This API is currently a no-op (in the sense that it has the same behavior as the already existing GetName()), but is meant long-term to provide a best-for-visualization version of the name of a function
It is still not hooked up to the command line 'bt' command, nor to the 'gui' mode, but I do have ideas on how to make that work going forward
rdar://21203242
llvm-svn: 241482
Previously we handled only a few opcode DWARF CFI opcode when we
were parsing the CIE. With this change we use mostly the same code
what we use for parsing the PDE which cover more opcode.
Differential revision: http://reviews.llvm.org/D10866
llvm-svn: 241332
The CFA offset shouldn't be irestored to the saved value in case
of a DW_CFA_restore_state opcode.
Differential revision: http://reviews.llvm.org/D10843
llvm-svn: 241331
A few extras were fixed
- Symbol::GetAddress() now returns an Address object, not a reference. There were places where people were accessing the address of a symbol when the symbol's value wasn't an address symbol. On MacOSX, undefined symbols have a value zero and some places where using the symbol's address and getting an absolute address of zero (since an Address object with no section and an m_offset whose value isn't LLDB_INVALID_ADDRESS is considered an absolute address). So fixing this required some changes to make sure people were getting what they expected.
- Since some places want to access the address as a reference, I added a few new functions to symbol:
Address &Symbol::GetAddressRef();
const Address &Symbol::GetAddressRef() const;
Linux test suite passes just fine now.
<rdar://problem/21494354>
llvm-svn: 240702
* Add and fix the emulation of several instruction.
* Disable frame pointer usage on Android.
* Specify return address register for the unwind plan instead of explict
tracking the value of RA.
* Replace prologue detection heuristics (unreliable in several cases)
with a logic to follow the branch instructions and restore the CFI
value based on them. The target address for a branch should have the
same CFI as the source address (if they are in the same function).
* Handle symbols in ELF files where the symbol size is not specified
with calcualting their size based on the next symbol (already done
in MachO files).
* Fix architecture in FuncUnwinders with filling up the inforamtion
missing from the object file with the architecture of the target.
* Add code to read register wehn the value is set to "IsSame" as it
meanse the value of a register in the parent frame is the same as the
value in the current frame.
Differential revision: http://reviews.llvm.org/D10447
llvm-svn: 240533
The problem is for lldb_private::Type instances that have encoding types (pointer/reference/const/volatile/restrict/typedef to type with user ID 0x123). If they started out with m_flags.clang_type_resolve_state being set to eResolveStateUnresolved (0), then when we would call Type::ResolveClangType(eResolveStateForward) we would complete the full type due to logic errors in the code.
We now only complete the type if clang_type_resolve_state is eResolveStateLayout or eResolveStateFull and we correctly upgrade the type's current completion state to eResolveStateForward after we make a forward delcaration to the pointer/reference/const/volatile/restrict/typedef type instead of leaving it set to eResolveStateUnresolved.
llvm-svn: 239752
Since interaction with the python interpreter is moving towards
being more isolated, we won't be able to include this header from
normal files anymore, all includes of it should be localized to
the python library which will live under source/bindings/API/Python
after a future patch.
None of the files that were including this header actually depended
on it anyway, so it was just a dead include in every single instance.
llvm-svn: 238581
If binding to port 0 is selected, the actual port is printed.
This improves the reliability of platform startup by ensuring that
a free port can be found.
TEST PLAN
./lldb-server platform --listen *:0
Listening for a connection from <port-number>...
Will appear on stdout (with other stuff potentially)
llvm-svn: 238173
Removed some unused variables, added some consts, changed some casts
to const_cast. I don't think any of these changes are very
controversial.
Differential Revision: http://reviews.llvm.org/D9674
llvm-svn: 237218
The ClangASTContext::getTargetInfo() will return NULL in this case and could cause us to crash if we don't check.
<rdar://problem/20543554>
llvm-svn: 236681
Summary:
GetEHFrameAugmentedUnwindPlan duplicated the work of GetEHFrameUnwindPlan in getting the original
plan from DWARF CFI. This changes the function to call GetEHFrameUnwindPlan instead of doing all
the work itself. A copy constructor is added to UnwindPlan to enable plan copying.
Test Plan: No regressions on linux test suite.
Reviewers: jasonmolenda, clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D9369
llvm-svn: 236607
compact unwind encodings for x86_64 / i386 omit-frame-pointer
code. It was possible for lldb to get the location of saved
registers incorrect for some of these functions.
<rdar://problem/20753264>
llvm-svn: 236478
Summary: Just what it says on the box.
Reviewers: jasonmolenda
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D9150
llvm-svn: 235493
module-loading support for the expression parser.
- It adds support for auto-loading modules referred
to by a compile unit. These references are
currently in the form of empty translation units.
This functionality is gated by the setting
target.auto-import-clang-modules (boolean) = false
- It improves and corrects support for loading
macros from modules, currently by textually
pasting all #defines into the user's expression.
The improvements center around including only those
modules that are relevant to the current context -
hand-loaded modules and the modules that are imported
from the current compile unit.
- It adds an "opt-in" mechanism for all of this
functionality. Modules have to be explicitly
imported (via @import) or auto-loaded (by enabling
the above setting) to enable any of this
functionality.
It also adds support to the compile unit and symbol
file code to deal with empty translation units that
indicate module imports, and plumbs this through to
the CompileUnit interface.
Finally, it makes the following changes to the test
suite:
- It adds a testcase that verifies that modules are
automatically loaded when the appropriate setting
is enabled (lang/objc/modules-auto-import); and
- It modifies lanb/objc/modules-incomplete to test
the case where a module #undefs something that is
#defined in another module.
<rdar://problem/20299554>
llvm-svn: 235313
context as the first thing we do. This prevents
crashes if some of the initial setup produces
messages or errors.
<rdar://problem/20457882>
llvm-svn: 234511
Summary:
If a struct type S has a member T that has a member that is a function that
returns a typedef of S* the respective field would be duplicated, which caused
an assert down the line in RecordLayoutBuilder. This patch adds a check that
removes the possibility of trying to resolve the same type twice within the
same callstack.
This commit also adds unit tests for these failures.
Fixes https://llvm.org/bugs/show_bug.cgi?id=20486.
Patch by Cristian Hancila.
Test Plan: Run unit tests.
Reviewers: clayborg spyffe
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D8561
llvm-svn: 234441
There were a couple of real bugs here regarding error checking and
signed/unsigned comparisons, but mostly these were just noise.
There was one class of bugs fixed here which is particularly
annoying, dealing with MSVC's non-standard behavior regarding
the underlying type of enums. See the comment in
lldb-enumerations.h for details. In short, from now on please use
FLAGS_ENUM and FLAGS_ANONYMOUS_ENUM when defining enums which
contain values larger than can fit into a signed integer.
llvm-svn: 233943
The underlying type of wchar_t is not defined by the standard. This CL
add logic to correctly use the type specified for the current target
based on TargetInfo.
llvm-svn: 233795
A char can have signed and unsigned encoding but previously lldb always
assumed it is signed. This CL adds a logic to detect the encoding of
'char' types based on the default encoding on the target architecture.
It fixes variable printing and expression evaluation on architectures
where 'char' is signed by default.
Differential revision: http://reviews.llvm.org/D8636
llvm-svn: 233682
Since ClangASTSource::layoutRecordType() was overriding a virtual
function in the base, this was inadvertently causing a new method
to be introduced rather than an override. To fix this all method
signatures are changed back to taking DenseMaps, and the `override`
keyword is added to make sure this type of error doesn't happen
again.
To keep the original fix intact, which is that fields and bases
must be added in offset order, the ImportOffsetMap() function
now copies the DenseMap into a vector and then sorts the vector
on the value type (e.g. the offset) before iterating over the
sorted vector and inserting the items.
llvm-svn: 233099
Summary:
This commit adds this alternate route only when parsing variable dies
corresponding to global or static variables. The motivation for this is that GCC
does not emit linkage names for functions and variables declared/defined in
anonymous namespaces. Having this alternate route fixes one part of
TestNamespace which fails when the test case is compiled with GCC.
An alternate route to get fully qualified names of functions whose linkage names
are missing will be added with a followup change. With that, the other failing
part of TestNamespace will also be fixed.
Test Plan: dotest.py -C gcc -p TestNamespace
Reviewers: clayborg
Reviewed By: clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D8569
llvm-svn: 233098
Prior to this patch, we would try to synthesize class types by
iterating over a DenseMap of FieldDecls and adding each one to
a CXXRecordDecl. Since a DenseMap doesn't provide a deterministic
ordering of the elements, this would not add the fields in
FieldOffset order, but rather in some random order determined by
the memory layout of the DenseMap.
This patch fixes the issue by changing DenseMaps to vectors. The
ability to lookup a value in the DenseMap was hardly being used,
and where it is sufficient to do a vector lookup.
Differential Revision: http://reviews.llvm.org/D8512
llvm-svn: 233090
So that we don't have to update every single #include in the entire
codebase to #include this new header (which used to get included by
lldb-private-log.h, we automatically #include "Logging.h" from
within "Log.h".
llvm-svn: 232653
This removes Host::Backtrace from the codebase, and changes all
call sites to use llvm::sys::PrintStackTrace(). This makes the
functionality available for all platforms, and even for platforms
which currently had a supported implementation of Host::Backtrace,
this patch should enable richer information in stack traces, such
as file and line number information, as well as giving it the
ability to unwind through inlined functions.
llvm-svn: 231511
- use a hardcoded formatter to match all vector types, and make it so that their element type is taken into account when doing default formatting
- special case a vector of char to display byte values instead of characters by default
Fixes the test failures Ilia was seeing
llvm-svn: 231504
Summary:
Symbols in ELF files can be versioned, but LLDB currently does not understand these. This problem
becomes apparent once one loads glibc with debug info. Here (in the .symtab section) the versions
are embedded in the name (name@VERSION), which causes issues when evaluating expressions
referencing memcpy for example (current glibc contains memcpy@@GLIBC_2.14 and
memcpy@GLIBC_2.2.5).
This problem was not evident without debug symbols as the .dynsym section
stores the bare names and the actual versions are present in a separate section (.gnu.version_d),
which LLDB ignores. This resulted in two definitions of memcpy in the symbol table.
This patch adds support for storing annotated names to the Symbol class. If
Symbol.m_contains_linker_annotations is true then this symbol is annotated. Unannotated name can
be obtained by calling StripLinkerAnnotations on the corresponding ObjectFile. ObjectFileELF
implements this to strip @VERSION suffixes when requested. Symtab uses this function to add the
bare name as well as the annotated name to the name lookup table.
To preserve the size of the Symbol class, I had to steal one bit from the m_type field.
Test Plan:
This fixes TestExprHelpExamples.py when run with a glibc with debug symbols. Writing
an environment agnostic test case would require building a custom shared library with symbol
versions and testing symbol resolution against that, which is somewhat challenging.
Reviewers: clayborg, jingham
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D8036
llvm-svn: 231228
Debugger.h is a huge file that gets included everywhere, and
FormatManager.h brings in a ton of unnecessary stuff and doesn't
even use anything from it in the header.
llvm-svn: 231161
When we have a debug map we have an executable with a bunch of STAB symbols and each source file has a N_SO symbol which scopes a bunch of symbols inside of it. We can use this to our advantage here when looking for the complete definition of an objective C class by looking for a symbol whose name matches the class name and whose type is eSymbolTypeObjCClass. If we find one, that symbol will be contained within a N_SO symbol. This symbol gets turned into a symbol whose type is eSymbolTypeSourceFile and that symbol will contain the eSymbolTypeObjCClass which helps us to locate the correct .o file and allows us to only look in that file.
To further accelerate things, if we are looking for the implementation, we can avoid looking at all .o files if we don't find a matching symbol because we have a debug map, which means the objective C symbol for the class can't have been stripped, so we can safely not search all remaining .o files. This will save us lots of time when trying to look for "NSObject" and any other AppKit and Foundation classes that we never have implementation definitions for.
<rdar://problem/19234225>
llvm-svn: 230562
Summary:
This patch enables evaluation of DWARF expressions setting the CFA during stack unwinding.
This makes TestSigtrampUnwind "almost" pass on linux. I am not enabling the test yet since the
symbol name for the signal trampoline does not get resolved properly due to a different bug, but
apart from that, the backtrace is sane.
I am unsure how this change affects Mac. I think it makes the unwinder prefer the DWARF unwind
plan instead of some custom platform-dependant plan. However, it does not affect the end result
- the stack unwinding works as expected.
Reviewers: jasonmolenda
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D7792
llvm-svn: 230211
Summary:
This change refactors UnwindPlan::Row to be able to store the fact that the CFA is value is set
by evaluating a dwarf expression (DW_CFA_def_cfa_expression). This is achieved by creating a new
class CFAValue and moving all CFA setting/getting code there. Note that code using the new
CFAValue::isDWARFExpression is not yet present and will be added in a follow-up patch. Therefore,
this patch should not change the functionality in any way.
Test Plan: Ran tests on Mac and Linux. No regressions detected.
Reviewers: jasonmolenda, clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D7755
llvm-svn: 230210
changing it was in r219544 - after living on that for a few
months, I wanted to take another crack at this.
The disassembly-format setting still exists and the old format
can be user specified with a setting like
${current-pc-arrow}${addr-file-or-load}{ <${function.name-without-args}${function.concrete-only-addr-offset-no-padding}>}:
This patch was discussed in http://reviews.llvm.org/D7578
<rdar://problem/19726421>
llvm-svn: 229186
There was a test in the test suite that was triggering the backtrace logging output that requested that the client pass an execution context. Sometimes we need the process for Objective C types because our static notion of the type might not align with the reality when being run in a live runtime.
Switched from an "ExecutionContext *" to an "ExecutionContextScope *" for greater ease of use.
llvm-svn: 228892
- you have a type that contains a typedef to a VectorType or an ExtVectorType
- that type is returned from an ARM function that LLDB steps over so we try to figure out the return type
- we try to determine if the type is a homogeneous aggregate type and we crash
We get not using getAs() when we should have been and using llvm::cast caused an assertion crash when the typedef type didn't return a valid VectorType or ExtVectorType.
<rdar://problem/19646550>
llvm-svn: 228771
The problem occurred when we had incorrect address ranges in the debug map that included the padding between functions causing the end address of a line table entry to fall into an inlinked (next function) address range.
<rdar://problem/19721144>
llvm-svn: 228707
This was causing code that opened multiple targets to try and get a path to debugserver from the GDB remote communication class, and it would get the LLDB path and some instances would return empty strings and it would cause debugserver to not be found.
<rdar://problem/18756927>
llvm-svn: 227935
And since enough of these are doing the right thing, add a test case to verify we are doing the right thing with freeze drying ObjC object types
Fixes rdar://18092770
llvm-svn: 227282
This is necessary because the byte size of an ObjC class type is not reliably statically knowable (e.g. because superclasses sit deep in frameworks that we have no debug info for)
The lack of reliable size info is a problem when trying to freeze-dry an ObjC instance (not the pointer, the pointee)
This commit lays the foundation for having language runtimes help in figuring out byte sizes, and having ClangASTType ask for runtime help
No feature change as no runtime actually implements the logic, and nowhere is an ExecutionContext passed in yet
llvm-svn: 227274
i386/x86_64 functions. The stack size was being multiplied by the
pointer size incorrectly. The register permutation placeholders
(UNWIND_X86_REG_NONE) were decrementing the stack offset of the
saved registers when it should not have been.
<rdar://problem/19570035>
llvm-svn: 226889
The refactor was motivated by some comments that Greg made
http://reviews.llvm.org/D6918
and also to break a dependency cascade that caused functions linking
in string->int conversion functions to pull in most of lldb
llvm-svn: 226199
This is done by adding a "Variable *" to SymbolContext and allowing SymbolFile::ResolveSymbolContext() so if an address is resolved into a symbol context, we can include the global or static variable for that address.
This means you can now find global variables that are merged globals when doing a "image lookup --verbose --address 0x1230000". Previously we would resolve a symbol and show "_MergedGlobals123 + 1234". But now we can show the global variable name.
The eSymbolContextEverything purposely does not include the new eSymbolContextVariable in its lookup since stack frame code does many lookups and we don't want it triggering the global variable lookups.
<rdar://problem/18945678>
llvm-svn: 226084
step through the complete function looking for any epilogue
instructions. If we find an epilogue sequence, re-instate
the correct unwind instructions if there is more code past
that epilogue -- this will correctly handle an x86 function
with multiple epilogues in it.
NB there is still a bug with the "eh_frame augmented"
UnwindPlans and mid-function epilogues. Looking at that next.
<rdar://problem/18863406>
llvm-svn: 225770
This completes the compact unwind support for x86 targets.
I'm still skipping the UNWIND_X86_64_MODE_STACK_IND encodings for
x86_64 right now because clang was emitting bad data for this form
until it was fixed in r217020 circa Sep 2014.
arm64 parsing still needs to be added.
llvm-svn: 224698
Most of the changes are to the FuncUnwinders class -- as we've added
more types of unwind information, the way this class was written was
making it a mess to maintain. Instead of trying to keep one
"non-call site" unwind plan and one "call site" unwind plan, track
all the different types of unwind plans we can possibly retrieve for
each function and have the call-site/non-call-site accessor methods
retrieve those.
Add a real "fast unwind plan" for x86_64 / i386 -- when doing an
unwind through a function, this only has to read the first 4 bytes
to tell if the function has a standard prologue sequence. If so,
we can use the architecture default unwind plan to backtrace
through this function. If we try to retrieve the save location for
other registers later on, a real unwind plan will be used. This
one is just for doing fast backtraces.
Change the compact unwind plan importer to fill in the valid address
range it is valid for.
Compact unwind, in theory, may have multiple entries for a single
function. The FuncUnwinders rewrite includes the start of supporting
this correctly. In practice compact unwind encodings are used for
the entire range of the function today -- in fact, sometimes the same
encoding is used for multiple functions that have the same unwind
rules. But I want to handle a single function that has multiple
different compact unwind UnwindPlans eventually.
llvm-svn: 224689
When lldb has a binary with protected section contents,
don't use the on-disk representation of that compact
uwnind -- read it only out of live memory where it has
been decrypted.
llvm-svn: 224670
The compact unwind importer is getting the wrong unwind info for one
case that I found. I haven't been able to fix the problem tonight
and I don't want to leave TOT behaving incorrectly, so just ignore
compact unwind until I can get to the bottom of this.
llvm-svn: 224321
section for x86_64 and i386 targets on Darwin systems. Currently only the
compact unwind encoding for normal frame-using functions is supported but it
will be easy handle frameless functions when I have a bit more free time to
test it. The LSDA and personality routines for functions are also retrieved
correctly for functions from the compact unwind section.
This new code is very fresh -- it passes the lldb testsuite and I've done
by-hand inspection of many functions and am getting correct behavior for all
of them. There may need to be some bug fixing over the next couple weeks as
I exercise and test it further. But I think it's fine right now so I'm
committing it.
<rdar://problem/13220837>
llvm-svn: 223625
encounter clang::ExternalASTSources that are not instances
of ClangExternalASTSourceCommon. We used to blithely
assume that all are, and so we could use static_cast<>.
That's no longer the case, so we have to have these AST
sources register themselves.
llvm-svn: 223560
support to LLDB. It includes the following:
- Changed DeclVendor to TypeVendor.
- Made the ObjCLanguageRuntime provide a DeclVendor
rather than a TypeVendor.
- Changed the consumers of TypeVendors to use
DeclVendors instead.
- Provided a few convenience functions on
ClangASTContext to make that easier.
llvm-svn: 223433
retrieves the personality routine addr and the
LSDA addr. Don't bother checking with the
"non-call site" unwind plan - this kind of
information is only going to come from the
call site unwind plan.
llvm-svn: 222226
eh_frame data. These two pieces of information are used in the
process of exception handler unwinding on SysV ABI systems.
This patch reads the data from the eh_frame section
(DWARFCallFrameInfo.cpp), allows for it to be saved & read out
of a given UnwindPlan (UnwindPlan.h, UnwindPlan.cpp) - as well
as printing the information in the UnwindPlan::Dump method - and
adds methods to the FuncUnwinders object so that higher levels
can query if a given function has an LSDA / personality routine
defined.
It's only lightly tested, but seems to be working correctly as long
as your have this information in eh_frame. Does not address getting
this information from compact unwind yet on Darwin systems.
<rdar://problem/18742797>
llvm-svn: 222214
Fixed include:
- Change Platform::ResolveExecutable(...) to take a ModuleSpec instead of a FileSpec + ArchSpec to help resolve executables correctly when we have just a path + UUID (no arch).
- Add the ability to set the listener in SBLaunchInfo and SBAttachInfo in case you don't want to use the debugger as the default listener.
- Modified all places that use the SBLaunchInfo/SBAttachInfo and the internal ProcessLaunchInfo/ProcessAttachInfo to not take a listener as a parameter since it is in the launch/attach info now
- Load a module's sections by default when removing a module from a target. Since we create JIT modules for expressions and helper functions, we could end up with stale data in the section load list if a module was removed from the target as the section load list would still have entries for the unloaded module. Target now has the following functions to help unload all sections a single or multiple modules:
size_t
Target::UnloadModuleSections (const ModuleList &module_list);
size_t
Target::UnloadModuleSections (const lldb::ModuleSP &module_sp);
llvm-svn: 222167
relative paths, like:
/whatever/llvm/lib/Sema/../../include/llvm/Sema/
That causes problems with our type uniquing, since we use the declaration file
and line as one component of the uniquing, and different ways of getting to the
same file will have different directory spellings, though they are functionally
equivalent. We end up with two copies of the exact same type because of this,
and that makes the expression parser give "duplicate type" errors.
I added a method to resolve paths with ../ in them and used that in the FileSpec::Equals,
for comparing Declarations and for doing Breakpoint compares as well, since they also
suffer from this if you specify breakpoints by full path (since nobody knows what
../'s to insert...)
<rdar://problem/18765814>
llvm-svn: 222075
Summary:
PowerPC handles the stack chain with the current stack pointer being a pointer
to the backchain (CFA). LLDB currently has no way of handling this, so this
adds a "CFA is dereferenced from a register" type.
Discussed with Jason Molenda, who also provided the initial patch for this.
Reviewers: jasonmolenda
Reviewed By: jasonmolenda
Subscribers: emaste, lldb-commits
Differential Revision: http://reviews.llvm.org/D6182
llvm-svn: 221788
let's let lldb try the arch default unwind every time but not destructively --
it doesn't permanently replace the main unwind method for that function from
now on.
This fix is for <rdar://problem/18683658>.
I tested it against Ryan Brown's go program test case and also a
collection of core files of tricky unwind scenarios
<rdar://problem/15664282> <rdar://problem/15835846>
<rdar://problem/15982682> <rdar://problem/16099440>
<rdar://problem/17364005> <rdar://problem/18556719>
that I've fixed over the last 6-9 months.
llvm-svn: 221238
output style can be customized. Change the built-in default to be
more similar to gdb's disassembly formatting.
The disassembly-format for a gdb-like output is
${addr-file-or-load} <${function.name-without-args}${function.concrete-only-addr-offset-no-padding}>:
The disassembly-format for the lldb style output is
{${function.initial-function}{${module.file.basename}`}{${function.name-without-args}}:\n}{${function.changed}\n{${module.file.basename}`}{${function.name-without-args}}:\n}{${current-pc-arrow} }{${addr-file-or-load}}:
The two backticks in the lldb style formatter triggers the sub-expression evaluation in
CommandInterpreter::PreprocessCommand() so you can't use that one as-is ... changing to
use ' characters instead of ` would work around that.
<rdar://problem/9885398>
llvm-svn: 219544
works, as do breakpoints, run and pause, display zeroth frame.
See
http://reviews.llvm.org/D5503
for a fuller description of the changes in this commit.
llvm-svn: 218596
For the Objective-C case, we do not have a "function type" notion, so we actually end up wrapping the clang ObjCMethodDecl in the Impl object, and ask function-y questions of it
In general, you can always ask for return type, number of arguments, and type of each argument using the TypeMemberFunction layer - but in the C++ case, you can also acquire a Type object for the function itself, which instead you can't do in the Objective-C case
llvm-svn: 218132
In practice, 64bit eh_frame is not used even for x86_64 binaries. The main reason is in eh_frame we almost always use pc-relative addressing, so addresses are within 32bits and gcc just sticks to 32bit eh_frame.
I generated 64bit eh_frame for Android Java runtime and unwind successfully in gdb, and in lldb with this patch.
Patch by Tong Shen.
llvm-svn: 216409
We decided to use assmbly profiler instead of eh_frame for frame 0 because for compiler generated code, eh_frame is usually synchronous(a.k.a. only valid at call site); and we have no way to tell if it's asynchronous or not.
But for x86 & x86_64 compiler generated code:
1. clang & GCC describes all prologue instructions in eh_frame;
2. mid-function stack pointer altering instructions can be easily detected.
So we can grab eh_frame, and use assembly profiler to augment it into asynchronous unwind table.
This change also benefits hand-written assembly; eh_frame for hand-written assembly is often asynchronous,so we have a much better chance to successfully unwind through them.
Change by Tong Shen.
llvm-svn: 216406
with binaries in the dyld shared cache (esp on iOS) where the file
address for the executable binary (maybe from memory, maybe from
an expanded copy of the dyld shared cache) is different from the
file address in the dSYM. In that case, ObjectFileMachO replaces
the file addresses from the original binary with the dSYM file
addresses (usually 0-based) -- lldb doesn't have a notion of two
file addresses for a given module so they need to agree.
There was a cache of file addresses over in the Symtab so I added
a method to the Module and the objects within to clear any file address
caches if they exist, and added an implementation in the Symtab
module to do that.
<rdar://problem/16929569>
llvm-svn: 216258
What it does:
- it introduces a concept of EncodingToType to the ObjCLanguageRuntime
The ObjC runtime has a "type encoding" feature that describes types as strings
The EncodingToType is a decoder for that format, making types out of type encoding strings
This feature already existed in some shape as we were using it to create method signatures out of the runtime, but this checkin extends the parser to support the full syntax, and moves things so that more parts of LLDB have access to this decoder
- it splits the ClassDescriptorV2 object to its own file, it was starting to grow too large
- it adds to the ClassDescriptor mechanism a notion of ivar storage; the ObjC runtime vends ivar information as well as method information
While ivar information is not ready for prime type (i.e. we don't want to add it to the runtime generated types for expression evaluator usage), there are potentially useful scenarios in which realizing ivar types could be useful. For now, the ClassDescriptor is going to hold ivar information directly. Existing code already allows describing ivars, this patch hooks those moving parts up so that one can actually ask a ClassDescriptor about ivars for the class it represents
and as a couple minor niceties:
- it makes it possible to retrieve the LLDB ClangASTContext that is associated to a clang::ASTContext
- it extends the ValueObject-to-ClassDescriptor API in the language runtime to deal correctly with base-class hierarchies
llvm-svn: 216026
we overrode debug information as the authoritative
source for type information, substituting types
from the Objective-C runtime. The runtime should
never be the primary source.
<rdar://problem/16065049>
llvm-svn: 214583
The fix adds a std::weak_ptr<Module> into the TypeImpl and fills in the weak pointer when possible. It also checks to make sure the module is still alive prior to using it which should make our API safer to use.
<rdar://problem/15455145>
llvm-svn: 212853
off_t is a type which is used for file offsets. Even more
specifically, it is only used by a limited number of C APIs that
deal with files. Any usage of off_t where the variable is not
intended to be used with one of these APIs is a bug, by definition.
This patch corrects some easy mis-uses of off_t, generally by
converting them to lldb::offset_t, but sometimes by using other
types such as size_t, when appropriate.
The use of off_t to represent these offsets has worked fine in
practice on linux-y platforms, since we used _FILE_OFFSET_64 to
guarantee that off_t was a uint64. On Windows, however,
_FILE_OFFSET_64 is unrecognized, and off_t will always be 32-bit.
So the usage of off_t on Windows actually leads to legitimate bugs.
Reviewed by: Greg Clayton
Differential Revision: http://reviews.llvm.org/D4358
llvm-svn: 212192
to modify the same UnwindTable object simultaneously. Fix
HistoryThread and HistoryUnwind's mutex lock acqusition to
retain the lock for the duration of the operation instead of
releasing the temporary immediately.
<rdar://problem/17055023>
llvm-svn: 211241
The compiler, when JIT'ing code, can emit illegal DWARF line tables (address is line table sequences must increase). This changes fixes that issue by replacing previous line entries whose start address is the same with the new line entry to avoid having multiple line entries with the same address. Since the address range of lines entries is determined by the delta between the current and next line entry, this shouldn't cause any issues.
llvm-svn: 211212
- Fix Xcode project to have source files for SBTypeEnumMember.h/SBTypeEnumMember.cpp in the right place
- Rename a member variable to inluce "_sp" suffix since it is a shared pointer
- Cleanup initialization code for TypeEnumMemberImpl to not warn about out of order initialization
llvm-svn: 210051
of the symbol itself rather than forcing clients to do
it. This simplifies the logic for the expression
parser a great deal.
<rdar://problem/16935324>
llvm-svn: 209494
and sharing it with all of its FuncUnwinders, have each FuncUnwinder
create an AssemblyProfiler on demand as needed. I was worried that
the cost of creating the llvm disassemblers would be high for this
approach but it's not supposed to be an expensive operation, and it
means we don't need to add locks around this section of code.
<rdar://problem/16992332>
llvm-svn: 209493
to its unwind assembly profiler to all of the FuncUnwinders (one
per symbol) under it. If lldb is running multiple targets, you
could get two different FuncUnwinders in the same Module trying
to use the same llvm disassembler simultaneously and that may be
a re-entrancy problem.
Instead, the UnwindTable has the unwind assembly profiler and when
the FuncUnwinders want to use it, they get exclusive access to
the assembly profiler until they're done using it.
<rdar://problem/16992332>
llvm-svn: 209488
Rationale:
Pretty simply, the idea is that sometimes type names are way too long and contain way too many details for the average developer to care about. For instance, a plain ol' vector of int might be shown as
std::__1::vector<int, std::__1::allocator<....
rather than the much simpler std::vector<int> form, which is what most developers would actually type in their code
Proposed solution:
Introduce a notion of "display name" and a corresponding API GetDisplayTypeName() to return such a crafted for visual representation type name
Obviously, the display name and the fully qualified (or "true") name are not necessarily the same - that's the whole point
LLDB could choose to pick the "display name" as its one true notion of a type name, and if somebody really needs the fully qualified version of it, let them deal with the problem
Or, LLDB could rename what it currently calls the "type name" to be the "display name", and add new APIs for the fully qualified name, making the display name the default choice
The choice that I am making here is that the type name will keep meaning the same, and people who want a type name suited for display will explicitly ask for one
It is the less risky/disruptive choice - and it should eventually make it fairly obvious when someone is asking for the wrong type
Caveats:
- for now, GetDisplayTypeName() == GetTypeName(), there is no logic to produce customized display type names yet.
- while the fully-qualified type name is still the main key to the kingdom of data formatters, if we start showing custom names to people, those should match formatters
llvm-svn: 209072
Enable data formatters to see-through pointers/references to typedefs
For instance, if Foo is a typedef to Bar, and there is a formatter for any/all of Bar*, Bar&, Bar&&, then Foo*, Foo&, and Foo&& should pick these up if Foo-specific formatters don't exist
llvm-svn: 205939
This is a purely mechanical change explicitly casting any parameters for printf
style conversion. This cleans up the warnings emitted by gcc 4.8 on Linux.
llvm-svn: 205607
This is a mechanical change addressing the various sign comparison warnings that
are identified by both clang and gcc. This helps cleanup some of the warning
spew that occurs during builds.
llvm-svn: 205390
These changes were written by Greg Clayton, Jim Ingham, Jason Molenda.
It builds cleanly against TOT llvm with xcodebuild. I updated the
cmake files by visual inspection but did not try a build. I haven't
built these sources on any non-Mac platforms - I don't think this
patch adds any code that requires darwin, but please let me know if
I missed something.
In debugserver, MachProcess.cpp and MachTask.cpp were renamed to
MachProcess.mm and MachTask.mm as they picked up some new Objective-C
code needed to launch processes when running on iOS.
llvm-svn: 205113
(lldb) b puts
(lldb) expr -g -i0 -- (int)puts("hello")
First we will stop at the entry point of the expression before it runs, then we can step over a few times and hit the breakpoint in "puts", then we can continue and finishing stepping and fininsh the expression.
Main features:
- New ObjectFileJIT class that can be easily created for JIT functions
- debug info can now be enabled when parsing expressions
- source for any function that is run throught the JIT is now saved in LLDB process specific temp directory and cleaned up on exit
- "expr -g --" allows you to single step through your expression function with source code
<rdar://problem/16382881>
llvm-svn: 204682
This is a mechanical cleanup of unused functions. In the case where the
functions are referenced (in comment form), I've simply commented out the
functions. A second pass to clean that up is warranted.
The functions which are otherwise unused have been removed. Some of these were
introduced in the initial commit and not in use prior to that point!
NFC
llvm-svn: 204310
Fix a bug where calling SBFrame::FindValue() would cause a copy of all variables in the block to be inserted in the frame's variable list, regardless of whether those same variables were there or not - which means one could end up with a frame with lots of duplicate copies of the same variables
llvm-svn: 201614
The many many benefits include:
1 - Input/Output/Error streams are now handled as real streams not a push style input
2 - auto completion in python embedded interpreter
3 - multi-line input for "script" and "expression" commands now allow you to edit previous/next lines using up and down arrow keys and this makes multi-line input actually a viable thing to use
4 - it is now possible to use curses to drive LLDB (please try the "gui" command)
We will need to deal with and fix any buildbot failures and tests and arise now that input/output and error are correctly hooked up in all cases.
llvm-svn: 200263
SBType SBType::GetTypedefedType();
Also added the ability to get a type by type ID from a SBModule:
SBType SBModule::GetTypeByID (lldb::user_id_t uid);
llvm-svn: 199939
210: test_with_dsym_and_run_command (Test-rdar-10642615.Radar10642615DataFormatterTestCase)
Test data formatter commands. ... Assertion failed: (isa<X>(Val) && "cast<Ty>() argument of incompatible type!"), function cast, file llvm/include/llvm/Support/Casting.h, line 239.
226: test_with_dsym_and_run_command (Test-rdar-13338477.Radar13338477DataFormatterTestCase)
Test that LLDB handles the clang typeclass Paren correctly. ... Assertion failed: (isa<X>(Val) && "cast<Ty>() argument of incompatible type!"), function cast, file llvm/include/llvm/Support/Casting.h, line 239.
llvm-svn: 198811
it is causing an llvm assert when run against
test/functionalities/data-formatter/rdar-10642615,
Assertion failed: (isa<X>(Val) && "cast<Ty>() argument of incompatible type!"), function cast, file llvm/include/llvm/Support/Casting.h, line 239.
llvm-svn: 198809
When determining the type of array members, do not see-through typedefs
For instance, in BOOL arr[4], we want the elements to be typed as BOOL, not signed char
llvm-svn: 198729
indications that the UnwindPlan is invalid -- for instance, a
complete lack of rows, or a row that fails to define a register to
base the CFA off of.
<rdar://problem/15246247>
llvm-svn: 196201
pure virtual base class and made StackFrame a subclass of that. As
I started to build on top of that arrangement today, I found that it
wasn't working out like I intended. Instead I'll try sticking with
the single StackFrame class -- there's too much code duplication to
make a more complicated class hierarchy sensible I think.
llvm-svn: 193983
defines a protocol that all subclasses will implement. StackFrame
is currently the only subclass and the methods that Frame vends are
nearly identical to StackFrame's old methods.
Update all callers to use Frame*/Frame& instead of pointers to
StackFrames.
This is almost entirely a mechanical change that touches a lot of
the code base so I'm committing it alone. No new functionality is
added with this patch, no new subclasses of Frame exist yet.
I'll probably need to tweak some of the separation, possibly moving
some of StackFrame's methods up in to Frame, but this is a good
starting point.
<rdar://problem/15314068>
llvm-svn: 193907
In almost all cases, the misuse is about "%lu" being used instead of the correct "%zu" (even though these are compatible on 64-bit platforms in practice). There are even a couple of cases where "%ld" (ie., signed int) is used instead of "%zu", and one where "%lu" is used instead of "%" PRIu64.
Fixes bug #17551.
Patch by "/dev/humancontroller"
llvm-svn: 193832
This commit reimplements the TypeImpl class (the class that backs SBType) in terms of a static,dynamic type pair
This is useful for those cases when the dynamic type of an ObjC variable can only be obtained in terms of an "hollow" type with no ivars
In that case, we could either go with the static type (+iVar information) or with the dynamic type (+inheritance chain)
With the new TypeImpl implementation, we try to combine these two sources of information in order to extract as much information as possible
This should improve the functionality of tools that are using the SBType API to do extensive dynamic type inspection
llvm-svn: 193564
Fixed an issue with reexported symbols on MacOSX by adding support for symbols re-exporting symbols. There is now a new symbol type eSymbolTypeReExported which contains a new name for the re-exported symbol and the new shared library. These symbols are only used when a symbol is re-exported as a symbol under a different name.
Modified the expression parser to be able to deal with finding the re-exported symbols and track down the actual symbol it refers to.
llvm-svn: 193101
ObjectFile::CopyData is used to copy a block of target memory to the
caller's buffer (e.g. for "memory read"). This should be a straight
memcpy, and not byte-swapped if the target and host have different
endianness.
Add a new DataExtractor::CopyData() method that performs this straight
copy and use it in ObjectFile::CopyData().
llvm-svn: 192323
to be explicit, to prevent horrid things like
std::string a = ConstString("foo")
from taking the path ConstString -> bool -> char
-> std::string.
This fixes, among other things, ClangFunction.
<rdar://problem/15137989>
llvm-svn: 191934
not have breakpoints set on it inserted into code that does have a valid line number. So allow
that line number, and the ThreadPlanStepRange should just continue stepping over 0 line ranges
as if they had the same line number as whatever we were previously stepping through.
llvm-svn: 191477
default-at-first-instruction UnwindPlan if we're at the beginning of a function and
the ABI can provide us with an UnwindPlan to get out of there before falling back
to the generic architectural default UnwindPlan (which usually assumes that the stack
has already been set up.)
Update the FuncUnwinders methods to gracefully handle the case where an assembly
profiler may not be available.
Fix a bug where FuncUnwinders::GetUnwindPlanArchitectureDefaultAtFunctionEntry was
returning the wrong UnwindPlan to its caller.
llvm-svn: 191262
a bunch of semicolons where the IndirectFieldDecls
were. These IndirectFieldDecls should have been
implicit.
<rdar://problem/14628784>
llvm-svn: 188247
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
- MachO files now correctly extract the UUID all the time
- More file size and offset verification done for universal mach-o files to watch for truncated files
- ObjectContainerBSDArchive now supports enumerating all objects in BSD archives (.a files)
- lldb_private::Module() can not be properly constructed using a ModuleSpec for a .o file in a .a file
- The BSD archive plug-in shares its cache for GetModuleSpecifications() and the create callback
- Improved printing for ModuleSpec objects
llvm-svn: 186211
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
- ObjectFile::GetSymtab() and ObjectFile::ClearSymtab() no longer takes any flags
- Module coordinates with the object files and contain a unified section list so that object file and symbol file can share sections when they need to, yet contain their own sections.
Other cleanups:
- Fixed Symbol::GetByteSize() to not have the symbol table compute the byte sizes on the fly
- Modified the ObjectFileMachO class to compute symbol sizes all at once efficiently
- Modified the Symtab class to store a file address lookup table for more efficient lookups
- Removed Section::Finalize() and SectionList::Finalize() as they did nothing
- Improved performance of the detection of symbol files that have debug maps by excluding stripped files and core files, debug files, object files and stubs
- Added the ability to tell if an ObjectFile has been stripped with ObjectFile::IsStripped() (used this for the above performance improvement)
llvm-svn: 185990
The script was able to point out and save 40 bytes in each lldb_private::Section by being very careful where we need to have virtual destructors and also by re-ordering members.
llvm-svn: 184364
//------------------------------------------------------------------
/// Get all types matching \a type_mask from debug info in this
/// module.
///
/// @param[in] type_mask
/// A bitfield that consists of one or more bits logically OR'ed
/// together from the lldb::TypeClass enumeration. This allows
/// you to request only structure types, or only class, struct
/// and union types. Passing in lldb::eTypeClassAny will return
/// all types found in the debug information for this module.
///
/// @return
/// A list of types in this module that match \a type_mask
//------------------------------------------------------------------
lldb::SBTypeList
SBModule::GetTypes (uint32_t type_mask)
//------------------------------------------------------------------
/// Get all types matching \a type_mask from debug info in this
/// compile unit.
///
/// @param[in] type_mask
/// A bitfield that consists of one or more bits logically OR'ed
/// together from the lldb::TypeClass enumeration. This allows
/// you to request only structure types, or only class, struct
/// and union types. Passing in lldb::eTypeClassAny will return
/// all types found in the debug information for this compile
/// unit.
///
/// @return
/// A list of types in this compile unit that match \a type_mask
//------------------------------------------------------------------
lldb::SBTypeList
SBCompileUnit::GetTypes (uint32_t type_mask = lldb::eTypeClassAny);
This lets you request types by filling out a mask that contains one or more bits from the lldb::TypeClass enumerations, so you can only get the types you really want.
llvm-svn: 184251
Modifying our data formatters matching algorithm to ensure that "const X*" is treated as equivalent to "X*"
Also, a couple improvements to the "lldb types" logging
llvm-svn: 184215
325,000 breakpoints for running "breakpoint set --func-regex ." on lldb itself (after hitting a breakpoint at main so that LLDB.framework is loaded) used to take up to an hour to set, now we are down under a minute. With warm file caches, we are at 40 seconds, and that is with setting 325,000 breakpoint through the GDB remote API. Linux and the native debuggers might be faster. I haven't timed what how much is debug info parsing and how much is the protocol traffic to/from GDB remote.
That there were many performance issues. Most of them were due to storing breakpoints in the wrong data structures, or using the wrong iterators to traverse the lists, traversing the lists in inefficient ways, and not optimizing certain function name lookups/symbol merges correctly.
Debugging after that is also now very efficient. There were issues with replacing the breakpoint opcodes in memory that was read, and those routines were also fixed.
llvm-svn: 183820
Modified the test programs to use floating point constants that always will display correctly. We had some numbers that were being rounded, and now that we are using clang, we no longer round them and we get more correct results.
llvm-svn: 183792
Name matching was working inconsistently across many places in LLDB. Anyone doing name lookups where you want to look for all types of names should used "eFunctionNameTypeAuto" as the sole name type mask. This will ensure that we get consistent "lookup function by name" results. We had many function calls using as mask like "eFunctionNameTypeBase | eFunctionNameTypeFull | eFunctionNameTypeMethod | eFunctionNameTypeSelector". This was due to the function lookup by name evolving over time, but as it stands today, use eFunctionNameTypeAuto when you want general name lookups. Either ModuleList::FindFunctions() or Module::FindFunctions() will figure out the right kinds of names to lookup and remove the "eFunctionNameTypeAuto" and replace it with the exact subset of what the name can be.
This checkin also changes eFunctionNameTypeAny over to use eFunctionNameTypeAuto to reflect this.
llvm-svn: 182179
"source list -n <func>" can now show more than one location that matches a function name. It will unique multiple of the same source locations so they don't get displayed. It also handles inline functions correctly.
llvm-svn: 182067
<rdar://problem/13594769>
Main changes in this patch include:
- cleanup plug-in interface and use ConstStrings for plug-in names
- Modfiied the BSD Archive plug-in to be able to pick out the correct .o file when .a files contain multiple .o files with the same name by using the timestamp
- Modified SymbolFileDWARFDebugMap to properly verify the timestamp on .o files it loads to ensure we don't load updated .o files and cause problems when debugging
The plug-in interface changes:
Modified the lldb_private::PluginInterface class that all plug-ins inherit from:
Changed:
virtual const char * GetPluginName() = 0;
To:
virtual ConstString GetPluginName() = 0;
Removed:
virtual const char * GetShortPluginName() = 0;
- Fixed up all plug-in to adhere to the new interface and to return lldb_private::ConstString values for the plug-in names.
- Fixed all plug-ins to return simple names with no prefixes. Some plug-ins had prefixes and most ones didn't, so now they all don't have prefixed names, just simple names like "linux", "gdb-remote", etc.
llvm-svn: 181631
This commit changes the ${function.name-with-args} prompt keyword to also tackle structs
Previously, since aggregates have no values, this would show up as foo=(null)
This checkin changes that to instead print foo=(Foo at 0x123) (i.e. typename at address)
There are other potential choices here (summary, one-liner printout of all members, ...) and I would love to hear feedback about better options, if any
llvm-svn: 181462
clang sugarcoats expressions of the sort *(int (*)[3])foo where foo is an int* saying that their type class is Paren
This checkin updates our lookup tables to properly desugar Paren into the actual type of interest
llvm-svn: 180938
in debug information more aggressive. Emitting
classes containing these methods causes crashes in
Clang when dealing with complex code bases.
<rdar://problem/12640887>
llvm-svn: 180895
std::string
Module::GetSpecificationDescription () const;
This returns the module as "/usr/lib/libfoo.dylib" for normal files (calls "std::string FileSpec::GetPath()" on m_file) but it also might include the object name in case the module is for a .o file in a BSD archive ("/usr/lib/libfoo.a(bar.o)"). Cleaned up necessary logging code to use it.
llvm-svn: 180717
There is a new static ObjectFile function you can call:
size_t
ObjectFile::GetModuleSpecifications (const FileSpec &file,
lldb::offset_t file_offset,
ModuleSpecList &specs)
This will fill in "specs" which the details of all the module specs (file + arch + UUID (if there is one) + object name (for BSD archive objects eventually) + file offset to the object in question).
This helps us when a user specifies a file that contains a single architecture, and also helps us when we are given a debug symbol file (like a dSYM file on MacOSX) that contains one or more architectures and we need to be able to match it up to an existing Module that has no debug info.
llvm-svn: 180224
Fixed LLDB to be able to correctly parse template parameters that have no name and no type. This can be triggered by the following LLVM/Clang code:
template <typename T, typename = void>
class SmallVectorTemplateCommon : public SmallVectorBase {
The “typename = void” was emitting DWARF with an empty DW_AT_name and no DW_AT_type. We now correctly infer that no DW_AT_type means “void” and that an empty name is ok.
This means you can now call functions on things that inherit from SmallVectorTemplateCommon.
llvm-svn: 180155
- Specify SC_Static given DIE attributes for static methods and operators.
Thanks to Wei Pan for his review and the help with root-causing.
llvm-svn: 179727
- Do not add symbols with no names
- Make sure that symbols from ELF symbol tables know that the byte size is correct. Previously the symbols would calculate their sizes by looking for the next symbol and take symbols that had zero size and make them have invalid sizes.
- Added the ability to dump raw ELF symbols by adding a Dump method to ELFSymbol
Also removed some unused code from lldb_private::Symtab.
llvm-svn: 179466
Now we can:
1 - see the return value for functions that return types that use the "ext_vector_size"
2 - dump values that use the vector attributes ("expr $ymm0")
3 - modified the DWARF parser to correctly parse GNU vector types from the DWARF by turning them into clang::Type::ExtVector types instead of just standard arrays
llvm-svn: 178924
Make lldb_private::RegularExpression thread safe everywhere. This was done by removing the m_matches array from the lldb_private::RegularExpression class and putting it into the new lldb_private::RegularExpression::Match class. When executing a regular expression you now have the option to create a lldb_private::RegularExpression::Match object and pass a pointer in if you want to get parenthesized matching. If you don't want any matching, you pass in NULL. The lldb_private::RegularExpression::Match object is initialized with the number of matches you desire. Any matching strings are now extracted from the lldb_private::RegularExpression::Match objects. This makes the regular expression objects thread safe and as a result many more regex objects were turned into static objects that end up using a local lldb_private::RegularExpression::Match object when executing.
llvm-svn: 178702
Symbol table function names should support lookups like symbols with debug info.
To fix this I:
- Gutted the way FindFunctions is used, there used to be way too much smarts only in the DWARF plug-in
- Made it more efficient by chopping the name up once and using simpler queries so that SymbolFile and Symtab plug-ins don't need to do as much
- Filter the results at a higher level
- Make the lldb_private::Symtab able to chop up C++ mangled names and make as much sense out of them as possible and also be able to search by basename, fullname, method name, and selector name.
llvm-svn: 178608
they are probably trivial. This means that we
don't confuse Clang about whether a class is
trivially copy constructible. It can figure
that out itself as long as we don't explicitly
feed it the constructors.
If the class is trivially copy-constructible,
this can change the ABI that Clang uses to call
functions that return that class (e.g., by making
the object be returned in a register), so this
is quite important for correctness.
<rdar://problem/13457741>
llvm-svn: 178411
ASTContexts that will not stay around. Before, we
did this in a very half-hearted way. Now we maintain
work queues of all Decls that need to be completed
before the source ASTContext can go away; we then
expunge their origins completely.
<rdar://problem/13511875>
llvm-svn: 178410
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
DWARFCallFrameInfo method which returns a RangeVector pre-size the
vector based on the number of entries it will be adding insted of
growing the vector as items are added.
llvm-svn: 177773
Clang requires them to have complete types, but
we were previously only completing them if they
were of tag or Objective-C object types.
I have implemented a method on the ASTImporter
whose job is to complete a type. It handles not
only the cases mentioned above, but also array
and atomic types.
<rdar://problem/13446777>
llvm-svn: 177672
This returns a vector of <file address, size> entries for all of
the functions in the module that have an eh_frame FDE.
Update ObjectFileMachO to use the eh_frame FDE function addresses if
the LC_FUNCTION_STARTS section is missing, to fill in the start
addresses of any symbols that have been stripped from the binary.
Generally speaking, lldb works best if it knows the actual start
address of every function in a module - it's especially important
for unwinding, where lldb inspects the instructions in the prologue
of the function. In a stripped binary, it is deprived of this
information and it reduces the quality of our unwinds and saved
register retrieval.
Other ObjectFile users may want to use the function addresses from
DWARFCallFrameInfo to fill in any stripped symbols like ObjectFileMachO
does already.
<rdar://problem/13365659>
llvm-svn: 177624
track the EH FDEs for the functions in a module to using a
RangeDataVector, a more light-weight data structure that only refers
to File addresses. Makes the initial FDE scan about 3x faster, uses
less memory.
<rdar://problem/13465650>
llvm-svn: 177585
Fixed a crasher in the new DWARF in .o files line table linking function where "back()" could end up being called on an empty std::vector.
llvm-svn: 177082
uninitialized memory, to getTrivialTypeSourceInfo,
which initializes its memory, when creating trivial
TypeSourceInfos.
<rdar://problem/13332253>
llvm-svn: 176899
counters for a variety of metrics associated
with expression parsing. This should give some
idea of how much work the expression parser is
doing on Clang's behalf, and help with hopefully
reducing that load over time.
<rdar://problem/13210748> Audit type search/import for expressions
llvm-svn: 176714
Make dynamic type detection faster by using the AST metadata to help out and allow us not to complete types when we don't need to.
After running "purge" on a MacOSX system, the Xcode variables view now populates more than 3x faster with this fix.
llvm-svn: 176676
Also added C++11 enum test cases to cover enums as int8_t, int16_t int32_t, int64_t, uint8_t, uint16_t, uint32_t, and uint64_t both for DWARF and dSYM cases. The DWARF being emitted by clang is missing the enum integer type, but the code is now ready to accept and deal with the integral type if it is supplied.
llvm-svn: 176548
LLDB wasn't printing the names for negative enums. Fixed the signed extraction of enumerators and how they were registered with clang's type system.
llvm-svn: 176533
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
- made sure we tell Clang not to try to
complete the type since it can't be
completed from its origin any more; and
- fixed a silly bug where we tried to
forget about the original decl's origins
rather than the deported decl's origin.
These produced some crashes in ptr_refs,
especially under libgmalloc.
<rdar://problem/13256150>
llvm-svn: 176233
StackFrame assumes m_sc is additive, but m_sc can lose its target. So now the SymbolContext::Clear() method takes a bool that indicates if the target should be cleared. Modified all existing code to properly set the bool argument.
llvm-svn: 175953
- generate-vers.pl has to be called by cmake to generate the version number
- parallel builds not yet supported; dependency on clang must be explicitly specified
Tested on Linux.
- Building on Mac will require code-signing logic to be implemented.
- Building on Windows will require OS-detection logic and some selective directory inclusion
Thanks to Carlo Kok (who originally prepared these CMakefiles for Windows) and Ben Langmuir
who ported them to Linux!
llvm-svn: 175795
changing the ClangASTSource to return a bool instead
of returning a list of results. Our testsuite mostly
works with this change, but some minor issues may
remain both on LLDB's side and on Clang's side.
llvm-svn: 174949
1 - A store off the end of a buffer in ValueObject.cpp
2 - DataExtractor had cases where bad offsets could cause invalid memory to be accessed.
llvm-svn: 174757
if it encountered bad debug information. This
debug information had an Objective-C method whose
selector disagreed with the true number of arguments
to that method.
<rdar://problem/12992864>
llvm-svn: 174557
lldb was mmap'ing archive files once per .o file it loads, now it correctly shares the archive between modules.
LLDB was also always mapping entire contents of universal mach-o files, now it maps just the slice that is required.
Added a new logging channel for "lldb" called "mmap" to help track future regressions.
Modified the ObjectFile and ObjectContainer plugin interfaces to take a data offset along with the file offset and size so we can implement the correct caching and efficient reading of parts of files without mmap'ing the entire file like we used to.
The current implementation still keeps entire .a files mmaped (once) and entire slices from universal files mmaped to ensure that if a client builds their binaries during a debug session we don't lose our data and get corrupt object file info and debug info.
llvm-svn: 174524
The first part of the fix for having LLDB handle LTO debugging when the DWARF is in the .o files. This part separates the object file's modules into a separate cache map that maps unique C strings for the N_OSO path to the ModuleSP since one object file might be mentioned more than once in LTO binaries.
llvm-svn: 174476
Fix in loading mach files from memory when using DynamicLoaderMacOSXDYLD.
Removed the uuid mismatch warning that could be spit out and any time during debugging and removed the test case that was looking for that. Currently the "add-dsym" or "target symbols add" command will report an error when the UUID's don't match.
Be more careful when checking and resolving section + offset addresses to make sure none of the base addresses are invalid.
llvm-svn: 174222
support reporting "this" as a templated class. The
expression parser wraps expressions in C++ methods
as methods with the signature
$__lldb_class::$__lldb_expr(...)
and previously responded to clang's queries about
$__lldb_class with the type of *this. This didn't
work if *this was a ClassTemplateSpecializationDecl
because ClassTemplateSpecializationDecls can't be
the result of simple name queries.
Instead what we do now is respond that $__lldb_class
is a typedef and that the target of the typedef is
the (potentially templated) type of *this. That is
much more robust.
Thanks to John McCall for key insights.
<rdar://problem/10987183>
llvm-svn: 174153
Cleaned up the objective C name parsing code to use a class.
Now breakpoints that are set by name that are objective C methods without the leading '+' or '-' will resolve. We do this by expanding all the objective C names for a given string. For example:
(lldb) b [MyString cStringUsingEncoding:]
Will set a breakpoint with multiple possible names:
-[MyString cStringUsingEncoding:]
+[MyString cStringUsingEncoding:]
Also if you have a category, it will strip the category and set a breakpoint in all variants:
(lldb) [MyString(my_category) cStringUsingEncoding:]
Will resolve to the following names:
-[MyString(my_category) cStringUsingEncoding:]
+[MyString(my_category) cStringUsingEncoding:]
-[MyString cStringUsingEncoding:]
+[MyString cStringUsingEncoding:]
Likewise when we have:
(lldb) b -[MyString(my_category) cStringUsingEncoding:]
It will resolve to two names:
-[MyString(my_category) cStringUsingEncoding:]
-[MyString cStringUsingEncoding:]
llvm-svn: 173858
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
Extending ValueObjectDynamicValue so that it stores a TypeAndOrName instead of a TypeSP.
This change allows us to reflect the notion that a ValueObject can have a dynamic type for which we have no debug information.
Previously, we would coalesce that to the static type of the object, potentially losing relevant information or even getting it wrong.
This fix ensures we can correctly report the class name for Cocoa objects whose types are hidden classes that we know nothing about (e.g. __NSArrayI for immutable arrays).
As a side effect, our --show-types argument to frame variable no longer needs to append custom dynamic type information.
llvm-svn: 173216
Enabling support for the wchar_t type.
Without the proper language option setup, clang's ASTContexts will be configured to have wchar_t == int
This patch enables the correct options to make sure that we report wchar_t as itself
Added a test case to make sure we do not regress
Adding files missing from the previous commit
llvm-svn: 172039
Setting breakpoints using "breakpoint set --selector <SEL>" previously didn't when there was no dSYM file.
Also fixed issues in the test suite that arose after fixing the bug.
Also fixed the log channels to properly ref count the log streams using weak pointers to the streams. This fixes a test suite problem that would happen when you specified a full path to the compiler with the "--compiler" option.
llvm-svn: 171816
Added SBTarget::EvaluateExpression() so expressions can be evaluated without needing a process.
Also fixed many functions that deal with clang AST types to be able to properly handle the clang::Type::Elaborated types ("struct foo", "class bar").
llvm-svn: 171476
The results from Clang name lookups changed to
be ArrayRefs, so I had to change the way we
check for the presence of a result and the way
we iterate across results.
llvm-svn: 170927
Fixed zero sized arrays to work correctly. This will only happen once we get a clang that emits correct debug info for zero sized arrays. For now I have marked the TestStructTypes.py as an expected failure.
llvm-svn: 169465
- add new header lldb-python.h to be included before other system headers
- short term fix (eventually python dependencies must be cleaned up)
Patch by Matt Kopec!
llvm-svn: 169341
Cleaned up the option parsing code to always pass around the short options as integers. Previously we cast this down to "char" and lost some information. I recently added an assert that would detect duplicate short character options which was firing during the test suite.
This fix does the following:
- make sure all short options are treated as "int"
- make sure that short options can be non-printable values when a short option is not required or when an option group is mixed into many commands and a short option is not desired
- fix the help printing to "do the right thing" in all cases. Previously if there were duplicate short character options, it would just not emit help for the duplicates
- fix option parsing when there are duplicates to parse options correctly. Previously the option parsing, when done for an OptionGroup, would just start parsing options incorrectly by omitting table entries and it would end up setting the wrong option value
llvm-svn: 169189
Simplify the logging on ObjectFile::~ObjectFile() to not access an classes above the object file (like the module) so we don't crash when logging object lifetimes. The log message contains the "this" pointer value which can be matched up with the constructor log.
llvm-svn: 168754
Allow the expression parser to see more than just data symbols. We now accept any symbol that has an address. We take precautions to only accept symbols by their mangled or demangled names only if the demangled name was not synthesized. If the demangled name is synthesized, then we now mark symbols accordingly and only compare against the mangled original name.
llvm-svn: 168668
Unnamed bitfields cause struct layout problems
Synthesize unnamed bitfields when required. Most compilers don't mention unnamed bitfields in the DWARF, so we need to create them to keep clang happy with the types we create from the DWARF. We currently can't do this for ObjC since the DW_AT_bit_offset value for any direct ivars of ObjC classes as the values for these attributes are bogus. A bug has been filed on Clang to fix this, and another bug has been filed on LLDB to make sure we fix the DWARF parser once the clang fix is in by looking the the DW_AT_producer in the compile unit attributes and finding the compiler version and only enabling it for newer versions of clang.
llvm-svn: 167424
LLDB now provides base class offsets (virtual and non virtual) to Clang's record layout. We previously were told this wasn't necessary, but it is when pragma pack gets involved.
llvm-svn: 167262
so it could hold this information, and then used it to look up unfound names in the object pointer
if it exists. This gets "frame var" to work for unqualified references to ivars captured in blocks.
But the expression parser is ignoring this information still.
llvm-svn: 166860
Full UnwindPlan is trying to do an impossible unwind; in that case
invalidate the Full UnwindPlan and replace it with the architecture
default unwind plan.
This is a scenario that happens occasionally with arm unwinds in
particular; the instruction analysis based full unwindplan can
mis-parse the functions and the stack walk stops prematurely. Now
we can do a simpleminded frame-chain walk to find the caller frame
and continue the unwind. It's not ideal but given the complicated
nature of analyzing the arm functions, and the lack of eh_frame
information on iOS, it is a distinct improvement and fixes some
long-standing problems with the unwinder on that platform.
This is fixing <rdar://problem/12091421>. I may re-use this
invalidate feature in the future if I can identify other cases where
the full unwindplan's unwind information is clearly incorrect.
This checkin also includes some cleanup for the volatile register
definition in the arm ABI plugin for <rdar://problem/10652166>
although work remains to be done for that bug.
llvm-svn: 166757
Allow type searches to specify a type keyword when searching for type. Currently supported type keywords are: struct, class, union, enum, and typedef.
So now you can search for types with a string like "struct foo".
llvm-svn: 166420
top-of-tree. Removed all local patches and llvm.zip.
The intent is that fron now on top-of-tree will
always build against LLVM/Clang top-of-tree, and
that problems building will be resolved as they
occur. Stable release branches of LLDB can be
constructed as needed and linked to specific release
branches of LLVM/Clang.
llvm-svn: 164563
not correctly store the contents of Objective-C
classes. This was due to a combination of
factors:
1) Types were only being completed if we were
looking inside them for specific ivars
(using FindExternalVisibleDeclsByName).
We now look the complete type up at every
FindExternalLexicalDecls.
2) Even if the types were completed properly,
ValueObjectConstResult overrode the type
of every ValueObject using the complete type
for its class from the debug information.
Superclasses of complete classes are not
guaranteed to be complete. Although "frame
variable" uses the debug information,
the expression parser does now piece together
complete types at every level (as described
in Bullet 1), so I provided a way for the
expression parser to prevent overriding.
3) Type sizes were being miscomputed by
ClangASTContext. It ignored the ISA pointer
and only counted fields. We now correctly
count the ISA in the size of an object.
<rdar://problem/12315386>
llvm-svn: 164333
it is unconditionally present now.
ObjectContainerBSDArchive::CreateInstance %z8.8x is not a valid printf arg specifier, %8.8zx would work
for size_t arg but this arg is addr_t. use %8.8llx and cast up to uint64_t.
ObjectFile::FindPlugin ditto.
DynamicRegisterInfo::SetRegisterInfo ifdef this function out if LLDB_DISABLE_PYTHON.
llvm-svn: 163599
Added a fix for incorrect dynamic typing. Before when asking if a C++ class could be dynamic, we would answer yes for incomplete C++ classes. This turned out to have issues where if a class was not virtual, yet had its first ivar be an instance of a virtual class, we would incorrectly say that a class was virtual and we would downcast it to be a pointer to the first ivar. We now ask the class to complete itself prior to answering the question. We need to test the effects on memory of this change prior to submission. It is the safest and best fix, but it does have a potential downside of higher memory consumption.
llvm-svn: 163014
Make breakpoint setting by file and line much more efficient by only looking for inlined breakpoint locations if we are setting a breakpoint in anything but a source implementation file. Implementing this complex for a many reasons. Turns out that parsing compile units lazily had some issues with respect to how we need to do things with DWARF in .o files. So the fixes in the checkin for this makes these changes:
- Add a new setting called "target.inline-breakpoint-strategy" which can be set to "never", "always", or "headers". "never" will never try and set any inlined breakpoints (fastest). "always" always looks for inlined breakpoint locations (slowest, but most accurate). "headers", which is the default setting, will only look for inlined breakpoint locations if the breakpoint is set in what are consudered to be header files, which is realy defined as "not in an implementation source file".
- modify the breakpoint setting by file and line to check the current "target.inline-breakpoint-strategy" setting and act accordingly
- Modify compile units to be able to get their language and other info lazily. This allows us to create compile units from the debug map and not have to fill all of the details in, and then lazily discover this information as we go on debuggging. This is needed to avoid parsing all .o files when setting breakpoints in implementation only files (no inlines). Otherwise we would need to parse the .o file, the object file (mach-o in our case) and the symbol file (DWARF in the object file) just to see what the compile unit was.
- modify the "SymbolFileDWARFDebugMap" to subclass lldb_private::Module so that the virtual "GetObjectFile()" and "GetSymbolVendor()" functions can be intercepted when the .o file contenst are later lazilly needed. Prior to this fix, when we first instantiated the "SymbolFileDWARFDebugMap" class, we would also make modules, object files and symbol files for every .o file in the debug map because we needed to fix up the sections in the .o files with information that is in the executable debug map. Now we lazily do this in the DebugMapModule::GetObjectFile()
Cleaned up header includes a bit as well.
llvm-svn: 162860
when you want to find the caller's saved pc, you look up the return address
register and use that. On arm, for instance, this would be the contents of
the link register (lr).
If the eh_frame CIE defines an RA, record that fact in the UnwindPlan.
When we're finding a saved register, if it's the pc, lok for the location
of the return address register instead.
<rdar://problem/12062310>
llvm-svn: 162167
Added new API to lldb::SBTypeMember for bitfields:
bool SBTypeMember::IsBitfield();
uint32_t SBTypeMember::GetBitfieldSizeInBits();
Also added new properties for easy access. Now SBTypeMember objects in python have a "fields" property for all type fields, "bases" for all direct bases, "vbases" for all virtual base classes and "members" for a combo of all three organized by bit offset. They all return a python list() of SBTypeMember objects. Usage:
(lldb) script
>>> t = lldb.target.FindFirstType("my_type")
>>> for field in t.fields:
... print field
>>> for vbase in t.vbases:
... print vbase
>>> for base in t.bases:
... print base
>>> for member in t.members:
... print member
Also added new "is_bitfield" property to the SBTypeMember objects that will return the result of SBTypeMember::IsBitfield(), and "bitfield_bit_size" which will return the result of SBTypeMember::GetBitfieldSizeInBits();
I also fixed "SBTypeMember::GetOffsetInBytes()" to return the correct byte offset.
llvm-svn: 161091
instructions, be sure to allocate new UnwindPlan::Row's each
time we push a row on to the UnwindPlan so we don't mutate
it any further.
(fallout from changing the UnwindPlan from having a vector
of Row's to having a vector of RowSP shared pointers.)
<rdar://problem/11997385>
llvm-svn: 161089
the state of the unwind instructions once the prologue has finished. If it hits an
early return epilogue in the middle of the function, re-instate the prologue after that
epilogue has completed so that we can still unwind for cases where the flow of control
goes past that early-return. <rdar://problem/11775059>
Move the UnwindPlan operator== definition into the .cpp file, expand the definition a bit.
Add some casts to a SBCommandInterpreter::HandleCompletion() log statement so it builds without
warning on 64- and 32-bit systems.
llvm-svn: 160337
a shared pointer to ease some memory management issues with a patch
I'm working on.
The main complication with using SPs for these objects is that most
methods that build up an UnwindPlan will construct a Row to a given
instruction point in a function, then add additional regsaves in
the next instruction point to that row and push it again. A little
care is needed to not mutate the previous instruction point's Row
once these are switched to being held behing shared pointers.
llvm-svn: 160214
UnwindPlans for a function. This specifically does not use any
previously-generated UnwindPlans so if any logging is performed
while creating the UnwindPlans, it will be repeated. This is
useful for when an lldb stack trace is not correct and you want
to gather diagnostic information from the user -- they can do
log enable -v lldb unwind, image show-unwind of the function, and
you'll get the full logging as the UnwindPlans are recreated.
llvm-svn: 160095
various other syntactic sugar work. Lambdas do
not due to some problems relocating code containing
lambdas. Rvalue references work when returned from
expressions, but need more testing.
llvm-svn: 156948
Add "--name" option to "image lookup" that will search both functions and symbols.
Also made all of the output from any of the "image lookup" commands be the same regardless of the lookup type (function name, symbol name, func or symbol, file and line, address, etc). The --verbose or -v option also will expand the results as needed and display things so they look the same.
llvm-svn: 156835
Fixed the DisassemblerLLVMC disassembler to parse more efficiently instead of parsing opcodes over and over. The InstructionLLVMC class now only reads the opcode in the InstructionLLVMC::Decode function. This can be done very efficiently for ARM and architectures that have fixed opcode sizes. For x64 it still calls the disassembler to get the byte size.
Moved the lldb_private::Instruction::Dump(...) function up into the lldb_private::Instruction class and it now uses the function that gets the mnemonic, operandes and comments so that all disassembly is using the same code.
Added StreamString::FillLastLineToColumn() to allow filling a line up to a column with a character (which is used by the lldb_private::Instruction::Dump(...) function).
Modified the Opcode::GetData() fucntion to "do the right thing" for thumb instructions.
llvm-svn: 156532
Switch over to the "*-apple-macosx" for desktop and "*-apple-ios" for iOS triples.
Also make the selection process for auto selecting platforms based off of an arch much better.
llvm-svn: 156354
<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
Pavel Labath