CHANGES:
- Thread locking switched from pthreads to C++11 standard library.
- Abstracted platform specific header includes into 'platform.h'.
- Create editline emulator for windows.
- Emulated various platform dependant functions on windows.
TODO:
- User input currently handled by gets_s(), work started on better handler:
see _WIP_INPUT_METHOD define blocks in 'ELWrapper.cpp'.
Aim is to handle 'tab' auto completion on windows.
- Tidy up 'getopt.inc' from lldbHostCommon to serve as LLDB Drivers getopt windows implementation.
llvm-svn: 192714
back in r173096 by Greg. When constructing a g packet or parsing a G packet,
and we're iterate over our register list, skip registers that are actually
just slices of other, real, registers. For instance, eax is 32-bits of rax
on x86_64.
<rdar://problem/15104187>
llvm-svn: 191802
the name of the remote gdb-protocol server, and get
a version number from it. This can be useful if lldb
needs to interoperate with a gdb-protocol server with
a known issue or bug.
llvm-svn: 191729
el_gets was using fflush to output it's string, but because we have our own filter running on the piped pty output, fflush only causes the prompt to be written into the pipe, and does not cause the filter code to run immediately.
The simplest fix is to manually block and wait for all editline output to be processed.
This fixes PR 14637.
llvm-svn: 191392
-S : Specifies a command file which will get sourced after the ~/.lldbinit but before file arguments are processed
-O : Specifies a single (one-line) command that will get ditto
and
-s : Specifies a command file which will get sourced after `pwd`/.lldbinit
-o : Specifies a command file which ditto
I also changed it so that by default these sourced commands will print their command result, but there's a
-q option to change that if you wish.
llvm-svn: 190734
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
- updated RNBDefs.h to allow version numbers to be passed in via preprocessor defines
- update libdebugserver.cpp to compile against latest DNBProcessKill signature
Review: http://llvm-reviews.chandlerc.com/D1331
llvm-svn: 188078
take for threads created while the program is running. Remove the testcase skips from TestConcurrentEvents.py,
since they all pass now, and fix TestWatchpointMultipleThreads.py - which should have caught this problem -
so it doesn't artificially break on new thread creation before the watchpoint triggers.
llvm.org/pr16566
<rdar://problem/14383244>
llvm-svn: 186132
The build system is currently miss-identifying GNU/kFreeBSD as FreeBSD.
This kind of simplification is sometimes useful, but in general it's not correct.
As GNU/kFreeBSD is an hybrid system, for kernel-related issues we want to match the
build definitions used for FreeBSD, whereas for userland-related issues we want to
match the definitions used for other systems with Glibc.
The current modification adjusts the build system so that they can be distinguished,
and explicitly adds GNU/kFreeBSD to the build checks in which it belongs.
Fixes bug #16446.
Patch by Robert Millan in the context of Debian.
llvm-svn: 185313
Found a race condition when killing an application where the state could be set to exited by the waitpid_thread() _before_ we call task resume (via MachProcess::PrivateResume()) in MachProcess::Kill().
llvm-svn: 185048
Match up with top’s implementation on recent Cab as API has changed a bit.
Tested the same binary running on Zin as well. Tested ARM binary on iOS as well.
llvm-svn: 185017
for any reason, use debugserver own's cputype as a best guess when
we reply to the debugger's qProcessInfo packet or when initializing
our register tables.
<rdar://problem/13406879>
llvm-svn: 184829
support files for debugserver to fix a build failure for arm. Also
remove some of the code used for software-driven single instruction
stepping; this is slowly being yanked out and these particular bits
overlap with the nub_break_t going away.
llvm-svn: 184828
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
you can now specify:
debugserver host:port
debugserver port
debugserver /path/to/file
When "host" is specified, we will only accept connections from that host. If host is not specified, we default to "localhost".
llvm-svn: 183457
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
- now, the output binary is called 'lldb-3.4' instead of 'lldb'
- a symlink 'lldb' -> 'lldb-3.4' is also created
- this fixes one of the problems preventing CMake from building Debian packages
llvm-svn: 182148
The "lldb" driver was interfering with STDOUT and STDERR if the output was over 1024 charcters long. The output was grabbing 1024 characters at a time, before it output the characters, it was writing characters to the screen to clear the current line. This has been fixed.
I also fixed the command interpreter from mixing the "(lldb) " prompt in with program output by always manually checking for program output. This was done by having the command interpreter know when it is in the middle of executing a command by setting a bool. This was needed since sometimes when a command would run the target, like with a command like 'expression (int)printf("hello\n")', the process would push a new input reader, and then pop it when it was done. This popping of the input reader would cause the command interpreter to get sent a reactivated message (from the private process state thread) and cause it to ask for another command, even though we were still in the middle of the command ('expression (int)printf("hello\n")'). Now we set a bool to true, run the command and set the bool to false. If we get reactivated while we are in the middle of a command, we don't say we are ready for a new command. This coupled with emitting the STDOUT/STDERR first after each command, followed by the command results, followed by then saying we are ready for a new command, should help cleanup the command line output on all platforms.
llvm-svn: 181807
If the user pressed ^D, that would bypass the exit confirmation mechanism
This checkin remedies that by making sure that users get prompted on exit when appropriate even if they use CTRL+D instead of typing quit at the prompt
llvm-svn: 181257
Most important was a new[] + delete mismatch in ScanFormatDescriptor()
and a couple of possible memory leaks in FileSpec::EnumerateDirectory().
llvm-svn: 181080
The sketch test case writes avg+stddev for all its metrics:
<key>fetch-frames</key>
<dict>
<key>description</key>
<string>time to dump backtrace for every frame in every thread</string>
<key>stddev</key>
<real>0.006270938361432314</real>
<key>value</key>
<real>0.011568079851851851</real>
</dict>
llvm-svn: 179550
Moving over to source-regex might be a good idea, but it’s not what we really want to test. Ideally, we would set a regex breakpoint to find the right line, then delete it and set one by file and line with the right line info from before.
llvm-svn: 179246
if we have an updated task_info call available; else fall back to getting
the default host-wide page size.
Update all uses of the vm page size to get it via MachVMMemory::PageSize().
<rdar://problem/13477763>, <rdar://problem/13498504>
llvm-svn: 178953
LLDB now can use a single dash for all long options for all commands form the command line and from the command interpreter. This involved just switching all calls from getopt_long() to getopt_long_only().
llvm-svn: 178789
Try and reap process when sending the "k" packet to avoid a race condition. We now wait for at most 1 second to reap the child process that we are killing.
llvm-svn: 178726
number in RNBRemote::HandlePacket_qProcessInfo -- add a new
GetCurrentThreadMachPort() so callers who need to make a mach
thred_get_state() call at the RNBRemote level will have a way to
get the port number.
llvm-svn: 178619
- make an overload of Launch() that takes an init list of const char* if all you need to tweak in the launch info are the command-line arguments
- make Run() return an int that you can use as an exit-code
- make dynamic values work properly when recursing in FetchVariables()
- make the po output more obvious in verbose mode
llvm-svn: 178578
To hook it up to individual test cases:
- define GetLongOptions() in your test case class to return something other than NULL (hopefully an array of options :-)
- implement ParseOption() to check for the short option char and do the right thing - return true at the end if you want more options to come your way or false if you don’t
- make sure that your Setup() call takes int& and char**& so that optind post-processing can happen - and call TestCase::Setup from your setup
llvm-svn: 178482
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
- memory delta and time for: target create
- memory delta and time for: setting breakpoint at main by name
- time to launch and hit bp at main
- overall memory of target create + bp main + run to main
- ovarall time of target create + bp main + run to main
llvm-svn: 177808
to in INADDR_LOOPBACK mode by default ("localhost only")
instead of INADDR_ANY ("accept connections from any system").
Add a new command line argument to debugserver, --open-connection
or -H which will enable the previous behavior. It would be used
if you were doing two-system debugging, with lldb running on one
system and debugserver running on the other. But it is a less
common workflow and should not be the default.
<rdar://problem/12583284>
llvm-svn: 177790
- Added new abtract Results class to keep CoreFoundation out of the tests. There are many subclasses for different settings:
Results::Result::Dictionary
Results::Result::Array
Results::Result::Unsigned
Results::Result::Double
Results::Result::String
- Gauge<T> can now write themselves out via a templatized write to results function:
template <class T>
Results::ResultSP GetResult (const char *description, T value);
- There are four specializations of this so far:
template <>
Results::ResultSP GetResult (const char *description, double value);
template <>
Results::ResultSP GetResult (const char *description, uint64_t value);
template <>
Results::ResultSP GetResult (const char *description, std::string value);
template <>
Results::ResultSP GetResult (const char *description, MemoryStats value);
- Don't emit the virtual memory reading from the task info call as it really doesn't mean much as it includes way too much (shared cache + other stuff we don't have control over)
- Fixed other test cases to build correctly and use the new classes
llvm-svn: 177696
This is especially useful to take measurements that span multiple test steps, or where you need to have different operations fall under the same measurement
An example of use is in the formatters perf test case
llvm-svn: 177597
- TestCase.m_thread is now filled in with the first thread that has a valid
stop reason. This eliminates the need for the SelectMyThread() functions.
- The first thread that stops for a reason is also set as the selected thread
in the process in case any command line commands are run.
- Changed launch over to take a SBLaunchInfo parameter so that the launch
function doesn't keep getting new arguments as they are needed.
- TestCase::Setup() and TestCase::Launch(SBLaunchInfo) now return bool to
indicate success of setup and launch.
- ActionWanted::Next(SBThread) was renamed to ActionWanted::StepOver(SBThread)
- ActionWanted::Finish(SBThread) was renamed to ActionWanted::StepOut(SBThread)
llvm-svn: 177376
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
- don't use preprocessor macros
- use switch statements
- don't put anything in the lldb namespace, use "lldb_perf" namespace.
- Pass the action struct into each TestStep() for each step fill in
- Modify the ActionWanted class to have accessors to make the continue, next, finish, kill instead of using preproc macros
llvm-svn: 177332
Added logging that will show up in the system console when we try to resume a process that is already running, or has an unexpected state.
llvm-svn: 176960
Drop the old f registers from debugserver's register list. Add the
NEON 128-bit q registers to debugserver, support reading and writing.
Add the new contains / invalidates mappings for the s, d, and q
registers so lldb will know what registers overlay what other registers.
Change the default format of s and d registers to be floating point
instead of hex. Remove some UTF-8 hyphen chars in comments in the ARM
register number definition headers.
<rdar://problem/13121797>
llvm-svn: 176915
This is a very basic implementation of a library that easily allows to drive LLDB.framework to write test cases for performance
This is separate from the LLDB testsuite in test/ in that:
a) this uses C++ instead of Python to avoid measures being affected by SWIG
b) this is in very early development and needs lots of tweaking before it can be considered functionally complete
c) this is not meant to test correctness but to help catch performance regressions
There is a sample application built against the library (in darwin/sketch) that uses the famous sample app Sketch as an inferior to measure certain basic parameters of LLDB's behavior.
The resulting output is a PLIST much like the following:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<array>
<dict>
<key>fetch-frames</key>
<real>0.13161715522222225</real>
</dict>
<dict>
<key>file-line-bkpt</key>
<real>0.029111678750000002</real>
</dict>
<dict>
<key>fetch-modules</key>
<real>0.00026376766666666668</real>
</dict>
<dict>
<key>fetch-vars</key>
<real>0.17820429311111111</real>
</dict>
<dict>
<key>run-expr</key>
<real>0.029676525769230768</real>
</dict>
</array>
</plist>
Areas for improvement:
- code cleanups (I will be out of the office for a couple days this coming week, but please keep ideas coming!)
- more metrics and test cases
- better error checking
This toolkit also comprises a simple event-loop-driven controller for LLDB, similar yet much simpler to what the Driver does to implement the lldb command-line tool.
llvm-svn: 176715
Make it configurable what to profile.
For Mac, we don't use the dirty page size yet and hence there is no need to gather that. This should be way better in not draining the battery since we are operating between 0% to 0.1% on the Mac after this change.
llvm-svn: 176451
own port namepsace) as the thread identifier to using the system-wide
globally unique thread id as the thread identifier number.
MachThread.cpp keeps both the unique id and the mach port number
for each thread. All layers outside MachThread class use the unique
id with three exceptions: (1) Mach exceptions come in with the port
number (thread_port) which needs to be translated, (2) any calls to
low-level thread_get_state/thread_set_state/thread_suspend etc need
to use the mach port number, (3) MachThreadList::UpdateThreadList
which creates the MachThread objects gets the unique id and passes
it to the MachThread ctor as an argument.
In general, any time nub_thread_t is used, it is now referring to a
unique thread id. Any time a thread_t is used, it is now referring
to a mach port number. There was some interchangability of these
types previously. nub_thread_t has also been changed to a 64-bit
type which necessitated some printf specification string changes.
I haven't been able to test these changes extensively yet but want
to checkpoint the work. The scenarios I've been testing are all
working correctly so while there may be some corner cases I haven't
hit yet, I think it is substantially correct.
<rdar://problem/12931414>
llvm-svn: 175870
Also added a TimeSpecTimeout class which can be used with any calls that take a "struct timespec *" as an argument. It is used by the KQueue class.
Also updated some project settings.
llvm-svn: 175377
hitting auto-continue signals while running a thread plan would cause us to lose control of the debug
session.
<rdar://problem/12993641>
llvm-svn: 174793
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
reply to be hex encoded, not decimal.
Fix the whitespace in the container-regs/invalidate-regs
documentation, fix one ambiguous hex/decimal number in an
example.
llvm-svn: 173225
Fixed the 32, 16, and 8 bit pseudo regs for x86_64 (real reg of "rax" which subvalues "eax", "ax", etc...) to correctly get updated when stepping. Also fixed it so actual registers can specify what other registers must be invalidated when a register is modified. Previously, only pseudo registers could invalidate other registers.
Modified the LLDB qRegisterInfo extension to the GDB remote interface to support specifying the containing registers with the new "container-regs" key whose value is a comma separated list of register numbers. Also added a "invalidate-regs" key whose value is also a comma separated list of register numbers.
Removed the hack GDBRemoteDynamicRegisterInfo::Addx86_64ConvenienceRegisters() function and modified "debugserver" to specify the registers correctly using the new "container-regs" and "invalidate-regs" keys.
llvm-svn: 173096
Swap in index ids for thread ids in GDBRemoteCommunicationClient. Besides dealing with the async logic, I have to take care of the situation when the inferior paused as well.
llvm-svn: 172869
Prevent profiling from working on older debugserver. Just a simple renaming since the caller is prepared to handle the ‘unimplemented’ answer.
llvm-svn: 172583
document some simple bourne shell to re-generate these from the DNBDefs.h
header file in case this needs to be done again in the future.
llvm-svn: 172494
1. Using mach port number, just like when inferior is paused.
2. Use key:value pair of thread used time instead of comma separated notation.
llvm-svn: 172012
Add unconditional logging messages to every place in debugserver
where we send a SIGKILL signal or do a ptrace PT_KILL call to
terminate the inferior process. When the debuggee is silently
killed off, the console logging from debugserver can disambiguate
whether debugserver killed off the process because it failed to
completely set it up, becuase it was told to (via the "k" packet),
or if some external daemon killed it.
llvm-svn: 171606
Update the debugserver "qProcessInfo" implementation to return the
cpu type, cpu subtype, OS and vendor information just like qHostInfo
does so lldb can create an ArchSpec based on the returned values.
Add a new GetProcessArchitecture to GDBRemoteCommunicationClient akin
to GetHostArchitecture. If the qProcessInfo packet is supported,
GetProcessArchitecture will return the cpu type / subtype of the
process -- e.g. a 32-bit user process running on a 64-bit x86_64 Mac
system.
Have ProcessGDBRemote set the Target's architecture based on the
GetProcessArchitecture when we've completed an attach/launch/connect.
llvm-svn: 170491
This can be used by lldb to ask for information
about the process debugserver is attached to/launched.
Particularly useful on a 64-bit x86 Mac system which
can run 32-bit or 64-bit user-land processes.
llvm-svn: 170409
This should be more consistent with the notion of command success/failure and avoids spewing warnings that the user might not care about
There will need to be an option to specify the level of verbosity desired (never show anything, only show failures, errors and warning, everything)
llvm-svn: 170167
Fixing an issue where errors in command files sourced as arguments to command-line lldb (e.g. ./lldb -s foo.cmd) would not be shown to the user
llvm-svn: 170146
- 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
- add new header lldb-python.h to be included before other system headers
- short term fix (eventually python dependencies must be cleaned up)
Patch by Matt Kopec!
llvm-svn: 169341
Cleaned up the option parsing code to always pass around the short options as integers. Previously we cast this down to "char" and lost some information. I recently added an assert that would detect duplicate short character options which was firing during the test suite.
This fix does the following:
- make sure all short options are treated as "int"
- make sure that short options can be non-printable values when a short option is not required or when an option group is mixed into many commands and a short option is not desired
- fix the help printing to "do the right thing" in all cases. Previously if there were duplicate short character options, it would just not emit help for the duplicates
- fix option parsing when there are duplicates to parse options correctly. Previously the option parsing, when done for an OptionGroup, would just start parsing options incorrectly by omitting table entries and it would end up setting the wrong option value
llvm-svn: 169189
Prevent async and sync calls to get profile data from stomping on each other.
At the same time, don't use '$' as end delimiter per chunk of profile data.
llvm-svn: 168948
DebugClang builds of LLDB to build a properly
codesigned debugserver. I did this by adding
a DebugClang configuration to debugserver that's
just a clone of the Debug configuration.
llvm-svn: 168746
Then make the Thread a Broadcaster, and get it to broadcast when the selected frame is changed (but only from the Command Line) and when Thread::ReturnFromFrame
changes the stack.
Made the Driver use this notification to print the new thread status rather than doing it in the command.
Fixed a few places where people were setting their broadcaster class by hand rather than using the static broadcaster class call.
<rdar://problem/12383087>
llvm-svn: 165640
Ed Maste