To make this work this patch extends LLDB to:
- Explicitly track the link_map address for each module. This is effectively the module handle, not sure why it wasn't already being stored off anywhere. As an extension later, it would be nice if someone were to add support for printing this as part of the modules list.
- Allow reading the per-thread data pointer via ptrace. I have added support for Linux here. I'll be happy to add support for FreeBSD once this is reviewed. OS X does not appear to have __thread variables, so maybe we don't need it there. Windows support should eventually be workable along the same lines.
- Make DWARF expressions track which module they originated from.
- Add support for the DW_OP_GNU_push_tls_address DWARF opcode, as generated by gcc and recent versions of clang. Earlier versions of clang (such as 3.2, which is default on Ubuntu right now) do not generate TLS debug info correctly so can not be supported here.
- Understand the format of the pthread DTV block. This is where it gets tricky. We have three basic options here:
1) Call "dlinfo" or "__tls_get_addr" on the inferior and ask it directly. However this won't work on core dumps, and generally speaking it's not a good idea for the debugger to call functions itself, as it has the potential to not work depending on the state of the target.
2) Use libthread_db. This is what GDB does. However this option requires having a version of libthread_db on the host cross-compiled for each potential target. This places a large burden on the user, and would make it very hard to cross-debug from Windows to Linux, for example. Trying to build a library intended exclusively for one OS on a different one is not pleasant. GDB sidesteps the problem and asks the user to figure it out.
3) Parse the DTV structure ourselves. On initial inspection this seems to be a bad option, as the DTV structure (the format used by the runtime to manage TLS data) is not in fact a kernel data structure, it is implemented entirely in useerland in libc. Therefore the layout of it's fields are version and OS dependent, and are not standardized.
However, it turns out not to be such a problem. All OSes use basically the same algorithm (a per-module lookup table) as detailed in Ulrich Drepper's TLS ELF ABI document, so we can easily write code to decode it ourselves. The only question therefore is the exact field layouts required. Happily, the implementors of libpthread expose the structure of the DTV via metadata exported as symbols from the .so itself, designed exactly for this kind of thing. So this patch simply reads that metadata in, and re-implements libthread_db's algorithm itself. We thereby get cross-platform TLS lookup without either requiring third-party libraries, while still being independent of the version of libpthread being used.
Test case included.
llvm-svn: 192922
the extra check introduces 22 new test failures with the LLDB clang buildbot.
Note that the unhandled DWARF_OP codes in DWARFExpression::Evaluate don't cause test failures if the check is ignored.
llvm-svn: 187480
in LLDB that load the canonical frame address rather than a location list.
- Handles the simple case where a CFA can be pulled from the current stack frame.
- Fixes more than one hundred failing tests with gcc 4.8!
TODO: Use UnwindPlan::GetRowForFunctionOffset if the DWARFExpression needs
to be evaluated in a context analogous to a virtual unwind (perhaps using RegisterContextLLDB).
- Also adds some comments to DWARFCallFrameInfo whenever I got confused.
llvm-svn: 187361
A long time ago we start with clang types that were created by the symbol files and there were many functions in lldb_private::ClangASTContext that helped. Later we create ClangASTType which contains a clang::ASTContext and an opauque QualType, but we didn't switch over to fully using it. There were a lot of places where we would pass around a raw clang_type_t and also pass along a clang::ASTContext separately. This left room for error.
This checkin change all type code over to use ClangASTType everywhere and I cleaned up the interfaces quite a bit. Any code that was in ClangASTContext that was type related, was moved over into ClangASTType. All code that used these types was switched over to use all of the new goodness.
llvm-svn: 186130
Show variables that were in the debug info but optimized out. Also display a good error message when one of these variables get used in an expression.
llvm-svn: 182066
LLDB is crashing when logging is enabled from lldb-perf-clang. This has to do with the global destructor chain as the process and its threads are being torn down.
All logging channels now make one and only one instance that is kept in a global pointer which is never freed. This guarantees that logging can correctly continue as the process tears itself down.
llvm-svn: 178191
DWARF with .o files now uses 40-60% less memory!
Big fixes include:
- Change line table internal representation to contain "file addresses". Since each line table is owned by a compile unit that is owned by a module, it makes address translation into lldb_private::Address easy to do when needed.
- Removed linked address members/methods from lldb_private::Section and lldb_private::Address
- lldb_private::LineTable can now relink itself using a FileRangeMap to make it easier to re-link line tables in the future
- Added ObjectFile::ClearSymtab() so that we can get rid of the object file symbol tables after we parse them once since they are not needed and kept memory allocated for no reason
- Moved the m_sections_ap (std::auto_ptr to section list) and m_symtab_ap (std::auto_ptr to the lldb_private::Symtab) out of each of the ObjectFile subclasses and put it into lldb_private::ObjectFile.
- Changed how the debug map is parsed and stored to be able to:
- Lazily parse the debug map for each object file
- not require the address map for a .o file until debug information is linked for a .o file
llvm-svn: 176454
Major fixed to allow reading files that are over 4GB. The main problems were that the DataExtractor was using 32 bit offsets as a data cursor, and since we mmap all of our object files we could run into cases where if we had a very large core file that was over 4GB, we were running into the 4GB boundary.
So I defined a new "lldb::offset_t" which should be used for all file offsets.
After making this change, I enabled warnings for data loss and for enexpected implicit conversions temporarily and found a ton of things that I fixed.
Any functions that take an index internally, should use "size_t" for any indexes and also should return "size_t" for any sizes of collections.
llvm-svn: 173463
Fixed an issue that could cause references the shared data for an object file to stay around longer than intended and could cause memory bloat when debugging multiple times.
llvm-svn: 161716
parser was creating malformed resuls. When the
location of a variable is computed by reading a
register and adding an offset, we shouldn't say
that the variable's value is located in that
register. This was confusing the expression
parser when trying to read a variable captured
by a block.
llvm-svn: 147668
Be better at detecting when DWARF changes and handle this more
gracefully than asserting and exiting.
Also fixed up a bunch of system calls that weren't properly checking
for EINTR.
llvm-svn: 147559
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
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
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
__attribute__ format so the compiler knows that this method takes
printf style formatter arguments and checks that it's being used
correctly. Fix a couple dozen incorrect SetErrorStringWithFormat()
calls throughout the sources.
llvm-svn: 140115
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
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
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
into some cleanup I have been wanting to do when reading/writing registers.
Previously all RegisterContext subclasses would need to implement:
virtual bool
ReadRegisterBytes (uint32_t reg, DataExtractor &data);
virtual bool
WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0);
There is now a new class specifically designed to hold register values:
lldb_private::RegisterValue
The new register context calls that subclasses must implement are:
virtual bool
ReadRegister (const RegisterInfo *reg_info, RegisterValue ®_value) = 0;
virtual bool
WriteRegister (const RegisterInfo *reg_info, const RegisterValue ®_value) = 0;
The RegisterValue class must be big enough to handle any register value. The
class contains an enumeration for the value type, and then a union for the
data value. Any integer/float values are stored directly in an appropriate
host integer/float. Anything bigger is stored in a byte buffer that has a length
and byte order. The RegisterValue class also knows how to copy register value
bytes into in a buffer with a specified byte order which can be used to write
the register value down into memory, and this does the right thing when not
all bytes from the register values are needed (getting a uint8 from a uint32
register value..).
All RegiterContext and other sources have been switched over to using the new
regiter value class.
llvm-svn: 131096
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
an issue with the way the UnwindLLDB was handing out RegisterContexts: it
was making shared pointers to register contexts and then handing out just
the pointers (which would get put into shared pointers in the thread and
stack frame classes) and cause double free issues. MallocScribble helped to
find these issues after I did some other cleanup. To help avoid any
RegisterContext issue in the future, all code that deals with them now
returns shared pointers to the register contexts so we don't end up with
multiple deletions. Also now that the RegisterContext class doesn't require
a stack frame, we patched a memory leak where a StackFrame object was being
created and leaked.
Made the RegisterContext class not have a pointer to a StackFrame object as
one register context class can be used for N inlined stack frames so there is
not a 1 - 1 mapping. Updates the ExecutionContextScope part of the
RegisterContext class to never return a stack frame to indicate this when it
is asked to recreate the execution context. Now register contexts point to the
concrete frame using a concrete frame index. Concrete frames are all of the
frames that are actually formed on the stack of a thread. These concrete frames
can be turned into one or more user visible frames due to inlining. Each
inlined stack frame has the exact same register context (shared via shared
pointers) as any parent inlined stack frames all the way up to the concrete
frame itself.
So now the stack frames and the register contexts should behave much better.
llvm-svn: 122976
RegisterContext* - normally this is retrieved from the ExecutionContext's
StackFrame but when we need to evaluate an expression while creating
the stack frame list this can be a little tricky.
Add DW_OP_deref_size, needed for the _sigtramp FDE expression.
Add support for processing DWARF expressions in RegisterContextLLDB.
Update callers to DWARFExpression::Evaluate.
llvm-svn: 119885
cases when getting the clang type:
- need only a forward declaration
- need a clang type that can be used for layout (members and args/return types)
- need a full clang type
This allows us to partially parse the clang types and be as lazy as possible.
The first case is when we just need to declare a type and we will complete it
later. The forward declaration happens only for class/union/structs and enums.
The layout type allows us to resolve the full clang type _except_ if we have
any modifiers on a pointer or reference (both R and L value). In this case
when we are adding members or function args or return types, we only need to
know how the type will be laid out and we can defer completing the pointee
type until we later need it. The last type means we need a full definition for
the clang type.
Did some renaming of some enumerations to get rid of the old "DC" prefix (which
stands for DebugCore which is no longer around).
Modified the clang namespace support to be almost ready to be fed to the
expression parser. I made a new ClangNamespaceDecl class that can carry around
the AST and the namespace decl so we can copy it into the expression AST. I
modified the symbol vendor and symbol file plug-ins to use this new class.
llvm-svn: 118976
don't crash if we disable logging when some code already has a copy of the
logger. Prior to this fix, logs were handed out as pointers and if they were
held onto while a log got disabled, then it could cause a crash. Now all logs
are handed out as shared pointers so this problem shouldn't happen anymore.
We are also using our new shared pointers that put the shared pointer count
and the object into the same allocation for a tad better performance.
llvm-svn: 118319
adding methods to C++ and objective C classes. In order to make methods, we
need the function prototype which means we need the arguments. Parsing these
could cause a circular reference that caused an assertion.
Added a new typedef for the clang opaque types which are just void pointers:
lldb::clang_type_t. This appears in lldb-types.h.
This was fixed by enabling struct, union, class, and enum types to only get
a forward declaration when we make the clang opaque qual type for these
types. When they need to actually be resolved, lldb_private::Type will call
a new function in the SymbolFile protocol to resolve a clang type when it is
not fully defined (clang::TagDecl::getDefinition() returns NULL). This allows
us to be a lot more lazy when parsing clang types and keeps down the amount
of data that gets parsed into the ASTContext for each module.
Getting the clang type from a "lldb_private::Type" object now takes a boolean
that indicates if a forward declaration is ok:
clang_type_t lldb_private::Type::GetClangType (bool forward_decl_is_ok);
So function prototypes that define parameters that are "const T&" can now just
parse the forward declaration for type 'T' and we avoid circular references in
the type system.
llvm-svn: 115012
debug map showed that the location lists in the .o files needed some
refactoring in order to work. The case that was failing was where a function
that was in the "__TEXT.__textcoal_nt" in the .o file, and in the
"__TEXT.__text" section in the main executable. This made symbol lookup fail
due to the way we were finding a real address in the debug map which was
by finding the section that the function was in in the .o file and trying to
find this in the main executable. Now the section list supports finding a
linked address in a section or any child sections. After fixing this, we ran
into issue that were due to DWARF and how it represents locations lists.
DWARF makes a list of address ranges and expressions that go along with those
address ranges. The location addresses are expressed in terms of a compile
unit address + offset. This works fine as long as nothing moves around. When
stuff moves around and offsets change between the remapped compile unit base
address and the new function address, then we can run into trouble. To deal
with this, we now store supply a location list slide amount to any location
list expressions that will allow us to make the location list addresses into
zero based offsets from the object that owns the location list (always a
function in our case).
With these fixes we can now re-link random address ranges inside the debugger
for use with our DWARF + debug map, incremental linking, and more.
Another issue that arose when doing the DWARF in the .o files was that GCC
4.2 emits a ".debug_aranges" that only mentions functions that are externally
visible. This makes .debug_aranges useless to us and we now generate a real
address range lookup table in the DWARF parser at the same time as we index
the name tables (that are needed because .debug_pubnames is just as useless).
llvm-gcc doesn't generate a .debug_aranges section, though this could be
fixed, we aren't going to rely upon it.
Renamed a bunch of "UINT_MAX" to "UINT32_MAX".
llvm-svn: 113829
complex inlined examples.
StackFrame classes don't have a "GetPC" anymore, they have "GetFrameCodeAddress()".
This is because inlined frames will have a PC value that is the same as the
concrete frame that owns the inlined frame, yet the code locations for the
frame can be different. We also need to be able to get the real PC value for
a given frame so that variables evaluate correctly. To get the actual PC
value for a frame you can use:
addr_t pc = frame->GetRegisterContext()->GetPC();
Some issues with the StackFrame stomping on its own symbol context were
resolved which were causing the information to change for a frame when the
stack ID was calculated. Also the StackFrame will now correctly store the
symbol context resolve flags for any extra bits of information that were
looked up (if you ask for a block only and you find one, you will alwasy have
the compile unit and function).
llvm-svn: 111964
which is now on by default. Frames are gotten from the unwinder as concrete
frames, then if inline frames are to be shown, extra information to track
and reconstruct these frames is cached with each Thread and exanded as needed.
I added an inline height as part of the lldb_private::StackID class, the class
that helps us uniquely identify stack frames. This allows for two frames to
shared the same call frame address, yet differ only in inline height.
Fixed setting breakpoint by address to not require addresses to resolve.
A quick example:
% cat main.cpp
% ./build/Debug/lldb test/stl/a.out
Current executable set to 'test/stl/a.out' (x86_64).
(lldb) breakpoint set --address 0x0000000100000d31
Breakpoint created: 1: address = 0x0000000100000d31, locations = 1
(lldb) r
Launching 'a.out' (x86_64)
(lldb) Process 38031 Stopped
* thread #1: tid = 0x2e03, pc = 0x0000000100000d31, where = a.out`main [inlined] std::string::_M_data() const at /usr/include/c++/4.2.1/bits/basic_string.h:280, stop reason = breakpoint 1.1, queue = com.apple.main-thread
277
278 _CharT*
279 _M_data() const
280 -> { return _M_dataplus._M_p; }
281
282 _CharT*
283 _M_data(_CharT* __p)
(lldb) bt
thread #1: tid = 0x2e03, stop reason = breakpoint 1.1, queue = com.apple.main-thread
frame #0: pc = 0x0000000100000d31, where = a.out`main [inlined] std::string::_M_data() const at /usr/include/c++/4.2.1/bits/basic_string.h:280
frame #1: pc = 0x0000000100000d31, where = a.out`main [inlined] std::string::_M_rep() const at /usr/include/c++/4.2.1/bits/basic_string.h:288
frame #2: pc = 0x0000000100000d31, where = a.out`main [inlined] std::string::size() const at /usr/include/c++/4.2.1/bits/basic_string.h:606
frame #3: pc = 0x0000000100000d31, where = a.out`main [inlined] operator<< <char, std::char_traits<char>, std::allocator<char> > at /usr/include/c++/4.2.1/bits/basic_string.h:2414
frame #4: pc = 0x0000000100000d31, where = a.out`main + 33 at /Volumes/work/gclayton/Documents/src/lldb/test/stl/main.cpp:14
frame #5: pc = 0x0000000100000d08, where = a.out`start + 52
Each inline frame contains only the variables that they contain and each inlined
stack frame is treated as a single entity.
llvm-svn: 111877
ClangExpressionVariables for found external variables
as well as for struct members, replacing the Tuple
and StructMember data structures.
llvm-svn: 111859
Richard Mitton