llvm-project/lldb/examples/synthetic/gnu_libstdcpp.py

481 lines
18 KiB
Python
Raw Normal View History

import re
import lldb.formatters.Logger
# C++ STL formatters for LLDB
# These formatters are based upon the version of the GNU libstdc++
# as it ships with Mac OS X 10.6.8 thru 10.8.0
# You are encouraged to look at the STL implementation for your platform
# before relying on these formatters to do the right thing for your setup
class StdListSynthProvider:
def __init__(self, valobj, dict):
logger = lldb.formatters.Logger.Logger()
self.valobj = valobj
self.count = None
logger >> "Providing synthetic children for a list named " + \
str(valobj.GetName())
def next_node(self, node):
logger = lldb.formatters.Logger.Logger()
return node.GetChildMemberWithName('_M_next')
def is_valid(self, node):
logger = lldb.formatters.Logger.Logger()
valid = self.value(self.next_node(node)) != self.node_address
if valid:
logger >> "%s is valid" % str(self.valobj.GetName())
else:
logger >> "synthetic value is not valid"
return valid
def value(self, node):
logger = lldb.formatters.Logger.Logger()
value = node.GetValueAsUnsigned()
logger >> "synthetic value for {}: {}".format(
str(self.valobj.GetName()), value)
return value
# Floyd's cycle-finding algorithm
# try to detect if this list has a loop
def has_loop(self):
global _list_uses_loop_detector
logger = lldb.formatters.Logger.Logger()
if not _list_uses_loop_detector:
logger >> "Asked not to use loop detection"
return False
slow = self.next
fast1 = self.next
fast2 = self.next
while self.is_valid(slow):
slow_value = self.value(slow)
fast1 = self.next_node(fast2)
fast2 = self.next_node(fast1)
if self.value(fast1) == slow_value or self.value(
fast2) == slow_value:
return True
slow = self.next_node(slow)
return False
def num_children(self):
logger = lldb.formatters.Logger.Logger()
if self.count is None:
# libstdc++ 6.0.21 added dedicated count field.
count_child = self.node.GetChildMemberWithName('_M_data')
if count_child and count_child.IsValid():
self.count = count_child.GetValueAsUnsigned(0)
if self.count is None:
self.count = self.num_children_impl()
return self.count
def num_children_impl(self):
logger = lldb.formatters.Logger.Logger()
try:
next_val = self.next.GetValueAsUnsigned(0)
prev_val = self.prev.GetValueAsUnsigned(0)
# After a std::list has been initialized, both next and prev will
# be non-NULL
if next_val == 0 or prev_val == 0:
return 0
if next_val == self.node_address:
return 0
if next_val == prev_val:
return 1
if self.has_loop():
return 0
size = 2
current = self.next
while current.GetChildMemberWithName(
'_M_next').GetValueAsUnsigned(0) != self.node_address:
size = size + 1
current = current.GetChildMemberWithName('_M_next')
return (size - 1)
except:
return 0
def get_child_index(self, name):
logger = lldb.formatters.Logger.Logger()
try:
return int(name.lstrip('[').rstrip(']'))
except:
return -1
def get_child_at_index(self, index):
logger = lldb.formatters.Logger.Logger()
logger >> "Fetching child " + str(index)
if index < 0:
return None
if index >= self.num_children():
return None
try:
offset = index
current = self.next
while offset > 0:
current = current.GetChildMemberWithName('_M_next')
offset = offset - 1
return current.CreateChildAtOffset(
'[' + str(index) + ']',
2 * current.GetType().GetByteSize(),
self.data_type)
except:
return None
def extract_type(self):
logger = lldb.formatters.Logger.Logger()
list_type = self.valobj.GetType().GetUnqualifiedType()
if list_type.IsReferenceType():
list_type = list_type.GetDereferencedType()
if list_type.GetNumberOfTemplateArguments() > 0:
data_type = list_type.GetTemplateArgumentType(0)
else:
data_type = None
return data_type
def update(self):
logger = lldb.formatters.Logger.Logger()
# preemptively setting this to None - we might end up changing our mind
# later
self.count = None
try:
impl = self.valobj.GetChildMemberWithName('_M_impl')
self.node = impl.GetChildMemberWithName('_M_node')
self.node_address = self.valobj.AddressOf().GetValueAsUnsigned(0)
self.next = self.node.GetChildMemberWithName('_M_next')
self.prev = self.node.GetChildMemberWithName('_M_prev')
self.data_type = self.extract_type()
self.data_size = self.data_type.GetByteSize()
except:
pass
def has_children(self):
return True
class StdVectorSynthProvider:
class StdVectorImplementation(object):
def __init__(self, valobj):
self.valobj = valobj
self.count = None
def num_children(self):
if self.count is None:
self.count = self.num_children_impl()
return self.count
def num_children_impl(self):
try:
start_val = self.start.GetValueAsUnsigned(0)
finish_val = self.finish.GetValueAsUnsigned(0)
end_val = self.end.GetValueAsUnsigned(0)
# Before a vector has been constructed, it will contain bad values
# so we really need to be careful about the length we return since
# uninitialized data can cause us to return a huge number. We need
# to also check for any of the start, finish or end of storage values
# being zero (NULL). If any are, then this vector has not been
# initialized yet and we should return zero
# Make sure nothing is NULL
if start_val == 0 or finish_val == 0 or end_val == 0:
return 0
# Make sure start is less than finish
if start_val >= finish_val:
return 0
# Make sure finish is less than or equal to end of storage
if finish_val > end_val:
return 0
# if we have a struct (or other data type that the compiler pads to native word size)
# this check might fail, unless the sizeof() we get is itself incremented to take the
# padding bytes into account - on current clang it looks like
# this is the case
num_children = (finish_val - start_val)
if (num_children % self.data_size) != 0:
return 0
else:
num_children = num_children / self.data_size
return num_children
except:
return 0
def get_child_at_index(self, index):
logger = lldb.formatters.Logger.Logger()
logger >> "Retrieving child " + str(index)
if index < 0:
return None
if index >= self.num_children():
return None
try:
offset = index * self.data_size
return self.start.CreateChildAtOffset(
'[' + str(index) + ']', offset, self.data_type)
except:
return None
def update(self):
# preemptively setting this to None - we might end up changing our
# mind later
self.count = None
try:
impl = self.valobj.GetChildMemberWithName('_M_impl')
self.start = impl.GetChildMemberWithName('_M_start')
self.finish = impl.GetChildMemberWithName('_M_finish')
self.end = impl.GetChildMemberWithName('_M_end_of_storage')
self.data_type = self.start.GetType().GetPointeeType()
self.data_size = self.data_type.GetByteSize()
# if any of these objects is invalid, it means there is no
# point in trying to fetch anything
if self.start.IsValid() and self.finish.IsValid(
) and self.end.IsValid() and self.data_type.IsValid():
self.count = None
else:
self.count = 0
except:
pass
return True
class StdVBoolImplementation(object):
def __init__(self, valobj, bool_type):
self.valobj = valobj
self.bool_type = bool_type
self.valid = False
def num_children(self):
if self.valid:
start = self.start_p.GetValueAsUnsigned(0)
finish = self.finish_p.GetValueAsUnsigned(0)
offset = self.offset.GetValueAsUnsigned(0)
if finish >= start:
return (finish - start) * 8 + offset
return 0
def get_child_at_index(self, index):
if index >= self.num_children():
return None
element_type = self.start_p.GetType().GetPointeeType()
element_bits = 8 * element_type.GetByteSize()
element_offset = (index / element_bits) * \
element_type.GetByteSize()
bit_offset = index % element_bits
element = self.start_p.CreateChildAtOffset(
'[' + str(index) + ']', element_offset, element_type)
bit = element.GetValueAsUnsigned(0) & (1 << bit_offset)
if bit != 0:
value_expr = "(bool)true"
else:
value_expr = "(bool)false"
return self.valobj.CreateValueFromExpression(
"[%d]" % index, value_expr)
def update(self):
try:
m_impl = self.valobj.GetChildMemberWithName('_M_impl')
self.m_start = m_impl.GetChildMemberWithName('_M_start')
self.m_finish = m_impl.GetChildMemberWithName('_M_finish')
self.start_p = self.m_start.GetChildMemberWithName('_M_p')
self.finish_p = self.m_finish.GetChildMemberWithName('_M_p')
self.offset = self.m_finish.GetChildMemberWithName('_M_offset')
self.valid = True
except:
self.valid = False
return True
def __init__(self, valobj, dict):
logger = lldb.formatters.Logger.Logger()
first_template_arg_type = valobj.GetType().GetTemplateArgumentType(0)
if str(first_template_arg_type.GetName()) == "bool":
self.impl = self.StdVBoolImplementation(
valobj, first_template_arg_type)
else:
self.impl = self.StdVectorImplementation(valobj)
logger >> "Providing synthetic children for a vector named " + \
str(valobj.GetName())
def num_children(self):
return self.impl.num_children()
def get_child_index(self, name):
try:
return int(name.lstrip('[').rstrip(']'))
except:
return -1
def get_child_at_index(self, index):
return self.impl.get_child_at_index(index)
def update(self):
return self.impl.update()
def has_children(self):
return True
class StdMapSynthProvider:
def __init__(self, valobj, dict):
logger = lldb.formatters.Logger.Logger()
self.valobj = valobj
self.count = None
logger >> "Providing synthetic children for a map named " + \
str(valobj.GetName())
# we need this function as a temporary workaround for rdar://problem/10801549
# which prevents us from extracting the std::pair<K,V> SBType out of the template
# arguments for _Rep_Type _M_t in the map itself - because we have to make up the
# typename and then find it, we may hit the situation were std::string has multiple
# names but only one is actually referenced in the debug information. hence, we need
# to replace the longer versions of std::string with the shorter one in order to be able
# to find the type name
def fixup_class_name(self, class_name):
logger = lldb.formatters.Logger.Logger()
if class_name == 'std::basic_string<char, std::char_traits<char>, std::allocator<char> >':
return 'std::basic_string<char>', True
if class_name == 'basic_string<char, std::char_traits<char>, std::allocator<char> >':
return 'std::basic_string<char>', True
if class_name == 'std::basic_string<char, std::char_traits<char>, std::allocator<char> >':
return 'std::basic_string<char>', True
if class_name == 'basic_string<char, std::char_traits<char>, std::allocator<char> >':
return 'std::basic_string<char>', True
return class_name, False
def update(self):
logger = lldb.formatters.Logger.Logger()
# preemptively setting this to None - we might end up changing our mind
# later
self.count = None
try:
# we will set this to True if we find out that discovering a node in the map takes more steps than the overall size of the RB tree
# if this gets set to True, then we will merrily return None for
# any child from that moment on
self.garbage = False
self.Mt = self.valobj.GetChildMemberWithName('_M_t')
self.Mimpl = self.Mt.GetChildMemberWithName('_M_impl')
self.Mheader = self.Mimpl.GetChildMemberWithName('_M_header')
map_type = self.valobj.GetType()
if map_type.IsReferenceType():
logger >> "Dereferencing type"
map_type = map_type.GetDereferencedType()
# Get the type of std::pair<key, value>. It is the first template
# argument type of the 4th template argument to std::map.
allocator_type = map_type.GetTemplateArgumentType(3)
self.data_type = allocator_type.GetTemplateArgumentType(0)
if not self.data_type:
# GCC does not emit DW_TAG_template_type_parameter for
# std::allocator<...>. For such a case, get the type of
# std::pair from a member of std::map.
rep_type = self.valobj.GetChildMemberWithName('_M_t').GetType()
self.data_type = rep_type.GetTypedefedType().GetTemplateArgumentType(1)
# from libstdc++ implementation of _M_root for rbtree
self.Mroot = self.Mheader.GetChildMemberWithName('_M_parent')
self.data_size = self.data_type.GetByteSize()
self.skip_size = self.Mheader.GetType().GetByteSize()
except:
pass
def num_children(self):
logger = lldb.formatters.Logger.Logger()
if self.count is None:
self.count = self.num_children_impl()
return self.count
def num_children_impl(self):
logger = lldb.formatters.Logger.Logger()
try:
root_ptr_val = self.node_ptr_value(self.Mroot)
if root_ptr_val == 0:
return 0
count = self.Mimpl.GetChildMemberWithName(
'_M_node_count').GetValueAsUnsigned(0)
logger >> "I have " + str(count) + " children available"
return count
except:
return 0
def get_child_index(self, name):
logger = lldb.formatters.Logger.Logger()
try:
return int(name.lstrip('[').rstrip(']'))
except:
return -1
def get_child_at_index(self, index):
logger = lldb.formatters.Logger.Logger()
logger >> "Being asked to fetch child[" + str(index) + "]"
if index < 0:
return None
if index >= self.num_children():
return None
if self.garbage:
logger >> "Returning None since we are a garbage tree"
return None
try:
offset = index
current = self.left(self.Mheader)
while offset > 0:
current = self.increment_node(current)
offset = offset - 1
# skip all the base stuff and get at the data
return current.CreateChildAtOffset(
'[' + str(index) + ']', self.skip_size, self.data_type)
except:
return None
# utility functions
def node_ptr_value(self, node):
logger = lldb.formatters.Logger.Logger()
return node.GetValueAsUnsigned(0)
def right(self, node):
logger = lldb.formatters.Logger.Logger()
return node.GetChildMemberWithName("_M_right")
def left(self, node):
logger = lldb.formatters.Logger.Logger()
return node.GetChildMemberWithName("_M_left")
def parent(self, node):
logger = lldb.formatters.Logger.Logger()
return node.GetChildMemberWithName("_M_parent")
# from libstdc++ implementation of iterator for rbtree
def increment_node(self, node):
logger = lldb.formatters.Logger.Logger()
max_steps = self.num_children()
if self.node_ptr_value(self.right(node)) != 0:
x = self.right(node)
max_steps -= 1
while self.node_ptr_value(self.left(x)) != 0:
x = self.left(x)
max_steps -= 1
logger >> str(max_steps) + " more to go before giving up"
if max_steps <= 0:
self.garbage = True
return None
return x
else:
x = node
y = self.parent(x)
max_steps -= 1
while(self.node_ptr_value(x) == self.node_ptr_value(self.right(y))):
x = y
y = self.parent(y)
max_steps -= 1
logger >> str(max_steps) + " more to go before giving up"
if max_steps <= 0:
self.garbage = True
return None
if self.node_ptr_value(self.right(x)) != self.node_ptr_value(y):
x = y
return x
def has_children(self):
return True
Part 1 of a series of fixes meant to improve reliability and increase ease of bug fixing for data formatter issues. We are introducing a new Logger class on the Python side. This has the same purpose, but is unrelated, to the C++ logging facility The Pythonic logging can be enabled by using the following scripting commands: (lldb) script Logger._lldb_formatters_debug_level = {0,1,2,...} 0 = no logging 1 = do log 2 = flush after logging each line - slower but safer 3 or more = each time a Logger is constructed, log the function that has created it more log levels may be added, each one being more log-active than the previous by default, the log output will come out on your screen, to direct it to a file: (lldb) script Logger._lldb_formatters_debug_filename = 'filename' that will make the output go to the file - set to None to disable the file output and get screen logging back Logging has been enabled for the C++ STL formatters and for Cocoa class NSData - more logging will follow synthetic children providers for classes list and map (both libstdcpp and libcxx) now have internal capping for safety reasons this will fix crashers where a malformed list or map would not ever meet our termination conditions to set the cap to a different value: (lldb) script {gnu_libstdcpp|libcxx}.{map|list}_capping_size = new_cap (by default, it is 255) you can optionally disable the loop detection algorithm for lists (lldb) script {gnu_libstdcpp|libcxx}.list_uses_loop_detector = False llvm-svn: 153676
2012-03-30 03:29:45 +08:00
_list_uses_loop_detector = True