The issue was when we called Debugger::RunIOHandler(), it would run the current IOHandler by activating it, and running it and then try to pop it and exit regardless of wether it was on top or not.
The new code will push the IOHandler that was passed in, and run the IOHandlers until the one passed in is successfully popped. This allows files for the "command source" to switch input handlers:
% cat /tmp/commands
br s -S alignLeftEdges:
br command add
bt
frame var
po self
DONE
b s -n main
br command add
bt
frame var
DONE
Note above we set a breakpoint, then add commands do it. The "br command add" will push the breakpoint comment gatherer until it sees "DONE" and then pop itself off the stack. The a new breakpoint will be set and it does the same thing again.
Now this file can be sourced from the command line:
% lldb -s /tmp/commands /path/to/a.out
And your breakpoints will be correctly setup!
<rdar://problem/17081650>
llvm-svn: 211329
lldb support. I'll be doing more testing & cleanup but I wanted to
get the initial checkin done.
This adds a new SBExpressionOptions::SetLanguage API for selecting a
language of an expression.
I added adds a new SBThread::GetInfoItemByPathString for retriving
information about a thread from that thread's StructuredData.
I added a new StructuredData class for representing
key-value/array/dictionary information (e.g. JSON formatted data).
Helper functions to read JSON and create a StructuredData object,
and to print a StructuredData object in JSON format are included.
A few Cocoa / Cocoa Touch data formatters were updated by Enrico
to track changes in iOS 8 / Yosemite.
Before we query a thread's extended information, the system runtime may
provide hints to the remote debug stub that it will use to retrieve values
out of runtime structures. I added a new SystemRuntime method
AddThreadExtendedInfoPacketHints which allows the SystemRuntime to add
key-value type data to the initial request that we send to the remote stub.
The thread-format formatter string can now retrieve values out of a thread's
extended info structured data. The default thread-format string picks up
two of these - thread.info.activity.name and thread.info.trace_messages.
I added a new "jThreadExtendedInfo" packet in debugserver; I will
add documentation to the lldb-gdb-remote.txt doc soon. It accepts
JSON formatted arguments (most importantly, "thread":threadnum) and
it returns a variety of information regarding the thread to lldb
in JSON format. This JSON return is scanned into a StructuredData
object that is associated with the thread; UI layers can query the
thread's StructuredData to see if key-values are present, and if
so, show them to the user. These key-values are likely to be
specific to different targets with some commonality among many
targets. For instance, many targets will be able to advertise the
pthread_t value for a thread.
I added an initial rough cut of "thread info" command which will print
the information about a thread from the jThreadExtendedInfo result.
I need to do more work to make this format reasonably.
Han Ming added calls into the pmenergy and pmsample libraries if
debugserver is run on Mac OS X Yosemite to get information about the
inferior's power use.
I added support to debugserver for gathering the Genealogy information
about threads, if it exists, and returning it in the jThreadExtendedInfo
JSON result.
llvm-svn: 210874
(lldb) file /bin/ls
(lldb) b malloc
(lldb) run
(lldb) process save-core /tmp/ls.core
Each ObjectFile plug-in now has the option to save core files by registering a new static callback.
llvm-svn: 210864
We preivously had two copies of ::BytesAvailable with only trivial
differences between them, and fixes have been applied to only one of
them.
Instead of duplicating the whole function, hide the FD_SET differences
behind a macro. This leaves only one small __APPLE__-specific #if
block, and fixes ^C on non-__APPLE__ platforms.
llvm-svn: 210592
Changes include:
- ObjectFileMachO can now determine if a binary is "*-apple-ios" or "*-apple-macosx" by checking the min OS and SDK load commands
- ArchSpec now says "<arch>-apple-macosx" is equivalent to "<arch>-apple-ios" since the simulator mixes and matches binaries (some from the system and most from the iOS SDK).
- Getting process inforamtion on MacOSX now correctly classifies iOS simulator processes so they have "*-apple-ios" architectures in the ProcessInstanceInfo
- PlatformiOSSimulator can now list iOS simulator processes correctly instead of showing nothing by using:
(lldb) platform select ios-simulator
(lldb) platform process list
- debugserver can now properly return "*-apple-ios" for the triple in the process info packets for iOS simulator executables
- GDBRemoteCommunicationClient now correctly passes along the triples it gets for process info by setting the OS in the llvm::Triple correctly
<rdar://problem/17060217>
llvm-svn: 209852
This fixes a number of trivial warnings in the Windows build. This is part of a larger effort to make the Windows build warning-free.
See http://reviews.llvm.org/D3914 for more details.
Change by Zachary Turner
llvm-svn: 209749
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
Add a callback that will allow an expression to be cancelled between the
expression evaluation stages (for the ClangUserExpressions.)
<rdar://problem/16790467>, <rdar://problem/16573440>
llvm-svn: 207944
- CTRL+C wasn't clearing the command in lldb
- CTRL+C doesn't work in python macros in lldb
- Ctrl+C no longer interrupts the running process that you attach to
<rdar://problem/15949205>
<rdar://problem/16778652>
<rdar://problem/16774411>
llvm-svn: 207816
Currently if you run _any_ python, python has the "lldb.debugger" global variable and it has a strong reference to a lldb_private::Debugger since it is a lldb::SBDebugger object with a shared pointer.
This makes sure that your LLDB command interpreter history is saved each time you quit command line LLDB.
llvm-svn: 207164
Set the correct FormatManager revision before starting to figure out the new formatters
This can avoid entering some corner cases where as part of figuring out formatters we try to figure out dynamic types, and in turn that causes us to go back in trying to fetch new formatters - it is not only a futile exercise, it's also prone to endless recursion
This would only cause a behavior change if getting this chain started would eventually cause something to run and alter the formatters, a very unlikely if at all possible sequence of events
llvm-svn: 205928
TIDs are conventionally shown as decimal values on FreeBSD and Linux.
Thus, use the ${thread.id%tid} format string to display the thread ID,
instead of a fixed hex format.
llvm.org/pr19380
llvm-svn: 205912
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
For some reason, the libc++ vector<bool> data formatter was essentially a costly no-up, doing everything required of it, except actually generating the child values!
This restores its functionality
llvm-svn: 205259
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
For small structs, the frame format now prints them as one-liners
This follows the same definition that frame variable does for deciding what a "small struct" is, and as such should be fairly consistent with the variable display in general
llvm-svn: 204762
Add a GetFoundationVersion() to AppleObjCRuntime
This API is used to return and cache the major version of Foundation.framework, which is potentially a useful piece of data to key off of to enable or disable certain ObjC related behaviors (especially in data formatters)
llvm-svn: 204756
(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
TestPromptFormats appears as though it may be a useful unit test.
Unfortunately, there is no invocation mechanism in place right now. It is
unclear how to add a unit test for this scenario to the existing tests. It
would be ideal to remove this entirely, but I am hopeful that this can/will be
pulled out into a test still since it uses a user accessible interface.
llvm-svn: 204309
for customizing "step-in" behavior (e.g. step-in doesn't step into code with no debug info), but also
the behavior of step-in/step-out and step-over when they step out of the frame they started in.
I also added as a proof of concept of this reworking a mode for stepping where stepping out of a frame
into a frame with no debug information will continue stepping out till it arrives at a frame that does
have debug information. This is useful when you are debugging callback based code where the callbacks
are separated from the code that initiated them by some library glue you don't care about, among other
things.
llvm-svn: 203747
read during materialization. First of all, report
if we can't read the data for some reason. Second,
consult the ValueObject's error and report that if
there's some problem.
<rdar://problem/16074201>
llvm-svn: 202552
I carefully reviewed exactly how the IOHandlers interact and found places where we weren't properly controlling things. There should be no overlapping prompts and all output should now come out in a controlled fashion.
<rdar://problem/16111293>
llvm-svn: 202525
Revert the spirit of r199857 - a convincing case can be made that overriding a summary's format markers behind its back is not the right thing to do
This commit reverts the behavior of the code to the previous model, and changes the test case to validate the opposite of what it was validating before
llvm-svn: 201455
ObjectFile::SetLoadAddress (Target &target,
lldb::addr_t value,
bool value_is_offset);
Now "value" is a slide if "value_is_offset" is true, and "value" is an image base address otherwise. All previous usage of this API was using slides.
Updated the ObjectFileELF and ObjectFileMachO SetLoadAddress methods to do the right thing.
Also updated the ObjectFileMachO::SetLoadAddress() function to not load __LINKEDIT when it isn't needed and to only load sections that belong to the executable object file.
llvm-svn: 201003
Move some code that was in DynamicLoaderPOSIXDLYD into the
base class DynamicLoader. In the case of UpdateLoadedSections(),
the test to see whether a file is loadable (its address is zero)
is not generally applicable so that test is changed to a more
universally applicable check for the SHF_ALLOC flag on the section.
Also make it explicit that the reading of the module_id in
DynamicLoaderPOSIXDYLD::GetThreadLocalData() is using a hardcoded
size (of module_id) of 4, which might not be appropriate on
big-endian 64-bit systems, leaving a FIXME comment in place.
llvm-svn: 200939
- empty lines in init files would repeat previous command and cause errors to be displayed
- all options to control showing the command, its output, if it should stop on error or continue, weren't being obeyed.
llvm-svn: 200860
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
GetU32 and GetU64, to use memcpy to copy bytes into a local buffer instead
of having a (uint64_t *) etc local variable, pointing to the address, and
dereferencing it. If compiled on a CPU where data alignment is required
(e.g. the LDM instruction on armv7) and we try to GetU64 out of a mmap'ed
DWARF file, that 8 byte quantity may not be world aligned and the program
can get an unaligned memory access fault.
<rdar://problem/15849231>
llvm-svn: 200069
ArchSpec now contains an optional distribution_id, with getters and
setters. Host::GetArchitecture () sets it on non-Apple platforms using
Host::GetDistributionId (). The distribution_id is ignored during
ArchSpec comparisons.
The gdb remote qHostInfo message transmits it, if set, via the
distribution_id={id-value} key/value pair. Updated gdb remote docs to
reflect this change.
As before, GetDistributionId () returns nothing on non-Linux platforms
at this time. On Linux, it is returned only if the lsb_platform
command is installed (in /bin or /usr/bin), and only if the
distributor id key is returned by 'lsb_platform -i'. This id is
lowercased, and whitespace is replaced with underscores.
llvm-svn: 199539
symbols correctly. There were a couple of pieces to this.
1) When a breakpoint location finds itself pointing to an Indirect symbol, when the site for it is created
it needs to resolve the symbol and actually set the site at its target.
2) Not all breakpoints want to do this (i.e. a straight address breakpoint should always set itself on the
specified address, so somem machinery was needed to specify that.
3) I added some info to the break list output for indirect symbols so you could see what was happening.
Also I made it clear when we re-route through re-exported symbols.
4) I moved ResolveIndirectFunction from ProcessPosix to Process since it works the exact same way on Mac OS X
and the other posix systems. If we find a platform that doesn't do it this way, they can override the
call in Process.
5) Fixed one bug in RunThreadPlan, if you were trying to run a thread plan after a "running" event had
been broadcast, the event coalescing would cause you to miss the ThreadPlan running event. So I added
a way to override the coalescing.
6) Made DynamicLoaderMacOSXDYLD::GetStepThroughTrampolinePlan handle Indirect & Re-exported symbols.
<rdar://problem/15280639>
llvm-svn: 198976
The "type format add" command gets a new flag --type (-t). If you pass -t <sometype>, upon fetching the value for an object of your type,
LLDB will display it as-if it was of enumeration type <sometype>
This is useful in cases of non-contiguous enums where there are empty gaps of unspecified values, and as such one cannot type their variables as the enum type,
but users would still like to see them as-if they were of the enum type (e.g. DWARF field types with their user-reserved ranges)
The SB API has also been improved to handle both types of formats, and a test case is added
llvm-svn: 198105
TypeFormatImpl used to just wrap a Format (and Flags for matching), and then ValueObject itself would do the printing deed
With this checkin, the responsibility of generating a value string is centralized in the data formatter (as it should, and already is for summaries)
This change is good practice per se, and should also enable us to extend the type format mechanism in a cleaner way
llvm-svn: 197874
Previously, an opcode set via SetOpcode32 (for example) was later
extracted via GetData() as a byte sequence in host order rather than
target order.
Review: http://llvm-reviews.chandlerc.com/D1838
llvm-svn: 196808
<rdar://problem/15314403>
This patch adds a new lldb_private::SectionLoadHistory class that tracks what shared libraries were loaded given a process stop ID. This allows us to keep a history of the sections that were loaded for a time T. Many items in history objects will rely upon the process stop ID in the future.
llvm-svn: 196557
This gets rid of our hacky "get_random_port()" which would grab a random port and tell debugserver to open that port. Now LLDB creates, binds, listens and accepts a connection by binding to port zero and sending the correctly bound port down as the host:port to connect back to.
Fixed the "ConnectionFileDescriptor" to be able to correctly listen for connections from a specified host, localhost, or any host. Prior to this fix "listen://" only accepted the following format:
listen://<port>
But now it can accept:
listen://<port> // Listen for connection from localhost on port <port>
listen://<host>:<port> // Listen for connection from <host> and <port>
listen://*:<port> // Listen for connection from any host on port <port>
llvm-svn: 196547
lldb_private::Debugger was #including some "lldb/API" header files which causes tools (lldb-platform and lldb-gdbserver) that link against the internals only (no API layer) to fail to link depending on which calls were being used.
Also fixed the current working directory so that it gets set correctly for remote test suite runs. Now the remote working directory is set to: "ARCH/TESTNUM/..." where ARCH is the current architecture name and "TESTNUM" is the current test number.
Fixed the "lldb-platform" and "lldb-gdbserver" to not warn about mismatched visibility settings by having each have their own exports file which contains nothing. This forces all symbols to not be exported, and also quiets the linker warnings.
llvm-svn: 196141
Rework data formatters matching algorithm
What happens now is that, for each category, the FormatNavigator generates all possible matches, and checks them one by one
Since the possible matches do not actually depend on the category (whether a match is accepted or not does, but that check can be shifted at a more convenient time),
it is actually feasible to generate every possible match upfront and then let individual categories just scan through those
This commit changes things by introducing a notion of formatters match candidate, and shifting responsibility for generating all of them given a (ValueObject,DynamicValueType) pair
from the FormatNavigator back to the FormatManager
A list of these candidates is then passed down to each category for matching
Candidates also need to remember whether they were generated by stripping pointers, references, typedefs, since this is something that individual formatters can choose to reject
This check, however, is conveniently only done once a "textual" match has been found, so that the list of candidates is truly category-independent
While the performance benefit is small (mostly, due to caching), this is much cleaner from a design perspective
llvm-svn: 195395
Example code:
remote_platform = lldb.SBPlatform("remote-macosx");
remote_platform.SetWorkingDirectory("/private/tmp")
debugger.SetSelectedPlatform(remote_platform)
connect_options = lldb.SBPlatformConnectOptions("connect://localhost:1111");
err = remote_platform.ConnectRemote(connect_options)
if err.Success():
print >> result, 'Connected to remote platform:'
print >> result, 'hostname: %s' % (remote_platform.GetHostname())
src = lldb.SBFileSpec("/Applications/Xcode.app/Contents/SharedFrameworks/LLDB.framework", False)
dst = lldb.SBFileSpec()
# copy src to platform working directory since "dst" is empty
err = remote_platform.Install(src, dst);
if err.Success():
print >> result, '%s installed successfully' % (src)
else:
print >> result, 'error: failed to install "%s": %s' % (src, err)
Implemented many calls needed in lldb-platform to be able to install a directory that contains symlinks, file and directories.
The remote lldb-platform can now launch GDB servers on the remote system so that remote debugging can be spawned through the remote platform when connected to a remote platform.
The API in SBPlatform is subject to change and will be getting many new functions.
llvm-svn: 195273
The demangler added in r193708 from cxa_demangle.cpp uses language features which are not supported by the latest visual studio. In order to preserve the msvc build, this patch restores the previous (non)functionality in windows under msvc by disabling the demangler.
llvm-svn: 195254
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
FreeBSD includes the elftoolchain project's demangler in the base system.
It does not handle some unusual mangled names, so use the inlined
libcxxabi one.
llvm-svn: 193776
Inlined a copy of cxa_demangle.cpp from:
http://llvm.org/svn/llvm-project/libcxxabi/trunk/src/cxa_demangle.cpp
For systems that don't have demangling built into the system, and for systems that don't want to use the version that is installed. Defining LLDB_USE_BUILTIN_DEMANGLER in your build system allows you to use the built in demangler. This setting is curently automatically enabled for Windows builds.
llvm-svn: 193708
Fixing a problem where ValueObject::GetPointeeData() would not accept "partial" valid reads (i.e. asking for 10 items and getting only 5 back)
While suboptimal, this situation is not a flat-out failure and could well be caused by legit scenarios, such as hitting a page boundary
Among others, this allows data formatters to print char* buffers allocated under libgmalloc
llvm-svn: 193704
One of the things that dynamic typing affects is the count of children a type has
Clear out the flag that makes us blindly believe the children count when a dynamic type change is detected
llvm-svn: 193663
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
Introduce a new boolean setting enable-auto-oneliner
This setting if set to false will force LLDB to not use the new compact one-line display
By default, one-line mode stays on, at least until we can be confident it works.
But now if it seriously impedes your workflow while it evolves/it works wonders but you still hate it, there's a way to turn it off
llvm-svn: 193450
Extend DummySyntheticProvider to actually use debug-info vended children as the source of information
Make Python synthetic children either be valid, or fallback to the dummy, like their C++ counterparts
This allows LLDB to actually stop bailing out upon encountering an invalid synthetic children provider front-end, and still displaying the non synthetized ivar info
llvm-svn: 192741
Fixed Module::ResolveSymbolContextForAddress() to be able to also look in the SymbolVendor's SymbolFile's ObjectFile for a more meaningful symbol when a symbol lookup finds a synthetic symbol from the main object file. This will help lookups on MacOSX as the main executable might be stripped, but the dSYM file always has a full symbol table.
llvm-svn: 192510
Based on the POSIX x86_64 register context. This is sufficient for opening
a mips64 (big endian) core file. Subsequent changes will connect the
disassembler, dynamic loader support, ABI, etc.
Review: http://llvm-reviews.chandlerc.com/D1873
llvm-svn: 192335
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
Constant ValueObjects should clear their description as well as their summary. Rationale being that both can depend on deeper-than-constified data
so both are subject to changes in "unpredictable" ways
To see this consider repeatedly po'ing a persistent variable of a type whose -description result changes at each invocation
llvm-svn: 192259
Formats (as in "type format") are now included in categories
The only bit missing is caching formats along with synthetic children and summaries, which might be now desirable
llvm-svn: 192217
that all clients use them explicitly. This will hopefully
prevent any future confusion where things get cast to types
we don't expect.
<rdar://problem/15146458>
llvm-svn: 191984
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
DumpValueObject() 2.0
This checkin restores pre-Xcode5 functionality to the "po" (expr -O) command:
- expr now has a new --description-verbosity (-v) argument, which takes either compact or full as a value (-v is the same as -vfull)
When the full mode is on, "po" will show the extended output with type name, persistent variable name and value, as in
(lldb) expr -O -v -- foo
(id) $0 = 0x000000010010baf0 {
1 = 2;
2 = 3;
}
When -v is omitted, or -vcompact is passed, the Xcode5-style output will be shown, as in
(lldb) expr -O -- foo
{
1 = 2;
2 = 3;
}
- for a non-ObjectiveC object, LLDB will still try to retrieve a summary and/or value to display
(lldb) po 5
5
-v also works in this mode
(lldb) expr -O -vfull -- 5
(int) $4 = 5
On top of that, this is a major refactoring of the ValueObject printing code. The functionality is now factored into a ValueObjectPrinter class for easier maintenance in the future
DumpValueObject() was turned into an instance method ValueObject::Dump() which simply calls through to the printer code, Dump_Impl has been removed
Test case to follow
llvm-svn: 191694
Now that SBValues can be setup to ignore synthetic values, this is no longer necessary, and so m_suppress_synthetic_value can go away
Another Hack Bites the Dust
llvm-svn: 191338
to build out the symbol table as addresses are used, and implements
the mechanism for ELF to add stripped symbols from eh_frame.
Uses this mechanism to allow disassembly for addresses corresponding
to stripped symbols for ELF, and provide hooks to implement this for
PE COFF.
Also removes eSymbolContextTailCall in favor of an option for
ResolveSymbolContextForAddress for consistency with the documentation
for eSymbolContextEverything. Essentially, this is just an option for
interpreting the so_addr.
llvm-svn: 191307
Specifically, allows the unwinder to handle the case where sc.function
gets resolved with a pc that is one past the address range of the function
(consistent with a tail call). However, there is no matching symbol.
Adds eSymbolContextTailCall to provide callers with control over the scope
of symbol resolution and to allow ResolveSymbolContextForAddress to handle
tail calls since this routine is common to unwind and disassembly.
llvm-svn: 191102
Targets and hosts today are little-endian (arm, x86), so this change
should be a no-op as they will not encounter the byte swapping cases.
Byte swapping will happen when cross debugging of big endian-targets
(e.g. MIPS, PPC) on a little-endian host (x86). Register- or word-
sized data copies need to be swapped, but calls to ExtractBytes or
CopyByteOrderedData that would invoke the swapping case are presumably
in error.
llvm-svn: 191005
We cannot use "GetMaxU64Bitfield" for non-power-of-two sizes, so just use
the same code that handles N > 8 for these.
Review: http://llvm-reviews.chandlerc.com/D1699
llvm-svn: 190873
for the frame is one past the address range of the calling function.
- Lowers the fix from RegisterContextLLDB for use with disassembly
- Fixes one of three issues in the disassembly test in TestInferiorAssert.py
Also adds documentation that explains the resolution depths and interface.
Note: This change affects the resolution scope for eSymbolContextFunction
without impacting the performance of eSymbolContextSymbol.
Thanks to Matt Kopec for his review.
llvm-svn: 190812
with prefer_file_cache == false. This is what we want to do when
the user is doing a disassemble command -- show the actual memory
contents in case the memory has been corrupted or something -- but
when we're profiling functions for stepping or unwinding
(ThreadPlanStepRange::GetInstructionsForAddress,
UnwindAssemblyInstEmulation::GetNonCallSiteUnwindP) we can read
__TEXT instructions directly out of the file, if it exists.
<rdar://problem/14397491>
llvm-svn: 190638
This allows the PC to be directly changed to a different line.
It's similar to the example python script in examples/python/jump.py, except implemented as a builtin.
Also this version will track the current function correctly even if the target line resolves to multiple addresses. (e.g. debugging a templated function)
llvm-svn: 190572
SVN r189964 provided a sample Python script to inspect unordered(multi){set|map} with synthetic children, contribued by Jared Grubb
This checkin converts that sample script to a C++ provider built into LLDB
A test case is also provided
llvm-svn: 190564
Since I renamed most of the LLVM Mach-O enums in r189314, I had to go fix
LLDB to use the new names. While I was here, I decided that a COFF
plugin really shouldn't be using Mach-O enums.
llvm-svn: 189316
Summary:
This merge brings in the improved 'platform' command that knows how to
interface with remote machines; that is, query OS/kernel information, push
and pull files, run shell commands, etc... and implementation for the new
communication packets that back that interface, at least on Darwin based
operating systems via the POSIXPlatform class. Linux support is coming soon.
Verified the test suite runs cleanly on Linux (x86_64), build OK on Mac OS
X Mountain Lion.
Additional improvements (not in the source SVN branch 'lldb-platform-work'):
- cmake build scripts for lldb-platform
- cleanup test suite
- documentation stub for qPlatform_RunCommand
- use log class instead of printf() directly
- reverted work-in-progress-looking changes from test/types/TestAbstract.py that work towards running the test suite remotely.
- add new logging category 'platform'
Reviewers: Matt Kopec, Greg Clayton
Review: http://llvm-reviews.chandlerc.com/D1493
llvm-svn: 189295
This means that "register read stmm0 --format f" actually works now.
This is a little messy but LLDB assumes 'long double' is portable, when it is not.
llvm-svn: 188698
address.
When loading a dSYM, and the file addresses of the dSYM Sections are
different than the executable binary Sections' file addresses, the
debug info won't be remapped to the actual load addresses correctly.
This only happens with binaries on the in-memory shared cache binaries
where their File addresses have been set to their actual load address
(outside an offset value) whereas the original executable and dSYM
have 0-based File addresses.
I think this patch will not be activated for other cases -- this is
the only case we know of where the dSYM and the executable's File
addresses differ -- but if this causes other problems we can restrict
it more carefully.
<rdar://problem/12335086>
llvm-svn: 188532
LLDB needs in memory module load level settings to control how much information is read from memory when loading in memory modules. This change adds a new setting:
(lldb) settings set target.memory-module-load-level [minimal|partial|complete]
minimal will load only sections (no symbols, or function bounds via function starts or EH frame)
partial will load sections + bounds
complete will load sections + bounds + symbols
llvm-svn: 188246
list have a shared pointer back to their DisassemblerLLVMC. This checkin force clears the InstructionList
in all the places we use the DisassemblerSP to stop the leaking for now. I'll go back and fix this
for real when I have time to do so.
<rdar://problem/14581918>
llvm-svn: 187473
ELF notes contain a 'name' field, which specifies a vendor who defines
the format of the note. Examples are 'FreeBSD' or 'GNU', or it may be
empty for generic notes.
Add a case for FreeBSD-specific notes, leaving Linux and GNU notes,
other vendor-specific notes, and generic notes to be handled by the
existing code for now.
Thanks to Samuel Jacob for reviewing and suggesting improvements.
llvm-svn: 186973
Added a setting to control timeout for kdp response packets. While I was at it, I also added a way to control the response timeout for gdb-remote packets.
KDP defaults to 5 seconds, and GDB defaults to 1 second. These were the default values that were in the code prior to adding these settings.
(lldb) settings set plugin.process.gdb-remote.packet-timeout 10
(lldb) settings set plugin.process.kdp-remote.packet-timeout 10
llvm-svn: 186360
- 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
There are two new classes:
lldb::SBModuleSpec
lldb::SBModuleSpecList
The SBModuleSpec wraps up a lldb_private::ModuleSpec, and SBModuleSpecList wraps up a lldb_private::ModuleSpecList.
llvm-svn: 185877
Both StopReadThread and the thread being stopped set the thread id to
0 after m_read_thread_enabled was set to false. If the thread being
stopped got there first then StopReadThread called pthread_join on an
invalid thread number. This is not a Good Thing,
Should fix a fairly regular segfault when quitting on Linux.
llvm-svn: 185107
Specifically, the ${target ${process ${thread and ${frame specifiers have been extended to allow a subkeyword .script:<fctName> (e.g. ${frame.script:FooFunction})
The functions are prototyped as
def FooFunction(Object,unused)
where object is of the respective SB-type (SBTarget for target.script, ... and so on)
This has not been implemented for ${var because it would be akin to a Python summary which is already well-defined in LLDB
llvm-svn: 184500
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
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
Add support for half-floats, as specified by IEEE-754-2008
With this checkin, you can now say:
(lldb) x/7hf foo
to read 7 half-floats at address foo
llvm-svn: 183716
level. Fixes a bug in "break set --source-pattern-regexp" when a shared library is
specified.
Also cleaned up the help text for --source-pattern-regexp so it is a little clearer.
<rdar://problem/14084261>
llvm-svn: 183476
Fixing an issue where formats would not propagate from parents to children in all cases
Details follow:
an SBValue has children and those are fetched along with their values
Now, one calls SBValue::SetFormat() on the parent
Technically, the format choices should propagate onto the children (see ValueObject::GetFormat())
But if the children values are already fetched, they won't notice the format change and won't update themselves
This commit fixes that by making ValueObject::GetValueAsCString() check if any format change intervened from the previous call to the current one
A test case is also added
llvm-svn: 183030
command script import now does reloads - for real
If you invoke command script import foo and it detects that foo has already been imported, it will
- invoke reload(foo) to reload the module in Python
- re-invoke foo.__lldb_init_module
This second step is necessary to ensure that LLDB does not keep cached copies of any formatter, command, ... that the module is providing
Usual caveats with Python imports persist. Among these:
- if you have objects lurking around, reloading the module won't magically update them to reflect changes
- if module A imports module B, reloading A won't reload B
These are Python-specific issues independent of LLDB that would require more extensive design work
The --allow-reload (-r) option is maintained for compatibility with existing scripts, but is clearly documented as redundant - reloading is always enabled whether you use it or not
llvm-svn: 182977
Which means "platform process list" should work and list the architecture.
We are now parsing the elf build-id if it exists, which should allow us to load stripped symbols (looking at that next).
llvm-svn: 182610
settings set use-color [false|true]
settings set prompt "${ansi.bold}${ansi.fg.green}(lldb)${ansi.normal} "
also "--no-use-colors" on the command prompt
llvm-svn: 182609
Lock the lldb_private::Module mutex while tearing down the module to make sure we don't get clients accessing the contents on a module as it is going away.
llvm-svn: 182511
Yet another implementation of the python in dSYM autoload :)
This time we are going with a ternary setting:
true - load, do not warn
false - do not load, do not warn
warn - do not load, warn (default)
llvm-svn: 182414
There are two settings:
target.load-script-from-symbol-file is a boolean that says load or no load (default: false)
target.warn-on-script-from-symbol-file is also a boolean, it says whether you want to be warned when a script file is not loaded due to security (default: true)
the auto loading on change for target.load-script-from-symbol-file is preserved
llvm-svn: 182336
This changes the setting target.load-script-from-symbol-file to be a ternary enum value:
default (the default value) will NOT load the script files but will issue a warning suggesting workarounds
yes will load the script files
no will not load the script files AND will NOT issue any warning
if you change the setting value from default to yes, that will then cause the script files to be loaded
(the assumption is you didn't know about the setting, got a warning, and quickly want to remedy it)
if you have a settings set command for this in your lldbinit file, be sure to change "true" or "false" into an appropriate "yes" or "no" value
llvm-svn: 182323
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
Provide a mechanism through which users can disable loading the Python scripts from dSYM files
This relies on a target setting: target.load-script-from-symbol-file which defaults to false ("do NOT load the script")
You need to set it to true before creating your target (or in your lldbinit file if you constantly rely on this feature) to allow the scripts to load
llvm-svn: 181709
<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
Most important was a new[] + delete mismatch in ScanFormatDescriptor()
and a couple of possible memory leaks in FileSpec::EnumerateDirectory().
llvm-svn: 181080
<rdar://problem/13723772>
Modified the lldb_private::Thread to work much better with the OperatingSystem plug-ins. Operating system plug-ins can now return have a "core" key/value pair in each thread dictionary for the OperatingSystemPython plug-ins which allows the core threads to be contained with memory threads. It also allows these memory threads to be stepped, resumed, and controlled just as if they were the actual backing threads themselves.
A few things are introduced:
- lldb_private::Thread now has a GetProtocolID() method which returns the thread protocol ID for a given thread. The protocol ID (Thread::GetProtocolID()) is usually the same as the thread id (Thread::GetID()), but it can differ when a memory thread has its own id, but is backed by an actual API thread.
- Cleaned up the Thread::WillResume() code to do the mandatory parts in Thread::ShouldResume(), and let the thread subclasses override the Thread::WillResume() which is now just a notification.
- Cleaned up ClearStackFrames() implementations so that fewer thread subclasses needed to override them
- Changed the POSIXThread class a bit since it overrode Thread::WillResume(). It is doing the wrong thing by calling "Thread::SetResumeState()" on its own, this shouldn't be done by thread subclasses, but the current code might rely on it so I left it in with a TODO comment with an explanation.
llvm-svn: 180886
UInts even if their contents were set as bytes.
This makes expressions using registers work
better, especially with core files.
<rdar://problem/13743427>
llvm-svn: 180810
Enabling LLDB to write to variables that are stored in registers
Previously, this would not work since the Value's Context loses the notion of the data being in a register
We now store an "original" context that comes out of DWARF parsing, and use that context's data when attempting a write
llvm-svn: 180803
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
variables in the ValueObject code:
- Report an error if the variable does not have
a valid address.
- Return the contents of the data to GetData(),
even if the value is constant.
<rdar://problem/13690855>
llvm-svn: 179876
lets a ValueObject's contents be set from raw
data. This has certain limitations (notably,
registers can only be set to data that is as
large as the register) but will be useful for
the new Materializer.
I also exposed this interface through SBValue.
I have added a testcase that exercises various
special cases of SBValue::SetData().
llvm-svn: 179437
This patch fixes the issue that we were using the C stack as a measure of depth of ValueObject hierarchies, in the sense that we were assuming that recursive ValueObject operations would never be deeper than the stack allows.
This assumption is easy to prove wrong, however.
For instance, after ~10k runs through this loop:
struct node
{
int value;
node* child;
node (int x)
{
value = x;
child = nullptr;
}
};
int main ()
{
node root(1);
node* ptr = &root;
int j = 2;
while (1)
{
ptr->child = new node(j++);
ptr = ptr->child;
}
return 0;
}
the deepmost child object will be deeper than the stack on most architectures, and we would be unable to display it
This checkin fixes the issue by introducing a notion of root of ValueObject hierarchies.
In a couple cases, we have to use an iterative algorithm instead of going to the root because we want to allow deeper customizations (e.g. formats, dynamic values).
While the patch passes our test suite without regressions, it is a good idea to keep eyes open for any unexpected behavior (recursion can be subtle..)
Also, I am hesitant to introduce a test case since failing at this will not just be marked as an "F", but most definitely crash LLDB.
llvm-svn: 179330
platform.plugin.darwin-kernel.kext-directories
platform.plugin.darwin-kernel.search-locally-for-kexts
and fix a few FileSpec handling issues for the kext-directories setting.
llvm-svn: 178920
Add two initial settings for the PlatformDarwinKernel plugin,
plugin.platform.darwin-kernel.search-locally-for-kexts [true|false]
plugin.platform.darwin-kernel.kext-directories [directory list]
llvm-svn: 178846
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
PC relative loads are missing disassembly comments when disassembled in a live process.
This issue was because some sections, like __TEXT and __DATA in libobjc.A.dylib, were being moved when they were put into the dyld shared cache. This could also affect any other system that slides sections individually.
The solution is to keep track of wether the bytes we will disassemble are from an executable file (file address), or from a live process (load address). We now do the right thing based off of this input in all cases.
llvm-svn: 178315
- Includes a stub for AVX support in the x86-64 register context and a failing test for register sets that are unavailable.
Thanks to Greg Clayton for his review feedback.
llvm-svn: 178252
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
ValueObjects themselves use DumpValueObjectOptions as the currency for the same purpose
The code to convert between these two units was replicated (to varying degrees of correctness) in several spots in the code
This checkin provides one and only one (and hopefully correct :-) entry point for this conversion
llvm-svn: 178044
Functions in "(anonymous namespace)" was causing LLDB to crash when trying to complete a type and it would also cause functions arguments to appear in wrong place in frame display when showing function arguments.
llvm-svn: 177965
Ensure that option -Y also works for expression as it does for frame variable
Also, if the user passes an explicit format specifier when printing a variable, override the summary's decision to hide the value.
This is required for scenarios like this to work:
(lldb) p/x c
(Class) $0 = 0x0000000100adb7f8 NSObject
Previously this would say:
(lldb) p/x c
(Class) $0 = NSObject
ignoring the explicit format specifier
llvm-svn: 177893
commands of the form
frame variable -f c-string foo
where foo is an arbitrary pointer (e.g. void*) now do the right thing, i.e. they deref the pointer and try to get a c-string at the pointed address instead of dumping the pointer bytes as a string. the old behavior is used as a fallback if things don’t go well
llvm-svn: 177799
Fixed a crasher in the SourceManager where it wasn't checking the m_target member variable for NULL.
In doing this fix, I hardened this class to have weak pointers to the debugger and target in case they do go away. I also changed SBSourceManager to hold onto weak pointers to the debugger and target so they don't keep objects alive by holding a strong reference to them.
llvm-svn: 177365
lldb remembers not-found source file, setting target.source-map doesn't make it re-check for it. Now this is fixed. Each time the source path remappings get updated, the modification ID in the PathMappingList gets bumped and then we know the re-check for sources.
llvm-svn: 177125
Made the "--reverse" option to "source list" also be able to use the "--count". This helps us implement support for regexp source list command:
(lldb) l -10
Which gets turned into:
(lldb) source list --reverse --count 10
Also simplified the code that is used to track showing more source from the last file and line.
llvm-svn: 176961
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
Calculate "can branch" using the MC API's rather than our hand-rolled regex'es.
As extra credit, allow setting the disassembly flavor for x86 based architectures to intel or att.
<rdar://problem/11319574>
<rdar://problem/9329275>
llvm-svn: 176392
in the Process destructor. Doing it there can be too late depending on what the internal state
and ProcessGDBRemote Async threads are doing.
<rdar://problem/13297536>
llvm-svn: 176203
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
Adding data formatters for iterators for std::map and std::vector (both libc++ and libstdcpp)
This does not include reverse iterators since they are both trickier (due to requirements the standard imposes on them) and much less useful
llvm-svn: 175787
Fixed a crasher when the ConnectionFileDescriptor was used in a process with over FD_SETSIZE (1024) files open. It would corrupt the stack and cause the stack checker to assert and kill the program.
The final fix was to "#define _DARWIN_UNLIMITED_SELECT" at the top of the one and only file that uses select () in the LLDB codebase and then make an array of "fd_set" objects so they can handle more than 1024 file descriptors. The new code can handle as many file descriptors as a process can create.
llvm-svn: 175378
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
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
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
Providing a compact display mode for "po" to use where the convenience variable name and the pointer value are both hidden.
This is for convenience when dealing with ObjC instances where the description often gets it right and the debugger-provided information is not useful to most people.
If you need either of these, "expr" will still show them.
llvm-svn: 173748
Data formatters now cache themselves.
This commit provides a new formatter cache mechanism. Upon resolving a formatter (summary or synthetic), LLDB remembers the resolution for later faster retrieval.
Also moved the data formatters subsystem from the core to its own group and folder for easier management, and done some code reorganization.
The ObjC runtime v1 now returns a class name if asked for the dynamic type of an object. This is required for formatters caching to work with the v1 runtime.
Lastly, this commit disposes of the old hack where ValueObjects had to remember whether they were queried for formatters with their static or dynamic type.
Now the ValueObjectDynamicValue class works well enough that we can use its dynamic value setting for the same purpose.
llvm-svn: 173728
Add the ability to give breakpoints a "kind" string, and have the StopInfoBreakpoint
print that in the brief description if set. Also print the kind - if set - in the breakpoint
listing.
Give kinds to a bunch of the internal breakpoints.
We were deleting the Mac OS X dynamic loader breakpoint as though the id we had stored away was
a breakpoint site ID, but in fact it was a breakpoint id, so we never actually deleted it. Fixed that.
llvm-svn: 173555
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
vector<bool> is specialized and the existing general summary for vectors would lie to the user.
Tackling libstdc++ and synthetic children is the following, less critical, part of this task
llvm-svn: 172671
AddCXXSummary in FormatManager::LoadSystemFormatters();
that function pulls in code that assumes python; can't
be used without the ifdef.
llvm-svn: 172300
Providing a data formatter for libc++ std::wstring
In the process, refactoring the std::string data formatter to be written in C++ so that commonalities between the two can be exploited
Also, providing a new API on the ValueObject to navigate a hierarchy by index-path
Lastly, an appropriate test case is included
llvm-svn: 172282
Fixed an issue with the auto loading of script resources in debug info files. Any platform can add support for this, and on MacOSX we allow dSYM files to contain python modules that get automatically loaded when a dSYM file is associated with an executable or shared library.
The modifications will now:
- Let the module locate the symbol file naturally instead of using a function that only works in certain cases. This helps us to locate the script resources as long as the dSYM file can be found.
- Don't try and do any of this if the script interpreter has scripting disabled.
- Allow more than one scripting resource to be found in a symbol file by returning the list
- Load the scripting resources when a symbol file is added via the "target symbols add" command.
- Be smarter about matching the dSYM mach-o file to an existing executable in the target images by stripping extensions on the symfile basname if needed.
llvm-svn: 172275
Fixed an issue where the platform auto select code was changing the architecture and causing the wrong architecture to be assigned to the target.
llvm-svn: 172251
Also modified the Value class so that you can evaluate expressions without a process, yet with some sections loaded in the target. This allows casting pointers that are in data sections to types and being able to evaluate expressions in the data. For example:
(lldb) target create a.out
(lldb) target modules load --file a.out --slide 0
... find address of something in data ...
(lldb) script
expr_opts = lldb.SBExpressionOptions()
v = lldb.target.EvaluateExpression('(foo *)0x1230000', expr_opts)
print v
vv = lldb.value(v)
print v.pt.x
Above we were able to cast a pointer to an address which was in a.out's data
section and print out entire structures and navigate to the child ivars of the expression.
llvm-svn: 172227
Providing data formatters for char16_t* and char32_t* C++11-style Unicode strings
Using this chance to refactor the UTF data reader used for data formatters for added generality
Added a relevant test case
llvm-svn: 172119
last source point listed.
Also fix the setting of the default file & line to the file containing main, when you do a plain "list".
<rdar://problem/12685226>
llvm-svn: 171945
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
Have the disassembler's Instruction::Dump always insert at least
one space character between an opcode and its arguments, don't let
a long opcode name abut the arguments.
llvm-svn: 171561
Supporting a compact display syntax for ObjC pointers where 0x00.....0 is replaced by a much more legible "nil"
e.g. this would show:
(NSArray *) $2 = nil
instead of:
(NSArray *) $2 = 0x0000000000000000 <nil>
llvm-svn: 170161
equality can be strict or loose and we want code to
explicitly choose one or the other.
Also renamed the Compare function to IsEqualTo, to
avoid confusion.
<rdar://problem/12856749>
llvm-svn: 170152
When displaying function.name-with-args format will now print "varname=<unavailable>" instead of omitting argument names and values when there is an error reading the value.
llvm-svn: 169781
Making MightHaveChildren() always return true regardless for our own data formatters
This is meant to optimize performance for common most-often-not-empty container classes
llvm-svn: 169759
Change the wording of NSNumber summary from absurd value to unexpected value when a tagged pointer shows up that does not match our knowledge of the internals
llvm-svn: 169751
- remove unused members
- add NO_PEDANTIC to selected Makefiles
- fix return values (removed NULL as needed)
- disable warning about four-char-constants
- remove unneeded const from operator*() declaration
- add missing lambda function return types
- fix printf() with no format string
- change sizeof to use a type name instead of variable name
- fix Linux ProcessMonitor.cpp to be 32/64 bit friendly
- disable warnings emitted by swig-generated C++ code
Patch by Matt Kopec!
llvm-svn: 169645
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
type of an Objective-C selector, don't bother
making the expression parser resolve it all over
again. Just send the message straight to the
object pointer as if it were an id, and cast the
result.
<rdar://problem/12799087>
llvm-svn: 169300
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
Make sure that ValueObjectDynamicValue clears itself when no dynamic type information can be found
This behavior was supposed to be already happening (as per the comment lines)
llvm-svn: 168743
This commit does three things:
(a) introduces a new notification model for adding/removing/changing modules to a ModuleList, and applies it to the Target's ModuleList, so that we make sure to always trigger the right set of actions
whenever modules come and go in a target. Certain spots in the code still need to "manually" notify the Target for several reasons, so this is a work in progress
(b) adds a new capability to the Platforms: locating a scripting resources associated to a module. A scripting resource is a Python file that can load commands, formatters, ... and any other action
of interest corresponding to the loading of a module. At the moment, this is only implemented on Mac OS X and only for files inside .dSYM bundles - the next step is going to be letting
the frameworks themselves hold their scripting resources. Implementors of platforms for other systems are free to implement "the right thing" for their own worlds
(c) hooking up items (a) and (b) so that targets auto-load the scripting resources as the corresponding modules get loaded in a target. This has a few caveats at the moment:
- the user needs to manually add the .py file to the dSYM (soon, it will also work in the framework itself)
- if two modules with the same name show up during the lifetime of an LLDB session, the second one won't be able to load its scripting resource, but will otherwise work just fine
llvm-svn: 167569
The operator== method is a synonym for IsExactMatch().
The essential difference between these two is that IsCompatibleMatch()
will say that armv7 and armv7s are compatible and return true.
IsExactMatch() will say that armv7 and armv7s are not a match.
An armv7s cpu can run either generic armv7 binaries or armv7s binaries
(the latter being tuned for it). When we're picking the slice of a
universal Mach-O file to load in an armv7s Target, we need to be able to
first look for an exact cpu subtype match (armv7s == armv7s) and failing
that, looking for a slice with a compatible architecture.
Update ObjectContainerUniversalMachO::GetObjectFile to prefer an exact
match of the cpu type, falling back to a compatible match if necessary.
<rdar://problem/12593515>
llvm-svn: 167365
statements instead of a switch for the size of the floating
point types; some architectures sizeof double and sizeof long
double are the same and that's invalid in a switch.
Fix the LLDB_DISABLE_PYTHON ifdef block in FormatManager::LoadObjCFormatters
so it builds on arm again.
llvm-svn: 167263
The attached patch adds eValueTypeVector to lldb_private::Value. The nested struct Vector is patterned after RegisterValue::m_data.buffer. This change to Value allows ClangExpressionDeclMap::LookupDecl to return vector register data for consumption by InterpreterStackFrame::ResolveValue. Note that ResolveValue was tweaked slightly to allocate enough memory for vector registers.
An immediate result of this patch is that "expr $xmm0" generates the same results on Linux as on the Mac, which is good enough for TestRegisters.py. In addition, the log of m_memory.PrintData(data_region.m_base, data_region.m_extent) shows that the register content has been resolved successfully. On the other hand, the output is glaringly empty:
runCmd: expr $xmm0
output: (unsigned char __attribute__((ext_vector_type(16)))) $0 = {}
Expecting sub string: vector_type
Matched
llvm-svn: 167033
This should delay initialization of Python until strictly necessary and speed-up debugger startup
Also, convert formatters for SEL and BOOL ObjC data-types from Python to C++, in order to reap more performance benefits from the above changes
llvm-svn: 166967
Removing the IsDynamic() and GetStaticValue() calls, so that they will default to the base class behavior:
- non-dynamic
- itself as the static value
This is in contrast with the previous behavior which could be confusing and could potentially cause issues when using those objects
llvm-svn: 166857
Adding the new has_children (or MightHaveChildren() in C++) for the existing synthetic children providers
In a few cases, the new call is going to be much more efficient than the previous num_children > 0 check
When the optimization was marginal (e.g. std::vector<>), the choice was to use num_children in order to keep
implementation details in one function instead of duplicating code
Next step is to provide test cases
llvm-svn: 166506
This commit enables the new HasChildren() feature for synthetic children providers
Namely, it hooks up the required bits and pieces so that individual synthetic children providers can implement a new (optional) has_children call
Default implementations have been provided where necessary so that any existing providers continue to work and behave correctly
Next steps are:
2) writing smart implementations of has_children for our providers whenever possible
3) make a test case
llvm-svn: 166495
Added a new API call to help efficiently determine if a SBValue could have children:
bool
SBValue::MightHaveChildren ();
This is inteneded to be used bui GUI programs that need to show if a SBValue needs a disclosure triangle when displaying a hierarchical type in a tree view without having to complete the type (by calling SBValue::GetNumChildren()) as completing the type is expensive.
llvm-svn: 166460
Given our implementation of ValueObjects we could have a scenario where a ValueObject has a dynamic type of Foo* at one point, and then its dynamic type changes to Bar*
If Bar* has synthetic children enabled, by the time we figure that out, our public API is already vending SBValues wrapping a DynamicVO, instead of a SyntheticVO and there was
no trivial way for us to change the SP inside an SBValue on the fly
This checkin reimplements SBValue in terms of a wrapper, ValueImpl, that allows this substitutions on-the-fly by overriding GetSP() to do The Right Thing (TM)
As an additional bonus, GetNonSyntheticValue() now works, and we can get rid of the ForceDisableSyntheticChildren idiom in ScriptInterpreterPython
Lastly, this checkin makes sure the synthetic VOs get the correct m_value and m_data from their parents (prevented summaries from working in some cases)
llvm-svn: 166426
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
plugin
dynamic-loader
macosx-kernel
(bool) disable-kext-loading
To settings can be set using:
(lldb) settings set plugin.dynamic-loader.macosx-kernel.disable-kext-loading true
I currently only hooked up the DynamicLoader plug-ins, but the code is very easy to duplicate when and if we need settings for other plug-ins.
llvm-svn: 166294
to work properly; when doing bare-boards rom debugging
force the OS to be one of those when initializing llvm.
<rdar://problem/12504138>
llvm-svn: 166057
Dynamic type code must be efficient and fast. Now it is.
Added ObjC v1 support for getting the complete list of ISA values.
The main flow of the AppleObjCRuntime subclasses is now they must override "virtual bool UpdateISAToDescriptorMap_Impl();". This function will update the complete list of ISA values and create ClassDescriptorSP objects for each one. Now we have the complete list of valid ISA values which we can use for verification when doing dynamic typing.
Refactored a bunch of stuff so that the AppleObjCRuntime subclasses don't have to implement as many functions as they used to.
llvm-svn: 165730
Also added a new option for "log enable" which is "--stack" which will print out a stack backtrace for each log line.
This was used to track down the leaking module issue I fixed last week.
llvm-svn: 165438
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
Shared libraries on MacOSX were not properly being removed from the shared
module list when re-running a debug session due to an error in:
Module::MatchesModuleSpec()
llvm-svn: 164991
This checkin adds the capability for LLDB to load plugins from external dylibs that can provide new commands
It exports an SBCommand class from the public API layer, and a new SBCommandPluginInterface
There is a minimal load-only plugin manager built into the debugger, which can be accessed via Debugger::LoadPlugin.
Plugins are loaded from two locations at debugger startup (LLDB.framework/Resources/PlugIns and ~/Library/Application Support/LLDB/PlugIns) and more can be (re)loaded via the "plugin load" command
For an example of how to make a plugin, refer to the fooplugin.cpp file in examples/plugins/commands
Caveats:
Currently, the new API objects and features are not exposed via Python.
The new commands can only be "parsed" (i.e. not raw) and get their command line via a char** parameter (we do not expose our internal Args object)
There is no unloading feature, which can potentially lead to leaks if you overwrite the commands by reloading the same or different plugins
There is no API exposed for option parsing, which means you may need to use getopt or roll-your-own
llvm-svn: 164865
We can now do:
Specify a path to a debug symbols file:
(lldb) add-dsym <path-to-dsym>
Go and download the dSYM file for the "libunc.dylib" module in your target:
(lldb) add-dsym --shlib libunc.dylib
Go and download the dSYM given a UUID:
(lldb) add-dsym --uuid <UUID>
Go and download the dSYM file for the current frame:
(lldb) add-dsym --frame
llvm-svn: 164806
Added the ability for OptionValueString objects to take flags. The only flag is currently for parsing escape sequences. Not the prompt string can have escape characters translate which will allow colors in the prompt.
Added functions to Args that will parse the escape sequences in a string, and also re-encode the escape sequences for display. This was looted from other parts of LLDB (the Debugger::FormatString() function).
llvm-svn: 163043
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
Fix a subtle ArchSpec::cores_match() logic issue which prevents the add-dsym command
to add a debug symbol file to one of the target's current modules.
llvm-svn: 162802
Added a new "interpreter" properties to encapsulate any properties for the command interpreter. Right now this contains only "expand-regex-aliases", so you can now enable (disabled by default) the echoing of the command that a regular expression alias expands to:
(lldb) b main
Breakpoint created: 1: name = 'main', locations = 1
Note that the expanded regular expression command wasn't shown by default. You can enable it if you want to:
(lldb) settings set interpreter.expand-regex-aliases true
(lldb) b main
breakpoint set --name 'main'
Breakpoint created: 1: name = 'main', locations = 1
Also enabled auto completion for enumeration option values (OptionValueEnumeration) and for boolean option values (OptionValueBoolean).
Fixed auto completion for settings names when nothing has been type (it should show all settings).
llvm-svn: 162418
- 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
the expression returns nothing. There is now a
setting, "notify-void." When the user enables
that setting, lldb prints (void) if an expression's
result is void. Otherwise, lldb is silent.
<rdar://problem/11225150>
llvm-svn: 161600
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
Use the built in demangler for Apple builds for now which has needed demangling fixes, and make the cxa_demangle.cpp use rtti in the Xcode project settings as it requires it be enabled.
llvm-svn: 161323
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
Improved the error message when we can find a function in the current program by printing the demangled name.
Also added the ability to create lldb_private::Mangled instances with a ConstString when we already have a ConstString for a mangled or demangled name. Also added the ability to call SetValue with a ConstString and also without a boolean to indicate if the string is mangled where we will now auto-detect if the string is mangled.
llvm-svn: 160450
Fixed a case where the python interpreter could end up holding onto a previous lldb::SBProcess (probably in lldb.process) when run under Xcode. Prior to this fix, the lldb::SBProcess held onto a shared pointer to a lldb_private::Process. This in turn could cause the process to still have a thread list with stack frames. The stack frames would have module shared pointers in the lldb_private::SymbolContext objects.
We also had issues with things staying in the shared module list too long when we found things by UUID (we didn't remove the out of date ModuleSP from the global module cache).
Now all of this is fixed and everything goes away between runs.
llvm-svn: 160140
% cat sp.cpp
#include <tr1/memory>
class A
{
public:
A (): m_i (12) {}
virtual ~A() {}
private:
int m_i;
};
int main (int argc, char const *argv[], char const *envp[])
{
A *a_pointers[2] = { NULL, NULL };
A a1;
A a2;
a_pointers[0] = &a1;
a_pointers[1] = &a2;
return 0;
}
And you stop at the "return 0", you can now read memory using the "address" format and see:
(lldb) memory read --format address `&a_pointers`
0x7fff5fbff870: 0x00007fff5fbff860 -> 0x00000001000010b0 vtable for A + 16
0x7fff5fbff878: 0x00007fff5fbff850 -> 0x00000001000010b0 vtable for A + 16
0x7fff5fbff880: 0x00007fff5fbff8d0
0x7fff5fbff888: 0x00007fff5fbff8c0
0x7fff5fbff890: 0x0000000000000001
0x7fff5fbff898: 0x36d54c275add2294
0x7fff5fbff8a0: 0x00007fff5fbff8b0
0x7fff5fbff8a8: 0x0000000100000bb4 a.out`start + 52
Note the extra dereference that was applied to 0x00007fff5fbff860 and 0x00007fff5fbff850 so we can see that these are "A" classes.
llvm-svn: 160085
Modified the heap.py to be able to correctly indentify the exact ivar for the "ptr_refs" command no matter how deep the ivar is in a class hierarchy. Also fixed the ability for the heap command to symbolicate the stack backtrace when MallocStackLogging is set in the environment and the "--stack" option was specified.
llvm-svn: 159883
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
The output of 'register read' should be prettier.
Modify RegisterValue::Dump() to take an additional parameter:
uint32_t reg_name_right_align_at
which defaults to 0 (i.e., no alignment at all). Update the 'register read' command impl to pass 8
as the alignment to RegisterValue::Dump() method. If more sophisticated scheme is desired, we will
need to introduce an additional command option to 'register read' later on.
llvm-svn: 158039
Fix confusing error message about "expression did not evaluate to an address" when doing 'watchpoint set expression".
Instead of using 0 as the fail_value when invoking ValueObject::GetValueAsUnsigned(), modify the API to take an addition
bool pointer (defaults to NULL) to indicate success/failure of value conversion.
llvm-svn: 158016
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
A local std::string was being filled in and then the function would return "s.c_str()".
A local StreamString (which contains a std::string) was being filled in, and essentially also returning the c string from the std::string, though it was in a the StreamString class.
The fix was to not do this by passing a stream object into StringList::Join() and fix the "arch_helper()" function to do what it should: cache the result in a global.
llvm-svn: 157519
Make 'help arch' return the list of supported architectures.
Add a convenience method StringList::Join(const char *separator) which is called from the help function for 'arch'.
Also add a simple test case.
llvm-svn: 157507
path on rerunning, evict the old module from the target module list, inform the breakpoints
about this so they can do something intelligent as well.
rdar://problem/11273043
llvm-svn: 157008
Make sure our debugger STDIN read thread shuts down quickly when we are done with it. We had a case where the owner of the file handle was not closing it and caused spins.
llvm-svn: 156879
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
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
Just call SBDebugger::SetTerminalWidth on the driver's SBDebugger, which does the same job, but no locks.
Also add the value checking to SetTerminalWidth you get with SetInternalVariable(..., "term-width", ...).
rdar://problem/11310563
llvm-svn: 155665
This worked correctly for 64-bit targets, but broke down data formatters in i386 mode. The formatters would try to read pointers out of the frozen-dried objects,
but were unable to do so because they would try fetching 8 bytes from a DataExtractor with only 4 bytes in it. This patch fixes the issue by always making the pointer-size
for a DataExtractor match the target setting.
llvm-svn: 155418
Fixed an issue where we get NULL compile units back from the symbol vendor. We need symbol vendors to be able to quickly give an estimate of the compile units that they have without having to fully vette them first, so anyone getting compile units from a module should be able to deal with a NULL compile unit being returned for a given index.
llvm-svn: 155398
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
Work around a deadlocking issue where "SBDebugger::MemoryPressureDetected ()" is being called and is causing a deadlock. We now just try and get the lock when trying to trim down the unique modules so we don't deadlock debugger GUI programs until we can find the root cause.
llvm-svn: 154339
ValueObject, and make sure that ValueObjects that
have null type names (because they have null types)
also have null qualified type names. This avoids
some potential crashes if
ValueObject::GetQualifiedTypeName tries to get the
name of their type by calling GetClangTypeImpl().
llvm-svn: 153718
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
A new setting enable-synthetic-value is provided on the target to disable this behavior.
There also is a new GetNonSyntheticValue() API call on SBValue to go back from synthetic to non-synthetic. There is no call to go from non-synthetic to synthetic.
The test suite has been changed accordingly.
Fallout from changes to type searching: an hack has to be played to make it possible to use maps that contain std::string due to the special name replacement operated by clang
Fixing a test case that was using libstdcpp instead of libc++ - caught as a consequence of said changes to type searching
llvm-svn: 153495
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
This is the feature that allowed the user to have things like:
class Base { ... };
class Derived : public Base { ... };
and have formatters defined for Base work automatically for Derived.
This feature turned out to be too expensive since it requires completing types.
This patch takes care of removing cascading (other than typedefs chain cascading), updating the test suite accordingly, and adding required Cocoa class names to keep the AppKit formatters working
llvm-svn: 153272
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
Changes to synthetic children:
- the update(self): function can now (optionally) return a value - if it returns boolean value True, ValueObjectSyntheticFilter will not clear its caches across stop-points
this should allow better performance for Python-based synthetic children when one can be sure that the child ValueObjects have not changed
- making a difference between a synthetic VO and a VO with a synthetic value: now a ValueObjectSyntheticFilter will not return itself as its own synthetic value, but will (correctly)
claim to itself be synthetic
- cleared up the internal synthetic children architecture to make a more consistent use of pointers and references instead of shared pointers when possible
- major cleanup of unnecessary #include, data and functions in ValueObjectSyntheticFilter itself
- removed the SyntheticValueType enum and replaced it with a plain boolean (to which it was equivalent in the first place)
Some clean ups to the summary generation code
Centralized the code that clears out user-visible strings and data in ValueObject
More efficient summaries for libc++ containers
llvm-svn: 153061
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
This has been done for those summaries where the difference is only cosmetic (e.g. naming things as items instead of values, ...)
The LLDB output style has been preserved when it provides more information (e.g. telling the type as well as the value of an NSNumber)
Test cases have been updated to reflect the updated output style where necessary
llvm-svn: 152592
std::string has a summary provider
std::vector std::list and std::map have both a summary and a synthetic children provider
Given the usage of a custom namespace (std::__1::classname) for the implementation of libc++, we keep both libstdcpp and libc++ formatters enabled at the same time since that raises no conflicts and enabled for seamless transition between the two
The formatters for libc++ reside in a libcxx category, and are loaded from libcxx.py (to be found in examples/synthetic)
The formatters-stl test cases have been divided to be separate for libcxx and libstdcpp. This separation is necessary because
(a) we need different compiler flags for libc++ than for libstdcpp
(b) libc++ inlines a lot more than libstdcpp and some code changes were required to accommodate this difference
llvm-svn: 152570
Don't show variable values in Xcode when they are out of scope. This allows Xcode to step a lot faster when there are many variables in the variables view.
llvm-svn: 152380
This solves an issue where a ValueObject was getting a wrong children count (usually, a huge value) and trying to resize the vector of children to fit that many ValueObject*
Added a loop detection algorithm to the synthetic children provider for std::list
Added a few more checks to the synthetic children provider for std::vector
Both std::list and std::vector's synthetic children providers now cache the count of children instead of recomputing it every time
std::map has a field that stores the count, so there is little need to cache it on our side
llvm-svn: 152371
shouldn't compile any of the XPC support code.
Update macosx/Host.mm to use that define.
Add a LLDB_DISABLE_PYTHON ifdef block around a new function in
Core/FormatManager.cpp.
<rdar://problem/10942125>
llvm-svn: 152293
This fix really needed to happen as a previous fix I had submitted for
calculating symbol sizes made many symbols appear to have zero size since
the function that was calculating the symbol size was calling another function
that would cause the calculation to happen again. This resulted in some symbols
having zero size when they shouldn't. This could then cause infinite stack
traces and many other side affects.
llvm-svn: 152244
expression command doesn't handle xmm or stmm registers...
o Update ClangASTContext::GetBuiltinTypeForEncodingAndBitSize() to now handle eEncodingVector.
o Modify RegisterValue::SetFromMemoryData() to fix the subtle error due to unitialized variables.
o Add a test file for "expr $xmm0".
llvm-svn: 152190
Several places in the ScriptInterpreter interface used StringList objects where an std::string would suffice - Fixed
Refactoring calls that generated special-purposes functions in the Python interpreter to use helper functions instead of duplicating blobs of code
llvm-svn: 152164
This was done in SBTarget:
lldb::SBInstructionList
lldb::SBTarget::ReadInstructions (lldb::SBAddress base_addr, uint32_t count);
Also cleaned up a few files in the LLDB.framework settings.
llvm-svn: 152152
fixed a few potential NULL-pointer derefs in ValueObject
we have a way to provide docstrings for properties we add to the SWIG layer - a few of these properties have a docstring already, more will come in future commits
added a new bunch of properties to SBData to make it more natural and Python-like to access the data they contain
llvm-svn: 151962
(b) fixes and improvements to the formatters for NSDate and NSString
(c) adding an introspection formatter for NSCountedSet
(d) making the Objective-C formatters test cases pass on both 64 and 32 bit
one of the test cases is marked as expected failure on i386 - support needs to be added to the LLDB core for it to pass
llvm-svn: 151826
2) providing an updated list of tagged pointers values for the objc_runtime module - hopefully this one is final
3) changing ValueObject::DumpValueObject to use an Options class instead of providing a bulky list of parameters to pass around
this change had been laid out previously, but some clients of DumpValueObject() were still using the old prototype and some arguments
were treated in a special way and passed in directly instead of through the Options class
4) providing new GetSummaryAsCString() and GetValueAsCString() calls in ValueObject that are passed a formatter object and a destination string
and fill the string by formatting themselves using the formatter argument instead of the default for the current ValueObject
5) removing the option to have formats and summaries stick to a variable for the current stoppoint
after some debate, we are going with non-sticky: if you say frame variable --format hex foo, the hex format will only be applied to the current command execution and not stick when redisplaying foo
the other option would be full stickiness, which means that foo would be formatted as hex for its whole lifetime
we are open to suggestions on what feels "natural" in this regard
llvm-svn: 151801
allocations by section. We install these sections
in the target process and inform the JIT of their
new locations.
Also removed some unused variable warnings.
llvm-svn: 151789
a) adds a Python summary provider for NSDate
b) changes the initialization for ScriptInterpreter so that we are not passing a bulk of Python-specific function pointers around
c) provides a new ScriptInterpreterObject class that allows for ref-count safe wrapping of scripting objects on the C++ side
d) contains much needed performance improvements:
1) the pointer to the Python function generating a scripted summary is now cached instead of looked up every time
2) redundant memory reads in the Python ObjC runtime wrapper are eliminated
3) summaries now use the m_summary_str in ValueObject to store their data instead of passing around ( == copying) an std::string object
e) contains other minor fixes, such as adding descriptive error messages for some cases of summary generation failure
llvm-svn: 151703
Fixed an error where if we tried to format a ValueObject using a format
that was incorrect for a variable, then it would set ValueObject::m_error
to an error state and stop the value from being able to be updated. We now
leave m_error alone and only let the update value code change that. Any errors
in formatting will return a valid value as C string that contains an error
string. This lets us then modify the format and redisplay without any issues.
llvm-svn: 151581
more of the local path, platform path, associated symbol file, UUID, arch,
object name and object offset. This allows many of the calls that were
GetSharedModule to reduce the number of arguments that were used in a call
to these functions. It also allows a module to be created with a ModuleSpec
which allows many things to be specified prior to any accessors being called
on the Module class itself.
I was running into problems when adding support for "target symbol add"
where you can specify a stand alone debug info file after debugging has started
where I needed to specify the associated symbol file path and if I waited until
after construction, the wrong symbol file had already been located. By using
the ModuleSpec it allows us to construct a module with as little or as much
information as needed and not have to change the parameter list.
llvm-svn: 151476
weak reference back to the Module. We were crashing when trying to make a
memory object file since it was trying to get the object in the Module
constructor before the "Module *" had been put into a shared pointer, and the
module was trying to initialize a weak pointer back to it.
llvm-svn: 151397
I started work on being able to add symbol files after a debug session
had started with a new "target symfile add" command and quickly ran into
problems with stale Address objects in breakpoint locations that had
lldb_private::Section pointers into modules that had been removed or
replaced. This also let to grabbing stale modules from those sections.
So I needed to thread harded the Address, Section and related objects.
To do this I modified the ModuleChild class to now require a ModuleSP
on initialization so that a weak reference can created. I also changed
all places that were handing out "Section *" to have them hand out SectionSP.
All ObjectFile, SymbolFile and SymbolVendors were inheriting from ModuleChild
so all of the find plug-in, static creation function and constructors now
require ModuleSP references instead of Module *.
Address objects now have weak references to their sections which can
safely go stale when a module gets destructed.
This checkin doesn't complete the "target symfile add" command, but it
does get us a lot clioser to being able to do such things without a high
risk of crashing or memory corruption.
llvm-svn: 151336
The formatter for NSString is an improved version of the one previously shipped as an example, the others are new in design and implementation.
A more robust and OO-compliant Objective-C runtime wrapper is provided for runtime versions 1 and 2 on 32 and 64 bit.
The formatters are contained in a category named "AppKit", which is not enabled at startup.
llvm-svn: 151299
Objective-C classes. This allows LLDB to find
ivars declared in class extensions in modules other
than where the debugger is currently stopped (we
already supported this when the debugger was
stopped in the same module as the definition).
This involved the following main changes:
- The ObjCLanguageRuntime now knows how to hunt
for the authoritative version of an Objective-C
type. It looks for the symbol indicating a
definition, and then gets the type from the
module containing that symbol.
- ValueObjects now report their type with a
potential override, and the override is set if
the type of the ValueObject is an Objective-C
class or pointer type that is defined somewhere
other than the original reported type. This
means that "frame variable" will always use the
complete type if one is available.
- The ClangASTSource now looks for the complete
type when looking for ivars. This means that
"expr" will always use the complete type if one
is available.
- I added a testcase that verifies that both
"frame variable" and "expr" work.
llvm-svn: 151214
is not available (LLDB_DISABLE_PYTHON is defined).
Change build-swig-Python.sh to emit an empty LLDBPythonWrap.cpp file if
this build is LLDB_DISABLE_PYTHON.
Change the "Copy to Xcode.app" shell script phase in the lldb.xcodeproj
to only do this copying for Mac native builds.
llvm-svn: 151035
objects for the backlink to the lldb_private::Process. The issues we were
running into before was someone was holding onto a shared pointer to a
lldb_private::Thread for too long, and the lldb_private::Process parent object
would get destroyed and the lldb_private::Thread had a "Process &m_process"
member which would just treat whatever memory that used to be a Process as a
valid Process. This was mostly happening for lldb_private::StackFrame objects
that had a member like "Thread &m_thread". So this completes the internal
strong/weak changes.
Documented the ExecutionContext and ExecutionContextRef classes so that our
LLDB developers can understand when and where to use ExecutionContext and
ExecutionContextRef objects.
llvm-svn: 151009
the lldb_private::StackFrame objects hold onto a weak pointer to the thread
object. The lldb_private::StackFrame objects the the most volatile objects
we have as when we are doing single stepping, frames can often get lost or
thrown away, only to be re-created as another object that still refers to the
same frame. We have another bug tracking that. But we need to be able to
have frames no longer be able to get the thread when they are not part of
a thread anymore, and this is the first step (this fix makes that possible
but doesn't implement it yet).
Also changed lldb_private::ExecutionContextScope to return shared pointers to
all objects in the execution context to further thread harden the internals.
llvm-svn: 150871
internals. The first part of this is to use a new class:
lldb_private::ExecutionContextRef
This class holds onto weak pointers to the target, process, thread and frame
and it also contains the thread ID and frame Stack ID in case the thread and
frame objects go away and come back as new objects that represent the same
logical thread/frame.
ExecutionContextRef objcets have accessors to access shared pointers for
the target, process, thread and frame which might return NULL if the backing
object is no longer available. This allows for references to persistent program
state without needing to hold a shared pointer to each object and potentially
keeping that object around for longer than it needs to be.
You can also "Lock" and ExecutionContextRef (which contains weak pointers)
object into an ExecutionContext (which contains strong, or shared pointers)
with code like
ExecutionContext exe_ctx (my_obj->GetExectionContextRef().Lock());
llvm-svn: 150801
Adding new API calls to SBValue to be able to retrieve the associated formatters
Some refactoring to FormatNavigator::Get() in order to shrink its size down to more manageable terms (a future, massive, refactoring effort will still be needed)
Test cases added for the above
llvm-svn: 150784
Greg Clayton