enabled after we'd found a few bugs that were caused by shadowed
local variables; the most important issue this turned up was
a common mistake of trying to obtain a mutex lock for the scope
of a code block by doing
Mutex::Locker(m_map_mutex);
This doesn't assign the lock object to a local variable; it is
a temporary that has its dtor called immediately. Instead,
Mutex::Locker locker(m_map_mutex);
does what is intended. For some reason -Wshadow happened to
highlight these as shadowed variables.
I also fixed a few obivous and easy shadowed variable issues
across the code base but there are a couple dozen more that
should be fixed when someone has a free minute.
<rdar://problem/12437585>
llvm-svn: 165269
remove the duplicates of this code in ProcessGDBRemote and ProcessKDP.
These two Process plugins will hardcode their DynamicLoader name to be
the DynamicLoaderDarwinKernel so the correct DynamicLoader is picked,
and return the kernel load address as the ImageInfosAddress.
<rdar://problem/12417038>
llvm-svn: 165080
the Symbols::LocateExecutableObjectFile method to locate kexts and
kernels instead of copying them out of the memory of the remote
system. This is the fix for <rdar://problem/12416384>.
Fix a variable shadowing problem in
Symbols::LocateMacOSXFilesUsingDebugSymbols which caused the symbol
rich executable binaries to not be found even if they were listed
in the dSYM Info.plist.
Change Symbols::DownloadObjectAndSymbolFile to ignore dsymForUUID's
negative cache - this is typically being called by the user and we
should try even if there's a incorrect entry in the negative cache.
llvm-svn: 165061
When attaching to a remote system that does not look like a typical vendor system, and no
executable binary was specified to lldb, check a couple of fixed locations where kernels
running in ASLR mode (slid in memory to a random address) store their load addr when booted
in debug mode, and relocate the symbols or load the kernel wholesale from the host computer
if we can find it.
<rdar://problem/7714201>
llvm-svn: 164888
loaded at a random offset).
To get the kernel's UUID and load address I need to send a kdp
packet so I had to implement the kernel relocation (and attempt to
find the kernel if none was provided to lldb already) in ProcessKDP
-- but this code really properly belongs in DynamicLoaderDarwinKernel.
I also had to add an optional Stream to ConnectRemote so
ProcessKDP::DoConnectRemote can print feedback about the remote kernel's
UUID, load address, and notify the user if we auto-loaded the kernel via
the UUID.
<rdar://problem/7714201>
llvm-svn: 164881
runtime, we read method signatures for both class
and instance methods out of the runtime data.
(lldb) fr var str
(NSString *) str = 0x0000000105000180 @"Hello from '/Volumes/Data/projects/lldb/test/lang/objc/foundation/a.out'"
(lldb) expr str.length
(unsigned long long) $0 = 72
(lldb) expr [NSString stringWithCString:"Hello world!" encoding:1]
(id) $1 = 0x0000000105100050
(lldb) po $1
$1 = 0x0000000105100050 Hello world!
(lldb) fr var array1
(NSArray *) array1 = 0x000000010010a6e0 @"3 objects"
(lldb) expr array1.count
(unsigned long long) $0 = 3
(lldb) expr [array1 objectAtIndex:2]
(id) $1 = 0x00000001000025d0
(lldb) po $1
$1 = 0x00000001000025d0 array1 object3
Notice that both regular and property-style notation
work. I still need to add explicit support for
properties with non-default setters/getters.
This information is only queried if an Objective-C
object does not have debug information for a complete
type available. Otherwise we query debug information
as usual.
llvm-svn: 164878
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
where the descriptor took a pointer to an object and
expected the Initialize function to dereference that
pointer and extract the isa value. This caused one
of our tests to fail.
llvm-svn: 164353
KDP -- now with rudimentary process control (continue only) and read + write registers (which means we can see stack frames) for x86_64, i386 and ARM.
llvm-svn: 164352
We can now read the relevant data structures for
the method list, and use a callback mechanism to
report their details to the AppleObjCTypeVendor,
which constructs appropriate Clang types.
llvm-svn: 164310
data structures more rapidly. Also added fields
for the other data structures in a class.
I also fixed a problem where I accidentally used
hasExternalLexicalStorage() instead of
hasExternalVisibleStorage() to mark an
incomplete object.
llvm-svn: 164197
populate Clang ObjCInterfaceDecls with their
ivars, methods, and properties. The default
implementation does nothing. I have also made
sure that AppleObjCRuntimeV2 creates
ObjCInterfaceDecls that actually get queried
appropriately.
llvm-svn: 164164
the dynamic and static runtime class tables to
construct our isa table. This is putting the runtime
in contact with unrealized classes, which we need
to deal with in order to get accurate information.
That's the next piece of work.
<rdar://problem/10986023>
llvm-svn: 163957
Partial fix for the above radar where we now resolve dsym mach-o files within the dSYM bundle when using "add-dsym" through the platform.
llvm-svn: 163676
information from the Objective-C runtime.
This patch takes the old AppleObjCSymbolVendor and
replaces it with an AppleObjCTypeVendor, which is
much more lightweight. Specifically, the SymbolVendor
needs to pretend that there is a backing symbol file
for the Types it vends, whereas a TypeVendor only
vends bare ClangASTTypes. These ClangASTTypes only
need to exist in an ASTContext.
The ClangASTSource now falls back to the runtime's
TypeVendor (if one exists) if the debug information
doesn't find a complete type for a particular
Objective-C interface. The runtime's TypeVendor
maintains an ASTContext full of types it knows about,
and re-uses the ISA-based type query information used
by the ValueObjects.
Currently, the runtime's TypeVendor doesn't provide
useful answers because we haven't yet implemented a
way to iterate across all ISAs contained in the target
process's runtime. That's the next step.
llvm-svn: 163651
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
The attached patch fixes a problem with performing an attach from the SBTarget API on Linux (and other systems that use ProcessPOSIX).
When Process::Attach was called from SBTarget, it resulted in a call to a form of the DoAttachWithID function that wasn't implemented in ProcessPOSIX, and so it fell back to the default implementation (which just returns an error). It didn't seem necessary to use the attach_info parameter for this case, so I just implemented it as a call to the simpler version of the function.
In debugging this problem, I also found that SBTarget wasn't checking the return value from the Attach call, causing it to hang when the attach fails.
llvm-svn: 163399
The attached patch adds support for debugging 32-bit processes when running a 64-bit lldb on an x86_64 Linux system.
Making this work required two basic changes:
1) Getting lldb to report that it could debug 32-bit processes
2) Changing an assumption about how ptrace works when debugging cross-platform
For the first change, I took a conservative approach and only enabled this for x86_64 Linux platforms. It may be that the change I made in Host.cpp could be extended to other 64-bit Linux platforms, but I'm not familiar enough with the other platforms to know for sure.
For the second change, the Linux ProcessMonitor class was assuming that ptrace(PTRACE_[PEEK|POKE]DATA...) would read/write a "word" based on the child process word size. However, the ptrace documentation says that the "word" size read or written is "determined by the OS variant." I verified experimentally that when ptracing a 32-bit child from a 64-bit parent a 64-bit word is read or written.
llvm-svn: 163398
on, basic inlined stepping works, including step-over of inlined functions. But for some as yet mysterious reason i386 debugging gets an
assert and dies immediately. So for now its off.
llvm-svn: 163044
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
Added code the initialize the register context in the OperatingSystemPython plug-in with the new PythonData classes, and added a test OperatingSystemPython module in lldb/examples/python/operating_system.py that we can use for testing.
llvm-svn: 162530
Previously we put a WatchpointSentry object within StopInfo.cpp to disable-and-then-enable the watchpoint itself
while we are performing the actions associated with the triggered watchpoint, which can cause the user-initiated
watchpoint disabling action to be negated.
Add a test case to verify that a watchpoint can be disabled during the callbacks.
llvm-svn: 162483
- no setting auto completion
- very manual and error prone way of getting/setting variables
- tons of code duplication
- useless instance names for processes, threads
Now settings can easily be defined like option values. The new settings makes use of the "OptionValue" classes so we can re-use the option value code that we use to set settings in command options. No more instances, just "does the right thing".
llvm-svn: 162366
'add-dsym' (aka 'target symbols add') should display error messages when dsym file is not found
or the dsym uuid does not match any existing modules. Add TestAddDsymCommand.py test file.
llvm-svn: 162332
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
return 0x0 as the read value instead of uninitialized
stack data so we get consistent behavior from the
emulator.
<rdar://problem/12058770>
llvm-svn: 161795
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
require an AddressClass, which is useless at this
point since it already knows the distinction between
32-bit Thumb opcodes and 32-bit ARM opcodes.
llvm-svn: 161382
keep a shared pointer to their disassembler. This
is important for the LLVM-C disassembler because
it needs to lock its parent in order to disassemble
itself.
This means that every interface that returned a
Disassembler* needs to return a DisassemblerSP, so
that the instructions and any external owners share
the same reference count on the object. I changed
all clients to use this shared pointer, which also
plugged a few leaks.
<rdar://problem/12002822>
llvm-svn: 161123
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
Convert from calling Halt in the lldb Driver.cpp's input reader's sigint handler to sending this AsyncInterrupt so it can be handled in the
event loop.
If you are attaching and get an async interrupt, abort the attach attempt.
Also remember to destroy the process if get interrupted while attaching.
Getting this to work also required handing the eBroadcastBitInterrupt in a few more places in Process WaitForEvent & friends.
<rdar://problem/10792425>
llvm-svn: 160903
calling functions. This is necessary on Mac OS X, since bad things can happen if you set
the registers of a thread that's sitting in a kernel trap.
<rdar://problem/11145013>
llvm-svn: 160756
Remove assertions and turn what used the be the assertion into a logged error with instructions on what to attach to a radar so we can track down why this is happening.
llvm-svn: 160392
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
frame pointer overwritten with the caller's fp value, return to
expressing the CFA in terms of the stack pointer.
<rdar://problem/11855862>
llvm-svn: 160150
a bit -- we're creating the UnwindPlan here, we can set the register set to
whatever is convenient for us, no need to handle different register sets.
A handful of small comment fixes I noticed while reading through the code.
llvm-svn: 159924
Fixed an issue where GDB servers that don't support the thread suffix could get registers states incorrectly due to an incorrect assumption that the current register thread (set using the "Hg%x" packet) will always be cached between runs. Now we clear the cached register thred when the process is resumed.
llvm-svn: 159603
is being run on iOS natively and we are examining a binary that is
in the shared-cache. The shared cache may be set up to not load the
symbol names in memory (and may be missing some local symbols entirely,
to boot) so we need to read the on-disk-but-not-mapped-into-memory cache
of symbol names/symbols before we start processing the in-memory nlist
entries.
This code needs to be reorganized into its own separate method, ideally
we'll find some way to not duplicate the nlist symbol handling. But
we need to handle this new format quickly and we'll clean up later.
Thanks for James McIlree for the patch. Fixes <rdar://problem/11639018>.
llvm-svn: 158891
than being given the pthread_mutex_t from the Mutex and locks that. That allows us to
track ownership of the Mutex better.
Used this to switch the LLDB_CONFIGURATION_DEBUG enabled assert when we can't get the
gdb-remote sequence mutex to assert when the thread that had the mutex releases it. This
is generally more useful information than saying just who failed to get it (since the
code that had it locked often had released it by the time the assert fired.)
llvm-svn: 158240
a cache of address ranges for child sections,
accelerating lookups. This cache is built during
object file loading, and is then set in stone once
the object files are done loading. (In Debug builds,
we ensure that the cache is never invalidated after
that.)
llvm-svn: 158188
m_interrupt_sent into account. Also don't reset m_interrupt_sent in SendInterrupt but do so in SendPacketAndWaitForResponse
when we know we've handled the interrupt.
Fix a code path through ProcessGDBRemote::DoDestroy where we were tearing down the debug session but
not setting the exit status.
llvm-svn: 158043
that automatically generated setters/getters only
get added to a class after explicitly declared (or
synthesized) getters/setters had the chance to be
added. This eliminates conflicts creating errors
of the form:
error: instance method '...' has incompatible result
types in different translation units ('X *' vs. 'id')
llvm-svn: 157956
Fixed an issue with the current type being set to DIE_IS_BEING_PARSED in the m_die_to_type map by making sure the type pointer is valid.
llvm-svn: 157836
Fixed a case where multiple threads can be asking to send a packet to the GDB server and one of three things will happen:
1 - everything works
2 - one thread will fail to send the packet due to not being able to get the sequence mutex
3 - one thread will try and interrupt the other packet sending and fail and not send the packet
Now the flow is a bit different. Prior to this fix we did:
if (try_get_sequence_mutex()) {
send_packet()
return success;
} else {
if (async_ok) {
interrupt()
send_packet()
resume()
return success;
}
}
return fail
The issue is that the call to "try_get_sequence_mutex()" could fail if another thread was sending a packet and could cause us to just not send the packet and an error would be returned.
What we really want is to try and get the sequence mutex, and if this succeeds, send the packet. Else check if we are running and if we are, do what we used to do. The big difference is when we aren't running, we wait for the sequence mutex so we don't drop packets. Pseudo code is:
if (try_get_sequence_mutex()) {
// Safe to send the packet right away
send_packet()
return success;
} else {
if (running) {
// We are running, interrupt and send async packet if ok to do so,
// else it is ok to fail
if (async_ok) {
interrupt()
send_packet()
resume()
return success;
}
}
else {
// Not running, wait for the sequence mutex so we don't drop packets
get_sequence_mutex()
send_packet()
return success;
}
}
return fail
llvm-svn: 157751
Fixed an issue with the symbol table parsing of files that have STAB entries in them where there are two N_SO entries where the first has a directory, and the second contains a full path:
[ 0] 00000002 64 (N_SO ) 00 0000 0000000000000000 '/Volumes/data/src/'
[ 1] 0000001e 64 (N_SO ) 00 0000 0000000000000000 '/Volumes/data/src/Source/main.m'
[ 2] 00000047 66 (N_OSO ) 09 0001 000000004fc642d2 '/tmp/main.o'
[ 3] 00000001 2e (N_BNSYM ) 01 0000 0000000000003864
[ 4] 000000bd 24 (N_FUN ) 01 0000 0000000000003864 '_main'
[ 5] 00000001 24 (N_FUN ) 00 0000 00000000000000ae
[ 6] 00000001 4e (N_ENSYM ) 01 0000 00000000000000ae
[ 7] 00000001 64 (N_SO ) 01 0000 0000000000000000
We now correctly combine entries 0 and 1 into a single entry.
llvm-svn: 157712
setting breakpoints. That's dangerous, since while we are setting a breakpoint,
the target might hit the dyld load notification, and start removing modules from
the list. This change adds a GetMutex accessor to the ModuleList class, and
uses it whenever we are accessing the target's ModuleList (as returned by GetImages().)
<rdar://problem/11552372>
llvm-svn: 157668
these functions will end in the sequence
mov %rbp, %rsp
ret
call __stack_chk_fail
instead of the usual mov, ret. The x86 assembly profiler only looked
for functions ending in 'ret' and added the Unwind row describing how to
set the CFA based on that -- the addition of the call insn (which is jumped
to earlier in the function body) threw off that inspection.
Resolves the need to "step" twice to get out of these functions when doing
source-level stepping.
<rdar://problem/11469705>
llvm-svn: 157454
Sending async packets can deadlock a program on darwin. We currently allow breakpoint packets and memory read/write packets (for software breakpoints) to be sent while a program is running. In the GDB remote plug-in, we will interrupt the run, send the async packet and resume (currently with the continue packet that caused the program to resume). If the GDB server supports the "vCont" packet, we might have initially continued with each thread stating it should continue. If new threads show up while we are stopped, which happend when running GCD, we can end up with new threads that we aren't mentioning in the continue list. So we start with a thread list of 1,2,3 and continue:
continue thread 1, continue thread 2, continue thread 3
Now we interrupt and set a breakpoint and we actually have threads 1,2,3,4 now when we are about to resume, yet we send:
continue thread 1, continue thread 2, continue thread 3
Any thread that isn't mentioned is currently going to stay suspended. This causes the deadlock.
llvm-svn: 157439
Add default Process::GetWatchpointSupportInfo() impl which returns an error of "not supported".
Add "qWatchpointSupportInfo" packet to the gdb communication layer to support this, and modify TestWatchpointCommands.py to test it.
llvm-svn: 157345
the value_regs field, which is useful for future expansion purposes. As of now, we have:
calculated_offset_of_eax = offset_of_rax + (offset_of_eax_from_the_descriptor which is 0)
llvm-svn: 157275
Add convenience registers eax, ebx, ecx, edx, edi, esi, ebp, esp to the 'register read' command for x86_64.
Add a GDBRemoteRegisterContext::Addx86_64ConvenienceRegisters() method called from ProcessGDBRemote::BuildDynamicRegisterInfo().
Servicing of eax, for example, is accomplished by delegating to rax with an adjusted offset into the register context.
llvm-svn: 157230
"break set" commands to set this per breakpoint. Also, some CreateBreakpoint API's in the lldb_private
namespace had "internal" first and "skip_prologue" second. "internal should always be last. Fixed that.
rdar://problem/11484729
llvm-svn: 157225
(actually, mainly just hooked up support that was already
there). Added a test case, although it's expected to fail
right now unless you're using top-of-tree LLVM.
llvm-svn: 157220
boxed expressions returning numbers and strings.
I also added boxed expressions to our testcases, and
enabled boxed expressions when libarclite is linked into
the inferior.
llvm-svn: 157026
that dynamically discovers remote register context information.
o GDBRemoteRegisterContext.h:
Change the prototype of HardcodeARMRegisters() to take a boolean flag, which now becomes
void
HardcodeARMRegisters(bool from_scratch);
o GDBRemoteRegisterContext.cpp:
HardcodeARMRegisters() now checks the from_scratch flag and decides whether to add composite registers to the already
existing primordial registers based on a table called g_composites which describes the composite registers.
o ProcessGDBRemote.cpp:
Modify the logic of ProcessGDBRemote::BuildDynamicRegisterInfo() to call m_register_info.HardcodeARMRegisters()
with the newly introduced 'bool from_scrach' flag.
rdar://problem/10652076
llvm-svn: 156773
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
the controlling plans so that they don't lose control.
Also change "ThreadPlanStepThrough" to take the return StackID for its backstop breakpoint as an argument
to the constructor rather than having it try to figure it out itself, since it might get it wrong whereas
the caller always knows where it is coming from.
rdar://problem/11402287
llvm-svn: 156529
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
No one was using it and Locker(pthread_mutex_t *) immediately asserts for
pthread_mutex_t's that don't come from a Mutex anyway. Rather than try to make
that work, we should maintain the Mutex abstraction and not pass around the
platform implementation...
Make Mutex::Locker::Lock take a Mutex & or a Mutex *, and remove the constructor
taking a pthread_mutex_t *. You no longer need to call Mutex::GetMutex to pass
your mutex to a Locker (you can't in fact, since I made it private.)
llvm-svn: 156221
us of its architecture, use that to set the Target's arch if it
doesn't already have one set.
In Process::CompleteAttach(), if the Target has a valid arch make
sure that the Platform we pick up is compatible with that arch; if
not, find a Platform that is compatible. Don't let the the default
platform override the Target's arch.
<rdar://problem/11185420>
llvm-svn: 156116
ObjCPlusPlus as Objective-C classes. Really the
compiler should say they have Objective-C runtime
class, but we should be a little more resilient
(we were refusing to find ivars in those classes
before).
Also added a test case.
llvm-svn: 155515
Fixed an issue that would happen when using debug map with DWARF in the .o files where we wouldn't ever track down the actual definition for a type when things were in namespaces. We now serialize the decl context information into an intermediate format which allows us to track down the correct definition for a type regardless of which DWARF symbol file it comes from. We do this by creating a "DWARFDeclContext" object that contains the DW_TAG + name for each item in a decl context which we can then use to veto potential accelerator table matches. For example, the accelerator tables store the basename of the type, so if you have "std::vector<int>", we would end up with an accelerator table entry for the type that contained "vector<int>", which we would then search for using a DWARFDeclContext object that contained:
[0] DW_TAG_class_type "vector<int>"
[1] DW_TAG_namespace "std"
This is currently used to track down forward declarations for things like "class a:🅱️:Foo;".
llvm-svn: 155488
class AnalysisResolver;
And we will look for it everywhere and find many many matches, but the decl context of those matching DIEs is "clang::AnalysisResolver", so we never match anything, yet we pull in waaayyy too much DWARF in the process.
To enable this logging enable the "lookups" category in the "dwarf" log channel:
(lldb) log enable dwarf lookups
llvm-svn: 155233
Fixed an issue where iOS debugging would trust the first file it found in the SDK regardless of the UUID not matching. Now we actually get smart and can find modules in ANY of the installed SDKs and remember which SDK is our fallback SDK.
llvm-svn: 155184
Also enabled PlatformRemoteiOS to select an SDK using the build number in case you have mutliple 5.0 SDKs installed:
(lldb) platform select remote-ios --build 11C123
llvm-svn: 154978
Error
Host::RunShellCommand (const char *command,
const char *working_dir,
int *status_ptr,
int *signo_ptr,
std::string *command_output_ptr,
uint32_t timeout_sec);
This will allow us to use this functionality in the host lldb_private::Platform, and also use it in our lldb-platform binary. It leverages the existing code in Host::LaunchProcess and ProcessLaunchInfo.
llvm-svn: 154730
Hello everyone,
please find the attached patch for TOT and lldb-platform-work branch, which provides the following changes:
- fixed a crash in the ProcessPOSIX constructor when an executable module object is not yet created.
- added support for the multi instanciated FreeBSD platform objects (the local host and remote as example).
- enabled the remote gdb plugin on FreeBSD.
llvm-svn: 154724
Enable logging the packet history when registers fail to read due to not getting the sequence mutex if "--verbose" is enabled on the log channel for the "gdb-remote" log category.
This will help us track down some issues.
llvm-svn: 154704
The less locks there are, the better. I removed the thread ID mutex and now just shared the m_thread_list's mutex to make sure we don't deadlock due to lock inversion.
llvm-svn: 154652
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
FunctionDecls into classes if it looked up a
method in a different DWARF context than the
one where it found the parent class's definition.
The symptom of this was, for a method A::B(),
1) LLDB finds A in context 1, creating a
CXXRecordDecl for A and marking it as needing
completion
2) LLDB looks up B in context 2, finds that its
parent A already has a CXXRecordDecl, but can't
find a CXXMethodDecl for B
3) Not finding a CXXMethodDecl for B, LLDB doesn't
set the flag indicating that B was resolved
4) Because the flag wasn't set, LLDB's fallthrough
code creates a FunctionDecl for B and sticks it
in the DeclContext -- in this case, A.
5) Clang crashes on finding a FunctionDecl inside a
CXXRecordDecl.
llvm-svn: 154627
for packet confirmation.
Also added a bit more logging.
Also, unlock the writer end of the run lock in Process.cpp on our way out of the private state
thread so that the Process can shut down cleanly.
<rdar://problem/11228538>
llvm-svn: 154601
Cleaned up the Mutex::Locker and the ReadWriteLock classes a bit.
Also cleaned up the GDBRemoteCommunication class to not have so many packet functions. Used the "NoLock" versions of send/receive packet functions when possible for a bit of performance.
llvm-svn: 154458
QListThreadsInStopReply
This GDB remote query command can enable added a "threads" key/value pair to all stop reply packets so that we always get a list of all threads in each stop reply packet. It increases performance if enabled (the reply to the "QListThreadsInStopReply" is "OK") by saving us from sending to command/reply pairs (the "qfThreadInfo" and "qsThreadInfo" packets), and also helps us keep the current process state up to date.
llvm-svn: 154380
The next step is to have our stop reply packets send the thread list in the actual stop reply packet to avoid a 2 packet overhead of sending the qfThreadInfo + response and qfThreadInfo + response.
llvm-svn: 154376
The current ProcessGDBRemote function that updates the threads could end up with an empty list if any other thread had the sequence mutex. We now don't clear the thread list when we can't access it, and we also have changed how lldb_private::Process handles the return code from the:
virtual bool
Process::UpdateThreadList (lldb_private::ThreadList &old_thread_list,
lldb_private::ThreadList &new_thread_list) = 0;
A bool is now returned to indicate if the list was actually updated or not and the lldb_private::Process class will only update the stop ID of the validity of the thread list if "true" is returned.
The ProcessGDBRemote also got an extra assertion that will hopefully assert when running debug builds so we can find the source of this issue.
llvm-svn: 154365
spin up a temporary "private state thread" that will respond to events from the lower level process plugins. This check-in should work to do
that, but it is still buggy. However, if you don't call functions on the private state thread, these changes make no difference.
This patch also moves the code in the AppleObjCRuntime step-through-trampoline handler that might call functions (in the case where the debug
server doesn't support the memory allocate/deallocate packet) out to a safe place to do that call.
llvm-svn: 154230
correctly if the setter/getter were not present
in the debug information. The fixes are as follows:
- We not only look for the method by its full name,
but also look for automatically-generated methods
when searching for a selector in an Objective-C
interface. This is necessary to find accessors.
- Extract the getter and setter name from the
DW_TAG_APPLE_Property declaration in the DWARF
if they are present; generate them if not.
llvm-svn: 154067
Fixed an issue where there were more than one way to get a CompileUnitSP created when using SymbolFileDWARF with SymbolFileDWARFDebugMap. This led to an assertion that would fire under certain conditions. Now there is only one way to create the compile unit and it will "do the right thing".
llvm-svn: 153908
(lldb) log enable --verbose lldb completion
This will print out backtraces for all type completion calls which will help us verify that we don't ever complete a type when we don't need to.
llvm-svn: 153787
Fixed an issue that could cause circular type parsing that will assert and kill LLDB.
Prior to this fix the DWARF parser would always create class types and not start their definitions (for both C++ and ObjC classes) until we were asked to complete the class later. When we had cases like:
class A
{
class B
{
};
};
We would alway try to complete A before specifying "A" as the decl context for B. Turns out we can just start the definition and still not complete the class since we can check the TagDecl::isCompleteDefinition() function. This only works for C++ types. This means we will not be pulling in the full definition of parent classes all the time and should help with our memory consumption and also reduce the amount of debug info we have to parse.
I also reduced redundant code that was checking in a lldb::clang_type_t was a possible C++ dynamic type since it was still completing the type, just to see if it was dynamic. This was fixed in another function that was checking for a type being dynamic as an ObjC or a C++ type, but there was dedicated fucntion for C++ that we missed.
llvm-svn: 153713
Symbol files (dSYM files on darwin) can now be specified during program execution:
(lldb) target symbols add /path/to/symfile/a.out.dSYM/Contents/Resources/DWARF/a.out
This command can be used when you have a debug session in progress and want to add symbols to get better debug info fidelity.
llvm-svn: 153693
for unbacked properties. We support two variants:
one in which the getter/setter are provided by
selector ("mySetter:") and one in which the
getter/setter are provided by signature
("-[MyClass mySetter:]").
llvm-svn: 153675
<rdar://problem/11051056>
Found a race condition when sending async packets in the ProcessGDBRemote.
A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like:
inst single step
inst single step
inst single step
inst single step
inst single step
step BP and run
inst single step
inst single step
inst single step
Now if we got an async packet while the program is running we get something like:
send --> continue
send --> interrupt
recv <-- interrupt stop reply packet
send --> async packet
recv <-- async response
send --> continue again and wait for actual stop
Problems arise when this was happening when single stepping a thread where we would get:
send --> step thread 123
send --> interrupt
send --> stop reply for thread 123 (from the step)
Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with:
send --> async packet (like memory read!)
recv <-- async response (which is the interrupt stop reply packet)
Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet...
To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted.
The fixes I checked in to two major things:
- watch for the extra stop reply if we need to
- make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished.
Needless to say this makes very fast stepping in Xcode much more reliable.
llvm-svn: 153629
indicates that the section is thread specific. Any functions the load a module
given a slide, will currently ignore any sections that are thread specific.
lldb_private::Section now has:
bool
Section::IsThreadSpecific () const
{
return m_thread_specific;
}
void
Section::SetIsThreadSpecific (bool b)
{
m_thread_specific = b;
}
The ELF plug-in has been modified to set this for the ".tdata" and the ".tbss"
sections.
Eventually we need to have each lldb_private::Thread subclass be able to
resolve a thread specific section, but for now they will just not resolve. The
code for that should be trivual to add, but the address resolving functions
will need to be changed to take a "ExecutionContext" object instead of just
a target so that thread specific sections can be resolved.
llvm-svn: 153537
1 - sections only get a valid VM size if they have SHF_ALLOC in the section flags
2 - symbol names are marked as mangled if they start with "_Z"
Also fixed the DWARF parser to correctly use the section file size when extracting the DWARF.
llvm-svn: 153496
Fixed type lookups to "do the right thing". Prior to this fix, looking up a type using "foo::bar" would result in a type list that contains all types that had "bar" as a basename unless the symbol file was able to match fully qualified names (which our DWARF parser does not).
This fix will allow type matches to be made based on the basename and then have the types that don't match filtered out. Types by name can be fully qualified, or partially qualified with the new "bool exact_match" parameter to the Module::FindTypes() method.
This fixes some issue that we discovered with dynamic type resolution as well as improves the overall type lookups in LLDB.
llvm-svn: 153482
- Addresses with no description were given
comments, leading to useless comments like
"; , "
- Addresses weren't resolved with respect
to the correct module.
llvm-svn: 153274
them both installed with the LLVM MC version being the default. I renamed the
name of the LLVM MC disassembler plug-in to "llvm-mc" and the LLVM enhanced
disassembly plug-in to "llvm-edis" and they can both be installed for now.
To use the "llvm-edis" disassembler, you can just specify it while disassembling:
(lldb) disassemble --plugin llvm-edis --name main
(lldb) disassemble --plugin llvm-mc --name main
This will allow us to compare the output of the two disassembler and eventually
deprecate the old one when the new one is ready. But it does use the new disassembler
by default so we continue to test it on a daily basis.
llvm-svn: 153231
the migration to ModuleSpec objects this broke and is now fixed.
Also fixed a case in the darwin kernel dynamic loader where we just need to
trust the load address of the kernel if we can't read it from memory.
llvm-svn: 153164
Each platform now knows if it can handle an architecture and a platform can be found using an architecture. Each platform can look at the arch, vendor and OS and know if it should be used or not.
llvm-svn: 153104
Fixed a case where the source path remappings on the module were too expensive to
use when we try to verify (stat the file system) that the remapped path points to
a valid file. Now we will use the lldb_private::Module path remappings (if any) when
parsing the debug info without verifying that the paths exist so we don't slow down
line table parsing speeds.
llvm-svn: 153059
http://llvm.org/bugs/show_bug.cgi?id=12232
Fixed a case where a missing "break" in a switch statement could cause an assertion to fire and kill the debug session.
The fix was derived from the findings of Andrea Bigagli, thanks Andrea.
llvm-svn: 152741
Fixed an issue with the FUNC_STARTS load command where we would get the
symbol size wrong and we would add all sorts of symbols due to bit zero being
set to indicate thumb.
llvm-svn: 152696
Simplify the locking strategy for Module and its owned objects to always use the Module's mutex to avoid A/B deadlocks. We had a case where a symbol vendor was locking itself and then calling a function that would try to get it's Module's mutex and at the same time another thread had the Module mutex that was trying to get the SymbolVendor mutex. Now any classes that inherit from ModuleChild should use the module lock using code like:
void
ModuleChildSubclass::Function
{
ModuleSP module_sp(GetModule());
if (module_sp)
{
lldb_private::Mutex::Locker locker(module_sp->GetMutex());
... do work here...
}
}
This will help avoid deadlocks by using as few locks as possible for a module and all its child objects and also enforce detecting if a module has gone away (the ModuleSP will be returned empty if the weak_ptr does refer to a valid object anymore).
llvm-svn: 152679
For EmulateInstructionARM::EmulatePUSH(), fix the logical branch for when pc is pushed to behave
like the other cases where:
context.SetRegisterToRegisterPlusOffset (reg_info, sp_reg, addr - sp);
is called to inform of the operation to set a register value to a memory location calculated from
a base register plus an offset.
llvm-svn: 152670
load notification for the first load) then we will set it the runtime to NULL and won't re-search for it.
Added a way for the dynamic loader to force a re-search, since it knows the world has changed.
llvm-svn: 152453
Get function boundaries from the LC_FUNCTION_STARTS load command. This helps to determine symbol sizes and also allows us to be able to debug stripped binaries.
If you have a stack backtrace that goes through a function that has been stripped from the symbol table, the variables for any functions above that stack frame will most likely be incorrect. It can also affect our ability to step in/out/through of a function.
llvm-svn: 152381
Jason Molenda