Training courses

Kernel and Embedded Linux

Bootlin training courses

Embedded Linux, kernel,
Yocto Project, Buildroot, real-time,
graphics, boot time, debugging...

Bootlin logo

Elixir Cross Referencer

#! /usr/bin/python2
import os
import sys
import shlex
import re

from headerutils import *
import Queue

file_list = list ()
usage = False

ignore_conditional = False

order = [
  "system.h",
  "coretypes.h",
  "backend.h",
  "target.h",
  "rtl.h",
  "c-family/c-target.h",
  "c-family/c-target-def.h",
  "tree.h",
  "cp/cp-tree.h",
  "c-family/c-common.h",  # these must come before diagnostic.h
  "c/c-tree.h",
  "fortran/gfortran.h",
  "gimple.h",
  "cfghooks.h",
  "df.h",
  "tm_p.h",
  "gimple-iterators.h",
  "ssa.h",
  "expmed.h",
  "optabs.h",
  "regs.h",
  "ira.h",
  "ira-int.h",
  "gimple-streamer.h"

]

exclude_special = [  "bversion.h", "obstack.h", "insn-codes.h", "hooks.h" ]

# includes is a dictionary indexed by a header files basename.
# it consists of a 2 element tuple:
# [0] - Name of header file which included this header.
# [1] - vector of header file names included by this file.

includes = { }

# when a header is included multiple times, indexing this dictionary will
# return a vector of all the headers which included it.
dups = { }

# When creating the master list, do not descend into these files for what 
# they include. Simply put the file itself in the list.  This is primarily
# required because the front end files inlcude orders tend to be at odds with
# the order of middle end files, and its impossible to synchronize them.\
# They are ordered such that everything resolves properly.
exclude_processing = [ "tree-vectorizer.h" , "c-target.h", "c-target-def.h", "cp-tree.h", "c-common.h", "c-tree.h", "gfortran.h" ]

master_list = list ()
# where include file comes from in src
h_from = { }

# create the master ordering list... this is the desired order of headers
def create_master_list (fn, verbose):
  if fn not in exclude_processing:
    for x in includes[fn][1]:
      create_master_list (x, verbose)
  if not fn in master_list:
    # Don't put diagnostic*.h into the ordering list. It is special since
    # various front ends have to set GCC_DIAG_STYLE before including it.
    # for each file, we'll tailor where it belongs by looking at the include
    # list and determine its position appropriately.
    if fn != "diagnostic.h" and fn != "diagnostic-core.h":
      master_list.append (fn)
      if (verbose):
        print fn + "      included by: " + includes[fn][0]



def print_dups ():
  if dups:
    print "\nduplicated includes"
  for i in dups:
    string =  "dup : " + i + " : "
    string += includes[i][0] 
    for i2 in dups[i]:
      string += ", "+i2
    print string


def process_known_dups ():
  # rtl.h gets tagged as a duplicate includer for all of coretypes.h, but that
  # is really for only generator files
  rtl_remove = includes["coretypes.h"][1] + ["statistics.h", "vec.h"]
  if dups:
    for i in rtl_remove:
      if dups[i] and "rtl.h" in dups[i]:
        dups[i].remove("rtl.h")
      if not dups[i]:
        dups.pop (i, None)

  # make sure diagnostic.h is the owner of diagnostic-core.h
  if includes["diagnostic-core.h"][0] != "diagnostic.h":
    dups["diagnostic-core.h"].append (includes["diagnostic-core.h"][0])
    includes["diagnostic-core.h"] = ("diagnostic.h", includes["diagnostic-core.h"][1])

# This function scans back thorugh the list of headers which included other
# headers to determine what file in HEADER_LIST brought 'HEADER' in.
def indirectly_included (header, header_list):
  nm = os.path.basename (header)
  while nm and includes.get(nm):
    if includes[nm][0] in header_list:
      return includes[nm][0]
    nm = includes[nm][0]

  # diagnostic.h and diagnostic-core.h may not show up because we removed them
  # from the header list to manually position in an appropriate place. They have
  # specific requirements that they need to occur after certain FE files which
  # may overide the definition of GCC_DIAG_STYLE.
  # Check the dup list for whete they may have been included from and return
  # that header.
  if header == "diagnostic-core.h":
    if dups.get("diagnostic-core.h"):
      for f in dups["diagnostic-core.h"]:
        if f in header_list:
          return f
    else:
      if header in header_list:
        return header
    # Now check if diagnostics is included indirectly anywhere
    header = "diagnostic.h"

  if header == "diagnostic.h":
    if dups.get("diagnostic.h"):
      for f in dups["diagnostic.h"]:
        if f in header_list:
          return f
    else:
      if header in header_list:
        return header 

  return ""


# This function will take a list of headers from a source file and return 
# the desired new new order of the canonical headers in DESIRED_ORDER. 
def get_new_order (src_h, desired_order):
  new_order = list ()
  for h in desired_order:
    if h in master_list:
      # Create the list of nested headers which included this file.
      iclist = list ()
      ib = includes[h][0]
      while ib:
        iclist.insert(0, ib)
        ib = includes[ib][0]
      if iclist:
        for x in iclist:
          # If header is in the source code, and we are allowed to look inside
          if x in src_h and x not in exclude_processing:
            if x not in new_order and x[:10] != "diagnostic" and h not in exclude_special:
              new_order.append (x)
              break;
      else:
        if h not in new_order:
          new_order.append (h)

  f = ""
  if "diagnostic.h" in src_h:
    f = "diagnostic.h"
  elif "diagnostic-core.h" in src_h:
    f = "diagnostic-core.h"

 
  # If either diagnostic header was directly included in the main file, check to
  # see if its already included indirectly, or whether we need to add it to the
  # end of the canonically orders headers.
  if f:
    ii = indirectly_included (f, src_h)
    if not ii or ii == f:
      new_order.append (f)

  return new_order
        
    

# stack of files to process
process_stack = list ()

def process_one (info):
  i = info[0]
  owner = info[1]
  name = os.path.basename(i)
  if os.path.exists (i):
    if includes.get(name) == None:
      l = find_unique_include_list (i)
      # create a list which has just basenames in it
      new_list = list ()
      for x in l:
        new_list.append (os.path.basename (x))
        process_stack.append((x, name))
      includes[name] = (owner, new_list)
    elif owner:
      if dups.get(name) == None:
        dups[name] = [ owner ]
      else:
        dups[name].append (owner)
  else:
    # seed tm.h with options.h since it is a build file and won't be seen. 
    if not includes.get(name):
      if name == "tm.h":
        includes[name] = (owner, [ "options.h" ])
        includes["options.h"] = ("tm.h", list ())
      else:
        includes[name] = (owner, list ())


show_master = False

for arg in sys.argv[1:]:
  if arg[0:1] == "-":
    if arg[0:2] == "-h":
      usage = True
    elif arg[0:2] == "-i":
      ignore_conditional = True
    elif arg[0:2] == "-v":
      show_master = True
    else:
      print "Error: unrecognized option " + arg
  elif os.path.exists(arg):
    file_list.append (arg)
  else:
    print "Error: file " + arg + " Does not exist."
    usage = True

if not file_list and not show_master:
  usage = True

if not usage and not os.path.exists ("coretypes.h"):
  usage = True
  print "Error: Must run command in main gcc source directory containing coretypes.h\n"

# process diagnostic.h first.. it's special since GCC_DIAG_STYLE can be
# overridden by languages, but must be done so by a file included BEFORE it.
# so make sure it isn't seen as included by one of those files by making it 
# appear to be included by the src file.
process_stack.insert (0, ("diagnostic.h", ""))

# Add the list of files in reverse order since it is processed as a stack later
for i in order:
  process_stack.insert (0, (i, "") )

# build up the library of what header files include what other files.
while process_stack:
  info = process_stack.pop ()
  process_one (info)

# Now create the master ordering list
for i in order:
  create_master_list (os.path.basename (i), show_master)

# handle warts in the duplicate list
process_known_dups ()
desired_order = master_list

if show_master:
  print " Canonical order of gcc include files: "
  for x in master_list:
    print x
  print " "

if usage:
  print "gcc-order-headers [-i] [-v] file1 [filen]"
  print "    Ensures gcc's headers files are included in a normalized form with"
  print "    redundant headers removed.  The original files are saved in filename.bak"
  print "    Outputs a list of files which changed."
  print " -i ignore conditional compilation."
  print "    Use after examining the file to be sure includes within #ifs are safe"
  print "    Any headers within conditional sections will be ignored."
  print " -v Show the canonical order of known headers"
  sys.exit(0)


didnt_do = list ()

for fn in file_list:
  nest = 0
  src_h = list ()
  src_line = { }

  master_list = list ()

  includes = { }
  dups = { }

  iinfo = process_ii_src (fn)
  src = ii_src (iinfo)
  include_list = ii_include_list (iinfo)

  if ii_include_list_cond (iinfo):
    if not ignore_conditional:
      print fn + ": Cannot process due to conditional compilation of includes"
      didnt_do.append (fn)
      src = list ()

  if not src:
    continue

  process_stack = list ()
  # prime the stack with headers in the main ordering list so we get them in
  # this order.
  for d in order:
    if d in include_list:
      process_stack.insert (0, (d, ""))

  for d in include_list:
      nm = os.path.basename(d)
      src_h.append (nm)
      iname = d
      iname2 = os.path.dirname (fn) + "/" + d
      if not os.path.exists (d) and os.path.exists (iname2):
        iname = iname2
      if iname not in process_stack:
        process_stack.insert (0, (iname, ""))
      src_line[nm] = ii_src_line(iinfo)[d]
      if src_line[nm].find("/*") != -1 and src_line[nm].find("*/") == -1:
        # this means we have a multi line comment, abort!'
        print fn + ": Cannot process due to a multi-line comment :"
        print "        " + src_line[nm]
        if fn not in didnt_do:
          didnt_do.append (fn)
        src = list ()

  if not src:
    continue

  # Now create the list of includes as seen by the source file.
  while process_stack:
    info = process_stack.pop ()
    process_one (info)
 
  for i in include_list:
    create_master_list (os.path.basename (i), False)

  new_src = list ()
  header_added = list ()
  new_order = list ()
  for line in src:
    d = find_pound_include (line, True, True)
    if not d or d[-2:] != ".h":
      new_src.append (line)
    else:
      if d == order[0] and not new_order:
        new_order = get_new_order (src_h, desired_order)
        for i in new_order:
          new_src.append (src_line[i])
          # if not seen, add it.
          if i not in header_added:
            header_added.append (i)
      else:
        nm = os.path.basename(d)
        if nm not in header_added:
          iby = indirectly_included (nm, src_h)
          if not iby:
            new_src.append (line)
            header_added.append (nm)

  if src != new_src:
    os.rename (fn, fn + ".bak")
    fl = open(fn,"w")
    for line in new_src:
      fl.write (line)
    fl.close ()
    print fn 

 
if didnt_do:
  print "\n\n Did not process the following files due to conditional dependencies:"
  str = ""
  for x in didnt_do:
    str += x + " "
  print str
  print "\n"
  print "Please examine to see if they are safe to process, and re-try with -i. "
  print "Safeness is determined by checking whether any of the reordered headers are"
  print "within a conditional and could be hauled out of the conditional, thus changing"
  print "what the compiler will see."
  print "Multi-line comments after a #include can also cause failuer, they must be turned"
  print "into single line comments or removed."