CLI documentation. See commits
d4024f58 ee4743ad 16bcf7d8 e0d802bc 54ccf226]

As a step before Doxygen, extract CLI-related struct initializers
from the code and parse that into a summary of the CLI commands
available with the provided help text, such as it is. At the moment
this only renders this into an indexed Markdown file that Doxygen
then picks up but later we can use this information to enrich the
existing VLIB_CLI_COMMAND macro documentor as well as provide
runtime documentation to VPP that is stored on disk outside the
binary image.

Additionally support a comment block immediately prior to
VLIB_CLI_COMMAND CLI command definitions in the form /*? ... ?*/
that can be used to include long-form documentation without having
it compiled into VPP.

Examples of documenting CLI commands can be found in
vlib/vlib/unix/cli.c which, whilst not perfect, should provide a
starting point. Screen captures of sample output can be seen at
https://chrisy.flirble.org/vpp/doxy-cli-example.png and
https://chrisy.flirble.org/vpp/doxy-cli-index.png .

Next, shift the Doxygen root makefile targets to their own Makefile.
The primary reason for this is that the siphon targets do dependency
tracking which means it needs to generate those dependencies whenever
make is run; that is pointless if we're not going to generate any
documentation. This includes the package dependencies since they since
they sometimes unnecessarily interfere with the code build in some cases
at the moment; later we will look to building a Python venv to host the
Python modules we use.

One final remark: In future we may consider deprecating .long_help
in the VLIB_CLI_COMMAND structure entirely but add perhaps .usage_help.
.short_help would be reserved for a summary of the command function
and .usage_help provide the syntax of that command. These changes would
provide great semantic value to the automaticly generated CLI
documentation. I could also see having .long_help replaced by a
mechanism that reads it from disk at runtime with a rudimentary
Markdown/Doxygen filter so that we can use the same text that is used in
the published documentation.

Change-Id: I8afccfd7fe2096411d8064ac954ca5a2cbca9ae7
Signed-off-by: Chris Luke <chrisy@flirble.org>

12 files changed, 1095 insertions, 81 deletions

diff --git a/.gitignore b/.gitignore
index 425261836aa..07668dafbc8 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,4 +1,5 @@
 *~
+*.sw[op]
 
 /build-root/.ccache
 /build-root/build-*/
@@ -66,6 +67,8 @@ GTAGS
 
 # Generated documentation
 /build-root/docs
+/build-root/.doxygen-bootstrap.ok
+/build-root/.doxygen-siphon.dep
 
 # indent backup files
 *.BAK
diff --git a/Makefile b/Makefile
index 4e3d65bcd02..20d88da4476 100644
--- a/Makefile
+++ b/Makefile
@@ -30,7 +30,7 @@ OS_VERSION_ID= $(shell grep '^VERSION_ID=' /etc/os-release | cut -f2- -d= | sed
 DEB_DEPENDS  = curl build-essential autoconf automake bison libssl-dev ccache
 DEB_DEPENDS += debhelper dkms git libtool libganglia1-dev libapr1-dev dh-systemd
 DEB_DEPENDS += libconfuse-dev git-review exuberant-ctags cscope
-DEB_DEPENDS += doxygen graphviz
+DEB_DEPENDS += python-dev
 ifeq ($(OS_VERSION_ID),14.04)
 	DEB_DEPENDS += openjdk-8-jdk-headless
 else
@@ -40,7 +40,7 @@ endif
 RPM_DEPENDS_GROUPS = 'Development Tools'
 RPM_DEPENDS  = redhat-lsb glibc-static java-1.8.0-openjdk-devel yum-utils
 RPM_DEPENDS += openssl-devel https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm apr-devel
-RPM_DEPENDS += doxygen graphviz
+RPM_DEPENDS += python-devel
 EPEL_DEPENDS = libconfuse-devel ganglia-devel
 
 ifneq ($(wildcard $(STARTUP_DIR)/startup.conf),)
@@ -54,7 +54,7 @@ endif
 
 .PHONY: help bootstrap wipe wipe-release build build-release rebuild rebuild-release
 .PHONY: run run-release debug debug-release build-vat run-vat pkg-deb pkg-rpm
-.PHONY: ctags cscope doxygen wipe-doxygen plugins plugins-release
+.PHONY: ctags cscope plugins plugins-release build-vpp-api
 
 help:
 	@echo "Make Targets:"
@@ -80,6 +80,7 @@ help:
 	@echo " gtags               - (re)generate gtags database"
 	@echo " cscope              - (re)generate cscope database"
 	@echo " doxygen             - (re)generate documentation"
+	@echo " bootstrap-doxygen   - setup Doxygen dependencies"
 	@echo " wipe-doxygen        - wipe all generated documentation"
 	@echo ""
 	@echo "Make Arguments:"
@@ -233,24 +234,21 @@ cscope: cscope.files
 # Build the documentation
 #
 
-DOXY_INPUT ?= \
-	README.md \
-	vppinfra \
-	svm \
-	vlib \
-	vlib-api \
-	vnet \
-	vpp \
-	vpp-api
+# Doxygen configuration and our utility scripts
+export DOXY_DIR ?= $(WS_ROOT)/doxygen
+
+define make-doxy
+	@OS_ID="$(OS_ID)" WS_ROOT="$(WS_ROOT)" BR="$(BR)" make -C $(DOXY_DIR) $@
+endef
+
+.PHONY: bootstrap-doxygen doxygen wipe-doxygen
+
+bootstrap-doxygen:
+	$(call make-doxy)
 
 doxygen:
-	@mkdir -p "$(BR)/docs"
-	ROOT="$(WS_ROOT)" \
-	     BUILD_ROOT="$(BR)" \
-	     INPUT="$(addprefix $(WS_ROOT)/,$(DOXY_INPUT))" \
-	     HTML=YES \
-	     VERSION="`git describe --tags --dirty`" \
-	     doxygen doxygen/doxygen.cfg
+	$(call make-doxy)
 
 wipe-doxygen:
-	rm -rf "$(BR)/docs"
+	$(call make-doxy)
+
diff --git a/doxygen/Makefile b/doxygen/Makefile
new file mode 100644
index 00000000000..471b6fd7f15
--- /dev/null
+++ b/doxygen/Makefile
@@ -0,0 +1,154 @@
+# Copyright (c) 2016 Comcast Cable Communications Management, LLC.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+#
+# Build the documentation
+#
+
+# These should be passed in by the root Makefile
+WS_ROOT ?= $(CURDIR)/..
+BR ?= $(WS_ROOT)/build-root
+OS_ID ?= $(shell grep '^ID=' /etc/os-release | cut -f2- -d= | sed -e 's/\"//g')
+
+# Package dependencies
+DOC_DEB_DEPENDS = doxygen graphviz python-pyparsing
+DOC_RPM_DEPENDS = doxygen graphviz pyparsing
+
+# Doxygen configuration and our utility scripts
+DOXY_DIR ?= $(WS_ROOT)/doxygen
+
+# Input directories and files
+DOXY_INPUT ?= \
+	README.md \
+	vppinfra \
+	svm \
+	vlib \
+	vlib-api \
+	vnet \
+	vpp \
+	vpp-api \
+	plugins
+
+# Files to exclude, from pre-Doxygen steps, eg because they're
+# selectively compiled.
+# Examples would be to exclude non-DPDK related sources when
+# there's a DPDK equivalent that conflicts.
+# This is specifically for the pre-Doxygen steps; Doxygen uses
+# @cond for this instead.
+DOXY_PRE_EXCLUDE ?= \
+	vlib/vlib/buffer.c
+
+# Generate a regexp for filenames to exclude
+DOXY_PRE_EXCLUDE_REGEXP = ($(subst .,\.,$(shell echo '$(strip $(DOXY_PRE_EXCLUDE))' | sed -e 's/ /|/g')))
+
+# Discover all the directories we might, possibly, maybe, have include files in
+DOXY_INCLUDE_PATH = $(shell set -e; cd $(WS_ROOT); for item in $(DOXY_INPUT); do find $$item -type d; done)
+
+# Discover if we have CPP available
+CPP ?= $(shell which cpp)
+ifneq ($(strip $(CPP)),)
+# Add whatever directories CPP normally includes
+DOXY_INCLUDE_PATH += $(shell set -e; $(CPP) -v </dev/null 2>&1 | grep -A 1000 '\#include' | awk '/^ /{print $$1}')
+endif
+
+# Target directory for doxygen output
+DOXY_OUTPUT ?= $(BR)/docs
+
+# Siphoned fragments end up in here
+SIPHON_INPUT ?= $(DOXY_OUTPUT)/siphon_fragments
+
+# Siphoned fragements are processed into here
+SIPHON_OUTPUT ?= $(DOXY_OUTPUT)/siphon_docs
+
+# Extra document inputs that are processed in addition to DOXY_INPUT
+EXTRA_DOXY_INPUT += $(SIPHON_OUTPUT)
+
+# All the siphon types we know about
+SIPHONS ?= clicmd
+
+SIPHON_FILES = $(addprefix $(SIPHON_INPUT)/,$(addsuffix .siphon,$(SIPHONS)))
+SIPHON_DOCS = $(addprefix $(SIPHON_OUTPUT)/,$(addsuffix .md,$(SIPHONS)))
+
+$(BR)/.doxygen-bootstrap.ok:
+ifeq ($(OS_ID),ubuntu)
+	@set -e; inst=; \
+		for i in $(DOC_DEB_DEPENDS); do \
+			dpkg-query --show $$i >/dev/null 2>&1 || inst="$$inst $$i"; \
+		done; \
+		if [ "$$inst" ]; then sudo apt-get $(CONFIRM) $(FORCE) install $$inst; fi
+else ifneq ("$(wildcard /etc/redhat-release)","")
+	@sudo yum install $(CONFIRM) $(DOC_RPM_DEPENDS)
+else
+	$(error "This option currently works only on Ubuntu or Centos systems")
+endif
+	@touch $@
+
+.PHONY: bootstrap-doxygen
+bootstrap-doxygen: $(BR)/.doxygen-bootstrap.ok
+
+.DELETE_ON_ERROR: $(BR)/.doxygen-siphon.dep
+$(BR)/.doxygen-siphon.dep: Makefile
+	set -e; rm -f "$@"; for input in $(DOXY_INPUT); do \
+		find "$(WS_ROOT)/$$input" -type f \
+			\( -name '*.[ch]' -or -name '*.dox' \) -print \
+			| grep -v -E '^$(WS_ROOT)/$(DOXY_PRE_EXCLUDE_REGEXP)$$' \
+			| sed -e "s/^/\$$(SIPHON_FILES): /" \
+			>> $@; \
+	done
+
+# Include the source -> siphon dependencies
+-include $(BR)/.doxygen-siphon.dep
+
+.NOTPARALLEL: $(SIPHON_FILES)
+$(SIPHON_FILES): $(DOXY_DIR)/siphon_generate.py $(BR)/.doxygen-bootstrap.ok
+	@rm -rf "$(SIPHON_INPUT)" "$(SIPHON_OUTPUT)"
+	@mkdir -p "$(SIPHON_INPUT)" "$(SIPHON_OUTPUT)"
+	@touch $(SIPHON_INPUT)/files
+	for input in $(DOXY_INPUT); do \
+		cd "$(WS_ROOT)"; \
+		find "$$input" -type f \
+			\( -name '*.[ch]' -or -name '*.dox' \) -print \
+			| grep -v -E '^$(DOXY_PRE_EXCLUDE_REGEXP)$$' \
+			>> $(SIPHON_INPUT)/files; \
+	done
+	set -e; cd "$(WS_ROOT)"; $(DOXY_DIR)/siphon_generate.py \
+		--output="$(SIPHON_INPUT)" \
+		"@$(SIPHON_INPUT)/files"
+
+
+.DELETE_ON_ERROR: $(SIPHON_DOCS)
+$(SIPHON_OUTPUT)/%.md: $(SIPHON_INPUT)/%.siphon $(DOXY_DIR)/siphon_process.py
+	set -e; cd "$(WS_ROOT)"; \
+		$(DOXY_DIR)/siphon_process.py --type=$(basename $(notdir $<)) \
+			--output="$(SIPHON_OUTPUT)" $< > $@
+
+# This target can be used just to generate the siphoned docs
+.PHONY: doxygen-siphon
+doxygen-siphon: $(SIPHON_DOCS)
+
+# Generate the doxygen docs
+doxygen: $(SIPHON_DOCS)
+	@mkdir -p "$(DOXY_OUTPUT)"
+	set -e; cd "$(WS_ROOT)"; \
+	    ROOT="$(WS_ROOT)" \
+	    BUILD_ROOT="$(BR)" \
+	    INPUT="$(addprefix $(WS_ROOT)/,$(DOXY_INPUT)) $(EXTRA_DOXY_INPUT)" \
+	    INCLUDE_PATH="$(DOXY_INCLUDE_PATH)" \
+	    HTML=YES \
+	    VERSION="`git describe --tags --dirty`" \
+	    doxygen $(DOXY_DIR)/doxygen.cfg
+
+wipe-doxygen:
+	rm -rf "$(BR)/docs" "$(BR)/.doxygen-siphon.d"
+
diff --git a/doxygen/dir.dox.sample b/doxygen/dir.dox.sample
index 41e84d7bee4..500fe595ae5 100644
--- a/doxygen/dir.dox.sample
+++ b/doxygen/dir.dox.sample
@@ -18,7 +18,7 @@
 
 /**
 @dir
-@brief Someone please fix this description
+@brief Someone please fix this description.
 @todo This directory needs a description.
 
 This is where you would document the contents of a directory.
@@ -26,3 +26,4 @@ This is where you would document the contents of a directory.
 This looks like a C file but it is not part of the build; it is purely
 for documentation.
 */
+/*? %%clicmd:group_label CLI section description%% ?*/
diff --git a/doxygen/doxygen.cfg b/doxygen/doxygen.cfg
index 0eadbd7397c..971a159502c 100644
--- a/doxygen/doxygen.cfg
+++ b/doxygen/doxygen.cfg
@@ -229,8 +229,22 @@ TAB_SIZE               = 8
 # newlines.
 
 ALIASES                =
+
+## Indexes VPP graph nodes
 ALIASES += "node=@xrefitem nodes \"Node Identifier\" \"Node Identifiers\" @c "
 
+## Formatting for CLI commands and output
+ALIASES += "cli{1}=<code><pre>\1</code></pre>"
+ALIASES += "clistart=<code><pre>"
+ALIASES += "cliend=</pre></code>"
+
+## Formatting for CLI example paragraphs
+ALIASES += "cliexpar=@par Example usage"
+ALIASES += "cliexcmd{1}=@clistart<b>vpp# <em>\1</em></b>@cliend"
+ALIASES += "cliexstart{1}=@cliexcmd{\1}@clistart"
+ALIASES += "cliexend=@cliend"
+
+
 # This tag can be used to specify a number of word-keyword mappings (TCL only).
 # A mapping has the form "name=value". For example adding "class=itcl::class"
 # will allow you to use the command class in the itcl::class meaning.
@@ -630,7 +644,7 @@ GENERATE_DEPRECATEDLIST= YES
 # sections, marked by \if <section_label> ... \endif and \cond <section_label>
 # ... \endcond blocks.
 
-ENABLED_SECTIONS       =
+ENABLED_SECTIONS       = DPDK
 
 # The MAX_INITIALIZER_LINES tag determines the maximum number of lines that the
 # initial value of a variable or macro / define can have for it to appear in the
@@ -901,6 +915,7 @@ INPUT_FILTER           =
 
 FILTER_PATTERNS        = \
 	*.c=$(ROOT)/doxygen/filter_c.py \
+	*.h=$(ROOT)/doxygen/filter_h.py \
 	*.api=$(ROOT)/doxygen/filter_api.py
 
 # If the FILTER_SOURCE_FILES tag is set to YES, the input filter (if set using
@@ -2022,7 +2037,7 @@ SEARCH_INCLUDES        = YES
 # preprocessor.
 # This tag requires that the tag SEARCH_INCLUDES is set to YES.
 
-INCLUDE_PATH           = $(INPUT)
+INCLUDE_PATH           = $(INCLUDE_PATH)
 
 # You can use the INCLUDE_FILE_PATTERNS tag to specify one or more wildcard
 # patterns (like *.h and *.hpp) to filter out the header-files in the
@@ -2046,7 +2061,8 @@ PREDEFINED             = \
 	__ORDER_LITTLE_ENDIAN__=1234 \
 	__BYTE_ORDER__=1234 \
 	__FLOAT_WORD_ORDER__=1234 \
-	DPDK=1
+	DPDK=1 \
+	always_inline:="static inline"
 	
 # If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then this
 # tag can be used to specify a list of macro names that should be expanded. The
diff --git a/doxygen/filter_c.py b/doxygen/filter_c.py
index db1681c9450..30b933ba79e 100755
--- a/doxygen/filter_c.py
+++ b/doxygen/filter_c.py
@@ -15,38 +15,73 @@
 
 # Filter for .c files to make various preprocessor tricks Doxygenish
 
-import sys, re
+import os, sys, re
 
 if len(sys.argv) < 2:
     sys.stderr.write("Usage: %s <filename>\n" % (sys.argv[0]))
     sys.exit(1)
 
-patterns = [
-    # Search for VLIB_CLI_COMMAND, extract its parameter and add a docblock for it
-    ( re.compile("(?P<m>VLIB_CLI_COMMAND)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"), r"/** @brief (@em constructor) \g<m> (\g<name>) */ vlib_cli_command_t \g<name>"),
+replace_patterns = [
+    # Search for VLIB_CLI_COMMAND, extract its parameters and add a docblock for it
+    ( re.compile("(?P<m>VLIB_CLI_COMMAND)\s*[(](?P<name>[a-zA-Z0-9_]+)[)]"),
+        r"/** @brief (@em constructor) \g<m> (\g<name>) */ vlib_cli_command_t \g<name>"),
+    ( re.compile("(?P<m>VLIB_CLI_COMMAND)\s*[(](?P<name>[a-zA-Z0-9_]+),\s*(?P<qual>[^)]*)[)]"),
+        r"/** @brief (@em constructor) \g<m> (\g<name>) */ \g<qual> vlib_cli_command_t \g<name>"),
 
-    # Search for VLIB_REGISTER_NODE, extract its parameter and add a docblock for it
-    ( re.compile("(?P<m>VLIB_REGISTER_NODE)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"), r"/** @brief (@em constructor) \g<m> (\g<name>) */ vlib_node_registration_t \g<name>"),
+    # Search for VLIB_REGISTER_NODE, extract its parameters and add a docblock for it
+    ( re.compile("(?P<m>VLIB_REGISTER_NODE)\s*[(](?P<name>[a-zA-Z0-9_]+)[)]"),
+        r"/** @brief (@em constructor) \g<m> (\g<name>) */ vlib_node_registration_t \g<name>"),
+    ( re.compile("(?P<m>VLIB_REGISTER_NODE)\s*[(](?P<name>[a-zA-Z0-9_]+),\s*(?P<qual>[^)]*)[)]"),
+        r"/** @brief (@em constructor) \g<m> (\g<name>) */ \g<qual> vlib_node_registration_t \g<name>"),
 
     # Search for VLIB_INIT_FUNCTION, extract its parameter and add a docblock for it
-    ( re.compile("(?P<m>VLIB_INIT_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)[)]"), r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ vlib_init_function_t * _vlib_init_function_\g<name>"),
-    ( re.compile("(?P<m>VLIB_DECLARE_INIT_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)[)]"), r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ vlib_init_function_t * _vlib_init_function_\g<name>"),
+    ( re.compile("(?P<m>VLIB_INIT_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)[)]"),
+        r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ vlib_init_function_t * _vlib_init_function_\g<name>"),
+    ( re.compile("(?P<m>VLIB_DECLARE_INIT_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)[)]"),
+        r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ vlib_init_function_t * _vlib_init_function_\g<name>"),
 
-    # Search for VLIB_LOOP_ENTER_FUNCTION, extract the 1st parameter (ignore any others) and add a docblock for it
-    ( re.compile("(?P<m>VLIB_MAIN_LOOP_ENTER_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"), r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ _vlib_main_loop_enter_\g<name>"),
-    ( re.compile("(?P<m>VLIB_MAIN_LOOP_EXIT_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"), r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ _vlib_main_loop_exit_\g<name>"),
+    # Search for VLIB_LOOP_ENTER_FUNCTION, extract the parameters and add a docblock for it
+    ( re.compile("(?P<m>VLIB_MAIN_LOOP_ENTER_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"),
+        r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ _vlib_main_loop_enter_\g<name>"),
+    ( re.compile("(?P<m>VLIB_MAIN_LOOP_EXIT_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"),
+        r"/** @brief (@em constructor) \g<m> (@ref \g<name>) */ _vlib_main_loop_exit_\g<name>"),
 
-    # Search for VLIB_CONFIG_FUNCTION, extract the 1st parameter (ignore any others) and add a docblock for it
-    ( re.compile("(?P<m>VLIB_CONFIG_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+),\s*(?P<n>\"[^\"]+\")(,[^)]*)?[)]"), r"/** @brief (@em constructor) \g<m> (\g<name>, \g<n>) */ vlib_config_function_runtime_t _vlib_config_function_\g<name>"),
-    ( re.compile("(?P<m>VLIB_EARLY_CONFIG_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+),\s*(?P<n>\"[^\"]+\")(,[^)]*)?[)]"), r"/** @brief (@em constructor) \g<m> (\g<name>, \g<n>) */ vlib_config_function_runtime_t _vlib_config_function_\g<name>"),
+    # Search for VLIB_CONFIG_FUNCTION, extract the parameters and add a docblock for it
+    ( re.compile("(?P<m>VLIB_CONFIG_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+),\s*(?P<n>\"[^\"]+\")(,[^)]*)?[)]"),
+        r"/** @brief (@em constructor) \g<m> (\g<name>, \g<n>) */ vlib_config_function_runtime_t _vlib_config_function_\g<name>"),
+    ( re.compile("(?P<m>VLIB_EARLY_CONFIG_FUNCTION)\s*[(](?P<name>[a-zA-Z0-9_]+),\s*(?P<n>\"[^\"]+\")(,[^)]*)?[)]"),
+        r"/** @brief (@em constructor) \g<m> (\g<name>, \g<n>) */ vlib_config_function_runtime_t _vlib_config_function_\g<name>"),
 
     # Search for "format_thing" and "unformat_thing" when used as a function pointer and add parens
-    ( re.compile("(?P<pre>(^|,)\s*)(?P<name>(un)?format_[a-zA-Z0-9_]+)(?P<post>\s*(,|$))") , r"\g<pre>\g<name>()\g<post>" ),
+    ( re.compile("(?P<pre>(^|,)\s*)(?P<name>(un)?format_[a-zA-Z0-9_]+)(?P<post>\s*(,|$))"),
+        r"\g<pre>\g<name>()\g<post>" ),
+
+    # Search for CLIB_PAD_FROM_TO(...); and replace with padding
+    # #define CLIB_PAD_FROM_TO(from,to) u8 pad_##from[(to) - (from)]
+    ( re.compile("(?P<m>CLIB_PAD_FROM_TO)\s*[(](?P<from>[^,]+),\s*(?P<to>[^)]+)[)]"),
+        r"/** Padding. */ u8 pad_\g<from>[(\g<to>) - (\g<from>)]" ),
+
 ]
 
-with open(sys.argv[1]) as fd:
+
+filename = sys.argv[1]
+cwd = os.getcwd()
+if filename[0:len(cwd)] == cwd:
+    filename = filename[len(cwd):]
+    if filename[0] == "/":
+        filename = filename[1:]
+
+with open(filename) as fd:
+    line_num = 0
+
     for line in fd:
+        line_num += 1
         str = line[:-1] # filter \n
-        for p in patterns:
+
+        # Look for search/replace patterns
+        for p in replace_patterns:
             str = p[0].sub(p[1], str)
+
         sys.stdout.write(str+"\n")
+
+# All done
diff --git a/doxygen/filter_h.py b/doxygen/filter_h.py
new file mode 100755
index 00000000000..967388d5743
--- /dev/null
+++ b/doxygen/filter_h.py
@@ -0,0 +1,53 @@
+#!/usr/bin/env python
+# Copyright (c) 2016 Comcast Cable Communications Management, LLC.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Filter for .c files to make various preprocessor tricks Doxygenish
+
+import os, sys, re
+
+if len(sys.argv) < 2:
+    sys.stderr.write("Usage: %s <filename>\n" % (sys.argv[0]))
+    sys.exit(1)
+
+replace_patterns = [
+    # Search for CLIB_PAD_FROM_TO(...); and replace with padding
+    # #define CLIB_PAD_FROM_TO(from,to) u8 pad_##from[(to) - (from)]
+    ( re.compile("(?P<m>CLIB_PAD_FROM_TO)\s*[(](?P<from>[^,]+),\s*(?P<to>[^)]+)[)]"),
+        r"/** Padding. */ u8 pad_\g<from>[(\g<to>) - (\g<from>)]" ),
+
+]
+
+
+filename = sys.argv[1]
+cwd = os.getcwd()
+if filename[0:len(cwd)] == cwd:
+    filename = filename[len(cwd):]
+    if filename[0] == "/":
+        filename = filename[1:]
+
+with open(filename) as fd:
+    line_num = 0
+
+    for line in fd:
+        line_num += 1
+        str = line[:-1] # filter \n
+
+        # Look for search/replace patterns
+        for p in replace_patterns:
+            str = p[0].sub(p[1], str)
+
+        sys.stdout.write(str+"\n")
+
+# All done
diff --git a/doxygen/siphon_generate.py b/doxygen/siphon_generate.py
new file mode 100755
index 00000000000..8b999114e52
--- /dev/null
+++ b/doxygen/siphon_generate.py
@@ -0,0 +1,322 @@
+#!/usr/bin/env python
+# Copyright (c) 2016 Comcast Cable Communications Management, LLC.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Looks for preprocessor macros with struct initializers and siphons them
+# off into another file for later parsing; ostensibly to generate
+# documentation from struct initializer data.
+
+import os, sys, re, argparse, json
+
+DEFAULT_OUTPUT = "build-root/docs/siphons"
+DEFAULT_PREFIX = os.getcwd()
+
+ap = argparse.ArgumentParser()
+ap.add_argument("--output", '-o', metavar="directory", default=DEFAULT_OUTPUT,
+        help="Output directory for .siphon files [%s]" % DEFAULT_OUTPUT)
+ap.add_argument("--input-prefix", metavar="path", default=DEFAULT_PREFIX,
+        help="Prefix to strip from input pathnames [%s]" % DEFAULT_PREFIX)
+ap.add_argument("input", nargs='+', metavar="input_file",
+        help="Input C source files")
+args = ap.parse_args()
+
+"""Patterns that match the start of code blocks we want to siphon"""
+siphon_patterns = [
+    ( re.compile("(?P<m>VLIB_CLI_COMMAND)\s*[(](?P<name>[a-zA-Z0-9_]+)(,[^)]*)?[)]"), "clicmd" ),
+]
+
+"""Matches a siphon comment block start"""
+siphon_block_start = re.compile("^\s*/\*\?\s*(.*)$")
+
+"""Matches a siphon comment block stop"""
+siphon_block_stop = re.compile("^(.*)\s*\?\*/\s*$")
+
+"""Siphon block directive delimiter"""
+siphon_block_delimiter = "%%"
+
+"""Matches a siphon block directive such as '%clicmd:group_label Debug CLI%'"""
+siphon_block_directive = re.compile("(%s)\s*([a-zA-Z0-9_:]+)\s+(.*)\s*(%s)" % \
+        (siphon_block_delimiter, siphon_block_delimiter))
+
+"""Matches the start of an initializer block"""
+siphon_initializer = re.compile("\s*=")
+
+"""
+count open and close braces in str
+return (0, index) when braces were found and count becomes 0.
+index indicates the position at which the last closing brace was
+found.
+return (-1, -1) if a closing brace is found before any opening one.
+return (count, -1) if not all opening braces are closed, count is the
+current depth
+"""
+def count_braces(str, count=0, found=False):
+    for index in range(0, len(str)):
+        if str[index] == '{':
+            count += 1;
+            found = True
+        elif str[index] == '}':
+            if count == 0:
+                # means we never found an open brace
+                return (-1, -1)
+            count -= 1;
+
+        if count == 0 and found:
+            return (count, index)
+
+    return (count, -1)
+
+# Collated output for each siphon
+output = {}
+
+# Build a list of known siphons
+known_siphons = []
+for item in siphon_patterns:
+	siphon = item[1]
+	if siphon not in known_siphons:
+		known_siphons.append(siphon)
+
+# Setup information for siphons we know about
+for siphon in known_siphons:
+	output[siphon] = {
+            "file": "%s/%s.siphon" % (args.output, siphon),
+            "global": {},
+            "items": [],
+        }
+
+# Pre-process file names in case they indicate a file with
+# a list of files
+files = []
+for filename in args.input:
+    if filename.startswith('@'):
+        with open(filename[1:], 'r') as fp:
+            lines = fp.readlines()
+            for line in lines:
+                files.append(line.strip())
+            lines = None
+    else:
+        files.append(filename)
+
+# Iterate all the input files we've been given
+for filename in files:
+    # Strip the current directory off the start of the
+    # filename for brevity
+    if filename[0:len(args.input_prefix)] == args.input_prefix:
+        filename = filename[len(args.input_prefix):]
+        if filename[0] == "/":
+            filename = filename[1:]
+
+    # Work out the abbreviated directory name
+    directory = os.path.dirname(filename)
+    if directory[0:2] == "./":
+        directory = directory[2:]
+    elif directory[0:len(args.input_prefix)] == args.input_prefix:
+        directory = directory[len(args.input_prefix):]
+    if directory[0] == "/":
+    	directory = directory[1:]
+
+    # Open the file and explore its contents...
+    sys.stderr.write("Siphoning from %s...\n" % filename)
+    directives = {}
+    with open(filename) as fd:
+        siphon = None
+        close_siphon = None
+        siphon_block = ""
+        in_block = False
+        line_num = 0
+        siphon_line = 0
+
+        for line in fd:
+            line_num += 1
+            str = line[:-1] # filter \n
+
+            """See if there is a block directive and if so extract it"""
+            def process_block_directive(str, directives):
+                m = siphon_block_directive.search(str)
+                if m is not None:
+                    k = m.group(2)
+                    v = m.group(3).strip()
+                    directives[k] = v
+                    # Return only the parts we did not match
+                    return str[0:m.start(1)] + str[m.end(4):]
+
+                return str
+
+            def process_block_prefix(str):
+                if str.startswith(" * "):
+                    str = str[3:]
+                elif str == " *":
+                    str = ""
+                return str
+                
+            if not in_block:
+                # See if the line contains the start of a siphon doc block
+                m = siphon_block_start.search(str)
+                if m is not None:
+                    in_block = True
+                    t = m.group(1)
+
+                    # Now check if the block closes on the same line
+                    m = siphon_block_stop.search(t)
+                    if m is not None:
+                        t = m.group(1)
+                        in_block = False
+
+                    # Check for directives
+                    t = process_block_directive(t, directives)
+
+                    # Filter for normal comment prefixes
+                    t = process_block_prefix(t)
+
+                    # Add what is left
+                    siphon_block += t
+
+                    # Skip to next line
+                    continue
+
+            else:
+                # Check to see if we have an end block marker
+                m = siphon_block_stop.search(str)
+                if m is not None:
+                    in_block = False
+                    t = m.group(1)
+                else:
+                    t = str
+
+                # Check for directives
+                t = process_block_directive(t, directives)
+
+                # Filter for normal comment prefixes
+                t = process_block_prefix(t)
+
+                # Add what is left
+                siphon_block += t + "\n"
+
+                # Skip to next line
+                continue
+
+
+            if siphon is None:
+                # Look for blocks we need to siphon
+                for p in siphon_patterns:
+                    if p[0].match(str):
+                        siphon = [ p[1], str + "\n", 0 ]
+                        siphon_line = line_num
+
+                        # see if we have an initializer
+                        m = siphon_initializer.search(str)
+                        if m is not None:
+                            # count the braces on this line
+                            (count, index) = count_braces(str[m.start():])
+                            siphon[2] = count
+                            # TODO - it's possible we have the initializer all on the first line
+                            # we should check for it, but also account for the possibility that
+                            # the open brace is on the next line
+                            #if count == 0:
+                            #    # braces balanced
+                            #    close_siphon = siphon
+                            #    siphon = None
+                        else:
+                            # no initializer: close the siphon right now
+                            close_siphon = siphon
+                            siphon = None
+            else:
+                # See if we should end the siphon here - do we have balanced
+                # braces?
+                (count, index) = count_braces(str, count=siphon[2], found=True)
+                if count == 0:
+                    # braces balanced - add the substring and close the siphon
+                    siphon[1] += str[:index+1] + ";\n"
+                    close_siphon = siphon
+                    siphon = None
+                else:
+                    # add the whole string, move on
+                    siphon[2] = count
+                    siphon[1] += str + "\n"
+
+            if close_siphon is not None:
+                # Write the siphoned contents to the right place
+                siphon_name = close_siphon[0]
+
+                # Copy directives for the file
+                details = {}
+                for key in directives:
+                    if ":" in key:
+                        (sn, label) = key.split(":")
+                        if sn == siphon_name:
+                            details[label] = directives[key]
+                    else:
+                        details[key] = directives[key]
+
+                # Copy details for this block
+                details['file'] = filename
+                details['line_start'] = siphon_line
+                details['line_end'] = line_num
+                details['siphon_block'] = siphon_block.strip()
+
+                # Some defaults
+                if "group" not in details:
+                    if "group_label" in details:
+                        # use the filename since group labels are mostly of file scope
+                        details['group'] = details['file']
+                    else:
+			details['group'] = directory
+
+                if "group_label" not in details:
+                    details['group_label'] = details['group']
+
+                details["block"] = close_siphon[1]
+
+                # Store the item
+                output[siphon_name]['items'].append(details)
+
+                # All done
+                close_siphon = None
+                siphon_block = ""
+
+        # Update globals
+        for key in directives.keys():
+            if ':' not in key:
+                continue
+
+            if filename.endswith("/dir.dox"):
+                # very special! use the parent directory name
+                l = directory
+            else:
+                l = filename
+
+            (sn, label) = key.split(":")
+
+            if sn not in output:
+                output[sn] = {}
+            if 'global' not in output[sn]:
+                output[sn]['global'] = {}
+            if l not in output[sn]['global']:
+                output[sn]['global'][l] = {}
+            if 'file' not in output[sn]:
+                output[sn]['file'] = "%s/%s.siphon" % (args.output, sn)
+            if 'items' not in output[sn]:
+                output[sn]['items'] = []
+
+            output[sn]['global'][l][label] = directives[key]
+
+
+# Write out the data
+for siphon in output.keys():
+    sys.stderr.write("Saving siphon %s...\n" % siphon)
+    s = output[siphon]
+    with open(s['file'], "a") as fp:
+        json.dump(s, fp, separators=(',', ': '), indent=4, sort_keys=True)
+
+# All done
diff --git a/doxygen/siphon_process.py b/doxygen/siphon_process.py
new file mode 100755
index 00000000000..82a166d31b5
--- /dev/null
+++ b/doxygen/siphon_process.py
@@ -0,0 +1,323 @@
+#!/usr/bin/env python
+# Copyright (c) 2016 Comcast Cable Communications Management, LLC.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Filter for .siphon files that are generated by other filters.
+# The idea is to siphon off certain initializers so that we can better
+# auto-document the contents of that initializer.
+
+import os, sys, re, argparse, cgi, json
+import pyparsing as pp
+
+import pprint
+
+DEFAULT_SIPHON ="clicmd"
+DEFAULT_OUTPUT = None
+DEFAULT_PREFIX = os.getcwd()
+
+siphon_map = {
+    'clicmd': "VLIB_CLI_COMMAND",
+}
+
+ap = argparse.ArgumentParser()
+ap.add_argument("--type", '-t', metavar="siphon_type", default=DEFAULT_SIPHON,
+        choices=siphon_map.keys(),
+        help="Siphon type to process [%s]" % DEFAULT_SIPHON)
+ap.add_argument("--output", '-o', metavar="directory", default=DEFAULT_OUTPUT,
+        help="Output directory for .md files [%s]" % DEFAULT_OUTPUT)
+ap.add_argument("--input-prefix", metavar="path", default=DEFAULT_PREFIX,
+        help="Prefix to strip from input pathnames [%s]" % DEFAULT_PREFIX)
+ap.add_argument("input", nargs='+', metavar="input_file",
+        help="Input .siphon files")
+args = ap.parse_args()
+
+if args.output is None:
+    sys.stderr.write("Error: Siphon processor requires --output to be set.")
+    sys.exit(1)
+
+
+def clicmd_index_sort(cfg, group, dec):
+    if group in dec and 'group_label' in dec[group]:
+        return dec[group]['group_label']
+    return group
+
+def clicmd_index_header(cfg):
+    s = "# CLI command index\n"
+    s += "\n[TOC]\n"
+    return s
+
+def clicmd_index_section(cfg, group, md):
+    return "\n@subpage %s\n\n" % md
+
+def clicmd_index_entry(cfg, meta, item):
+    v = item["value"]
+    return "* [%s](@ref %s)\n" % (v["path"], meta["label"])
+
+def clicmd_sort(cfg, meta, item):
+    return item['value']['path']
+
+def clicmd_header(cfg, group, md, dec):
+    if group in dec and 'group_label' in dec[group]:
+        label = dec[group]['group_label']
+    else:
+        label = group
+    return "\n@page %s %s\n" % (md, label)
+
+def clicmd_format(cfg, meta, item):
+    v = item["value"]
+    s = "\n@section %s %s\n" % (meta['label'], v['path'])
+
+    # The text from '.short_help = '.
+    # Later we should split this into short_help and usage_help
+    # since the latter is how it is primarily used but the former
+    # is also needed.
+    if "short_help" in v:
+        tmp = v["short_help"].strip()
+
+        # Bit hacky. Add a trailing period if it doesn't have one.
+        if tmp[-1] != ".":
+            tmp += "."
+
+        s += "### Summary/usage\n    %s\n\n" % tmp
+
+    # This is seldom used and will likely be deprecated
+    if "long_help" in v:
+        tmp = v["long_help"]
+
+        s += "### Long help\n    %s\n\n" % tmp
+
+    # Extracted from the code in /*? ... ?*/ blocks
+    if "siphon_block" in item["meta"]:
+        sb = item["meta"]["siphon_block"]
+
+        if sb != "":
+            # hack. still needed?
+            sb = sb.replace("\n", "\\n")
+            try:
+                sb = json.loads('"'+sb+'"')
+                s += "### Description\n%s\n\n" % sb
+            except:
+                pass
+
+    # Gives some developer-useful linking
+    if "item" in meta or "function" in v:
+        s += "### Declaration and implementation\n\n"
+
+        if "item" in meta:
+            s += "Declaration: @ref %s (%s:%d)\n\n" % \
+                (meta['item'], meta["file"], int(item["meta"]["line_start"]))
+
+        if "function" in v:
+            s += "Implementation: @ref %s.\n\n" % v["function"]
+
+    return s
+
+
+siphons = {
+    "VLIB_CLI_COMMAND": {
+        "index_sort_key": clicmd_index_sort,
+        "index_header": clicmd_index_header,
+        "index_section": clicmd_index_section,
+        "index_entry": clicmd_index_entry,
+        'sort_key': clicmd_sort,
+        "header": clicmd_header,
+        "format": clicmd_format,
+    }
+}
+
+
+# PyParsing definition for our struct initializers which look like this:
+# VLIB_CLI_COMMAND (show_sr_tunnel_command, static) = {
+#    .path = "show sr tunnel",
+#    .short_help = "show sr tunnel [name <sr-tunnel-name>]",
+#    .function = show_sr_tunnel_fn,
+#};
+def getMacroInitializerBNF():
+    cs = pp.Forward()
+    ident = pp.Word(pp.alphas + "_", pp.alphas + pp.nums + "_")
+    intNum = pp.Word(pp.nums)
+    hexNum = pp.Literal("0x") + pp.Word(pp.hexnums)
+    octalNum = pp.Literal("0") + pp.Word("01234567")
+    integer = (hexNum | octalNum | intNum) + \
+        pp.Optional(pp.Literal("ULL") | pp.Literal("LL") | pp.Literal("L"))
+    floatNum = pp.Regex(r'\d+(\.\d*)?([eE]\d+)?') + pp.Optional(pp.Literal("f"))
+    char = pp.Literal("'") + pp.Word(pp.printables, exact=1) + pp.Literal("'")
+    arrayIndex = integer | ident
+
+    lbracket = pp.Literal("(").suppress()
+    rbracket = pp.Literal(")").suppress()
+    lbrace = pp.Literal("{").suppress()
+    rbrace = pp.Literal("}").suppress()
+    comma = pp.Literal(",").suppress()
+    equals = pp.Literal("=").suppress()
+    dot = pp.Literal(".").suppress()
+    semicolon = pp.Literal(";").suppress()
+
+    # initializer := { [member = ] (variable | expression | { initializer } ) }
+    typeName = ident
+    varName = ident
+
+    typeSpec = pp.Optional("unsigned") + \
+               pp.oneOf("int long short float double char u8 i8 void") + \
+               pp.Optional(pp.Word("*"), default="")
+    typeCast = pp.Combine( "(" + ( typeSpec | typeName ) + ")" ).suppress()
+
+    string = pp.Combine(pp.OneOrMore(pp.QuotedString(quoteChar='"',
+        escChar='\\', multiline=True)), adjacent=False)
+    literal = pp.Optional(typeCast) + (integer | floatNum | char | string)
+    var = pp.Combine(pp.Optional(typeCast) + varName + pp.Optional("[" + arrayIndex + "]"))
+
+    expr = (literal | var) # TODO
+
+
+    member = pp.Combine(dot + varName + pp.Optional("[" + arrayIndex + "]"), adjacent=False)
+    value = (expr | cs)
+
+    entry = pp.Group(pp.Optional(member + equals, default="") + value)
+    entries = (pp.ZeroOrMore(entry + comma) + entry + pp.Optional(comma)) | \
+              (pp.ZeroOrMore(entry + comma))
+
+    cs << (lbrace + entries + rbrace)
+
+    macroName = ident
+    params = pp.Group(pp.ZeroOrMore(expr + comma) + expr)
+    macroParams = lbracket + params + rbracket
+
+    mi = macroName + pp.Optional(macroParams) + equals + pp.Group(cs) + semicolon
+    mi.ignore(pp.cppStyleComment)
+    return mi
+
+
+mi = getMacroInitializerBNF()
+
+# Parse the input file into a more usable dictionary structure
+cmds = {}
+line_num = 0
+line_start = 0
+for filename in args.input:
+    sys.stderr.write("Parsing items in file \"%s\"...\n" % filename)
+    data = None
+    with open(filename, "r") as fd:
+        data = json.load(fd)
+
+    cmds['_global'] = data['global']
+
+    # iterate the items loaded and regroup it
+    for item in data["items"]:
+        try:
+            o = mi.parseString(item['block']).asList()
+        except:
+            sys.stderr.write("Exception parsing item: %s\n%s\n" \
+                    % (json.dumps(item, separators=(',', ': '), indent=4),
+                        item['block']))
+            raise
+
+        group = item['group']
+        file = item['file']
+        macro = o[0]
+        param = o[1][0]
+
+        if group not in cmds:
+            cmds[group] = {}
+
+        if file not in cmds[group]:
+            cmds[group][file] = {}
+
+        if macro not in cmds[group][file]:
+            cmds[group][file][macro] = {}
+
+        c = {
+            'params': o[2],
+            'meta': {},
+            'value': {},
+        }
+
+        for key in item:
+            if key == 'block':
+                continue
+            c['meta'][key] = item[key]
+
+        for i in c['params']:
+            c['value'][i[0]] = cgi.escape(i[1])
+
+        cmds[group][file][macro][param] = c
+
+
+# Write the header for this siphon type
+cfg = siphons[siphon_map[args.type]]
+sys.stdout.write(cfg["index_header"](cfg))
+contents = ""
+
+def group_sort_key(item):
+    if "index_sort_key" in cfg:
+        return cfg["index_sort_key"](cfg, item, cmds['_global'])
+    return item
+
+# Iterate the dictionary and process it
+for group in sorted(cmds.keys(), key=group_sort_key):
+    if group.startswith('_'):
+        continue
+
+    sys.stderr.write("Processing items in group \"%s\"...\n" % group)
+
+    cfg = siphons[siphon_map[args.type]]
+    md = group.replace("/", "_").replace(".", "_")
+    sys.stdout.write(cfg["index_section"](cfg, group, md))
+
+    if "header" in cfg:
+        dec = cmds['_global']
+        contents += cfg["header"](cfg, group, md, dec)
+
+    for file in sorted(cmds[group].keys()):
+        if group.startswith('_'):
+            continue
+
+        sys.stderr.write("- Processing items in file \"%s\"...\n" % file)
+
+        for macro in sorted(cmds[group][file].keys()):
+            if macro != siphon_map[args.type]:
+                continue
+            sys.stderr.write("-- Processing items in macro \"%s\"...\n" % macro)
+            cfg = siphons[macro]
+
+            meta = {
+                "group": group,
+                "file": file,
+                "macro": macro,
+                "md": md,
+            }
+
+            def item_sort_key(item):
+                if "sort_key" in cfg:
+                    return cfg["sort_key"](cfg, meta, cmds[group][file][macro][item])
+                return item
+
+            for param in sorted(cmds[group][file][macro].keys(), key=item_sort_key):
+                sys.stderr.write("--- Processing item \"%s\"...\n" % param)
+
+                meta["item"] = param
+
+                # mangle "md" and the item to make a reference label
+                meta["label"] = "%s___%s" % (meta["md"], param)
+
+                if "index_entry" in cfg:
+                    s = cfg["index_entry"](cfg, meta, cmds[group][file][macro][param])
+                    sys.stdout.write(s)
+
+                if "format" in cfg:
+                    contents += cfg["format"](cfg, meta, cmds[group][file][macro][param])
+
+sys.stdout.write(contents)
+
+# All done
diff --git a/vlib/vlib/dir.dox b/vlib/vlib/dir.dox
index 8ca47cd79ef..4806e7a91c6 100644
--- a/vlib/vlib/dir.dox
+++ b/vlib/vlib/dir.dox
@@ -19,3 +19,5 @@
 @dir
 @brief VLIB application library source.
 */
+/*? %%clicmd:group_label VLIB application library%% ?*/
+
diff --git a/vlib/vlib/unix/cli.c b/vlib/vlib/unix/cli.c
index 92bb8bc3945..bf09ee0cce2 100644
--- a/vlib/vlib/unix/cli.c
+++ b/vlib/vlib/unix/cli.c
@@ -42,6 +42,7 @@
  * Provides a command line interface so humans can interact with VPP.
  * This is predominantly a debugging and testing mechanism.
  */
+/*? %%clicmd:group_label Debug CLI %% ?*/
 
 #include <vlib/vlib.h>
 #include <vlib/unix/unix.h>
@@ -146,9 +147,13 @@ typedef struct
      CLI process. */
   u8 *input_vector;
 
+  /** This session has command history. */
   u8 has_history;
+  /** Array of vectors of commands in the history. */
   u8 **command_history;
+  /** The command currently pointed at by the history cursor. */
   u8 *current_command;
+  /** How far from the end of the history array the user has browsed. */
   i32 excursion;
 
   /** Maximum number of history entries this session will store. */
@@ -157,7 +162,12 @@ typedef struct
   /** Current command line counter */
   u32 command_number;
 
+  /** The string being searched for in the history. */
   u8 *search_key;
+  /** If non-zero then the CLI is searching in the history array.
+   * - @c -1 means search backwards.
+   * - @c 1 means search forwards.
+   */
   int search_mode;
 
   /** Position of the insert cursor on the current input line */
@@ -232,41 +242,41 @@ unix_cli_file_free (unix_cli_file_t * f)
 /** CLI actions */
 typedef enum
 {
-  UNIX_CLI_PARSE_ACTION_NOACTION = 0,  /**< No action */
-  UNIX_CLI_PARSE_ACTION_CRLF,	       /**< Carriage return, newline or enter */
-  UNIX_CLI_PARSE_ACTION_TAB,	       /**< Tab key */
-  UNIX_CLI_PARSE_ACTION_ERASE,	       /**< Erase cursor left */
-  UNIX_CLI_PARSE_ACTION_ERASERIGHT,    /**< Erase cursor right */
-  UNIX_CLI_PARSE_ACTION_UP,	       /**< Up arrow */
-  UNIX_CLI_PARSE_ACTION_DOWN,	       /**< Down arrow */
-  UNIX_CLI_PARSE_ACTION_LEFT,
-  UNIX_CLI_PARSE_ACTION_RIGHT,
-  UNIX_CLI_PARSE_ACTION_HOME,
-  UNIX_CLI_PARSE_ACTION_END,
-  UNIX_CLI_PARSE_ACTION_WORDLEFT,
-  UNIX_CLI_PARSE_ACTION_WORDRIGHT,
-  UNIX_CLI_PARSE_ACTION_ERASELINELEFT,
-  UNIX_CLI_PARSE_ACTION_ERASELINERIGHT,
-  UNIX_CLI_PARSE_ACTION_CLEAR,
-  UNIX_CLI_PARSE_ACTION_REVSEARCH,
-  UNIX_CLI_PARSE_ACTION_FWDSEARCH,
-  UNIX_CLI_PARSE_ACTION_YANK,
-  UNIX_CLI_PARSE_ACTION_TELNETIAC,
-
-  UNIX_CLI_PARSE_ACTION_PAGER_CRLF,
-  UNIX_CLI_PARSE_ACTION_PAGER_QUIT,
-  UNIX_CLI_PARSE_ACTION_PAGER_NEXT,
-  UNIX_CLI_PARSE_ACTION_PAGER_DN,
-  UNIX_CLI_PARSE_ACTION_PAGER_UP,
-  UNIX_CLI_PARSE_ACTION_PAGER_TOP,
-  UNIX_CLI_PARSE_ACTION_PAGER_BOTTOM,
-  UNIX_CLI_PARSE_ACTION_PAGER_PGDN,
-  UNIX_CLI_PARSE_ACTION_PAGER_PGUP,
-  UNIX_CLI_PARSE_ACTION_PAGER_REDRAW,
-  UNIX_CLI_PARSE_ACTION_PAGER_SEARCH,
-
-  UNIX_CLI_PARSE_ACTION_PARTIALMATCH,
-  UNIX_CLI_PARSE_ACTION_NOMATCH
+  UNIX_CLI_PARSE_ACTION_NOACTION = 0,	/**< No action */
+  UNIX_CLI_PARSE_ACTION_CRLF,		/**< Carriage return, newline or enter */
+  UNIX_CLI_PARSE_ACTION_TAB,		/**< Tab key */
+  UNIX_CLI_PARSE_ACTION_ERASE,		/**< Erase cursor left */
+  UNIX_CLI_PARSE_ACTION_ERASERIGHT,	/**< Erase cursor right */
+  UNIX_CLI_PARSE_ACTION_UP,		/**< Up arrow */
+  UNIX_CLI_PARSE_ACTION_DOWN,		/**< Down arrow */
+  UNIX_CLI_PARSE_ACTION_LEFT,		/**< Left arrow */
+  UNIX_CLI_PARSE_ACTION_RIGHT,		/**< Right arrow */
+  UNIX_CLI_PARSE_ACTION_HOME,		/**< Home key (jump to start of line) */
+  UNIX_CLI_PARSE_ACTION_END,		/**< End key (jump to end of line) */
+  UNIX_CLI_PARSE_ACTION_WORDLEFT,	/**< Jump cursor to start of left word */
+  UNIX_CLI_PARSE_ACTION_WORDRIGHT,	/**< Jump cursor to start of right word */
+  UNIX_CLI_PARSE_ACTION_ERASELINELEFT,	/**< Erase line to left of cursor */
+  UNIX_CLI_PARSE_ACTION_ERASELINERIGHT,	/**< Erase line to right & including cursor */
+  UNIX_CLI_PARSE_ACTION_CLEAR,		/**< Clear the terminal */
+  UNIX_CLI_PARSE_ACTION_REVSEARCH,	/**< Search backwards in command history */
+  UNIX_CLI_PARSE_ACTION_FWDSEARCH,	/**< Search forwards in command history */
+  UNIX_CLI_PARSE_ACTION_YANK,		/**< Undo last erase action */
+  UNIX_CLI_PARSE_ACTION_TELNETIAC,	/**< Telnet control code */
+
+  UNIX_CLI_PARSE_ACTION_PAGER_CRLF,	/**< Enter pressed (CR, CRLF, LF, etc) */
+  UNIX_CLI_PARSE_ACTION_PAGER_QUIT,	/**< Exit the pager session */
+  UNIX_CLI_PARSE_ACTION_PAGER_NEXT,	/**< Scroll to next page */
+  UNIX_CLI_PARSE_ACTION_PAGER_DN,	/**< Scroll to next line */
+  UNIX_CLI_PARSE_ACTION_PAGER_UP,	/**< Scroll to previous line */
+  UNIX_CLI_PARSE_ACTION_PAGER_TOP,	/**< Scroll to first line */
+  UNIX_CLI_PARSE_ACTION_PAGER_BOTTOM,	/**< Scroll to last line */
+  UNIX_CLI_PARSE_ACTION_PAGER_PGDN,	/**< Scroll to next page */
+  UNIX_CLI_PARSE_ACTION_PAGER_PGUP,	/**< Scroll to previous page */
+  UNIX_CLI_PARSE_ACTION_PAGER_REDRAW,	/**< Clear and redraw the page on the terminal */
+  UNIX_CLI_PARSE_ACTION_PAGER_SEARCH,	/**< Search the pager buffer */
+
+  UNIX_CLI_PARSE_ACTION_PARTIALMATCH,	/**< Action parser found a partial match */
+  UNIX_CLI_PARSE_ACTION_NOMATCH		/**< Action parser did not find any match */
 } unix_cli_parse_action_t;
 
 /** @brief Mapping of input buffer strings to action values.
@@ -485,6 +495,9 @@ unix_cli_match_action (unix_cli_parse_actions_t * a,
 }
 
 
+/** Add bytes to the output vector and then flagg the I/O system that bytes
+ * are available to be sent.
+ */
 static void
 unix_cli_add_pending_output (unix_file_t * uf,
 			     unix_cli_file_t * cf,
@@ -502,6 +515,9 @@ unix_cli_add_pending_output (unix_file_t * uf,
     }
 }
 
+/** Delete all bytes from the output vector and flag the I/O system
+ * that no more bytes are available to be sent.
+ */
 static void
 unix_cli_del_pending_output (unix_file_t * uf,
 			     unix_cli_file_t * cf, uword n_bytes)
@@ -983,13 +999,13 @@ unix_vlib_cli_output (uword cli_file_index, u8 * buffer, uword buffer_bytes)
 
 /** Identify whether a terminal type is ANSI capable.
  *
- * Compares the string given in @term with a list of terminal types known
+ * Compares the string given in @c term with a list of terminal types known
  * to support ANSI escape sequences.
  *
  * This list contains, for example, @c xterm, @c screen and @c ansi.
  *
  * @param term A string with a terminal type in it.
- * @param len The length of the string in @term.
+ * @param len The length of the string in @c term.
  *
  * @return @c 1 if the terminal type is recognized as supporting ANSI
  *         terminal sequences; @c 0 otherwise.
@@ -2059,6 +2075,10 @@ done:
     goto more;
 }
 
+/** Destroy a CLI session.
+ * @note If we destroy the @c stdin session this additionally signals
+ *       the shutdown of VPP.
+ */
 static void
 unix_cli_kill (unix_cli_main_t * cm, uword cli_file_index)
 {
@@ -2088,6 +2108,7 @@ unix_cli_kill (unix_cli_main_t * cm, uword cli_file_index)
   pool_put (cm->cli_file_pool, cf);
 }
 
+/** Handle system events. */
 static uword
 unix_cli_process (vlib_main_t * vm,
 		  vlib_node_runtime_t * rt, vlib_frame_t * f)
@@ -2130,6 +2151,8 @@ done:
   return 0;
 }
 
+/** Called when a CLI session file descriptor can be written to without
+ * blocking. */
 static clib_error_t *
 unix_cli_write_ready (unix_file_t * uf)
 {
@@ -2152,6 +2175,7 @@ unix_cli_write_ready (unix_file_t * uf)
   return /* no error */ 0;
 }
 
+/** Called when a CLI session file descriptor has data to be read. */
 static clib_error_t *
 unix_cli_read_ready (unix_file_t * uf)
 {
@@ -2482,8 +2506,8 @@ unix_cli_config (vlib_main_t * vm, unformat_input_t * input)
 
 VLIB_CONFIG_FUNCTION (unix_cli_config, "unix-cli");
 
-/** Called when VPP is shutting down, this resets the system
- * terminal state, if previously saved.
+/** Called when VPP is shutting down, this restores the system
+ * terminal state if previously saved.
  */
 static clib_error_t *
 unix_cli_exit (vlib_main_t * vm)
@@ -2500,7 +2524,7 @@ unix_cli_exit (vlib_main_t * vm)
 VLIB_MAIN_LOOP_EXIT_FUNCTION (unix_cli_exit);
 
 /** Set the CLI prompt.
- * @param The C string to set the prompt to.
+ * @param prompt The C string to set the prompt to.
  * @note This setting is global; it impacts all current
  *       and future CLI sessions.
  */
@@ -2531,6 +2555,12 @@ unix_cli_quit (vlib_main_t * vm,
   return 0;
 }
 
+/*?
+ * Terminates the current CLI session.
+ *
+ * If VPP is running in @em interactive mode and this is the console session
+ * (that is, the session on @c stdin) then this will also terminate VPP.
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (unix_cli_quit_command, static) = {
   .path = "quit",
@@ -2597,6 +2627,13 @@ done:
   return error;
 }
 
+/*?
+ * Executes a sequence of CLI commands which are read from a file.
+ *
+ * If a command is unrecognised or otherwise invalid then the usual CLI
+ * feedback will be generated, however execution of subsequent commands
+ * from the file will continue.
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (cli_exec, static) = {
   .path = "exec",
@@ -2706,6 +2743,9 @@ unix_cli_show_history (vlib_main_t * vm,
   return 0;
 }
 
+/*?
+ * Displays the command history for the current session, if any.
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (cli_unix_cli_show_history, static) = {
   .path = "history",
@@ -2755,6 +2795,24 @@ unix_cli_show_terminal (vlib_main_t * vm,
   return 0;
 }
 
+/*?
+ * Displays various information about the state of the current terminal
+ * session.
+ *
+ * @cliexpar
+ * @cliexstart{show terminal}
+ * Terminal name:   unix-cli-stdin
+ * Terminal mode:   char-by-char
+ * Terminal width:  123
+ * Terminal height: 48
+ * ANSI capable:    yes
+ * History enabled: yes
+ * History limit:   50
+ * Pager enabled:   yes
+ * Pager limit:     100000
+ * CRLF mode:       LF
+ * @cliexend
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (cli_unix_cli_show_terminal, static) = {
   .path = "show terminal",
@@ -2799,6 +2857,13 @@ unix_cli_set_terminal_pager (vlib_main_t * vm,
   return 0;
 }
 
+/*?
+ * Enables or disables the terminal pager for this session. Generally
+ * this defaults to enabled.
+ *
+ * Additionally allows the pager buffer size to be set; though note that
+ * this value is set globally and not per session.
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (cli_unix_cli_set_terminal_pager, static) = {
   .path = "set terminal pager",
@@ -2850,6 +2915,13 @@ unix_cli_set_terminal_history (vlib_main_t * vm,
   return 0;
 }
 
+/*?
+ * Enables or disables the command history function of the current
+ * terminal. Generally this defaults to enabled.
+ *
+ * This command also allows the maximum size of the history buffer for
+ * this session to be altered.
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (cli_unix_cli_set_terminal_history, static) = {
   .path = "set terminal history",
@@ -2880,6 +2952,14 @@ unix_cli_set_terminal_ansi (vlib_main_t * vm,
   return 0;
 }
 
+/*?
+ * Enables or disables the use of ANSI control sequences by this terminal.
+ * The default will vary based on terminal detection at the start of the
+ * session.
+ *
+ * ANSI control sequences are used in a small number of places to provide,
+ * for example, color text output and to control the cursor in the pager.
+?*/
 /* *INDENT-OFF* */
 VLIB_CLI_COMMAND (cli_unix_cli_set_terminal_ansi, static) = {
   .path = "set terminal ansi",
diff --git a/vlib/vlib/unix/dir.dox b/vlib/vlib/unix/dir.dox
new file mode 100644
index 00000000000..cdded0f19d3
--- /dev/null
+++ b/vlib/vlib/unix/dir.dox
@@ -0,0 +1,27 @@
+/*
+ * Copyright (c) 2016 Comcast Cable Communications Management, LLC.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/* Doxygen directory documentation */
+
+/**
+@dir
+@brief VLIB Unix interface
+
+VLIB application library Unix interface layer.
+
+*/
+/*? %%clicmd:group_label VLIB Unix stuff%% ?*/
+