I started work on being able to add symbol files after a debug session
had started with a new "target symfile add" command and quickly ran into
problems with stale Address objects in breakpoint locations that had
lldb_private::Section pointers into modules that had been removed or
replaced. This also let to grabbing stale modules from those sections.
So I needed to thread harded the Address, Section and related objects.
To do this I modified the ModuleChild class to now require a ModuleSP
on initialization so that a weak reference can created. I also changed
all places that were handing out "Section *" to have them hand out SectionSP.
All ObjectFile, SymbolFile and SymbolVendors were inheriting from ModuleChild
so all of the find plug-in, static creation function and constructors now
require ModuleSP references instead of Module *.
Address objects now have weak references to their sections which can
safely go stale when a module gets destructed.
This checkin doesn't complete the "target symfile add" command, but it
does get us a lot clioser to being able to do such things without a high
risk of crashing or memory corruption.
llvm-svn: 151336
Objective-C classes. This allows LLDB to find
ivars declared in class extensions in modules other
than where the debugger is currently stopped (we
already supported this when the debugger was
stopped in the same module as the definition).
This involved the following main changes:
- The ObjCLanguageRuntime now knows how to hunt
for the authoritative version of an Objective-C
type. It looks for the symbol indicating a
definition, and then gets the type from the
module containing that symbol.
- ValueObjects now report their type with a
potential override, and the override is set if
the type of the ValueObject is an Objective-C
class or pointer type that is defined somewhere
other than the original reported type. This
means that "frame variable" will always use the
complete type if one is available.
- The ClangASTSource now looks for the complete
type when looking for ivars. This means that
"expr" will always use the complete type if one
is available.
- I added a testcase that verifies that both
"frame variable" and "expr" work.
llvm-svn: 151214
indicate whether inline functions are desired.
This allows the expression parser, for instance,
to filter out inlined functions when looking for
functions it can call.
llvm-svn: 150279
enable us to track the depth of parsing and what is being parsed. This
helps when trying to track down difficult type parsing issues and is only
enabled in non-production builds.
llvm-svn: 150203
working, but not functions). I need to check on a few things to make sure
I am registering everything correctly in the right order and in the right
contexts.
llvm-svn: 149858
interface (.i) files for each class.
Changed the FindFunction class from:
uint32_t
SBTarget::FindFunctions (const char *name,
uint32_t name_type_mask,
bool append,
lldb::SBSymbolContextList& sc_list)
uint32_t
SBModule::FindFunctions (const char *name,
uint32_t name_type_mask,
bool append,
lldb::SBSymbolContextList& sc_list)
To:
lldb::SBSymbolContextList
SBTarget::FindFunctions (const char *name,
uint32_t name_type_mask = lldb::eFunctionNameTypeAny);
lldb::SBSymbolContextList
SBModule::FindFunctions (const char *name,
uint32_t name_type_mask = lldb::eFunctionNameTypeAny);
This makes the API easier to use from python. Also added the ability to
append a SBSymbolContext or a SBSymbolContextList to a SBSymbolContextList.
Exposed properties for lldb.SBSymbolContextList in python:
lldb.SBSymbolContextList.modules => list() or all lldb.SBModule objects in the list
lldb.SBSymbolContextList.compile_units => list() or all lldb.SBCompileUnits objects in the list
lldb.SBSymbolContextList.functions => list() or all lldb.SBFunction objects in the list
lldb.SBSymbolContextList.blocks => list() or all lldb.SBBlock objects in the list
lldb.SBSymbolContextList.line_entries => list() or all lldb.SBLineEntry objects in the list
lldb.SBSymbolContextList.symbols => list() or all lldb.SBSymbol objects in the list
This allows a call to the SBTarget::FindFunctions(...) and SBModule::FindFunctions(...)
and then the result can be used to extract the desired information:
sc_list = lldb.target.FindFunctions("erase")
for function in sc_list.functions:
print function
for symbol in sc_list.symbols:
print symbol
Exposed properties for the lldb.SBSymbolContext objects in python:
lldb.SBSymbolContext.module => lldb.SBModule
lldb.SBSymbolContext.compile_unit => lldb.SBCompileUnit
lldb.SBSymbolContext.function => lldb.SBFunction
lldb.SBSymbolContext.block => lldb.SBBlock
lldb.SBSymbolContext.line_entry => lldb.SBLineEntry
lldb.SBSymbolContext.symbol => lldb.SBSymbol
Exposed properties for the lldb.SBBlock objects in python:
lldb.SBBlock.parent => lldb.SBBlock for the parent block that contains
lldb.SBBlock.sibling => lldb.SBBlock for the sibling block to the current block
lldb.SBBlock.first_child => lldb.SBBlock for the first child block to the current block
lldb.SBBlock.call_site => for inline functions, return a lldb.declaration object that gives the call site file, line and column
lldb.SBBlock.name => for inline functions this is the name of the inline function that this block represents
lldb.SBBlock.inlined_block => returns the inlined function block that contains this block (might return itself if the current block is an inlined block)
lldb.SBBlock.range[int] => access the address ranges for a block by index, a list() with start and end address is returned
lldb.SBBlock.ranges => an array or all address ranges for this block
lldb.SBBlock.num_ranges => the number of address ranges for this blcok
SBFunction objects can now get the SBType and the SBBlock that represents the
top scope of the function.
SBBlock objects can now get the variable list from the current block. The value
list returned allows varaibles to be viewed prior with no process if code
wants to check the variables in a function. There are two ways to get a variable
list from a SBBlock:
lldb::SBValueList
SBBlock::GetVariables (lldb::SBFrame& frame,
bool arguments,
bool locals,
bool statics,
lldb::DynamicValueType use_dynamic);
lldb::SBValueList
SBBlock::GetVariables (lldb::SBTarget& target,
bool arguments,
bool locals,
bool statics);
When a SBFrame is used, the values returned will be locked down to the frame
and the values will be evaluated in the context of that frame.
When a SBTarget is used, global an static variables can be viewed without a
running process.
llvm-svn: 149853
Fixed "target modules list" (aliased to "image list") to output more information
by default. Modified the "target modules list" to have a few new options:
"--header" or "-h" => show the image header address
"--offset" or "-o" => show the image header address offset from the address in the file (the slide applied to the shared library)
Removed the "--symfile-basename" or "-S" option, and repurposed it to
"--symfile-unique" "-S" which will show the symbol file if it differs from
the executable file.
ObjectFile's can now be loaded from memory for cases where we don't have the
files cached locally in an SDK or net mounted root. ObjectFileMachO can now
read mach files from memory.
Moved the section data reading code into the ObjectFile so that the object
file can get the section data from Process memory if the file is only in
memory.
lldb_private::Module can now load its object file in a target with a rigid
slide (very common operation for most dynamic linkers) by using:
bool
Module::SetLoadAddress (Target &target, lldb::addr_t offset, bool &changed)
lldb::SBModule() now has a new constructor in the public interface:
SBModule::SBModule (lldb::SBProcess &process, lldb::addr_t header_addr);
This will find an appropriate ObjectFile plug-in to load an image from memory
where the object file header is at "header_addr".
llvm-svn: 149804
LLVM/Clang. This brings in several fixes, including:
- Improvements in the Just-In-Time compiler's
allocation of memory: the JIT now allocates
memory in chunks of sections, improving its
ability to generate relocations. I have
revamped the RecordingMemoryManager to reflect
these changes, as well as to get the memory
allocation and data copying out fo the
ClangExpressionParser code. Jim Grosbach wrote
the updates to the JIT on the LLVM side.
- A new ExternalASTSource interface to allow LLDB to
report accurate structure layout information to
Clang. Previously we could only report the sizes
of fields, not their offsets. This meant that if
data structures included field alignment
directives, we could not communicate the necessary
alignment to Clang and accesses to the data would
fail. Now we can (and I have update the relevant
test case). Thanks to Doug Gregor for implementing
the Clang side of this fix.
- The way Objective-C interfaces are completed by
Clang has been made consistent with RecordDecls;
with help from Doug Gregor and Greg Clayton I have
ensured that this still works.
- I have eliminated all local LLVM and Clang patches,
committing the ones that are still relevant to LLVM
and Clang as needed.
I have tested the changes extensively locally, but
please let me know if they cause any trouble for you.
llvm-svn: 149775
a type when we have a forward declaration. We always have found a
type by basename, but now we also compare the decl context of the
die we are trying to complete with the matches we find from the accelerator
tables to ensure we get the right one.
llvm-svn: 149593
due to RTTI worries since llvm and clang don't use RTTI, but I was able to
switch back with no issues as far as I can tell. Once the RTTI issue wasn't
an issue, we were looking for a way to properly track weak pointers to objects
to solve some of the threading issues we have been running into which naturally
led us back to std::tr1::weak_ptr. We also wanted the ability to make a shared
pointer from just a pointer, which is also easily solved using the
std::tr1::enable_shared_from_this class.
The main reason for this move back is so we can start properly having weak
references to objects. Currently a lldb_private::Thread class has a refrence
to its parent lldb_private::Process. This doesn't work well when we now hand
out a SBThread object that contains a shared pointer to a lldb_private::Thread
as this SBThread can be held onto by external clients and if they end up
using one of these objects we can easily crash.
So the next task is to start adopting std::tr1::weak_ptr where ever it makes
sense which we can do with lldb_private::Debugger, lldb_private::Target,
lldb_private::Process, lldb_private::Thread, lldb_private::StackFrame, and
many more objects now that they are no longer using intrusive ref counted
pointer objects (you can't do std::tr1::weak_ptr functionality with intrusive
pointers).
llvm-svn: 149207
will ask ExternalASTSource objects to help laying out a type. This is needed
because the DWARF typically doesn't contain alignement or packing attribute
values, and we need to be able to match up types that the compiler uses
in expressions.
llvm-svn: 149160
be fetched too many times and the DisassemblerLLVM was appending to strings
when the opcode, mnemonic and comment accessors were called multiple times
and if any of the strings were empty.
Also fixed the test suite failures from recent Objective C modifications.
llvm-svn: 148460
for each ObjCInterfaceDecl was imposing performance
penalties for Objective-C apps. Instead, we now use
the normal function query mechanisms, which use the
relevant accelerator tables.
This fix also includes some modifications to the
SymbolFile which allow us to find Objective-C methods
and report their Clang Decls correctly.
llvm-svn: 148457
objective C class names when extracting the class name, selector and
name without category for objective C full class and instance method
names.
llvm-svn: 148435
much smarter by extracting search results more efficiently and by properly obeying the
must_be_implementation bool in the SymbolFileDWARF::FindCompleteObjCDefinitionTypeForDIE()
function.
llvm-svn: 148413
Fixed the new __apple_types to be able to accept a DW_TAG_structure_type
forward declaration and then find a DW_TAG_class_type definition, or vice
versa.
llvm-svn: 148097
are made up from the ObjC runtime symbols. For now the latter contain nothing but the fact that the name
describes an ObjC class, and so are not useful for things like dynamic types.
llvm-svn: 148059
Fix DWARF parsing issue we can run into when using llvm-gcc based dSYM files.
Also fix the parsing of objective C built-in types (Class, id and SEL) so
they don't parse more information that is not needed due to the way they
are represented in DWARF.
llvm-svn: 148016
so that we don't have "fprintf (stderr, ...)" calls sprinkled everywhere.
Changed all needed locations over to using this.
For non-darwin, we log to stderr only. On darwin, we log to stderr _and_
to ASL (Apple System Log facility). This will allow GUI apps to have a place
for these error and warning messages to go, and also allows the command line
apps to log directly to the terminal.
llvm-svn: 147596
Be better at detecting when DWARF changes and handle this more
gracefully than asserting and exiting.
Also fixed up a bunch of system calls that weren't properly checking
for EINTR.
llvm-svn: 147559
Fixed an issue where our new accelerator tables could cause a crash
when we got a full 32 bit hash match, yet a C string mismatch.
We had a member variable in DWARFMappedHash::Prologue named
"min_hash_data_byte_size" the would compute the byte size of HashData
so we could skip hash data efficiently. It started out with a byte size
value of 4. When we read the table in from disk, we would clear the
atom array and read it from disk, and the byte size would still be set
to 4. We would then, as we read each atom from disk, increment this count.
So the byte size of the HashData was off, which means when we get a lookup
whose 32 bit hash does matches, but the C string does NOT match (which is
very very rare), then we try and skip the data for that hash and we would
add an incorrect offset and get off in our parsing of the hash data and
cause this crash.
To fix this I added a few safeguards:
1 - I now correctly clear the hash data size when we reset the atom array using the new DWARFMappedHash::Prologue::ClearAtoms() function.
2 - I now correctly always let the AppendAtom() calculate the byte size of the hash (before we were doing things manually some times, which was correct, but not good)
3 - I also track if the size of each HashData is a fixed byte size or not, and "do the right thing" when we need to skip the data.
4 - If we do get off in the weeds, then I make sure to return an error and stop any further parsing from happening.
llvm-svn: 147334
<rdar://problem/10561406>
Stopped the SymbolFileDWARF::FindFunctions (...) from always calculating
the line table entry for all functions that were found. This can slow down
the expression parser if it ends up finding a bunch of matches. Fixed the
places that were relying on the line table entry being filled in.
Discovered a recursive stack blowout that happened when "main" didn't have
line info for it and there was no line information for "main"
llvm-svn: 146330
An assertion was firing when parsing types due to trying to complete parent
class decl contenxt types too often.
Also, relax where "dsymutil" binary can come from in the Makefile.rules.
llvm-svn: 146310
that if we prefer the current compile unit, followed by any compile units that
already had their DIEs parsed, followed by the rest of the matches, that we
might save some memory. This turned out not to help much. The code is commented
out, but I want to check it in so I don't lose the code in case it could help
later.
Added the ability to efficiently find the objective C class implementation
when using the new .apple_types acclerator tables with the type flags. If the
type flags are not available, we default back to what we were doing before.
llvm-svn: 146250
in the context in which it was originally found, the
expression parser now goes hunting for it in all modules
(in the appropriate namespace, if applicable). This means
that forward-declared types that exist in another shared
library will now be resolved correctly.
Added a test case to cover this. The test case also tests
"frame variable," which does not have this functionality
yet.
llvm-svn: 146204
that is in a class from the expression parser, and it was causing an
assertion. There is now a function that will correctly resolve a type
even if it is in a class.
llvm-svn: 146141
take a SymbolFile reference and a lldb::user_id_t and be used in objects
which represent things in debug symbols that have types where we don't need
to know the true type yet, such as in lldb_private::Variable objects. This
allows us to defer resolving the type until something is used. More specifically
this allows us to get 1000 local variables from the current function, and if
the user types "frame variable argc", we end up _only_ resolving the type for
"argc" and not for the 999 other local variables. We can expand the use of this
as needed in the future.
Modified the DWARFMappedHash class to be able to read the HashData that has
more than just the DIE offset. It currently will read the atoms in the header
definition and read the data correctly. Currently only the DIE offset and
type flags are supported. This is needed for adding type flags to the
.apple_types hash accelerator tables.
Fixed a assertion crash that would happen if we have a variable that had a
DW_AT_const_value instead of a location where "location.LocationContains_DW_OP_addr()"
would end up asserting when it tried to parse the variable location as a
DWARF opcode list.
Decreased the amount of memory that LLDB would use when evaluating an expression
by 3x - 4x for clang. There was a place in the namespace lookup code that was
parsing all namespaces with a certain name in a DWARF file instead of stopping
when it found the first match. This was causing all of the compile units with
a matching namespace to get parsed into memory and causing unnecessary memory
bloat.
Improved "Target::EvaluateExpression(...)" to not try and find a variable
when the expression contains characters that would certainly cause an expression
to need to be evaluated by the debugger.
llvm-svn: 146130
for now to fix testcases. Once we have a valid use
for the function information (i.e., once properties
returning UnknownAnyTy are allowed, once we read
return type information from the runtime, among
other uses) I will re-enable this.
llvm-svn: 146129
and CompleteTagDeclarationDefinition() on Objective-C
interfaces populated by SymbolFileSymtab::FindTypes(),
we should mark the interface as forward-declared when
we create it.
llvm-svn: 145825
add them to a fast lookup map. lldb_private::Symtab now export the following
public typedefs:
namespace lldb_private {
class Symtab {
typedef std::vector<uint32_t> IndexCollection;
typedef UniqueCStringMap<uint32_t> NameToIndexMap;
};
}
Clients can then find symbols by name and or type and end up with a
Symtab::IndexCollection that is filled with indexes. These indexes can then
be put into a name to index lookup map and control if the mangled and
demangled names get added to the map:
bool add_demangled = true;
bool add_mangled = true;
Symtab::NameToIndexMap name_to_index;
symtab->AppendSymbolNamesToMap (indexes, add_demangled, add_mangled, name_to_index).
This can be repeated as many times as needed to get a lookup table that
you are happy with, and then this can be sorted:
name_to_index.Sort();
Now name lookups can be done using a subset of the symbols you extracted from
the symbol table. This is currently being used to extract objective C types
from object files when there is no debug info in SymbolFileSymtab.
Cleaned up how the objective C types were being vended to be more efficient
and fixed some errors in the regular expression that was being used.
llvm-svn: 145777
class. The thing with Objective C classes is the debug info might have a
definition that isn't just a forward decl, but it is incomplete. So we need to
look and see if we can find the complete definition and avoid recursing a lot
due to the fact that our accelerator tables will have many versions of the
type, but only one complete one. We might not also have the complete type
and we need to deal with this correctly.
llvm-svn: 145759
Objective-C, making symbol lookups for various raw
Objective-C symbols work correctly. The IR interpreter
makes these lookups because Clang has emitted raw
symbol references for ivars and classes.
Also improved performance in SymbolFiles, caching the
result of asking for SymbolFile abilities.
llvm-svn: 145758
Fixed an issue where if we have the DWARF equivalent of:
struct foo;
class foo { ... };
Or vice versa, we wouldn't be able to find the complete type. Since many
compilers allow forward declarations to have struct and definitions to have
class, we need to be able to deal with both cases. This commit fixes this in
the DWARF parser.
llvm-svn: 145733
to use ConstStrings. The const char*s were assumed to
be from ConstStrings before, but since storing a full-on
ConstString is no more expensive than storing a const
char* it makes better sense to enforce uniqueness with
the type checker.
llvm-svn: 145688
to find Objective-C class types by looking in the
symbol tables for the individual object files.
I did this as follows:
- I added code to SymbolFileSymtab that vends
Clang types for symbols matching the pattern
"_OBJC_CLASS_$_NSMyClassName," making them
appear as Objective-C classes. This only occurs
in modules that do not have debug information,
since otherwise SymbolFileDWARF would be in
charge of looking up types.
- I made a new SymbolVendor subclass for the
Apple Objective-C runtime that is in charge of
making global lookups of Objective-C types. It
currently just sends out type lookup requests to
the appropriate SymbolFiles, but in the future we
will probably extend it to query the runtime more
completely.
I also modified a testcase whose behavior is changed
by the fact that we now actually return an Objective-C
type for __NSCFString.
llvm-svn: 145526
to launch a process for debugging. Since this isn't supported on all platforms,
we need to do what we used to do if this isn't supported. I added:
bool
Platform::CanDebugProcess ();
This will get checked before trying to launch a process for debugging and then
fall back to launching the process through the current host debugger. This
should solve the issue for linux and keep the platform code clean.
Centralized logging code for logging errors, warnings and logs when reporting
things for modules or symbol files. Both lldb_private::Module and
lldb_private::SymbolFile now have the following member functions:
void
LogMessage (Log *log, const char *format, ...);
void
ReportWarning (const char *format, ...);
void
ReportError (const char *format, ...);
These will all output the module name and object (if any) such as:
"error: lldb.so ...."
"warning: my_archive.a(foo.o) ...."
This will keep the output consistent and stop a lot of logging calls from
having to try and output all of the information that uniquely identifies
a module or symbol file. Many places in the code were grabbing the path to the
object file manually and if the module represented a .o file in an archive, we
would see log messages like:
error: foo.a - some error happened
llvm-svn: 145219
to 30% of memory. The size doubling was killing us and we ended up with up to
just under 50% of empty capacity. Cleaning this up saves us a ton of memory.
llvm-svn: 145086
1 - the DIE collections no longer have the NULL tags which saves up to 25%
of the memory on typical C++ code
2 - faster parsing by not having to run the SetDIERelations() function anymore
it is done when parsing the DWARF very efficiently.
llvm-svn: 144983
we say that the vectors of DWARFDebugInfoEntry objects were the highest on the
the list.
With these changes we cut our memory usage by 40%!!! I did this by reducing
the size of the DWARFDebugInfoEntry from a previous:
uint32_t offset
uint32_t parent_idx
uint32_t sibling_idx
Abbrev * abbrev_ptr
which was 20 bytes, but rounded up to 24 bytes due to alignment. Now we have:
uint32_t offset
uint32_t parent_idx
uint32_t sibling_idx
uint32_t abbr_idx:15, // 32767 possible abbreviation codes
has_children:1, // 0 = no children, 1 = has children
tag:16; // DW_TAG_XXX value
This gets us down to 16 bytes per DIE. I tested some VERY large DWARF files
(900MB) and found there were only ~700 unique abbreviations, so 32767 should
be enough for any sane compiler. If it isn't there are built in assertions
that will fire off and tell us.
llvm-svn: 144975
This is the actual fix for the above radar where global variables that weren't
initialized were not being shown correctly when leaving the DWARF in the .o
files. Global variables that aren't intialized have symbols in the .o files
that specify they are undefined and external to the .o file, yet document the
size of the variable. This allows the compiler to emit a single copy, but makes
it harder for our DWARF in .o files with the executable having a debug map
because the symbol for the global in the .o file doesn't exist in a section
that we can assign a fixed up linked address to, and also the DWARF contains
an invalid address in the "DW_OP_addr" location (always zero). This means that
the DWARF is incorrect and actually maps all such global varaibles to the
first file address in the .o file which is usually the first function. So we
can fix this in either of two ways: make a new fake section in the .o file
so that we have a file address in the .o file that we can relink, or fix the
the variable as it is created in the .o file DWARF parser and actually give it
the file address from the executable. Each variable contains a
SymbolContextScope, or a single pointer that helps us to recreate where the
variables came from (which module, file, function, etc). This context helps
us to resolve any file addresses that might be in the location description of
the variable by pointing us to which file the file address comes from, so we
can just replace the SymbolContextScope and also fix up the location, which we
would have had to do for the other case as well, and update the file address.
Now globals display correctly.
The above changes made it possible to determine if a variable is a global
or static variable when parsing DWARF. The DWARF emits a DW_TAG_variable tag
for each variable (local, global, or static), yet DWARF provides no way for
us to classify these variables into these categories. We can now detect when
a variable has a simple address expressions as its location and this will help
us classify these correctly.
While making the above changes I also noticed that we had two symbol types:
eSymbolTypeExtern and eSymbolTypeUndefined which mean essentially the same
thing: the symbol is not defined in the current object file. Symbol objects
also have a bit that specifies if a symbol is externally visible, so I got
rid of the eSymbolTypeExtern symbol type and moved all code locations that
used it to use the eSymbolTypeUndefined type.
llvm-svn: 144489
string to avoid possible later crashes.
Modified the locations that do set the crash description to NULL out the
string when they are done doing their tasks.
llvm-svn: 144297
Fixed an issue where if you had an initialized global variable, we would not
link it up correctly in the debug info if the .o file had the symbols as
UNDF + EXT (undefined external). We now properly link the globals.
llvm-svn: 144259
generated special member functions (constructors,
destructors, etc.) for classes that don't really have
them. We needed to mark these as artificial to reflect
the debug information; this bug does that for
constructors and destructors.
The "etc." case (certain assignment operators, mostly)
remains to be fixed.
llvm-svn: 143526
method as __attribute__ ((used)) when adding it to a
class. This functionality is useful when stopped in
anonymous namespaces: expressions attached to classes
in anonymous namespaces are typically elided by Clang's
CodeGen because they have no namespaces are intended
not to be externally visible. __attribute__ ((used))
forces CodeGen to emit the function.
Right now, __attribute__ ((used)) causes the JIT not to
emit the function, so we're not enabling it until we
fix that.
llvm-svn: 143469
in the same hashed format as the ".apple_names", but they map objective C
class names to all of the methods and class functions. We need to do this
because in the DWARF the methods for Objective C are never contained in the
class definition, they are scattered about at the translation unit level and
they don't even have attributes that say the are contained within the class
itself.
Added 3 new formats which can be used to display data:
eFormatAddressInfo
eFormatHexFloat
eFormatInstruction
eFormatAddressInfo describes an address such as function+offset and file+line,
or symbol + offset, or constant data (c string, 2, 4, 8, or 16 byte constants).
The format character for this is "A", the long format is "address".
eFormatHexFloat will print out the hex float format that compilers tend to use.
The format character for this is "X", the long format is "hex float".
eFormatInstruction will print out disassembly with bytes and it will use the
current target's architecture. The format character for this is "i" (which
used to be being used for the integer format, but the integer format also has
"d", so we gave the "i" format to disassembly), the long format is
"instruction".
Mate the lldb::FormatterChoiceCriterion enumeration private as it should have
been from the start. It is very specialized and doesn't belong in the public
API.
llvm-svn: 143114
Fixed an issue where bad DWARF from clang would get recycled from DWARF back
into types and cause clang to assert and die, killing the lldb binary, when
it tried to used the type in an expression.
llvm-svn: 142897
tables (like the .apple_namespaces) and it would cause us to index DWARF that
didn't need to be indexed.
Updated the MappedHash.h (generic Apple accelerator table) and the DWARF
specific one (HashedNameToDIE.h) to be up to date with the latest and
greatest hash table format.
llvm-svn: 142627
process IDs, and thread IDs, but was mainly needed for for the UserID's for
Types so that DWARF with debug map can work flawlessly. With DWARF in .o files
the type ID was the DIE offset in the DWARF for the .o file which is not
unique across all .o files, so now the SymbolFileDWARFDebugMap class will
make the .o file index part (the high 32 bits) of the unique type identifier
so it can uniquely identify the types.
llvm-svn: 142534
so we don't have to lookup types in a type list by ID.
Changed the DWARF parser to remove the "can externally complete myself" bits
from the type when we are in the process of completing the type itself to
avoid an onslaught of external visible decl requests from the
clang::ExternalASTSource.
llvm-svn: 142461
which had previously been commented out while I tested
it. It's not fully working yet, but it doesn't break
our testsuite and it's an important piece of
functionality.
Also added some logging to SymbolFileDWARF to help
diagnose entities that are found in a symbol file,
but do not reside in the expected namespace.
llvm-svn: 141894
context object. Having it populated and registered
within a single FindExternalVisibleDecls call worked
fine when there was only one call (i.e., when we were
just looking in the global namespace).
However, now FindExternalVisibleDecls is called for
nested namespaces as well, which means that it is
called not once but many times (once per module in
which the parent namespace appears). This means that
the namespace mapping is built up across many calls
to the inferior FindExternalVisibleDecls, so I moved
it into a data structure (the search context) that is
shared by all calls.
I also added some logging to make it easier to see
what is happening during a namespace search, and
cleaned up some existing logging.
llvm-svn: 141888
down through Module and SymbolVendor into SymbolFile.
Added checks to SymbolFileDWARF that restrict symbol
searches when a namespace is passed in.
llvm-svn: 141847
SymbolFIle (it was done mostly in the BreakpointResolverName resolver before.) Then
tailor our searches to the way the indexed maps are laid out. This removes a bunch
of test case failures using indexed dSYM's.
llvm-svn: 141428
Fixed up DWARFDebugAranges to use the new range classes.
Fixed the enumeration parsing to take a lldb_private::Error to avoid a lot of duplicated code. Now when an invalid enumeration is supplied, an error will be returned and that error will contain a list of the valid enumeration values.
llvm-svn: 141382
(lldb) log enable dwarf lookups
This allows us to see when lookups are being done on functions, addresses,
and types by both name and regular expresssion.
llvm-svn: 141259
index the DWARF. Also fixed an issue with memory accelerator tables with a
size of 1 where we would loop infinitely.
Added support for parsing the new .apple_namespaces section which gives us a
memory hash table for looking up namespaces.
llvm-svn: 141128
Also reduce the size of the lldb_private::Symbol objects by removing the
lldb_private::Function pointer that was in each symbol. Running Instruments
has shown that when debugging large applications with DWARF in .o files that
lldb_private::Symbol objects are one of the highest users of memory. No one
was using the Symbol::GetFunction() call anyway.
llvm-svn: 140881
information generated for it. Say we have a concrete function "foo" which
has inlined function "a" which calls another inlined function "b":
foo
1 {
2 {
a ()
3 {
b ()
4 {
}
}
}
}
Sometimes we see the compiler generate an address range in the DWARF for "foo"
(block 1 above) as say [0x1000-0x1100). Then the range for "a" is something
like [0x1050-0x1060) (note that it is correctly scoped within the "foo"
address range). And then we get "b" which is a child of "a", yet the debug
info says it has a range of [0x1060-0x1080) (not contained within "a"). We now
detect this issue when making our blocks and add an extra range to "a".
Also added a new "lldb" logging category named "symbol" where we can find out
about symbol file errors and warnings.
llvm-svn: 140822
are available, we currently will still index the DWARF ourselves
and assert if the name lookups differ. This will help us transition
to the new accelerator tables and make sure they are workng before
we switch over entirely.
llvm-svn: 140788
etc to specific source files.
Added SB API's to specify these source files & also more than one module.
Added an "exact" option to CompileUnit's FindLineEntry API.
llvm-svn: 140362
Fix the RegularExpression class so it has a real copy constructor.
Fix the breakpoint setting with multiple shared libraries so it makes
one breakpoint not one per shared library.
Add SBFileSpecList, to be used to expose the above to the SB interface (not done yet.)
llvm-svn: 140225
stdarg formats to use __attribute__ format so the compiler can flag
incorrect uses. Fix all incorrect uses. Most of these are innocuous,
a few were resulting in crashes.
llvm-svn: 140185
used to do this because we needed to find the shared pointer for a .o
file when the .o file's module was needed in a SymbolContext since the
module in a symbol context was a shared pointer. Now that we are using
intrusive pointers we don't have this limitation anymore since any
instrusive shared pointer can be made from a pointer to an object
all on its own.
Also switched over to having the Module and SymbolVendor use shared
pointers to their object files as had a leak on MacOSX when the
SymbolVendor's object file wasn't the same as the Module's (debug info
in a stand along file (dSYM file)). Now everything will correctly clean
itself up when the module goes away after an executable gets rebuilt.
Now we correctly get rid of .o files that are used with the DWARF with
debug map executables on subsequent runs since the only shared pointer
to the object files in from the DWARF symbol file debug map parser, and
when the module gets replaced, it destroys to old one along with all .o
files.
Also added a small optimization when using BSD archives where we will
remove old BSD containers from the shared list when they are outdated.
llvm-svn: 140002
ModuleSP
Module::GetSP();
Since we are now using intrusive ref counts, we can easily turn any
pointer to a module into a shared pointer just by assigning it.
llvm-svn: 139984
Address ranges are now split up into two different tables:
- one in DWARFDebugInfo that is compile unit specific
- one in each DWARFCompileUnit that has exact function DIE offsets
This helps keep the size of the aranges down since the main table will get
uniqued and sorted and have consecutive ranges merged. We then only parse the
compile unit one on demand once we have determined that a compile unit contains
the address in question. We also now use the .debug_aranges section if there
is one instead of always indexing the DWARF manually.
NameToDIE now uses a UniqueCStringMap<dw_offset> map instead of a std::map.
std::map is very bulky as each node has 3 pointers and the key and value types.
This gets our NameToDIE entry down to 12 bytes each instead of 48 which saves
us a lot of memory when we have very large DWARF.
DWARFDebugAranges now has a smaller footprint for each range it contains to
save on memory.
llvm-svn: 139557
One fixes a trailing comma bug (g++ doesn't like them)
The other gets the Error from the result of an expression evaluation and uses it as the error for the Process::LoadImage() method.
llvm-svn: 139336
Set the default Source File & line to main (if it can be found.) at startup. Selecting the current thread & or frame resets
the current source file & line, and "source list" as well as the breakpoint command "break set -l <NUM>" will use the
current source file.
llvm-svn: 139323
register names when dumping variable locations and location lists. Also did
some cleanup where "int" types were being used for "lldb::RegisterKind"
values.
llvm-svn: 138988
DWARF accelerator table sections to the DWARF parser. These sections are similar
to the .debug_pubnames and .debug_pubtypes, but they are designed to be hash tables
that are saved to disc in a way that the sections can just be loaded into memory
and used without any work on the debugger side. The .debug_pubnames and .debug_pubtypes
sections are not ordered, contain a copy of the name in the section itself which
makes these sections quite large, they only include publicly exported names (so no
static functions, no types defined inside functions), many compilers put different
information in them making them very unreliable so most debugger ignore these sections
and parse the DWARF on their own. The tables must also be parsed and sorted in order
to be used effectively. The new sections can be quickly loaded and very efficiently be used
to do name to DIE lookups with very little up front work. The format of these new
sections will be changing while we work out the bugs, but we hope to have really
fast name to DIE lookups soon.
llvm-svn: 138979
C++ methods for a function depending on how the DWARF was
created. Now we parse the class type from the definition,
and all methods that use DW_AT_specification or DW_AT_abstract_origin
attributes to point to the definition, now won't create
duplicate entries. This is in response to how clang++ creates
much different DWARF than gcc.
llvm-svn: 137737
was failing if the DWARF was laid out in a certain way. The way
we detect C++ classes is now more robust so that a class method
can be defined outside of the class and refer to a definition inside
the class with a DW_AT_specification or DW_AT_abstract_origin attribute.
Fixed a case in Thread.cpp where we were looking up info in the frame
when we didn't need to. This was from some changes to support external
editors. Now the info is only looked up if needed.
llvm-svn: 137436
that detects what context the current expression is
meant to execute in. LLDB now properly consults
the method declaration in the debug information
rather than trying to hunt down the "this" or "self"
pointer by name, which can be misleading.
Other fixes include:
- LLDB now properly detects that it is inside
an inlined C++ member function.
- LLDB now allows access to non-const members when
in const code.
- The functions in SymbolFile that locate the
DeclContext containing a DIE have been renamed
to reflect what they actually do. I have added
new functions that find the DeclContext for the
DIE itself.
I have also introduced testcases for C++ and
Objective-C.
llvm-svn: 136999
appropriately between C++ static methods and non-static
methods. This bug made it impossible to call most static
methods, either because Clang did not recognize that a
method could be called without providing a "this"
parameter, or because Clang did not properly mangle the
name of the method when searching for it in the target.
Also added a testcase.
llvm-svn: 136733
completes the support in the LLDB expression parser
for incomplete types. Clang now imports types
lazily, and we complete those types as necessary.
Changes include:
- ClangASTSource now supports three APIs which it
passes to ClangExpressionDeclMap. CompleteType
completes a TagDecl or an ObjCInterfaceDecl when
needed; FindExternalVisibleDecls finds named
entities that are visible in the expression's
scope; and FindExternalLexicalDecls performs a
(potentially restricted) search for entities
inside a lexical scope like a namespace. These
changes mean that entities in namespaces should
work normally.
- The SymbolFileDWARF code for searching a context
for a specific name is now more general, and can
search arbitrary contexts.
- We are continuing to adapt our calls into LLVM
from interfaces that take start and end iterators
when accepting multiple items to interfaces that
use ArrayRef.
- I have cleaned up some code, especially our use
of namespaces.
This change is neutral for our testsuite and greatly
improves correctness for large programs (like Clang)
with complicated type systems. It should also lay
the groundwork for improving the expression parser's
performance as we are lazier and lazier about
providing type information.
llvm-svn: 136555
by name by adding an extra parameter to the lldb_private::Target breakpoint
setting functions.
Added a function in the DWARF symbol file plug-in that can dump errors
and prints out which DWARF file the error is happening in so we can track
down what used to be assertions easily.
Fixed the MacOSX kernel plug-in to properly read the kext images and set
the kext breakpoint to watch for kexts as they are loaded.
llvm-svn: 134990
variables prior to running your binary. Zero filled sections now get
section data correctly filled with zeroes when Target::ReadMemory
reads from the object file section data.
Added new option groups and option values for file lists. I still need
to hook up all of the options to "target variable" to allow more complete
introspection by file and shlib.
Added the ability for ValueObjectVariable objects to be created with
only the target as the execution context. This allows them to be read
from the object files through Target::ReadMemory(...).
Added a "virtual Module * GetModule()" function to the ValueObject
class. By default it will look to the parent variable object and
return its module. The module is needed when we have global variables
that have file addresses (virtual addresses that are specific to
module object files) and in turn allows global variables to be displayed
prior to running.
Removed all of the unused proxy object support that bit rotted in
lldb_private::Value.
Replaced a lot of places that used "FileSpec::Compare (lhs, rhs) == 0" code
with the more efficient "FileSpec::Equal (lhs, rhs)".
Improved logging in GDB remote plug-in.
llvm-svn: 134579
"struct ", "class ", and "union " from the start of any type names that are
extracted from clang QualType objects. I had to fix test suite cases that
were expecting the struct/union/class prefix to be there.
llvm-svn: 134132
inspection of namespaces in the expression parser.
ClangExpressionDeclMap hitherto reported that namespaces had
been completely imported, even though the namespaces are
returned empty. To deal with this situation, ClangASTSource
was recently extended with an API to complete incomplete type
definitions, and, for greater efficiency, to complete these
definitions partially, returning only those objects that have
a given name.
This commit supports these APIs on LLDB's side, and uses it
to provide information on types resident in namespaces.
Namespaces are now imported as they were -- that is to say,
empty -- but with minimal import mode on. This means that
Clang will come back and request their contents by name as
needed. We now respond with information on the contained
types; this will be followed soon by information on functions
and variables.
llvm-svn: 133852
issue in the way block variables are marked as parsed. In the DWARF parser we
always parse all blocks for a function at once, so we can mark all blocks as
having all variables parsed and avoid recursive function calls to try and
reparse things that have already been handled.
Fixed an issue with how variables get scoped into blocks. The DWARF parser can
now handle abtract class definitions that contain concrete static variables.
When the concrete instance of the class functions get instantiated, they will
track down the concrete block for the abtract block and add the variable to
each block.
llvm-svn: 133302
darwin (not sure about other platforms).
Modified the communication and connection classes to not require the
BytesAvailable function. Now the "Read(...)" function has a timeout in
microseconds.
Fixed a lot of assertions that were firing off in certain cases and replaced
them with error output and code that can deal with the assertion case.
llvm-svn: 133224
In DWARFCompileUnit::ExtractDIEsIfNeeded we are relying on a compilation units
DIEs to be terminated by a null entry. I think the standard is fairly clear
that all sibling chains are to be terminated by null, but at least gcc 4.5.2
disagrees -- the top level chain drops the final entry. This results in us
interpreting the next compilation unit header as a DIE.
Regardless of whether gcc is right or wrong, we should not overstep a
compilation units extent. This patch ensures that we do not attempt to extract
a DIE beyond the length specified for a given DWARFCompileUnit by ensuring our
current offset is strictly less than the start of the next CU.
llvm-svn: 131721
DWARFDebugAranges::Sort() calls std::stable_sort() over a set of address ranges
and then proceeds to collapse neighboring ranges together.
One problem with the current implementation is that it does an incomplete job.
When a pair of ranges are merged the next pair considered does not include the
just-merged range. IOW, three consecutive ranges are never collapsed into one.
Another problem is that for each range merged we are calling
std::vector::erase() which "shifts" all remaining elements of the vector by one
position on every merge. The end result (in the worst case) is a quadratic
algorithm -- not good when the input vector is large.
The following patch merges all consecutive ranges and removes the quadratic
behavior. The implementation uses an auxiliary vector of indices in order to
remember all ranges that can be dropped, then performs the coalescing of ranges
in a single pass.
Patch from Stephen Wilson with some minor modification by me.
llvm-svn: 129595
Modified the OptionGroupOptions to be able to specify only some of the options
that should be appended by using the usage_mask in the group defintions and
also provided a way to remap them to a new usage mask after the copy. This
allows options to be re-used and also targetted for specific option groups.
Modfied the CommandArgumentType to have a new eArgTypePlatform enumeration.
Taught the option parser to be able to automatically use the appropriate
auto completion for a given options if nothing is explicitly specified
in the option definition. So you don't have to specify it in the option
definition tables.
Renamed the default host platform name to "host", and the default platform
hostname to be "localhost".
Modified the "file" and "platform select" commands to make sure all options
and args are good prior to creating a new platform. Also defer the computation
of the architecture in the file command until all options are parsed and the
platform has either not been specified or reset to a new value to avoid
computing the arch more than once.
Switch the PluginManager code over to using llvm::StringRef for string
comparisons and got rid of all the AccessorXXX functions in lieu of the newer
mutex + collection singleton accessors.
llvm-svn: 129483
Something changed in commit r129112 where a few standard headers vanished from
the include chain when building on Linux. Fix up by including limits.h for
INT_MAX and PATH_MAX where needed, and stdio.h for printf().
llvm-svn: 129130
const data, etc, and also for SBAddress objects to classify their type of
section they are in and also getting the module for a section offset address.
lldb::SymbolType SBSymbol::GetType();
lldb::SectionType SBAddress::GetSectionType ();
lldb::SBModule SBAddress::GetModule ();
llvm-svn: 128602
public types and public enums. This was done to keep the SWIG stuff from
parsing all sorts of enums and types that weren't needed, and allows us to
abstract our API better.
llvm-svn: 128239
static archive that can be linked against. LLDB.framework/lldb.so
exports a very controlled API. Splitting the API into a static
library allows other tools (debugserver for now) to use the power
of the LLDB debugger core, yet not export it as its API is not
portable or maintainable. The Host layer and many of the other
internal only APIs can now be statically linked against.
Now LLDB.framework/lldb.so links against "liblldb-core.a" instead
of compiling the .o files only for the shared library. This fix
is only for compiling with Xcode as the Makefile based build already
does this.
The Xcode projecdt compiler has been changed to LLVM. Anyone using
Xcode 3 will need to manually change the compiler back to GCC 4.2,
or update to Xcode 4.
llvm-svn: 127963
for templatized types that could cause parts of a std::vector (and I am sure
other STL types) to be incorrectly uniqued to each other wreaking havoc on
variable display for types within the same executable module.
llvm-svn: 127662
an interface to a local or remote debugging platform. By default each host OS
that supports LLDB should be registering a "default" platform that will be
used unless a new platform is selected. Platforms are responsible for things
such as:
- getting process information by name or by processs ID
- finding platform files. This is useful for remote debugging where there is
an SDK with files that might already or need to be cached for debug access.
- getting a list of platform supported architectures in the exact order they
should be selected. This helps the native x86 platform on MacOSX select the
correct x86_64/i386 slice from universal binaries.
- Connect to remote platforms for remote debugging
- Resolving an executable including finding an executable inside platform
specific bundles (macosx uses .app bundles that contain files) and also
selecting the appropriate slice of universal files for a given platform.
So by default there is always a local platform, but remote platforms can be
connected to. I will soon be adding a new "platform" command that will support
the following commands:
(lldb) platform connect --name machine1 macosx connect://host:port
Connected to "machine1" platform.
(lldb) platform disconnect macosx
This allows LLDB to be well setup to do remote debugging and also once
connected process listing and finding for things like:
(lldb) process attach --name x<TAB>
The currently selected platform plug-in can now auto complete any available
processes that start with "x". The responsibilities for the platform plug-in
will soon grow and expand.
llvm-svn: 127286
clang_type_t
GetClangFullType(); // Get a completely defined clang type
clang_type_t
GetClangLayoutType(); // Get a clang type that can be used for type layout
clang_type_t
GetClangForwardType(); // A type that can be completed if needed, but is more efficient.
llvm-svn: 125691
ArchDefaultUnwindPlan plug-in interfaces are now cached per architecture
instead of being leaked for every frame.
Split the ArchDefaultUnwindPlan_x86 into ArchDefaultUnwindPlan_x86_64 and
ArchDefaultUnwindPlan_i386 interfaces.
There were sporadic crashes that were due to something leaking or being
destroyed when doing stack crawls. This patch should clear up these issues.
llvm-svn: 125541
module's AST context. Prior to this fix, with gcc binaries, we end up with
a full class definition for any used classes in each compile unit due to the
one definition rule. This would result in us making N copies of class T, where
N is the number of compile units that use class T, in the module AST. When
an expression would then try and use any types that were duplicated, it would
quickly confuse clang and make expression evaluation fail due to all of the
duplicate types that got copied over. This is now fixed by making a map of
types in the DWARF that maps type names to a collection of types + declaration
(file + line number) + DIE. Then later when we find a type we look in this
module map and find any already cached types that we can just use.
8935777
llvm-svn: 125207
in the DWARF + debug map symbol file parser.
Also cleaned up the "image lookup --address ADDR" output when we it results
in something that is in an inlined function. Now we correctly dump out the
full inlined call stack.
llvm-svn: 125072
LLDB plugin directory and a user LLDB plugin directory. We currently still
need to work out at what layer the plug-ins will be, but at least we are
prepared for plug-ins. Plug-ins will attempt to be loaded from the
"/Developer/Library/PrivateFrameworks/LLDB.framework/Resources/Plugins"
folder, and from the "~/Library/Application Support/LLDB/Plugins" folder on
MacOSX. Each plugin will be scanned for:
extern "C" bool LLDBPluginInitialize(void);
extern "C" void LLDBPluginTerminate(void);
If at least LLDBPluginInitialize is found, the plug-in will be loaded. The
LLDBPluginInitialize function returns a bool that indicates if the plug-in
should stay loaded or not (plug-ins might check the current OS, current
hardware, or anything else and determine they don't want to run on the current
host). The plug-in is uniqued by path and added to a static loaded plug-in
map. The plug-in scanning happens during "lldb_private::Initialize()" which
calls to the PluginManager::Initialize() function. Likewise with termination
lldb_private::Terminate() calls PluginManager::Terminate(). The paths for the
plug-in directories is fetched through new Host calls:
bool Host::GetLLDBPath (ePathTypeLLDBSystemPlugins, dir_spec);
bool Host::GetLLDBPath (ePathTypeLLDBUserPlugins, dir_spec);
This way linux and other systems can define their own appropriate locations
for plug-ins to be loaded.
To allow dynamic shared library loading, the Host layer has also been modified
to include shared library open, close and get symbol:
static void *
Host::DynamicLibraryOpen (const FileSpec &file_spec,
Error &error);
static Error
Host::DynamicLibraryClose (void *dynamic_library_handle);
static void *
Host::DynamicLibraryGetSymbol (void *dynamic_library_handle,
const char *symbol_name,
Error &error);
lldb_private::FileSpec also has been modified to support directory enumeration
in an attempt to abstract the directory enumeration into one spot in the code.
The directory enumertion function is static and takes a callback:
typedef enum EnumerateDirectoryResult
{
eEnumerateDirectoryResultNext, // Enumerate next entry in the current directory
eEnumerateDirectoryResultEnter, // Recurse into the current entry if it is a directory or symlink, or next if not
eEnumerateDirectoryResultExit, // Exit from the current directory at the current level.
eEnumerateDirectoryResultQuit // Stop directory enumerations at any level
};
typedef FileSpec::EnumerateDirectoryResult (*EnumerateDirectoryCallbackType) (void *baton,
FileSpec::FileType file_type,
const FileSpec &spec);
static FileSpec::EnumerateDirectoryResult
FileSpec::EnumerateDirectory (const char *dir_path,
bool find_directories,
bool find_files,
bool find_other,
EnumerateDirectoryCallbackType callback,
void *callback_baton);
This allow clients to specify the directory to search, and specifies if only
files, directories or other (pipe, symlink, fifo, etc) files will cause the
callback to be called. The callback also gets to return with the action that
should be performed after this directory entry. eEnumerateDirectoryResultNext
specifies to continue enumerating through a directory with the next entry.
eEnumerateDirectoryResultEnter specifies to recurse down into a directory
entry, or if the file is not a directory or symlink/alias to a directory, then
just iterate to the next entry. eEnumerateDirectoryResultExit specifies to
exit the current directory and skip any entries that might be remaining, yet
continue enumerating to the next entry in the parent directory. And finally
eEnumerateDirectoryResultQuit means to abort all directory enumerations at
all levels.
Modified the Declaration class to not include column information currently
since we don't have any compilers that currently support column based
declaration information. Columns support can be re-enabled with the
additions of a #define.
Added the ability to find an EmulateInstruction plug-in given a target triple
and optional plug-in name in the plug-in manager.
Fixed a few cases where opendir/readdir was being used, but yet not closedir
was being used. Soon these will be deprecated in favor of the new directory
enumeration call that was added to the FileSpec class.
llvm-svn: 124716
lldb_private::Function objects. Previously the SymbolFileSymtab subclass
would return lldb_private::Symbol objects when it was asked to find functions.
The Module::FindFunctions (...) now take a boolean "bool include_symbols" so
that the module can track down functions and symbols, yet functions are found
by the SymbolFile plug-ins (through the SymbolVendor class), and symbols are
gotten through the ObjectFile plug-ins.
Fixed and issue where the DWARF parser might run into incomplete class member
function defintions which would make clang mad when we tried to make certain
member functions with invalid number of parameters (such as an operator=
operator that had no parameters). Now we just avoid and don't complete these
incomplete functions.
llvm-svn: 124359
types. What was happening was the DWARF parser was almost ignoring definitions
(DIEs with the DW_AT_declaration set to 1). It wasn't ignoring declarations
that had _some_ children. When this happened, we would treat the declaration
as a complete type. Often we would have a declaration of a type with just
some enum definitions inside and nothing else. Now we correctly ignore these
definitions, and also I added some changes to allow us to figure out what
decl context these special declarations actually point to.
llvm-svn: 124179
going to actually be used as the symbol file plug-in by looking only for suitable
N_OSO symbols and avoiding sorting function (N_FUN) and global/static (N_GSYM/N_STSYM)
symbols when there are no suitable N_OSO objects.
llvm-svn: 123889
the way LLDB lazily gets complete definitions for types within the debug info.
When we run across a class/struct/union definition in the DWARF, we will only
parse the full definition if we need to. This works fine for top level types
that are assigned directly to variables and arguments, but when we have a
variable with a class, lets say "A" for this example, that has a member:
"B *m_b". Initially we don't need to hunt down a definition for this class
unless we are ever asked to do something with it ("expr m_b->getDecl()" for
example). With my previous approach to lazy type completion, we would be able
to take a "A *a" and get a complete type for it, but we wouldn't be able to
then do an "a->m_b->getDecl()" unless we always expanded all types within a
class prior to handing out the type. Expanding everything is very costly and
it would be great if there were a better way.
A few months ago I worked with the llvm/clang folks to have the
ExternalASTSource class be able to complete classes if there weren't completed
yet:
class ExternalASTSource {
....
virtual void
CompleteType (clang::TagDecl *Tag);
virtual void
CompleteType (clang::ObjCInterfaceDecl *Class);
};
This was great, because we can now have the class that is producing the AST
(SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources
and the object that creates the forward declaration types can now also
complete them anywhere within the clang type system.
This patch makes a few major changes:
- lldb_private::Module classes now own the AST context. Previously the TypeList
objects did.
- The DWARF parsers now sign up as an external AST sources so they can complete
types.
- All of the pure clang type system wrapper code we have in LLDB (ClangASTContext,
ClangASTType, and more) can now be iterating through children of any type,
and if a class/union/struct type (clang::RecordType or ObjC interface)
is found that is incomplete, we can ask the AST to get the definition.
- The SymbolFileDWARFDebugMap class now will create and use a single AST that
all child SymbolFileDWARF classes will share (much like what happens when
we have a complete linked DWARF for an executable).
We will need to modify some of the ClangUserExpression code to take more
advantage of this completion ability in the near future. Meanwhile we should
be better off now that we can be accessing any children of variables through
pointers and always be able to resolve the clang type if needed.
llvm-svn: 123613
Anytime we had a valid python list that was trying to go from Python down into
our C++ API, it was allocating too little memory and it ended up smashing
whatever was next to the allocated memory.
Added typemap conversions for "void *, size_t" so we can get
SBProcess::ReadMemory() working. Also added a typemap for "const void *, size_t"
so we can get SBProcess::WriteMemory() to work.
Fixed an issue in the DWARF parser where we weren't correctly calculating the
DeclContext for all types and classes. We now should be a lot more accurate.
Fixes include: enums should now be setting their parent decl context correctly.
We saw a lot of examples where enums in classes were not being properly
namespace scoped. Also, classes within classes now get properly scoped.
Fixed the objective C runtime pointer checkers to let "nil" pointers through
since these are accepted by compiled code. We also now don't call "abort()"
when a pointer doesn't validate correctly since this was wreaking havoc on
the process due to the way abort() works. We now just dereference memory
which should give us an exception from which we can easily and reliably
recover.
llvm-svn: 123428
if two functions had the same demangled names (constructors where
we have the in charge and not in charge version) we could end up
mixing the two up when making the function in the DWARF. This was
because we need to lookup the symbol by name and we need to use the
mangled name if there is one. This ensures we get the correct address
and that we resolve the linked addresses correctly for DWARf with debug
map.
llvm-svn: 121116
wrongly as the target of a pointer rather than the
SEL pointer itself. This caused incorrect behavior
when dealing with Objective-C selector variables.
llvm-svn: 121048
RegisterContext* - normally this is retrieved from the ExecutionContext's
StackFrame but when we need to evaluate an expression while creating
the stack frame list this can be a little tricky.
Add DW_OP_deref_size, needed for the _sigtramp FDE expression.
Add support for processing DWARF expressions in RegisterContextLLDB.
Update callers to DWARFExpression::Evaluate.
llvm-svn: 119885
can too. So now the lldb_private::Variable class has support for this.
Variables now have support for having a basename ("i"), and a mangled name
("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i").
Nowwhen searching for a variable by name, users might enter the fully qualified
name, or just the basename. So new test functions were added to the Variable
and Mangled classes as:
bool NameMatches (const ConstString &name);
bool NameMatches (const RegularExpression ®ex);
I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search
for global variables that are not in the current file scope by first starting
with the current module, then moving on to all modules.
Fixed an issue in the DWARF parser that could cause a varaible to get parsed
more than once. Now, once we have parsed a VariableSP for a DIE, we cache
the result even if a variable wasn't made so we don't do any re-parsing. Some
DW_TAG_variable DIEs don't have locations, or are missing vital info that
stops a debugger from being able to display anything for it, we parse a NULL
variable shared pointer for these DIEs so we don't keep trying to reparse it.
llvm-svn: 119085
breakpoints on inlined functions by name. This involved fixing the DWARF parser
to correctly back up and parse the concrete function when we find inlined
functions by name, then grabbing any appropriate inlined blocks and returning
symbol contexts with the block filled in. After this was fixed, the breakpoint
by name resolver needed to correctly deal with symbol contexts that had the
inlined block filled in in the symbol contexts.
llvm-svn: 119017
a debug map with DWARF in the .o files due to the attemted shortcut that was
being taken where the global variables were being searched for by looking in
the symbol table. The problem with the symbols in the symbol table is we don't
break apart the symbol names for symbols when they are mangled into basename
and the fully mangled name since this would take a lot of CPU time to chop up
the mangled names and try and find the basenames. The DWARF info typically has
this broken up for us where the basename of the variable is in a the DW_AT_name
attribute, and the mangled name is in the DW_AT_MIPS_linkage_name attribute.
Now we correctly find globals by searching all OSO's for the information so we
can take advantage of this split information.
llvm-svn: 119012
cases when getting the clang type:
- need only a forward declaration
- need a clang type that can be used for layout (members and args/return types)
- need a full clang type
This allows us to partially parse the clang types and be as lazy as possible.
The first case is when we just need to declare a type and we will complete it
later. The forward declaration happens only for class/union/structs and enums.
The layout type allows us to resolve the full clang type _except_ if we have
any modifiers on a pointer or reference (both R and L value). In this case
when we are adding members or function args or return types, we only need to
know how the type will be laid out and we can defer completing the pointee
type until we later need it. The last type means we need a full definition for
the clang type.
Did some renaming of some enumerations to get rid of the old "DC" prefix (which
stands for DebugCore which is no longer around).
Modified the clang namespace support to be almost ready to be fed to the
expression parser. I made a new ClangNamespaceDecl class that can carry around
the AST and the namespace decl so we can copy it into the expression AST. I
modified the symbol vendor and symbol file plug-ins to use this new class.
llvm-svn: 118976
namespaces by name given an optional symbol context. I might end up
dressing up the "clang::NamespaceDecl" into a lldb_private::Namespace
class if we need to do more than is currenlty required of namespaces.
Currently we only need to be able to lookup a namespace by name when
parsing expressions, so I kept it simple for now. The idea here is
even though we are passing around a "clang::NamespaceDecl *", that
we always have it be an opaque pointer (it is forward declared inside
of "lldb/Core/ClangForward.h") and we only use clang::NamespaceDecl
implementations inside of ClangASTContext, or ClangASTType when we need
to extract information from the namespace decl object.
llvm-svn: 118737
comes from by using a virtual function to provide it from the Module's
SymbolVendor by default. This allows the DWARF parser, when being used to
parse DWARF in .o files with a parent DWARF + debug map parser, to get its
type list from the DWARF + debug map parser so when we go and find full
definitions for types (that might come from other .o files), we can use the
type list from the debug map parser. Otherwise we ended up mixing clang types
from one .o file (say a const pointer to a forward declaration "class A") with
the a full type from another .o file. This causes expression parsing, when
copying the clang types from those parsed by the DWARF parser into the
expression AST, to fail -- for good reason. Now all types are created in the
same list.
Also added host support for crash description strings that can be set before
doing a piece of work. On MacOSX, this ties in with CrashReporter support
that allows a string to be dispalyed when the app crashes and allows
LLDB.framework to print a description string in the crash log. Right now this
is hookup up the the CommandInterpreter::HandleCommand() where each command
notes that it is about to be executed, so if we crash while trying to do this
command, we should be able to see the command that caused LLDB to exit. For
all other platforms, this is a nop.
llvm-svn: 118672
Fixed the DWARF plug-in such that when it gets all attributes for a DIE, that
it omits the DW_AT_sibling and DW_AT_declaration when getting attributes
from a DW_AT_abstract_origin or DW_AT_specification DIE.
llvm-svn: 118654
to not get resolved.
Fixed the "void **isa_ptr" variable inside the objective C verifier to start
with a '$' character so we don't go looking for it in our program.
Moved the lookup for "$__lldb_class" into the part that knows we are looking
for internal types that start with a '$'.
llvm-svn: 118488
where the DWARF is in the .o files so they can track down the actual type for
a forward declaration. This was working before for just DWARF files, but not
for DWARF in .o files where the actual definition was in another .o file.
Modified the main thread name in the driver to be more consistent with the
other LLDB thread names.
llvm-svn: 118383
Objective-C member accessibility restrictions.
Ultimately Clang should have a flag that ignores
accessibility for Objective-C like it has one
for C++.
llvm-svn: 117768
than just the entire log channel.
Add checks, where appropriate, to make sure a log channel/category has
not been disabled before attempting to write to it.
llvm-svn: 117715
by type ID (the most common type of type lookup).
Changed the API logging a bit to always show the objects in the OBJECT(POINTER)
format so it will be easy to locate all instances of an object or references
to it when looking at logs.
llvm-svn: 117641
all of the calls inlined in the header file for better performance.
Fixed the summary for C string types (array of chars (with any combo if
modifiers), and pointers to chars) work in all cases.
Fixed an issue where a forward declaration to a clang type could cause itself
to resolve itself more than once if, during the resolving of the type itself
it caused something to try and resolve itself again. We now remove the clang
type from the forward declaration map in the DWARF parser when we start to
resolve it and avoid this additional call. This should stop any duplicate
members from appearing and throwing all the alignment of structs, unions and
classes.
llvm-svn: 117437
So the issue here was that we have lldb_private::FileSpec that by default was
always resolving a path when using the:
FileSpec::FileSpec (const char *path);
and in the:
void FileSpec::SetFile(const char *pathname, bool resolve = true);
This isn't what we want in many many cases. One example is you have "/tmp" on
your file system which is really "/private/tmp". You compile code in that
directory and end up with debug info that mentions "/tmp/file.c". Then you
type:
(lldb) breakpoint set --file file.c --line 5
If your current working directory is "/tmp", then "file.c" would be turned
into "/private/tmp/file.c" which won't match anything in the debug info.
Also, it should have been just a FileSpec with no directory and a filename
of "file.c" which could (and should) potentially match any instances of "file.c"
in the debug info.
So I removed the constructor that just takes a path:
FileSpec::FileSpec (const char *path); // REMOVED
You must now use the other constructor that has a "bool resolve" parameter that you must always supply:
FileSpec::FileSpec (const char *path, bool resolve);
I also removed the default parameter to SetFile():
void FileSpec::SetFile(const char *pathname, bool resolve);
And fixed all of the code to use the right settings.
llvm-svn: 116944
since we can't parse DW_TAG_namespace DIEs as types. They are only decls in
clang. All of the types we handle right now have both clang "XXXType" classes
to go with the "XXXDecl" classes which means they can be used within the
lldb_private::Type class. I need to check to see which other decls that don't
have associated type objects need to float around the debugger and possibly
make a lldb_private::Decl class to manage them.
llvm-svn: 116558
artifical members (like the vtable pointer member that shows up in the DWARF).
We were adding this to each class which was making all member variables be off
by a pointer size.
Added a test case so we can track this with "test/forward".
Fixed the type name index in DWARF to include all the types after finding
some types were being omitted due to the DW_AT_specification having the
DW_AT_declaration attribute which was being read into the real type instances
when there were forward declarations in the DWARF, causing the type to be
omitted. We now check to make sure any DW_AT_declaration values are only
respected when parsing types if the attribute is from the current DIE.
After fixing the missing types, we ran into some issues with the expression
parser finding duplicate entries for __va_list_tag since they are built in
types and would result in a "duplicate __va_list_tag definition" error. We
are now just ignoring this name during lookup, but we will need to see if
we can get the name lookup function to not get called in these cases.
Fixed an issue that would cause an assertion where DW_TAG_subroutine_types
that had no children, would not properly make a clang function type of:
"void (*) (void)".
llvm-svn: 116392
being chopped up correctly). The DWARF plug-in also keeps a map of the ObjC
class names to selectors for easy parsing of all class selectors when we parse
the class type.
llvm-svn: 116290
Added the start of Host specific launch services, though it currently isn't
hookup up to anything. We want to be able to launch a process and use the
native launch services to launch an app like it would be launched by the
user double clicking on the app. We also eventually want to be able to run
a command line app in a newly spawned terminal to avoid terminal sharing.
Fixed an issue with the new DWARF forward type declaration stuff. A crasher
was found that was happening when trying to properly expand the forward
declarations.
llvm-svn: 115213
adding methods to C++ and objective C classes. In order to make methods, we
need the function prototype which means we need the arguments. Parsing these
could cause a circular reference that caused an assertion.
Added a new typedef for the clang opaque types which are just void pointers:
lldb::clang_type_t. This appears in lldb-types.h.
This was fixed by enabling struct, union, class, and enum types to only get
a forward declaration when we make the clang opaque qual type for these
types. When they need to actually be resolved, lldb_private::Type will call
a new function in the SymbolFile protocol to resolve a clang type when it is
not fully defined (clang::TagDecl::getDefinition() returns NULL). This allows
us to be a lot more lazy when parsing clang types and keeps down the amount
of data that gets parsed into the ASTContext for each module.
Getting the clang type from a "lldb_private::Type" object now takes a boolean
that indicates if a forward declaration is ok:
clang_type_t lldb_private::Type::GetClangType (bool forward_decl_is_ok);
So function prototypes that define parameters that are "const T&" can now just
parse the forward declaration for type 'T' and we avoid circular references in
the type system.
llvm-svn: 115012
interface in ClangASTContext. Also added two bool returning functions that
indicated if an opaque clang qual type is a CXX class type, and if it is an
ObjC class type.
Objective C classes now will get their methods added lazily as they are
encountered. The reason for this is currently, unlike C++, the
DW_TAG_structure_type and owns the ivars, doesn't not also contain the
member functions. This means when we parse the objective C class interface
we either need to find all functions whose names start with "+[CLASS_NAME"
or "-[CLASS_NAME" and add them all to the class, or when we parse each objective
C function, we slowly add it to the class interface definition. Since objective
C's class doesn't change internal bits according to whether it has certain types
of member functions (like C++ does if it has virtual functions, or if it has
user ctors/dtors), I currently chose to lazily populate the class when each
functions is parsed. Another issue we run into with ObjC method declarations
is the "self" and "_cmd" implicit args are not marked as artificial in the
DWARF (DW_AT_artifical), so we currently have to look for the parameters by
name if we are trying to omit artificial function args if the language of the
compile unit is ObjC or ObjC++.
llvm-svn: 114722
- Sema is now exported (and there was much rejoicing.)
- Storage classes are now centrally defined.
Also fixed some bugs that the new LLVM picked up.
llvm-svn: 114622
- All single character options will now be printed together
- Changed all options that contains underscores to contain '-' instead
- Made the help come out a little flatter by showing the long and short
option on the same line.
- Modified the short character for "--ignore-count" options to "-i"
llvm-svn: 114265
find the hotspots in our code when indexing the DWARF. A combination of
using SmallVector to avoid collection allocations, using fixed form
sizes when possible, and optimizing the hot loops contributed to the
speedup.
llvm-svn: 113961
Added a "bool show_fullpaths" to many more objects that were
previously always dumping full paths.
Fixed a few places where the DWARF was not indexed when we
we needed it to be when making queries. Also fixed an issue
where the DWARF in .o files wasn't searching all .o files
for the types.
Fixed an issue with the output from "image lookup --type <TYPENAME>"
where the name and byte size might not be resolved and might not
display. We now call the accessors so we end up seeing all of the
type info.
llvm-svn: 113951
all types in all compile units. I added a new kind of accelerator table to
the DWARF that allows us to index the DWARF compile units and DIEs in a way
that doesn't require the data to stay loaded. Currently when indexing the
DWARF we check if the compile unit had parsed its DIEs and if it hasn't we
index the data and free all of the DIEs so we can reparse later when we need
to after using one of our complete accelerator tables to determine we need
to reparse some DWARF. If the DIEs had already been parsed we leave them
loaded. The new accelerator table uses the "const char *" pointers from our
ConstString class as the keys, and NameToDIE::Info as the value. This info
contains the compile unit index and the DIE index which means we are pointed
right to the DIE we need unlike the other DWARF accelerator tables that often
just point us to the compile unit we would find our answer in.
llvm-svn: 113933
debug map showed that the location lists in the .o files needed some
refactoring in order to work. The case that was failing was where a function
that was in the "__TEXT.__textcoal_nt" in the .o file, and in the
"__TEXT.__text" section in the main executable. This made symbol lookup fail
due to the way we were finding a real address in the debug map which was
by finding the section that the function was in in the .o file and trying to
find this in the main executable. Now the section list supports finding a
linked address in a section or any child sections. After fixing this, we ran
into issue that were due to DWARF and how it represents locations lists.
DWARF makes a list of address ranges and expressions that go along with those
address ranges. The location addresses are expressed in terms of a compile
unit address + offset. This works fine as long as nothing moves around. When
stuff moves around and offsets change between the remapped compile unit base
address and the new function address, then we can run into trouble. To deal
with this, we now store supply a location list slide amount to any location
list expressions that will allow us to make the location list addresses into
zero based offsets from the object that owns the location list (always a
function in our case).
With these fixes we can now re-link random address ranges inside the debugger
for use with our DWARF + debug map, incremental linking, and more.
Another issue that arose when doing the DWARF in the .o files was that GCC
4.2 emits a ".debug_aranges" that only mentions functions that are externally
visible. This makes .debug_aranges useless to us and we now generate a real
address range lookup table in the DWARF parser at the same time as we index
the name tables (that are needed because .debug_pubnames is just as useless).
llvm-gcc doesn't generate a .debug_aranges section, though this could be
fixed, we aren't going to rely upon it.
Renamed a bunch of "UINT_MAX" to "UINT32_MAX".
llvm-svn: 113829
no elements so that they at least have 1 element.
Added the ability to show the declaration location of variables to the
"frame variables" with the "--show-declaration" option ("-c" for short).
Changed the "frame variables" command over to use the value object code
so that we use the same code path as the public API does when accessing and
displaying variable values.
llvm-svn: 113733
union, or class that contained an enumeration type. When I was creating
the clang enumeration decl, I wasn't calling "EnumDecl::setIntegerType (QualType)"
which means that if the enum decl was ever asked to figure out it's bit width
(getTypeInfo()) it would crash. We didn't run into this with enum types that
weren't inside classes because the DWARF already told us how big the type was
and when we printed an enum we would never need to calculate the size, we
would use the pre-cached byte size we got from the DWARF. When the enum was
in a struct/union/class and we tried to layout the struct, the layout code
would attempt to get the type info and segfault.
llvm-svn: 113729
They will now be represented as:
eSymbolTypeFunction: eSymbolTypeCode with IsDebug() == true
eSymbolTypeGlobal: eSymbolTypeData with IsDebug() == true and IsExternal() == true
eSymbolTypeStatic: eSymbolTypeData with IsDebug() == true and IsExternal() == false
This simplifies the logic when dealing with symbols and allows for symbols
to be coalesced into a single symbol most of the time.
Enabled the minimal symbol table for mach-o again after working out all the
kinks. We now get nice concise symbol tables and debugging with DWARF in the
.o files with a debug map in the binary works well again. There were issues
where the SymbolFileDWARFDebugMap symbol file parser was using symbol IDs and
symbol indexes interchangeably. Now that all those issues are resolved
debugging is working nicely.
llvm-svn: 113678
to spawn a thread for each process that is being monitored. Previously
LLDB would spawn a single thread that would wait for any child process which
isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on
linux). The old single thread used to call wait4() with a pid of -1 which
could cause it to reap child processes that it shouldn't have.
Re-wrote the way Function blocks are handles. Previously I attempted to keep
all blocks in a single memory allocation (in a std::vector). This made the
code somewhat efficient, but hard to work with. I got rid of the old BlockList
class, and went to a straight parent with children relationship. This new
approach will allow for partial parsing of the blocks within a function.
llvm-svn: 111706
the resulting function from the .o file DWARF didn't make it into the final
executable. I recently changed the way FindFunctions() worked in the DWARF
with debug map case that caused regressions in our test suite for dead
stripped functions. The previous changes allowed us to leverage the powerful
searching added to the DWARF parser (search by full name, basename, selector,
or method name), without having to chop up the symbol names from the symbol
table and do any special parsing of the names to extract the basename,
selector or method. Previously we would look through the symbol table for
matches first, then try and find the .o file with DWARF for that symbol and
only search those .o files. Now we let the DWARF for the .o file search using
the new search styles, and filter out any functions that didn't make it.
llvm-svn: 111322
Sean Callanan