<rdar://problem/11285931>
Use the DWARRF end prologue markers when trying to skip prologue instructions instead of blindly using the second line table address entry.
llvm-svn: 155600
the debug information individual Decls came from.
We've had a metadata infrastructure for a while,
which was intended to solve a problem we've since
dealt with in a different way. (It was meant to
keep track of which definition of an Objective-C
class was the "true" definition, but we now find
it by searching the symbols for the class symbol.)
The metadata is attached to the ExternalASTSource,
which means it has a one-to-one correspondence with
AST contexts.
I've repurposed the metadata infrastructure to
hold the object file and DIE offset for the DWARF
information corresponding to a Decl. There are
methods in ClangASTContext that get and set this
metadata, and the ClangASTImporter is capable of
tracking down the metadata for Decls that have been
copied out of the debug information into the
parser's AST context without using any additional
memory.
To see the metadata, you just have to enable the
expression log:
-
(lldb) log enable lldb expr
-
and watch the import messages. The high 32 bits
of the metadata indicate the index of the object
file in its containing DWARFDebugMap; I have also
added a log which you can use to track that mapping:
-
(lldb) log enable dwarf map
-
This adds 64 bits per Decl, which in my testing
hasn't turned out to be very much (debugging Clang
produces around 6500 Decls in my tests). To track
how much data is being consumed, I've also added a
global variable g_TotalSizeOfMetadata which tracks
the total number of Decls that have metadata in all
active AST contexts.
Right now this metadata is enormously useful for
tracking down bugs in the debug info parser. In the
future I also want to use this information to provide
more intelligent error messages instead of printing
empty source lines wherever Clang refers to the
location where something is defined.
llvm-svn: 154634
correctly if the setter/getter were not present
in the debug information. The fixes are as follows:
- We not only look for the method by its full name,
but also look for automatically-generated methods
when searching for a selector in an Objective-C
interface. This is necessary to find accessors.
- Extract the getter and setter name from the
DW_TAG_APPLE_Property declaration in the DWARF
if they are present; generate them if not.
llvm-svn: 154067
Fixed an issue where there were more than one way to get a CompileUnitSP created when using SymbolFileDWARF with SymbolFileDWARFDebugMap. This led to an assertion that would fire under certain conditions. Now there is only one way to create the compile unit and it will "do the right thing".
llvm-svn: 153908
Fixed an issue that could cause circular type parsing that will assert and kill LLDB.
Prior to this fix the DWARF parser would always create class types and not start their definitions (for both C++ and ObjC classes) until we were asked to complete the class later. When we had cases like:
class A
{
class B
{
};
};
We would alway try to complete A before specifying "A" as the decl context for B. Turns out we can just start the definition and still not complete the class since we can check the TagDecl::isCompleteDefinition() function. This only works for C++ types. This means we will not be pulling in the full definition of parent classes all the time and should help with our memory consumption and also reduce the amount of debug info we have to parse.
I also reduced redundant code that was checking in a lldb::clang_type_t was a possible C++ dynamic type since it was still completing the type, just to see if it was dynamic. This was fixed in another function that was checking for a type being dynamic as an ObjC or a C++ type, but there was dedicated fucntion for C++ that we missed.
llvm-svn: 153713
Line tables when using DWARF in .o files can be wrong when two entries get moved around by the compiler. This was due to incorrect logic in the line entry comparison operator.
llvm-svn: 153685
for unbacked properties. We support two variants:
one in which the getter/setter are provided by
selector ("mySetter:") and one in which the
getter/setter are provided by signature
("-[MyClass mySetter:]").
llvm-svn: 153675
1 - sections only get a valid VM size if they have SHF_ALLOC in the section flags
2 - symbol names are marked as mangled if they start with "_Z"
Also fixed the DWARF parser to correctly use the section file size when extracting the DWARF.
llvm-svn: 153496
Fixed type lookups to "do the right thing". Prior to this fix, looking up a type using "foo::bar" would result in a type list that contains all types that had "bar" as a basename unless the symbol file was able to match fully qualified names (which our DWARF parser does not).
This fix will allow type matches to be made based on the basename and then have the types that don't match filtered out. Types by name can be fully qualified, or partially qualified with the new "bool exact_match" parameter to the Module::FindTypes() method.
This fixes some issue that we discovered with dynamic type resolution as well as improves the overall type lookups in LLDB.
llvm-svn: 153482
Adding a test case that checks that we do not complete types before due time. This should help us track cases similar to the cascading data formatters.
llvm-svn: 153363
Fixed a performance regression when dynamic types are enable where we would ask a C++ type if it can possibly be dynamic. Previously we would force the type to complete itself and then anwwer the question definitively. Now we ask the type if it is already complete and only definitively answer the question for completed types and just say "yes" for non-complete C++ types. We also always now answer yes for Objective C classes and do not complete those types either.
llvm-svn: 153284
LLDB can match incorrect line table entries when an address is between two valid line entries (in the gap between the valid debug info), now it doesn't!
llvm-svn: 153077
Simplify the locking strategy for Module and its owned objects to always use the Module's mutex to avoid A/B deadlocks. We had a case where a symbol vendor was locking itself and then calling a function that would try to get it's Module's mutex and at the same time another thread had the Module mutex that was trying to get the SymbolVendor mutex. Now any classes that inherit from ModuleChild should use the module lock using code like:
void
ModuleChildSubclass::Function
{
ModuleSP module_sp(GetModule());
if (module_sp)
{
lldb_private::Mutex::Locker locker(module_sp->GetMutex());
... do work here...
}
}
This will help avoid deadlocks by using as few locks as possible for a module and all its child objects and also enforce detecting if a module has gone away (the ModuleSP will be returned empty if the weak_ptr does refer to a valid object anymore).
llvm-svn: 152679
This fix really needed to happen as a previous fix I had submitted for
calculating symbol sizes made many symbols appear to have zero size since
the function that was calculating the symbol size was calling another function
that would cause the calculation to happen again. This resulted in some symbols
having zero size when they shouldn't. This could then cause infinite stack
traces and many other side affects.
llvm-svn: 152244
expression command doesn't handle xmm or stmm registers...
o Update ClangASTContext::GetBuiltinTypeForEncodingAndBitSize() to now handle eEncodingVector.
o Modify RegisterValue::SetFromMemoryData() to fix the subtle error due to unitialized variables.
o Add a test file for "expr $xmm0".
llvm-svn: 152190
This was done in SBTarget:
lldb::SBInstructionList
lldb::SBTarget::ReadInstructions (lldb::SBAddress base_addr, uint32_t count);
Also cleaned up a few files in the LLDB.framework settings.
llvm-svn: 152152
so that the expression parser can look up members
of anonymous structs correctly. This meant creating
all the proper IndirectFieldDecls in each Record
after it has been completely populated with members.
llvm-svn: 151868
allocations by section. We install these sections
in the target process and inform the JIT of their
new locations.
Also removed some unused variable warnings.
llvm-svn: 151789
more of the local path, platform path, associated symbol file, UUID, arch,
object name and object offset. This allows many of the calls that were
GetSharedModule to reduce the number of arguments that were used in a call
to these functions. It also allows a module to be created with a ModuleSpec
which allows many things to be specified prior to any accessors being called
on the Module class itself.
I was running into problems when adding support for "target symbol add"
where you can specify a stand alone debug info file after debugging has started
where I needed to specify the associated symbol file path and if I waited until
after construction, the wrong symbol file had already been located. By using
the ModuleSpec it allows us to construct a module with as little or as much
information as needed and not have to change the parameter list.
llvm-svn: 151476
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
objects for the backlink to the lldb_private::Process. The issues we were
running into before was someone was holding onto a shared pointer to a
lldb_private::Thread for too long, and the lldb_private::Process parent object
would get destroyed and the lldb_private::Thread had a "Process &m_process"
member which would just treat whatever memory that used to be a Process as a
valid Process. This was mostly happening for lldb_private::StackFrame objects
that had a member like "Thread &m_thread". So this completes the internal
strong/weak changes.
Documented the ExecutionContext and ExecutionContextRef classes so that our
LLDB developers can understand when and where to use ExecutionContext and
ExecutionContextRef objects.
llvm-svn: 151009
the lldb_private::StackFrame objects hold onto a weak pointer to the thread
object. The lldb_private::StackFrame objects the the most volatile objects
we have as when we are doing single stepping, frames can often get lost or
thrown away, only to be re-created as another object that still refers to the
same frame. We have another bug tracking that. But we need to be able to
have frames no longer be able to get the thread when they are not part of
a thread anymore, and this is the first step (this fix makes that possible
but doesn't implement it yet).
Also changed lldb_private::ExecutionContextScope to return shared pointers to
all objects in the execution context to further thread harden the internals.
llvm-svn: 150871
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
user space programs. The core file support is implemented by making a process
plug-in that will dress up the threads and stack frames by using the core file
memory.
Added many default implementations for the lldb_private::Process functions so
that plug-ins like the ProcessMachCore don't need to override many many
functions only to have to return an error.
Added new virtual functions to the ObjectFile class for extracting the frozen
thread states that might be stored in object files. The default implementations
return no thread information, but any platforms that support core files that
contain frozen thread states (like mach-o) can make a module using the core
file and then extract the information. The object files can enumerate the
threads and also provide the register state for each thread. Since each object
file knows how the thread registers are stored, they are responsible for
creating a suitable register context that can be used by the core file threads.
Changed the process CreateInstace callbacks to return a shared pointer and
to also take an "const FileSpec *core_file" parameter to allow for core file
support. This will also allow for lldb_private::Process subclasses to be made
that could load crash logs. This should be possible on darwin where the crash
logs contain all of the stack frames for all of the threads, yet the crash
logs only contain the registers for the crashed thrad. It should also allow
some variables to be viewed for the thread that crashed.
llvm-svn: 150154
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
instead of the __repr__. __repr__ is a function that should return an
expression that can be used to recreate an python object and we were using
it to just return a human readable string.
Fixed a crasher when using the new implementation of SBValue::Cast(SBType).
Thread hardened lldb::SBValue and lldb::SBWatchpoint and did other general
improvements to the API.
Fixed a crasher in lldb::SBValue::GetChildMemberWithName() where we didn't
correctly handle not having a target.
llvm-svn: 149743
uint32_t
SBType::GetNumberOfTemplateArguments ();
lldb::SBType
SBType::GetTemplateArgumentType (uint32_t idx);
lldb::TemplateArgumentKind
SBType::GetTemplateArgumentKind (uint32_t idx);
Some lldb::TemplateArgumentKind values don't have a corresponding SBType
that will be returned from SBType::GetTemplateArgumentType(). This will
help our data formatters do their job by being able to find out the
type of template params and do smart things with those.
llvm-svn: 149658
You can now access a frame in a thread using:
lldb.SBThread.frame[int] -> lldb.SBFrame object for a frame in a thread
Where "int" is an integer index. You can also access a list object with all of
the frames using:
lldb.SBThread.frames => list() of lldb.SBFrame objects
All SB objects that give out SBAddress objects have properties named "addr"
lldb.SBInstructionList now has the following convenience accessors for len() and
instruction access using an index:
insts = lldb.frame.function.instructions
for idx in range(len(insts)):
print insts[idx]
Instruction lists can also lookup an isntruction using a lldb.SBAddress as the key:
pc_inst = lldb.frame.function.instructions[lldb.frame.addr]
lldb.SBProcess now exposes:
lldb.SBProcess.is_alive => BOOL Check if a process is exists and is alive
lldb.SBProcess.is_running => BOOL check if a process is running (or stepping):
lldb.SBProcess.is_running => BOOL check if a process is currently stopped or crashed:
lldb.SBProcess.thread[int] => lldb.SBThreads for a given "int" zero based index
lldb.SBProcess.threads => list() containing all lldb.SBThread objects in a process
SBInstruction now exposes:
lldb.SBInstruction.mnemonic => python string for instruction mnemonic
lldb.SBInstruction.operands => python string for instruction operands
lldb.SBInstruction.command => python string for instruction comment
SBModule now exposes:
lldb.SBModule.uuid => uuid.UUID(), an UUID object from the "uuid" python module
lldb.SBModule.symbol[int] => lldb.Symbol, lookup symbol by zero based index
lldb.SBModule.symbol[str] => list() of lldb.Symbol objects that match "str"
lldb.SBModule.symbol[re] => list() of lldb.Symbol objecxts that match the regex
lldb.SBModule.symbols => list() of all symbols in a module
SBAddress objects can now access the current load address with the "lldb.SBAddress.load_addr"
property. The current "lldb.target" will be used to try and resolve the load address.
Load addresses can also be set using this accessor:
addr = lldb.SBAddress()
addd.load_addr = 0x123023
Then you can check the section and offset to see if the address got resolved.
SBTarget now exposes:
lldb.SBTarget.module[int] => lldb.SBModule from zero based module index
lldb.SBTarget.module[str] => lldb.SBModule by basename or fullpath or uuid string
lldb.SBTarget.module[uuid.UUID()] => lldb.SBModule whose UUID matches
lldb.SBTarget.module[re] => list() of lldb.SBModule objects that match the regex
lldb.SBTarget.modules => list() of all lldb.SBModule objects in the target
SBSymbol now exposes:
lldb.SBSymbol.name => python string for demangled symbol name
lldb.SBSymbol.mangled => python string for mangled symbol name or None if there is none
lldb.SBSymbol.type => lldb.eSymbolType enum value
lldb.SBSymbol.addr => SBAddress object that represents the start address for this symbol (if there is one)
lldb.SBSymbol.end_addr => SBAddress for the end address of the symbol (if there is one)
lldb.SBSymbol.prologue_size => pythin int containing The size of the prologue in bytes
lldb.SBSymbol.instructions => SBInstructionList containing all instructions for this symbol
SBFunction now also has these new properties in addition to what is already has:
lldb.SBFunction.addr => SBAddress object that represents the start address for this function
lldb.SBFunction.end_addr => SBAddress for the end address of the function
lldb.SBFunction.instructions => SBInstructionList containing all instructions for this function
SBFrame now exposes the SBAddress for the frame:
lldb.SBFrame.addr => SBAddress which is the section offset address for the current frame PC
These are all in addition to what was already added. Documentation and website
updates coming soon.
llvm-svn: 149489
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
We should ultimately introduce GetAs...Type
functions in all cases where we have Is...Type
functions that know how to look inside typedefs.
llvm-svn: 148512
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
master AST importer imports types.
- First, before importing the definition of a
Decl from its source, notify the underlying
importer of the source->destination mapping.
Especially for anonymous strucutres that are
otherwise hard to unique in the target AST
context, this hint is very helpful.
- When deporting a type or Decl from one
ASTContext to another (deporting occurs in
the case of moving result types from the
parser's AST context to the result AST
context), don't forget their origin if the
origin is the original debug information.
llvm-svn: 148152
debug info, call it anonymous. This isn't
perfect, because Clang actually considers the
following struct not to be anonymous:
–
struct {
int x;
int y;
} g_foo;
-
but DWARF doesn't make the distinction.
llvm-svn: 148145
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
mmap() the entire object file contents into memory with MAP_PRIVATE.
We do this because object file contents can change on us and currently
this helps alleviate this situation. It also make the code for accessing
object file data much easier to manage and we don't end up opening the
file, reading some data and closing the file over and over.
llvm-svn: 148017
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
Switch from GetReturnValue, which was hardly ever used, to GetReturnValueObject
which is much more convenient.
Return the "return value object" as a persistent variable if requested.
llvm-svn: 147157
types that have been imported multiple times.
The discussion below uses this diagram:
ASTContext A B C
Decl Da Db Dc
ASTImporter \-Iab-/\-Iac-/
\-----Iac----/
When a Decl D is imported from ASTContext A to
ASTContext B, the ASTImporter Iab records the
pair <Da, Db> in a DenseMap. That way, if Iab
ever encounters Da again (for example, as the
DeclContext for another Decl), it can use the
imported version. This is not an optimization,
it is critical: if I import the field "st_dev"
as part of importing "struct stat," the field
must have DeclContext equal to the parent
structure or we end up with multiple different
Decls containing different parts of "struct
stat." "struct stat" is imported once and
recorded in the DenseMap; then the ASTImporter
finds that same version when looking for the
DeclContext of "st_dev."
The bug arises when Db is imported into another
ASTContext C and ASTContext B goes away. This
often occurs when LLDB produces result variables
for expressions. Ibc is aware of the transport
of Db to Dc, but a brand new ASTImporter, Iac,
is responsible for completing Dc from its source
upon request. That ASTImporter has no mappings,
so it will produce a clone of Dc when attempting
to import its children. That means that type
completion operations on Dc will fail.
The solution is to create Iac as soon as Ibc
imports D from B to C, and inform Iac of the
mapping between Da and Dc. This allows type
completion to happen correctly.
llvm-svn: 147016
we handle Objective-C method calls. Currently,
LLDB treats the result of an Objective-C method
as unknown if the type information doesn't have
the method's signature. Now Clang can cast the
result to id if it isn't explicitly cast.
I also added a test case for this, as well as a
fix for a type import problem that this feature
exposed.
llvm-svn: 146756
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
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
from symbols more accessible, I have added a second
map to the ClangASTImporter: the ObjCInterfaceMetaMap.
This map keeps track of all type definitions found for
a particular Objective-C interface, allowing the
ClangASTSource to refer to all possible sources when
looking for method definitions.
There is a bug in lookup that I still need to figure out,
but after that we should be able to report full method
information for Objective-C classes shown in symbols.
Also fixed some errors I ran into when enabling the maps
for the persistent type store. The persistent type store
previously did not use the ClangASTImporter to import
types, instead using ASTImporters that got allocated each
time a type needed copying. To support the requirements
of the persistent type store -- namely, that types must be
copied, completed, and then completely severed from their
origin in the parser's AST context (which will go away) --
I added a new function called DeportType which severs all
these connections.
llvm-svn: 145914
methods. The Clang dump is now much more verbose,
but when somebody types "target modules lookup -t"
that is typically what they're looking for.
llvm-svn: 145892
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
for all our external AST sources that lets us associate
arbitrary flags with the types we put into the AST
contexts. Also added an API on ClangASTContext that
allows access to these flags given only an ASTContext
and a type.
Because we don't have access to RTTI, and because at
some point in the future we might encounter external
AST sources that we didn't make (so they don't subclass
ClangExternalASTSourceCommon) I added a magic number
that we check before doing anything else, so that we
can catch that problem as soon as it appears.
llvm-svn: 145748
object file can correctly make these symbols which will abstract us from the
file format and ABI and we can then ask for the objective C class symbol for
a class and find out which object file it was defined in.
llvm-svn: 145744
Fixed templates with NonTypeTemplateParmDecl objects. For example:
template <unsigned N>
....
This fixes SmallVector and all of the other classes that have template params
that are non types.
llvm-svn: 145667
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
management of what allocations remain after an
expression finishes executing. This saves around
2.5KiB per expression for simple expressions.
llvm-svn: 145342
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 allow variables in the persistent variable store to know
how to complete themselves from debug information. That
fixes a variety of bugs during dematerialization of
expression results and also makes persistent variable and
result variables ($foo, $4, ...) more useful.
I have also added logging improvements that make it much
easier to figure out how types are moving from place to
place, and made some checking a little more aggressive.
The commit includes patches to Clang which are currently being
integrated into Clang proper; once these fixes are in Clang
top-of-tree, these patches will be removed. The patches don't
fix API; rather, they fix some internal bugs in Clang's
ASTImporter that were exposed when LLDB was moving types from
place to place multiple times.
llvm-svn: 144969
completion information between different AST
contexts. It works like this:
- If a Decl is imported from a context that
has completion metadata, then that Decl
is associated with the same completion
information (possibly none) as the Decl
it was imported from.
- If a Decl is imported from a context that
does not have completion metadata, then it
is marked as completable by consulting the
Decl and context it was imported from.
llvm-svn: 144838
for each AST context it knows about in a single
object. This makes it faster to look up the
appropriate ASTImpoter for a given ASTContext
pair and also makes it much easier to delete all
metadata for a given AST context.
In the future, this fix will allow the
ClangASTImporter to propagate completion
information between the metadata for different
AST contexts as its minions move AST objects
around.
llvm-svn: 144835
rather than individually on behalf of each
ASTContext. This allows the ASTImporter to know
about all containers of types, which will let it
be smarter about forwarding information about
type origins. That means that the following
sequence of steps will be possible (after a few
more changes):
- Import a type from a Module's ASTContext into
an expression parser ASTContext, tracking its
origin information -- this works now.
- Because the result of the expression uses that
type, import it from the expression parser
ASTContext into the Target's scratch AST
context, forwarding the origin information --
this needs to be added.
- For a later expression that uses the result,
import the type from the Target's scratch AST
context, still forwarding origin information
-- this also needs to be added.
- Use the intact origin information to complete
the type as needed -- this works now if the
origin information is present.
To this end, I made the following changes:
- ASTImporter top-level copy functions now
require both a source and a destination AST
context parameter.
- The ASTImporter now knows how to purge
records related to an ASTContext that is
going away.
- The Target now owns and creates the ASTImporter
whenever the main executable changes or (in the
absence of a main executable) on demand.
llvm-svn: 144802
of problems with Objective-C object completion. To go
along with the LLVM/Clang-side fixes, we have a variety
of Objective-C improvements.
Fixes include:
- It is now possible to run expressions when stopped in
an Objective-C class method and have "self" act just
like "self" would act in the class method itself (i.e.,
[self classMethod] works without casting the return
type if debug info is present). To accomplish this,
the expression masquerades as a class method added by
a category.
- Objective-C objects can now provide methods and
properties and methods to Clang on demand (i.e., the
ASTImporter sets hasExternalVisibleDecls on Objective-C
interface objects).
- Objective-C built-in types, which had long been a bone
of contention (should we be using "id"? "id*"?), are
now fetched correctly using accessor functions on
ClangASTContext. We inhibit searches for them in the
debug information.
There are also a variety of logging fixes, and I made two
changes to the test suite:
- Enabled a test case for Objective-C properties in the
current translation unit.
- Added a test case for calling Objective-C class methods
when stopped in a class method.
llvm-svn: 144607
Fixed an issues with the SBType and SBTypeMember classes:
- Fixed SBType to be able to dump itself from python
- Fixed SBType::GetNumberOfFields() to return the correct value for objective C interfaces
- Fixed SBTypeMember to be able to dump itself from python
- Fixed the SBTypeMember ability to get a field offset in bytes (the value
being returned was wrong)
- Added the SBTypeMember ability to get a field offset in bits
Cleaned up a lot of the Stream usage in the SB API files.
llvm-svn: 144493
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
interfaces. This allows us to pull in Objective-C
method types on demand, which is also now implemented.
Also added a minor fix to prevent multiple-definition
errors for "Class" and "id".
llvm-svn: 144405
lookups for Objective-C methods by selector.
Right now all it does is print log information.
Also improved the logging for imported TagDecls
to indicate whether or not the definition for
the imported TagDecl is complete.
llvm-svn: 144203
C++ vtables, fixing a record layout problem in the
expression parser.
Also fixed various problems with the generation
and unpacking of llvm.zip given our new better
handling of multiple architectures in the LLVM
build.
(And added a log message that will hopefully catch
record layout problems in the future.)
llvm-svn: 143741
- If you download and build the sources in the Xcode project, x86_64 builds
by default using the "llvm.zip" checkpointed LLVM.
- If you delete the "lldb/llvm.zip" and the "lldb/llvm" folder, and build the
Xcode project will download the right LLVM sources and build them from
scratch
- If you have a "lldb/llvm" folder already that contains a "lldb/llvm/lib"
directory, we will use the sources you have placed in the LLDB directory.
Python can now be disabled for platforms that don't support it.
Changed the way the libllvmclang.a files get used. They now all get built into
arch specific directories and never get merged into universal binaries as this
was causing issues where you would have to go and delete the file if you wanted
to build an extra architecture slice.
llvm-svn: 143678
allows us to set __attribute__ ((used)) on expressions
that masquerade as methods. When we are stopped in
classes in anonymous namespaces, this fix (and enabling
__attribute__ ((used)) on the method) will allow
expressions to run.
llvm-svn: 143560
Fixed an issue where the DWARF might mention that a class has a constructor
(default, copy or move), destructor, or an assignment operator (copy or move)
and it might not have an actual implementation in your code. Then you try and
use this struct or class in an expression and the JIT would ask for the
address of these methods that were in the declaration, yet there are none.
We now "do the right thing" for trivial ctors, dtors and assignment operators
by telling the methods that they are are defaulted and trivial, and clang will
then just do all of the work with builtins!
llvm-svn: 143528
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
of reference types. Previously, such variables were
materialized as references to those references, which
caused undesried behavior in Clang and was useless anyway
(the benefit of using references to variables is that it
allows expressions to modify variables in place, but for
references that's not required).
Now we just materialize the references directly, which
fixes a variety of expressions that use references.
llvm-svn: 143137
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
be set if linking against an LLVM compiled with
NDEBUG off. If it is set, we do not enable NDEBUG
in any place where we include LLVM headers.
llvm-svn: 143036
linked against a debug LLVM, runs a variety of
functions -- currently just one -- that verify
that the Decls we create are valid.
ClangASTContext now calls this verifier whenever
it adds a Decl to a DeclContext, and the verifier
checks that the AccessSpecifier is sane.
llvm-svn: 143000
permits a namespace map to be created and populated
when the namespace is imported, not just when it is
requested via FindExternalVisibleDecls().
llvm-svn: 142690
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
FindExternalVisibleDecls and FindExternalLexicalDecls
are marked and given unique IDs, so that all logging
done as part of their execution can be traced back to
the proper call.
Also there was some logging that really wasn't helpful
in most cases so I disabled it unless verbose logging
(log enable -v lldb expr) is enabled.
llvm-svn: 141987
Specifically, the expression parser used to use
functions attached to SymbolContext to do lookups,
but nowadays it searches a ModuleList or Module
directly instead. These functions had no
remaining clients so I removed them to prevent
bit rot.
I also removed a stray callback function from
ClangExpressionDeclMap.
llvm-svn: 141899
down through Module and SymbolVendor into SymbolFile.
Added checks to SymbolFileDWARF that restrict symbol
searches when a namespace is passed in.
llvm-svn: 141847
of namespaces (only in the modules where they've
been found) for entities inside those namespaces.
For each NamespaceDecl that has been imported into
the parser, we maintain a map containing
[ModuleSP, ClangNamespaceDecl] pairs in the ASTImporter.
This map has one entry for each module in which the
namespace has been found. When we later scan for an
entity inside a namespace, we search only the modules
in which that namespace was found.
Also made a small whitespace fix in
ClangExpressionParser.cpp.
llvm-svn: 141748
This involved minor changes to the way we report Objective-C
methods, as well as cosmetic changes and added parameters
for a variety of Clang APIs.
llvm-svn: 141437
Fixed an assertion that was causing a crash. The bug describes a case where we have an inlined block that doesn't contain the frame PC that was used to lookup the symbol context in the first place. This really shouldn't happen, so
now we log if we run into this and don't assert.
llvm-svn: 141338
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
symbol context that represents an inlined function. This function has been
renamed internally to:
bool
SymbolContext::GetParentOfInlinedScope (const Address &curr_frame_pc,
SymbolContext &next_frame_sc,
Address &next_frame_pc) const;
And externally to:
SBSymbolContext
SBSymbolContext::GetParentOfInlinedScope (const SBAddress &curr_frame_pc,
SBAddress &parent_frame_addr) const;
The correct blocks are now correctly calculated.
Switched the stack backtracing engine (in StackFrameList) and the address
context printing over to using the internal SymbolContext::GetParentOfInlinedScope(...)
so all inlined callstacks will match exactly.
llvm-svn: 140910
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
The old way of storing blocks used to use the sibling pointer, but now all
blocks contain a collection of shared pointers to blocks so this isn't required
anymore and a parent can be asked to find the sibling block for a child block.
llvm-svn: 140808
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
symbolication. Also improved the SBInstruction API to allow
access to the instruction opcode name, mnemonics, comment and
instruction data.
Added the ability to edit SBLineEntry objects (change the file,
line and column), and also allow SBSymbolContext objects to be
modified (set module, comp unit, function, block, line entry
or symbol).
The SymbolContext and SBSymbolContext can now generate inlined
call stack infomration for symbolication much easier using the
SymbolContext::GetParentInlinedFrameInfo(...) and
SBSymbolContext::GetParentInlinedFrameInfo(...) methods.
llvm-svn: 140518
contents starting at an offset (2 separate methods). This helps
the scripting interface stay more natural by allowing both from
Python.
Added the ability to dump data with address annotations when
call SBData::GetDescription().
Hooked up the SBSection to the __repr__ so you can print section
objects from within python.
Improved the dumping of symbols from python.
Fixed the .i interface references which were set to "Relative to this Group"
which somehow included Jim's "lldb-clean" root directory in the path. The
interfaces are now in a folder called "interfaces" withing the Xcode API
subfolder.
llvm-svn: 140451
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
shared pointers.
Changed the ExecutionContext over to use shared pointers for
the target, process, thread and frame since these objects can
easily go away at any time and any object that was holding onto
an ExecutionContext was running the risk of using a bad object.
Now that the shared pointers for target, process, thread and
frame are just a single pointer (they all use the instrusive
shared pointers) the execution context is much safer and still
the same size.
Made the shared pointers in the the ExecutionContext class protected
and made accessors for all of the various ways to get at the pointers,
references, and shared pointers.
llvm-svn: 140298
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
We had some cases where getting the shared pointer for a module from
the global module list was causing a performance issue when debugging
with DWARF in .o files. Now that the module uses intrusive ref counts,
we can easily convert any pointer to a shared pointer.
llvm-svn: 139983
to execute expressions even in the absence of a process.
This allows expressions to run in situations where the
target cannot run -- e.g., to perform calculations based
on type information, or to inspect a binary's static
data.
This modification touches the following files:
lldb-private-enumerations.h
Introduce a new enum specifying the policy for
processing an expression. Some expressions should
always be JITted, for example if they are functions
that will be used over and over again. Some
expressions should always be interpreted, for
example if the target is unsafe to run. For most,
it is acceptable to JIT them, but interpretation
is preferable when possible.
Target.[h,cpp]
Have EvaluateExpression now accept the new enum.
ClangExpressionDeclMap.[cpp,h]
Add support for the IR interpreter and also make
the ClangExpressionDeclMap more robust in the
absence of a process.
ClangFunction.[cpp,h]
Add support for the new enum.
IRInterpreter.[cpp,h]
New implementation.
ClangUserExpression.[cpp,h]
Add support for the new enum, and for running
expressions in the absence of a process.
ClangExpression.h
Remove references to the old DWARF-based method
of evaluating expressions, because it has been
superseded for now.
ClangUtilityFunction.[cpp,h]
Add support for the new enum.
ClangExpressionParser.[cpp,h]
Add support for the new enum, remove references
to DWARF, and add support for checking whether
the expression could be evaluated statically.
IRForTarget.[h,cpp]
Add support for the new enum, and add utility
functions to support the interpreter.
IRToDWARF.cpp
Removed
CommandObjectExpression.cpp
Remove references to the obsolete -i option.
Process.cpp
Modify calls to ClangUserExpression::Evaluate
to pass the correct enum (for dlopen/dlclose)
SBValue.cpp
Add support for the new enum.
SBFrame.cpp
Add support for he new enum.
BreakpointOptions.cpp
Add support for the new enum.
llvm-svn: 139772
UnwindPlan for unwinding from the first instruction of an otherwise
unknown function call (GetUnwindPlanArchitectureDefaultAtFunctionEntry()).
Update RegisterContextLLDB::GetFullUnwindPlanForFrame() to detect the
case of a frame 0 at address 0x0 which indicates that we jumped through
a NULL function pointer. Use the ABI's FunctionEntryUnwindPlan to
find the caller frame.
These changes make it so lldb can identify the calling frame correctly
in code like
int main ()
{
void (*f)(void) = 0;
f();
}
llvm-svn: 139760
Fixed up many API calls to not be "const" as const doesn't mean anything to
most of our lldb::SB objects since they contain a shared pointer, auto_ptr, or
pointer to the types which circumvent the constness anyway.
llvm-svn: 139428
- introduced two new classes ValueObjectConstResultChild and ValueObjectConstResultImpl: the first one is a ValueObjectChild obtained from
a ValueObjectConstResult, the second is a common implementation backend for VOCR and VOCRCh of method calls meant to read through pointers stored
in frozen objects ; now such reads transparently move from host to target as required
- as a consequence of the above, removed code that made target-memory copies of expression results in several places throughout LLDB, and also
removed code that enabled to recognize an expression result VO as such
- introduced a new GetPointeeData() method in ValueObject that lets you read a given amount of objects of type T from a VO
representing a T* or T[], and doing dereferences transparently
in private layer it returns a DataExtractor ; in public layer it returns an instance of a newly created lldb::SBData
- as GetPointeeData() does the right thing for both frozen and non-frozen ValueObject's, reimplemented ReadPointedString() to use it
en lieu of doing the raw read itself
- introduced a new GetData() method in ValueObject that lets you get a copy of the data that backs the ValueObject (for pointers,
this returns the address without any previous dereferencing steps ; for arrays it actually reads the whole chunk of memory)
in public layer this returns an SBData, just like GetPointeeData()
- introduced a new CreateValueFromData() method in SBValue that lets you create a new SBValue from a chunk of data wrapped in an SBData
the limitation to remember for this kind of SBValue is that they have no address: extracting the address-of for these objects (with any
of GetAddress(), GetLoadAddress() and AddressOf()) will return invalid values
- added several tests to check that "p"-ing objects (STL classes, char* and char[]) will do the right thing
Solved a bug where global pointers to global variables were not dereferenced correctly for display
New target setting "max-string-summary-length" gives the maximum number of characters to show in a string when summarizing it, instead of the hardcoded 128
Solved a bug where the summary for char[] and char* would not be shown if the ValueObject's were dumped via the "p" command
Removed m_pointers_point_to_load_addrs from ValueObject. Introduced a new m_address_type_of_children, which each ValueObject can set to tell the address type
of any pointers and/or references it creates. In the current codebase, this is load address most of the time (the only notable exception being file
addresses that generate file address children UNLESS we have a live process)
Updated help text for summary-string
Fixed an issue in STL formatters where std::stlcontainer::iterator would match the container's synthetic children providers
Edited the syntax and help for some commands to have proper argument types
llvm-svn: 139160
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
revision and adding a patch that fixes an AsmParser
crash on ARM.
One feature that we unfortunately lost (for the
moment!) is the ability to cast unknown code symbols
to arbitrary function types and put the resulting
function pointer in a result variable. This feature
will be back, though.
llvm-svn: 138036
cause extra shared pointer references to one or more modules to be leaked.
This would cause many object files to stay around the life of LLDB, so after
a recompile and rexecution, we would keep adding more and more memory. After
fixing the leak, we found many cases where leaked stack frames were still
being used and causing crashes in the test suite. These are now all resolved.
llvm-svn: 137516
ability to dump more information about modules in "target modules list". We
can now dump the shared pointer reference count for modules, the pointer to
the module itself (in case performance tools can help track down who has
references to said pointer), and the modification time.
Added "target delete [target-idx ...]" to be able to delete targets when they
are no longer needed. This will help track down memory usage issues and help
to resolve when module ref counts keep getting incremented. If the command gets
no arguments, the currently selected target will be deleted. If any arguments
are given, they must all be valid target indexes (use the "target list"
command to get the current target indexes).
Took care of a bunch of "no newline at end of file" warnings.
TimeValue objects can now dump their time to a lldb_private::Stream object.
Modified the "target modules list --global" command to not error out if there
are no targets since it doesn't require a target.
Fixed an issue in the MacOSX DYLD dynamic loader plug-in where if a shared
library was updated on disk, we would keep using the older one, even if it was
updated.
Don't allow the ModuleList::GetSharedModule(...) to return an empty module.
Previously we could specify a valid path on disc to a module, and specify an
architecture that wasn't contained in that module and get a shared pointer to
a module that wouldn't be able to return an object file or a symbol file. We
now make sure an object file can be extracted prior to adding the shared pointer
to the module to get added to the shared list.
llvm-svn: 137196
if your datatype provides synthetic children, "frame variable object[index]" should now do the right thing
in cases where the above syntax would have been rejected before, i.e.
object is not a pointer nor an array (frame variable ignores potential overload of [])
object is a pointer to an Objective-C class (which cannot be dereferenced)
expression will still run operator[] if available and complain if it cannot do so
synthetic children by name do not work yet
llvm-svn: 137097
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
a native architecture that doesn't match the universal
slice that is being used for all executables, we weren't
correctly descending through the platform architectures
and resolving the binaries.
llvm-svn: 136980
the SBType implementation classes.
Fixed LLDB core and the test suite to not use deprecated SBValue APIs.
Added a few new APIs to SBValue:
int64_t
SBValue::GetValueAsSigned(int64_t fail_value=0);
uint64_t
SBValue::GetValueAsUnsigned(uint64_t fail_value=0)
llvm-svn: 136829
Fixed a bug where Objective-C variables coming out of the expression parser could crash the Python synthetic providers:
- expression parser output has a "frozen data" component, which is a byte-exact copy of the value (in host memory),
if trying to read into memory based on the host address, LLDB would crash. we are now passing the correct (target)
pointer to the Python code
Objective-C "id" variables are now formatted according to their dynamic type, if the -d option to frame variable is used:
- Code based on the Objective-C 2.0 runtime is used to obtain this information without running code on the target
llvm-svn: 136695
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
- Completely new implementation of SBType
- Various enhancements in several other classes
Python synthetic children providers for std::vector<T>, std::list<T> and std::map<K,V>:
- these return the actual elements into the container as the children of the container
- basic template name parsing that works (hopefully) on both Clang and GCC
- find them in examples/synthetic and in the test suite in functionalities/data-formatter/data-formatter-python-synth
New summary string token ${svar :
- the syntax is just the same as in ${var but this new token lets you read the values
coming from the synthetic children provider instead of the actual children
- Python providers above provide a synthetic child len that returns the number of elements
into the container
Full bug fix for the issue in which getting byte size for a non-complete type would crash LLDB
Several other fixes, including:
- inverted the order of arguments in the ClangASTType constructor
- EvaluationPoint now only returns SharedPointer's to Target and Process
- the help text for several type subcommands now correctly indicates argument-less options as such
llvm-svn: 136504
- Summaries for char*, const char* and char[] are loaded at startup as
system-wide summaries. This means you cannot delete them unless you use
the -a option to type summary delete/clear
- You can add your own system-wide summaries by using the -w option to type
summary add
Several code improvements for the Python summaries feature
llvm-svn: 135326
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
- a new --name option for "type summary add" lets you give a name to a summary
- a new --summary option for "frame variable" lets you bind a named summary to one or more variables
${var%s} now works for printing the value of 0-terminated CStrings
type format test case now tests for cascading
- this is disabled on GCC because GCC may end up stripping typedef chains, basically breaking cascading
new design for the FormatNavigator class
new template class CleanUp2 meant to support cleanup routines with 1 additional parameter beyond resource handle
llvm-svn: 134943
with the "target modules lookup --address <addr>" command. The variable
ID's, names, types, location for the address, and declaration is
displayed.
This can really help with crash logs since we get, on MacOSX at least,
the registers for the thread that crashed so it is often possible to
figure out some of the variable contents.
llvm-svn: 134886
use lldb_private::Target::ReadMemory(...) to allow constant strings
to be displayed in global variables prior on in between process
execution.
Centralized the variable declaration dumping into:
bool
Variable::DumpDeclaration (Stream *s, bool show_fullpaths, bool show_module);
Fixed an issue if you used "target variable --regex <regex>" where the
variable name would not be displayed, but the regular expression would.
Fixed an issue when viewing global variables through "target variable"
might not display correctly when doing DWARF in object files.
llvm-svn: 134878
Made it so that you can create synthetic children of array
value objects. This is for creating array members when the
array index is out of range. This comes in handy when you have
a structure definition like:
struct Collection
{
uint32_t count;
Item array[0];
};
"array" has 1 item, but many times in practice there are more
items in "item_array".
This allows you to do:
(lldb) target variable g_collection.array[3]
To implement this, the get child at index has been modified
to have a "ignore_array_bounds" boolean that can be set to true.
llvm-svn: 134846
variable where we were using a static cast when we shouldn't have been.
Fixed another array related issue where we weren't looking through a typedef.
llvm-svn: 134823
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
objective C or C++ methods when "self" or "this" were in scope, but had
invalid locations in a DWARF location list. The lack of a valid value caused
us to use an invalid type value and then we tried to import that invalid
value and we would crash.
llvm-svn: 134518
- ${*expr} now simply means to dereference expr before actually using it
- bitfields, array ranges and pointer ranges now work in a (hopefully) more natural and language-compliant way
a new class TypeHierarchyNavigator replicates the behavior of the FormatManager in going through type hierarchies
when one-lining summary strings, children's summaries can be used as well as values
llvm-svn: 134458
- type names can now be regular expressions (exact matching is done first, and is faster)
- integral (and floating) types can be printed as bitfields, i.e. ${var[low-high]} will extract bits low thru high of the value and print them
- array subscripts are supported, both for arrays and for pointers. the syntax is ${*var[low-high]}, or ${*var[]} to print the whole array (the latter only works for statically sized arrays)
- summary is now printed by default when a summary string references a variable. if that variable's type has no summary, value is printed instead. to force value, you can use %V as a format specifier
- basic support for ObjectiveC:
- ObjectiveC inheritance chains are now walked through
- %@ can be specified as a summary format, to print the ObjectiveC runtime description for an object
- some bug fixes
llvm-svn: 134293
"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
implements three commands:
type summary add <format> <typename1> [<typename2> ...]
type summary delete <typename1> [<typename2> ...]
type summary list [<typename1> [<typename2>] ...]
type summary clear
This allows you to specify the default format that will be used to display
summaries for variables, shown when you use "frame variable" or "expression", or the SBValue classes.
Examples:
type summary add "x = ${var.x}" Point
type summary list
type summary add --one-liner SimpleType
llvm-svn: 134108
level in the public API.
Also modified the ValueObject values to be able to display global variables
without having a valid running process. The globals will read themselves from
the object file section data if there is no process, and from the process if
there is one.
Also fixed an issue where modifications for dynamic types could cause child
values of ValueObjects to not show up if the value was unable to evaluate
itself (children of NULL pointer objects).
llvm-svn: 134102
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
This commit adds a new top level command named "type". Currently this command
implements three commands:
type format add <format> <typename1> [<typename2> ...]
type format delete <typename1> [<typename2> ...]
type format list [<typename1> [<typename2>] ...]
This allows you to specify the default format that will be used to display
types when you use "frame variable" or "expression", or the SBValue classes.
Examples:
// Format uint*_t as hex
type format add x uint16_t uint32_t uint64_t
// Format intptr_t as a pointer
type format add p intptr_t
The format characters are the same as "printf" for the most part with many
additions. These format character specifiers are also used in many other
commands ("frame variable" for one). The current list of format characters
include:
a - char buffer
b - binary
B - boolean
c - char
C - printable char
d - signed decimal
e - float
f - float
g - float
i - signed decimal
I - complex integer
o - octal
O - OSType
p - pointer
s - c-string
u - unsigned decimal
x - hex
X - complex float
y - bytes
Y - bytes with ASCII
llvm-svn: 133728
operator counts due to bad debug DWARF debug info. We now verify the operator
has a valid number of params using the clang operator tables.
llvm-svn: 133375
This us useful because sometomes you have to show a single character as: 'a'
(using eFormatChar) and other times you might have an array of single
charcters for display as: 'a' 'b' 'c', and other times you might want to
show the contents of buffer of characters that can contain non printable
chars: "\0\x22\n123".
This also fixes an issue that currently happens when you have a single character
C string (const char *a = "a"; or char b[1] = { 'b' };) that was being output
as "'a'" incorrectly due to the way the eFormatChar format output worked.
llvm-svn: 133316
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
representing variables whose type must be inferred
from the way they are used. Functions without debug
information now return UnknownAnyTy and must be cast.
Variables with no debug information are not yet using
UnknownAnyTy; instead they are assumed to be void*.
Support for variables of unknown type is coming (and,
in fact, some relevant support functions are included
in this commit) but will take a bit of extra effort.
The testsuite has also been updated to reflect the new
requirement that the result of printf be cast, i.e.
expr (int) printf("Hello world!")
llvm-svn: 131263
respective ABI plugins as they were plug-ins that supplied ABI specfic info.
Also hookep up the UnwindAssemblyInstEmulation so that it can generate the
unwind plans for ARM.
Changed the way ABI plug-ins are handed out when you get an instance from
the plug-in manager. They used to return pointers that would be mananged
individually by each client that requested them, but now they are handed out
as shared pointers since there is no state in the ABI objects, they can be
shared.
llvm-svn: 131193
looked up. Queries for global types were made
too specific -- including the current module
and compile unit in the query was limiting the
search when we wanted a truly global search.
llvm-svn: 131145
Switch the EmulateInstruction to use the standard RegisterInfo structure
that is defined in the lldb private types intead of passing the reg kind and
reg num everywhere. EmulateInstruction subclasses also need to provide
RegisterInfo structs given a reg kind and reg num. This eliminates the need
for the GetRegisterName() virtual function and allows more complete information
to be passed around in the read/write register callbacks. Subclasses should
always provide RegiterInfo structs with the generic register info filled in as
well as at least one kind of register number in the RegisterInfo.kinds[] array.
llvm-svn: 130256
inline contexts when the deepest most block is not inlined.
Added source path remappings to the lldb_private::Target class that allow it
to remap paths found in debug info so we can find source files that are elsewhere
on the current system.
Fixed disassembly by function name to disassemble inline functions that are
inside other functions much better and to show enough context before the
disassembly output so you can tell where things came from.
Added the ability to get more than one address range from a SymbolContext
class for the case where a block or function has discontiguous address ranges.
llvm-svn: 130044
line tables specify breakpoints can be set in the source. When dumping the
source, the number of breakpoints that can be set on a source line are shown
as a prefix:
(lldb) source list -f test.c -l1 -c222 -b
1 #include <stdio.h>
2 #include <sys/fcntl.h>
3 #include <unistd.h>
4 int
5 sleep_loop (const int num_secs)
[2] 6 {
7 int i;
[1] 8 for (i=0; i<num_secs; ++i)
9 {
[1] 10 printf("%d of %i - sleep(1);\n", i, num_secs);
[1] 11 sleep(1);
12 }
13 return 0;
[1] 14 }
15
16 int
17 main (int argc, char const* argv[])
[1] 18 {
[1] 19 printf("Process: %i\n\n", getpid());
[1] 20 puts("Press any key to continue..."); getchar();
[1] 21 sleep_loop (20);
22 return 12;
[1] 23 }
Above we can see there are two breakpoints for line 6 and one breakpoint for
lines 8, 10, 11, 14, 18, 19, 20, 21 and 23. All other lines have no line table
entries for them. This helps visualize the data provided in the debug
information without having to manually dump all line tables. It also includes
all inline breakpoint that may result for a given file which can also be very
handy to see.
llvm-svn: 129747
threads, and stack frame down in the lldb_private::Process,
lldb_private::Thread, lldb_private::StackFrameList and the
lldb_private::StackFrame classes. We had some command line
commands that had duplicate versions of the process status
output ("thread list" and "process status" for example).
Removed the "file" command and placed it where it should
have been: "target create". Made an alias for "file" to
"target create" so we stay compatible with GDB commands.
We can now have multple usable targets in lldb at the
same time. This is nice for comparing two runs of a program
or debugging more than one binary at the same time. The
new command is "target select <target-idx>" and also to see
a list of the current targets you can use the new "target list"
command. The flow in a debug session can be:
(lldb) target create /path/to/exe/a.out
(lldb) breakpoint set --name main
(lldb) run
... hit breakpoint
(lldb) target create /bin/ls
(lldb) run /tmp
Process 36001 exited with status = 0 (0x00000000)
(lldb) target list
Current targets:
target #0: /tmp/args/a.out ( arch=x86_64-apple-darwin, platform=localhost, pid=35999, state=stopped )
* target #1: /bin/ls ( arch=x86_64-apple-darwin, platform=localhost, pid=36001, state=exited )
(lldb) target select 0
Current targets:
* target #0: /tmp/args/a.out ( arch=x86_64-apple-darwin, platform=localhost, pid=35999, state=stopped )
target #1: /bin/ls ( arch=x86_64-apple-darwin, platform=localhost, pid=36001, state=exited )
(lldb) bt
* thread #1: tid = 0x2d03, 0x0000000100000b9a a.out`main + 42 at main.c:16, stop reason = breakpoint 1.1
frame #0: 0x0000000100000b9a a.out`main + 42 at main.c:16
frame #1: 0x0000000100000b64 a.out`start + 52
Above we created a target for "a.out" and ran and hit a
breakpoint at "main". Then we created a new target for /bin/ls
and ran it. Then we listed the targest and selected our original
"a.out" program, so we showed two concurent debug sessions
going on at the same time.
llvm-svn: 129695
the CommandInterpreter where it was always being used.
Make sure that Modules can track their object file offsets correctly to
allow opening of sub object files (like the "__commpage" on darwin).
Modified the Platforms to be able to launch processes. The first part of this
move is the platform soon will become the entity that launches your program
and when it does, it uses a new ProcessLaunchInfo class which encapsulates
all process launching settings. This simplifies the internal APIs needed for
launching. I want to slowly phase out process launching from the process
classes, so for now we can still launch just as we used to, but eventually
the platform is the object that should do the launching.
Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able
to launch processes with all of the new eLaunchFlag settings. Modified any
code that was manually launching processes to use the Host::LaunchProcess
functions.
Fixed an issue where lldb_private::Args had implicitly defined copy
constructors that could do the wrong thing. This has now been fixed by adding
an appropriate copy constructor and assignment operator.
Make sure we don't add empty ModuleSP entries to a module list.
Fixed the commpage module creation on MacOSX, but we still need to train
the MacOSX dynamic loader to not get rid of it when it doesn't have an entry
in the all image infos.
Abstracted many more calls from in ProcessGDBRemote down into the
GDBRemoteCommunicationClient subclass to make the classes cleaner and more
efficient.
Fixed the default iOS ARM register context to be correct and also added support
for targets that don't support the qThreadStopInfo packet by selecting the
current thread (only if needed) and then sending a stop reply packet.
Debugserver can now start up with a --unix-socket (-u for short) and can
then bind to port zero and send the port it bound to to a listening process
on the other end. This allows the GDB remote platform to spawn new GDB server
instances (debugserver) to allow platform debugging.
llvm-svn: 129351
This allows you to have a platform selected, then specify a triple using
"i386" and have the remaining triple items (vendor, os, and environment) set
automatically.
Many interpreter commands take the "--arch" option to specify an architecture
triple, so now the command options needed to be able to get to the current
platform, so the Options class now take a reference to the interpreter on
construction.
Modified the build LLVM building in the Xcode project to use the new
Xcode project level user definitions:
LLVM_BUILD_DIR - a path to the llvm build directory
LLVM_SOURCE_DIR - a path to the llvm sources for the llvm that will be used to build lldb
LLVM_CONFIGURATION - the configuration that lldb is built for (Release,
Release+Asserts, Debug, Debug+Asserts).
I also changed the LLVM build to not check if "lldb/llvm" is a symlink and
then assume it is a real llvm build directory versus the unzipped llvm.zip
package, so now you can actually have a "lldb/llvm" directory in your lldb
sources.
llvm-svn: 129112
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
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
platform status -- gets status information for the selected platform
platform create <platform-name> -- creates a new instance of a remote platform
platform list -- list all available platforms
platform select -- select a platform instance as the current platform (not working yet)
When using "platform create" it will create a remote platform and make it the
selected platform. For instances for iPhone OS debugging on Mac OS X one can
do:
(lldb) platform create remote-ios --sdk-version=4.0
Remote platform: iOS platform
SDK version: 4.0
SDK path: "/Developer/Platforms/iPhoneOS.platform/DeviceSupport/4.0"
Not connected to a remote device.
(lldb) file ~/Documents/a.out
Current executable set to '~/Documents/a.out' (armv6).
(lldb) image list
[ 0] /Volumes/work/gclayton/Documents/devb/attach/a.out
[ 1] /Developer/Platforms/iPhoneOS.platform/DeviceSupport/4.0/Symbols/usr/lib/dyld
[ 2] /Developer/Platforms/iPhoneOS.platform/DeviceSupport/4.0/Symbols/usr/lib/libSystem.B.dylib
Note that this is all happening prior to running _or_ connecting to a remote
platform. Once connected to a remote platform the OS version might change which
means we will need to update our dependecies. Also once we run, we will need
to match up the actualy binaries with the actualy UUID's to files in the
SDK, or download and cache them locally.
This is just the start of the remote platforms, but this modification is the
first iteration in getting the platforms really doing something.
llvm-svn: 127934
Still need to add "in methods of a class" to the specifiers, and the ability to write the stop hooks in the Scripting language as well as in the Command Language.
llvm-svn: 127457
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
of Stephen Wilson's idea (thanks for the input Stephen!). What I ended up
doing was:
- Got rid of ArchSpec::CPU (which was a generic CPU enumeration that mimics
the contents of llvm::Triple::ArchType). We now rely upon the llvm::Triple
to give us the machine type from llvm::Triple::ArchType.
- There is a new ArchSpec::Core definition which further qualifies the CPU
core we are dealing with into a single enumeration. If you need support for
a new Core and want to debug it in LLDB, it must be added to this list. In
the future we can allow for dynamic core registration, but for now it is
hard coded.
- The ArchSpec can now be initialized with a llvm::Triple or with a C string
that represents the triple (it can just be an arch still like "i386").
- The ArchSpec can still initialize itself with a architecture type -- mach-o
with cpu type and subtype, or ELF with e_machine + e_flags -- and this will
then get translated into the internal llvm::Triple::ArchSpec + ArchSpec::Core.
The mach-o cpu type and subtype can be accessed using the getter functions:
uint32_t
ArchSpec::GetMachOCPUType () const;
uint32_t
ArchSpec::GetMachOCPUSubType () const;
But these functions are just converting out internal llvm::Triple::ArchSpec
+ ArchSpec::Core back into mach-o. Same goes for ELF.
All code has been updated to deal with the changes.
This should abstract us until later when the llvm::TargetSpec stuff gets
finalized and we can then adopt it.
llvm-svn: 126278
a Stream, and then added GetOutputData & GetErrorData to get the accumulated data.
- Added a StreamTee that will tee output to two provided lldb::StreamSP's.
- Made the CommandObjectReturn use this so you can Tee the results immediately to
the debuggers output file, as well as saving up the results to return when the command
is done executing.
- HandleCommands now uses this so that if you have a set of commands that continue the target
you will see the commands come out as they are processed.
- The Driver now uses this to output the command results as you go, which makes the interface
more reactive seeming.
llvm-svn: 126015
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
now, in addition to cpu type/subtype and architecture flavor, contains:
- byte order (big endian, little endian)
- address size in bytes
- llvm::Triple for true target triple support and for more powerful plug-in
selection.
llvm-svn: 125602
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
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
Greg Clayton