mirror of
git://git.yoctoproject.org/poky.git
synced 2025-07-19 21:09:03 +02:00
89d1788412
It turns out the codeparser cache is the bottleneck I've been observing when running bitbake commands, particularly as it grows. There are some things we can do about this: * We were processing the cache with "intern()" at save time. Its actually much more memory efficient to do this at creation time. * Use hashable objects such as frozenset rather than set so that we can compare objects * De-duplicate the cache objects, link duplicates to the same object saving memory and disk usage and improving speed * Using custom setstate/getstate to avoid the overhead of object attribute names in the cache file To make this work, a global cache was needed for the list of set objects as this was the only way I could find to get the data in at setstate object creation time :(. Parsing shows a modest improvement with these changes, cache load time is significantly better, cache save time is reduced since there is now no need to reprocess the data and cache is much smaller. We can drop the compress_keys() code and internSet code from the shared cache core since its no longer used and replaced by codeparser specific pieces. (Bitbake rev: 4aaf56bfbad4aa626be8a2f7a5f70834c3311dd3) Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
407 lines
14 KiB
Python
407 lines
14 KiB
Python
import ast
|
|
import codegen
|
|
import logging
|
|
import os.path
|
|
import bb.utils, bb.data
|
|
from itertools import chain
|
|
from pysh import pyshyacc, pyshlex, sherrors
|
|
from bb.cache import MultiProcessCache
|
|
|
|
|
|
logger = logging.getLogger('BitBake.CodeParser')
|
|
|
|
try:
|
|
import cPickle as pickle
|
|
except ImportError:
|
|
import pickle
|
|
logger.info('Importing cPickle failed. Falling back to a very slow implementation.')
|
|
|
|
|
|
def check_indent(codestr):
|
|
"""If the code is indented, add a top level piece of code to 'remove' the indentation"""
|
|
|
|
i = 0
|
|
while codestr[i] in ["\n", "\t", " "]:
|
|
i = i + 1
|
|
|
|
if i == 0:
|
|
return codestr
|
|
|
|
if codestr[i-1] == "\t" or codestr[i-1] == " ":
|
|
return "if 1:\n" + codestr
|
|
|
|
return codestr
|
|
|
|
|
|
# Basically pickle, in python 2.7.3 at least, does badly with data duplication
|
|
# upon pickling and unpickling. Combine this with duplicate objects and things
|
|
# are a mess.
|
|
#
|
|
# When the sets are originally created, python calls intern() on the set keys
|
|
# which significantly improves memory usage. Sadly the pickle/unpickle process
|
|
# doesn't call intern() on the keys and results in the same strings being duplicated
|
|
# in memory. This also means pickle will save the same string multiple times in
|
|
# the cache file.
|
|
#
|
|
# By having shell and python cacheline objects with setstate/getstate, we force
|
|
# the object creation through our own routine where we can call intern (via internSet).
|
|
#
|
|
# We also use hashable frozensets and ensure we use references to these so that
|
|
# duplicates can be removed, both in memory and in the resulting pickled data.
|
|
#
|
|
# By playing these games, the size of the cache file shrinks dramatically
|
|
# meaning faster load times and the reloaded cache files also consume much less
|
|
# memory. Smaller cache files, faster load times and lower memory usage is good.
|
|
#
|
|
# A custom getstate/setstate using tuples is actually worth 15% cachesize by
|
|
# avoiding duplication of the attribute names!
|
|
|
|
class SetCache(object):
|
|
def __init__(self):
|
|
self.setcache = {}
|
|
|
|
def internSet(self, items):
|
|
|
|
new = []
|
|
for i in items:
|
|
new.append(intern(i))
|
|
s = frozenset(new)
|
|
if hash(s) in self.setcache:
|
|
return self.setcache[hash(s)]
|
|
self.setcache[hash(s)] = s
|
|
return s
|
|
|
|
codecache = SetCache()
|
|
|
|
class pythonCacheLine(object):
|
|
def __init__(self, refs, execs, contains):
|
|
self.refs = codecache.internSet(refs)
|
|
self.execs = codecache.internSet(execs)
|
|
self.contains = {}
|
|
for c in contains:
|
|
self.contains[c] = codecache.internSet(contains[c])
|
|
|
|
def __getstate__(self):
|
|
return (self.refs, self.execs, self.contains)
|
|
|
|
def __setstate__(self, state):
|
|
(refs, execs, contains) = state
|
|
self.__init__(refs, execs, contains)
|
|
def __hash__(self):
|
|
l = (hash(self.refs), hash(self.execs))
|
|
for c in sorted(self.contains.keys()):
|
|
l = l + (c, hash(self.contains[c]))
|
|
return hash(l)
|
|
|
|
class shellCacheLine(object):
|
|
def __init__(self, execs):
|
|
self.execs = codecache.internSet(execs)
|
|
|
|
def __getstate__(self):
|
|
return (self.execs)
|
|
|
|
def __setstate__(self, state):
|
|
(execs) = state
|
|
self.__init__(execs)
|
|
def __hash__(self):
|
|
return hash(self.execs)
|
|
|
|
class CodeParserCache(MultiProcessCache):
|
|
cache_file_name = "bb_codeparser.dat"
|
|
CACHE_VERSION = 7
|
|
|
|
def __init__(self):
|
|
MultiProcessCache.__init__(self)
|
|
self.pythoncache = self.cachedata[0]
|
|
self.shellcache = self.cachedata[1]
|
|
self.pythoncacheextras = self.cachedata_extras[0]
|
|
self.shellcacheextras = self.cachedata_extras[1]
|
|
|
|
# To avoid duplication in the codeparser cache, keep
|
|
# a lookup of hashes of objects we already have
|
|
self.pythoncachelines = {}
|
|
self.shellcachelines = {}
|
|
|
|
def newPythonCacheLine(self, refs, execs, contains):
|
|
cacheline = pythonCacheLine(refs, execs, contains)
|
|
h = hash(cacheline)
|
|
if h in self.pythoncachelines:
|
|
return self.pythoncachelines[h]
|
|
self.pythoncachelines[h] = cacheline
|
|
return cacheline
|
|
|
|
def newShellCacheLine(self, execs):
|
|
cacheline = shellCacheLine(execs)
|
|
h = hash(cacheline)
|
|
if h in self.shellcachelines:
|
|
return self.shellcachelines[h]
|
|
self.shellcachelines[h] = cacheline
|
|
return cacheline
|
|
|
|
def init_cache(self, d):
|
|
MultiProcessCache.init_cache(self, d)
|
|
|
|
# cachedata gets re-assigned in the parent
|
|
self.pythoncache = self.cachedata[0]
|
|
self.shellcache = self.cachedata[1]
|
|
|
|
def create_cachedata(self):
|
|
data = [{}, {}]
|
|
return data
|
|
|
|
codeparsercache = CodeParserCache()
|
|
|
|
def parser_cache_init(d):
|
|
codeparsercache.init_cache(d)
|
|
|
|
def parser_cache_save(d):
|
|
codeparsercache.save_extras(d)
|
|
|
|
def parser_cache_savemerge(d):
|
|
codeparsercache.save_merge(d)
|
|
|
|
Logger = logging.getLoggerClass()
|
|
class BufferedLogger(Logger):
|
|
def __init__(self, name, level=0, target=None):
|
|
Logger.__init__(self, name)
|
|
self.setLevel(level)
|
|
self.buffer = []
|
|
self.target = target
|
|
|
|
def handle(self, record):
|
|
self.buffer.append(record)
|
|
|
|
def flush(self):
|
|
for record in self.buffer:
|
|
self.target.handle(record)
|
|
self.buffer = []
|
|
|
|
class PythonParser():
|
|
getvars = (".getVar", ".appendVar", ".prependVar")
|
|
containsfuncs = ("bb.utils.contains", "base_contains", "oe.utils.contains", "bb.utils.contains_any")
|
|
execfuncs = ("bb.build.exec_func", "bb.build.exec_task")
|
|
|
|
def warn(self, func, arg):
|
|
"""Warn about calls of bitbake APIs which pass a non-literal
|
|
argument for the variable name, as we're not able to track such
|
|
a reference.
|
|
"""
|
|
|
|
try:
|
|
funcstr = codegen.to_source(func)
|
|
argstr = codegen.to_source(arg)
|
|
except TypeError:
|
|
self.log.debug(2, 'Failed to convert function and argument to source form')
|
|
else:
|
|
self.log.debug(1, self.unhandled_message % (funcstr, argstr))
|
|
|
|
def visit_Call(self, node):
|
|
name = self.called_node_name(node.func)
|
|
if name and name.endswith(self.getvars) or name in self.containsfuncs:
|
|
if isinstance(node.args[0], ast.Str):
|
|
varname = node.args[0].s
|
|
if name in self.containsfuncs and isinstance(node.args[1], ast.Str):
|
|
if varname not in self.contains:
|
|
self.contains[varname] = set()
|
|
self.contains[varname].add(node.args[1].s)
|
|
else:
|
|
self.references.add(node.args[0].s)
|
|
else:
|
|
self.warn(node.func, node.args[0])
|
|
elif name in self.execfuncs:
|
|
if isinstance(node.args[0], ast.Str):
|
|
self.var_execs.add(node.args[0].s)
|
|
else:
|
|
self.warn(node.func, node.args[0])
|
|
elif name and isinstance(node.func, (ast.Name, ast.Attribute)):
|
|
self.execs.add(name)
|
|
|
|
def called_node_name(self, node):
|
|
"""Given a called node, return its original string form"""
|
|
components = []
|
|
while node:
|
|
if isinstance(node, ast.Attribute):
|
|
components.append(node.attr)
|
|
node = node.value
|
|
elif isinstance(node, ast.Name):
|
|
components.append(node.id)
|
|
return '.'.join(reversed(components))
|
|
else:
|
|
break
|
|
|
|
def __init__(self, name, log):
|
|
self.var_execs = set()
|
|
self.contains = {}
|
|
self.execs = set()
|
|
self.references = set()
|
|
self.log = BufferedLogger('BitBake.Data.PythonParser', logging.DEBUG, log)
|
|
|
|
self.unhandled_message = "in call of %s, argument '%s' is not a string literal"
|
|
self.unhandled_message = "while parsing %s, %s" % (name, self.unhandled_message)
|
|
|
|
def parse_python(self, node):
|
|
h = hash(str(node))
|
|
|
|
if h in codeparsercache.pythoncache:
|
|
self.references = set(codeparsercache.pythoncache[h].refs)
|
|
self.execs = set(codeparsercache.pythoncache[h].execs)
|
|
self.contains = {}
|
|
for i in codeparsercache.pythoncache[h].contains:
|
|
self.contains[i] = set(codeparsercache.pythoncache[h].contains[i])
|
|
return
|
|
|
|
if h in codeparsercache.pythoncacheextras:
|
|
self.references = set(codeparsercache.pythoncacheextras[h].refs)
|
|
self.execs = set(codeparsercache.pythoncacheextras[h].execs)
|
|
self.contains = {}
|
|
for i in codeparsercache.pythoncacheextras[h].contains:
|
|
self.contains[i] = set(codeparsercache.pythoncacheextras[h].contains[i])
|
|
return
|
|
|
|
code = compile(check_indent(str(node)), "<string>", "exec",
|
|
ast.PyCF_ONLY_AST)
|
|
|
|
for n in ast.walk(code):
|
|
if n.__class__.__name__ == "Call":
|
|
self.visit_Call(n)
|
|
|
|
self.execs.update(self.var_execs)
|
|
|
|
codeparsercache.pythoncacheextras[h] = codeparsercache.newPythonCacheLine(self.references, self.execs, self.contains)
|
|
|
|
class ShellParser():
|
|
def __init__(self, name, log):
|
|
self.funcdefs = set()
|
|
self.allexecs = set()
|
|
self.execs = set()
|
|
self.log = BufferedLogger('BitBake.Data.%s' % name, logging.DEBUG, log)
|
|
self.unhandled_template = "unable to handle non-literal command '%s'"
|
|
self.unhandled_template = "while parsing %s, %s" % (name, self.unhandled_template)
|
|
|
|
def parse_shell(self, value):
|
|
"""Parse the supplied shell code in a string, returning the external
|
|
commands it executes.
|
|
"""
|
|
|
|
h = hash(str(value))
|
|
|
|
if h in codeparsercache.shellcache:
|
|
self.execs = set(codeparsercache.shellcache[h].execs)
|
|
return self.execs
|
|
|
|
if h in codeparsercache.shellcacheextras:
|
|
self.execs = set(codeparsercache.shellcacheextras[h].execs)
|
|
return self.execs
|
|
|
|
self._parse_shell(value)
|
|
self.execs = set(cmd for cmd in self.allexecs if cmd not in self.funcdefs)
|
|
|
|
codeparsercache.shellcacheextras[h] = codeparsercache.newShellCacheLine(self.execs)
|
|
|
|
return self.execs
|
|
|
|
def _parse_shell(self, value):
|
|
try:
|
|
tokens, _ = pyshyacc.parse(value, eof=True, debug=False)
|
|
except pyshlex.NeedMore:
|
|
raise sherrors.ShellSyntaxError("Unexpected EOF")
|
|
|
|
for token in tokens:
|
|
self.process_tokens(token)
|
|
|
|
def process_tokens(self, tokens):
|
|
"""Process a supplied portion of the syntax tree as returned by
|
|
pyshyacc.parse.
|
|
"""
|
|
|
|
def function_definition(value):
|
|
self.funcdefs.add(value.name)
|
|
return [value.body], None
|
|
|
|
def case_clause(value):
|
|
# Element 0 of each item in the case is the list of patterns, and
|
|
# Element 1 of each item in the case is the list of commands to be
|
|
# executed when that pattern matches.
|
|
words = chain(*[item[0] for item in value.items])
|
|
cmds = chain(*[item[1] for item in value.items])
|
|
return cmds, words
|
|
|
|
def if_clause(value):
|
|
main = chain(value.cond, value.if_cmds)
|
|
rest = value.else_cmds
|
|
if isinstance(rest, tuple) and rest[0] == "elif":
|
|
return chain(main, if_clause(rest[1]))
|
|
else:
|
|
return chain(main, rest)
|
|
|
|
def simple_command(value):
|
|
return None, chain(value.words, (assign[1] for assign in value.assigns))
|
|
|
|
token_handlers = {
|
|
"and_or": lambda x: ((x.left, x.right), None),
|
|
"async": lambda x: ([x], None),
|
|
"brace_group": lambda x: (x.cmds, None),
|
|
"for_clause": lambda x: (x.cmds, x.items),
|
|
"function_definition": function_definition,
|
|
"if_clause": lambda x: (if_clause(x), None),
|
|
"pipeline": lambda x: (x.commands, None),
|
|
"redirect_list": lambda x: ([x.cmd], None),
|
|
"subshell": lambda x: (x.cmds, None),
|
|
"while_clause": lambda x: (chain(x.condition, x.cmds), None),
|
|
"until_clause": lambda x: (chain(x.condition, x.cmds), None),
|
|
"simple_command": simple_command,
|
|
"case_clause": case_clause,
|
|
}
|
|
|
|
for token in tokens:
|
|
name, value = token
|
|
try:
|
|
more_tokens, words = token_handlers[name](value)
|
|
except KeyError:
|
|
raise NotImplementedError("Unsupported token type " + name)
|
|
|
|
if more_tokens:
|
|
self.process_tokens(more_tokens)
|
|
|
|
if words:
|
|
self.process_words(words)
|
|
|
|
def process_words(self, words):
|
|
"""Process a set of 'words' in pyshyacc parlance, which includes
|
|
extraction of executed commands from $() blocks, as well as grabbing
|
|
the command name argument.
|
|
"""
|
|
|
|
words = list(words)
|
|
for word in list(words):
|
|
wtree = pyshlex.make_wordtree(word[1])
|
|
for part in wtree:
|
|
if not isinstance(part, list):
|
|
continue
|
|
|
|
if part[0] in ('`', '$('):
|
|
command = pyshlex.wordtree_as_string(part[1:-1])
|
|
self._parse_shell(command)
|
|
|
|
if word[0] in ("cmd_name", "cmd_word"):
|
|
if word in words:
|
|
words.remove(word)
|
|
|
|
usetoken = False
|
|
for word in words:
|
|
if word[0] in ("cmd_name", "cmd_word") or \
|
|
(usetoken and word[0] == "TOKEN"):
|
|
if "=" in word[1]:
|
|
usetoken = True
|
|
continue
|
|
|
|
cmd = word[1]
|
|
if cmd.startswith("$"):
|
|
self.log.debug(1, self.unhandled_template % cmd)
|
|
elif cmd == "eval":
|
|
command = " ".join(word for _, word in words[1:])
|
|
self._parse_shell(command)
|
|
else:
|
|
self.allexecs.add(cmd)
|
|
break
|