intel/llvm
1
0
Fork 0
mirror of https://github.com/intel/llvm.git synced 2026-01-20 10:58:11 +08:00
Code Issues Packages Projects Releases Wiki Activity
Files
41febc658b62f15983301ebff759c99909f2ecbf
llvm/lldb/source/Plugins/Process/gdb-remote/ProcessGDBRemote.cpp

2700 lines
98 KiB
C++
Raw Normal View History

Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
//===-- ProcessGDBRemote.cpp ------------------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
// C Includes
#include <errno.h>
#include <spawn.h>
#include <sys/types.h>
#include <sys/stat.h>
// C++ Includes
#include <algorithm>
#include <map>
// Other libraries and framework includes
#include "lldb/Breakpoint/WatchpointLocation.h"
Move Args.{cpp,h} and Options.{cpp,h} to Interpreter where they really belong. llvm-svn: 106034
2010-06-15 19:49:27 +00:00
#include "lldb/Interpreter/Args.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
#include "lldb/Core/ArchSpec.h"
#include "lldb/Core/Debugger.h"
#include "lldb/Core/ConnectionFileDescriptor.h"
Moved FileSpec into the Host layer since it will vary from host to host. We have a common unix implementation in lldb/source/Host/common/FileSpec.cpp. llvm-svn: 125078
2011-02-08 05:05:52 +00:00
#include "lldb/Host/FileSpec.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
#include "lldb/Core/InputReader.h"
#include "lldb/Core/Module.h"
#include "lldb/Core/PluginManager.h"
#include "lldb/Core/State.h"
#include "lldb/Core/StreamString.h"
#include "lldb/Core/Timer.h"
#include "lldb/Host/TimeValue.h"
#include "lldb/Symbol/ObjectFile.h"
#include "lldb/Target/DynamicLoader.h"
#include "lldb/Target/Target.h"
#include "lldb/Target/TargetList.h"
Move source/Utility/PseudoTerminal.h into include/lldb/Utility. The top of the header file seems to indicate that this was intended to be over at include/lldb/Core but we should be in line with the .cpp file's location so it's include/lldb/Utility for now. llvm-svn: 105753
2010-06-09 21:28:42 +00:00
#include "lldb/Utility/PseudoTerminal.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// Project includes
#include "lldb/Host/Host.h"
Merged Eli Friedman's linux build changes where he added Makefile files that enabled LLVM make style building and made this compile LLDB on Mac OS X. We can now iterate on this to make the build work on both linux and macosx. llvm-svn: 108009
2010-07-09 20:39:50 +00:00
#include "Utility/StringExtractorGDBRemote.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
#include "GDBRemoteRegisterContext.h"
#include "ProcessGDBRemote.h"
#include "ProcessGDBRemoteLog.h"
#include "ThreadGDBRemote.h"
Abtracted the old "lldb_private::Thread::StopInfo" into an abtract class. This will allow debugger plug-ins to make any instance of "lldb_private::StopInfo" that can completely describe any stop reason. It also provides a framework for doing intelligent things with the stop info at important times in the lifetime of the inferior. Examples include the signal stop info in StopInfoUnixSignal. It will check with the process to see that the current action is for the signal. These actions include wether to stop for the signal, wether the notify that the signal was hit, and wether to pass the signal along to the inferior process. The StopInfoUnixSignal class overrides the "ShouldStop()" method of StopInfo and this allows the stop info to determine if it should stop at the signal or continue the process. StopInfo subclasses must override the following functions: virtual lldb::StopReason GetStopReason () const = 0; virtual const char * GetDescription () = 0; StopInfo subclasses can override the following functions: // If the subclass returns "false", the inferior will resume. The default // version of this function returns "true" which means the default stop // info will stop the process. The breakpoint subclass will check if // the breakpoint wants us to stop by calling any installed callback on // the breakpoint, and also checking if the breakpoint is for the current // thread. Signals will check if they should stop based off of the // UnixSignal settings in the process. virtual bool ShouldStop (Event *event_ptr); // Sublasses can state if they want to notify the debugger when "ShouldStop" // returns false. This would be handy for breakpoints where you want to // log information and continue and is also used by the signal stop info // to notify that a signal was received (after it checks with the process // signal settings). virtual bool ShouldNotify (Event *event_ptr) { return false; } // Allow subclasses to do something intelligent right before we resume. // The signal class will figure out if the signal should be propagated // to the inferior process and pass that along to the debugger plug-ins. virtual void WillResume (lldb::StateType resume_state) { // By default, don't do anything } The support the Mach exceptions was moved into the lldb/source/Plugins/Process/Utility folder and now doesn't polute the lldb_private::Thread class with platform specific code. llvm-svn: 110184
2010-08-04 01:40:35 +00:00
#include "StopInfoMachException.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
#define DEBUGSERVER_BASENAME "debugserver"
using namespace lldb;
using namespace lldb_private;
static inline uint16_t
get_random_port ()
{
return (arc4random() % (UINT16_MAX - 1000u)) + 1000u;
}
const char *
ProcessGDBRemote::GetPluginNameStatic()
{
return "process.gdb-remote";
}
const char *
ProcessGDBRemote::GetPluginDescriptionStatic()
{
return "GDB Remote protocol based debugging plug-in.";
}
void
ProcessGDBRemote::Terminate()
{
PluginManager::UnregisterPlugin (ProcessGDBRemote::CreateInstance);
}
Process*
ProcessGDBRemote::CreateInstance (Target &target, Listener &listener)
{
return new ProcessGDBRemote (target, listener);
}
bool
ProcessGDBRemote::CanDebug(Target &target)
{
// For now we are just making sure the file exists for a given module
ModuleSP exe_module_sp(target.GetExecutableModule());
if (exe_module_sp.get())
return exe_module_sp->GetFileSpec().Exists();
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
// However, if there is no executable module, we return true since we might be preparing to attach.
return true;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
//----------------------------------------------------------------------
// ProcessGDBRemote constructor
//----------------------------------------------------------------------
ProcessGDBRemote::ProcessGDBRemote(Target& target, Listener &listener) :
Process (target, listener),
m_dynamic_loader_ap (),
m_flags (0),
m_stdio_mutex (Mutex::eMutexTypeRecursive),
m_gdb_comm(),
m_debugserver_pid (LLDB_INVALID_PROCESS_ID),
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
m_debugserver_thread (LLDB_INVALID_HOST_THREAD),
Merged Eli Friedman's linux build changes where he added Makefile files that enabled LLVM make style building and made this compile LLDB on Mac OS X. We can now iterate on this to make the build work on both linux and macosx. llvm-svn: 108009
2010-07-09 20:39:50 +00:00
m_last_stop_packet (),
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
m_register_info (),
m_async_broadcaster ("lldb.process.gdb-remote.async-broadcaster"),
m_async_thread (LLDB_INVALID_HOST_THREAD),
Merged Eli Friedman's linux build changes where he added Makefile files that enabled LLVM make style building and made this compile LLDB on Mac OS X. We can now iterate on this to make the build work on both linux and macosx. llvm-svn: 108009
2010-07-09 20:39:50 +00:00
m_curr_tid (LLDB_INVALID_THREAD_ID),
m_curr_tid_run (LLDB_INVALID_THREAD_ID),
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
m_z0_supported (1),
Added the ability to detect which vCont packets (using the "vCont?") packet are supported by the remote GDB target. We can also now deal with the lack of vCont support and send packets that the remote GDB stub can use. We also error out of the continue if LLDB tries to do something too complex when vCont isn't supported. llvm-svn: 125433
2011-02-12 06:28:37 +00:00
m_continue_c_tids (),
m_continue_C_tids (),
m_continue_s_tids (),
m_continue_S_tids (),
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
m_dispatch_queue_offsets_addr (LLDB_INVALID_ADDRESS),
Merged Eli Friedman's linux build changes where he added Makefile files that enabled LLVM make style building and made this compile LLDB on Mac OS X. We can now iterate on this to make the build work on both linux and macosx. llvm-svn: 108009
2010-07-09 20:39:50 +00:00
m_packet_timeout (1),
m_max_memory_size (512),
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
m_waiting_for_attach (false),
Added an interface for noticing new thread creation. At this point, I only turn it on when we are requesting a single thread to run. May seem like a silly thing to do, but the kernel on MacOS X will inject new threads into a program willy-nilly, and I would like to keep them from running if I can. llvm-svn: 124018
2011-01-22 01:33:44 +00:00
m_local_debugserver (true),
m_thread_observation_bps()
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
}
//----------------------------------------------------------------------
// Destructor
//----------------------------------------------------------------------
ProcessGDBRemote::~ProcessGDBRemote()
{
Fixed a crasher where when a ProcessGDBRemote class was being destroyed, it would eventually destroy the dynamic loader (when the lldb_private::Process::m_dynamic_loader_ap destroys itself in the object member destructor chain). The dynamic loader was calling a pure virtual method in Process which was causing a crash. The quick fix is to reset the auto pointer in the ProcessGDBRemote destructor when ProcessGDBRemote is still a valid object with all its pure virtual functions intact. llvm-svn: 121704
2010-12-13 18:11:18 +00:00
m_dynamic_loader_ap.reset();
Patch that allows for thread_t to be something more complex than an integer. Modified patch from Kirk Beitz. llvm-svn: 125067
2011-02-08 01:34:25 +00:00
if (IS_VALID_LLDB_HOST_THREAD(m_debugserver_thread))
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
{
Host::ThreadCancel (m_debugserver_thread, NULL);
thread_result_t thread_result;
Host::ThreadJoin (m_debugserver_thread, &thread_result, NULL);
m_debugserver_thread = LLDB_INVALID_HOST_THREAD;
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// m_mach_process.UnregisterNotificationCallbacks (this);
Clear();
}
//----------------------------------------------------------------------
// PluginInterface
//----------------------------------------------------------------------
const char *
ProcessGDBRemote::GetPluginName()
{
return "Process debugging plug-in that uses the GDB remote protocol";
}
const char *
ProcessGDBRemote::GetShortPluginName()
{
return GetPluginNameStatic();
}
uint32_t
ProcessGDBRemote::GetPluginVersion()
{
return 1;
}
void
ProcessGDBRemote::GetPluginCommandHelp (const char *command, Stream *strm)
{
strm->Printf("TODO: fill this in\n");
}
Error
ProcessGDBRemote::ExecutePluginCommand (Args &command, Stream *strm)
{
Error error;
error.SetErrorString("No plug-in commands are currently supported.");
return error;
}
Log *
ProcessGDBRemote::EnablePluginLogging (Stream *strm, Args &command)
{
return NULL;
}
void
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
ProcessGDBRemote::BuildDynamicRegisterInfo (bool force)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
if (!force && m_register_info.GetNumRegisters() > 0)
return;
char packet[128];
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
m_register_info.Clear();
StringExtractorGDBRemote::Type packet_type = StringExtractorGDBRemote::eResponse;
uint32_t reg_offset = 0;
uint32_t reg_num = 0;
for (; packet_type == StringExtractorGDBRemote::eResponse; ++reg_num)
{
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
const int packet_len = ::snprintf (packet, sizeof(packet), "qRegisterInfo%x", reg_num);
assert (packet_len < sizeof(packet));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
StringExtractorGDBRemote response;
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
if (m_gdb_comm.SendPacketAndWaitForResponse(packet, packet_len, response, 2, false))
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
packet_type = response.GetType();
if (packet_type == StringExtractorGDBRemote::eResponse)
{
std::string name;
std::string value;
ConstString reg_name;
ConstString alt_name;
ConstString set_name;
RegisterInfo reg_info = { NULL, // Name
NULL, // Alt name
0, // byte size
reg_offset, // offset
eEncodingUint, // encoding
eFormatHex, // formate
{
LLDB_INVALID_REGNUM, // GCC reg num
LLDB_INVALID_REGNUM, // DWARF reg num
LLDB_INVALID_REGNUM, // generic reg num
The first part of an lldb native stack unwinder. The Unwind and RegisterContext subclasses still need to be finished; none of this code is used by lldb at this point (unless you call into it by hand). The ObjectFile class now has an UnwindTable object. The UnwindTable object has a series of FuncUnwinders objects (Function Unwinders) -- one for each function in that ObjectFile we've backtraced through during this debug session. The FuncUnwinders object has a few different UnwindPlans. UnwindPlans are a generic way of describing how to find the canonical address of a given function's stack frame (the CFA idea from DWARF/eh_frame) and how to restore the caller frame's register values, if they have been saved by this function. UnwindPlans are created from different sources. One source is the eh_frame exception handling information generated by the compiler for unwinding an exception throw. Another source is an assembly language inspection class (UnwindAssemblyProfiler, uses the Plugin architecture) which looks at the instructions in the funciton prologue and describes the stack movements/register saves that are done. Two additional types of UnwindPlans that are worth noting are the "fast" stack UnwindPlan which is useful for making a first pass over a thread's stack, determining how many stack frames there are and retrieving the pc and CFA values for each frame (enough to create StackFrameIDs). Only a minimal set of registers is recovered during a fast stack walk. The final UnwindPlan is an architectural default unwind plan. These are provided by the ArchDefaultUnwindPlan class (which uses the plugin architecture). When no symbol/function address range can be found for a given pc value -- when we have no eh_frame information and when we don't have a start address so we can't examine the assembly language instrucitons -- we have to make a best guess about how to unwind. That's when we use the architectural default UnwindPlan. On x86_64, this would be to assume that rbp is used as a stack pointer and we can use that to find the caller's frame pointer and pc value. It's a last-ditch best guess about how to unwind out of a frame. There are heuristics about when to use one UnwindPlan versues the other -- this will all happen in the still-begin-written UnwindLLDB subclass of Unwind which runs the UnwindPlans. llvm-svn: 113581
2010-09-10 07:49:16 +00:00
reg_num, // GDB reg num
reg_num // native register number
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
};
while (response.GetNameColonValue(name, value))
{
if (name.compare("name") == 0)
{
reg_name.SetCString(value.c_str());
}
else if (name.compare("alt-name") == 0)
{
alt_name.SetCString(value.c_str());
}
else if (name.compare("bitsize") == 0)
{
reg_info.byte_size = Args::StringToUInt32(value.c_str(), 0, 0) / CHAR_BIT;
}
else if (name.compare("offset") == 0)
{
uint32_t offset = Args::StringToUInt32(value.c_str(), UINT32_MAX, 0);
Applied PluginManager.cpp patch from Jean-Daniel Dupas. Fixed problem Jean-Daniel Dupas found in ProcessGDBRemote.cpp. llvm-svn: 105857
2010-06-11 23:44:18 +00:00
if (reg_offset != offset)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
reg_offset = offset;
}
}
else if (name.compare("encoding") == 0)
{
if (value.compare("uint") == 0)
reg_info.encoding = eEncodingUint;
else if (value.compare("sint") == 0)
reg_info.encoding = eEncodingSint;
else if (value.compare("ieee754") == 0)
reg_info.encoding = eEncodingIEEE754;
else if (value.compare("vector") == 0)
reg_info.encoding = eEncodingVector;
}
else if (name.compare("format") == 0)
{
if (value.compare("binary") == 0)
reg_info.format = eFormatBinary;
else if (value.compare("decimal") == 0)
reg_info.format = eFormatDecimal;
else if (value.compare("hex") == 0)
reg_info.format = eFormatHex;
else if (value.compare("float") == 0)
reg_info.format = eFormatFloat;
else if (value.compare("vector-sint8") == 0)
reg_info.format = eFormatVectorOfSInt8;
else if (value.compare("vector-uint8") == 0)
reg_info.format = eFormatVectorOfUInt8;
else if (value.compare("vector-sint16") == 0)
reg_info.format = eFormatVectorOfSInt16;
else if (value.compare("vector-uint16") == 0)
reg_info.format = eFormatVectorOfUInt16;
else if (value.compare("vector-sint32") == 0)
reg_info.format = eFormatVectorOfSInt32;
else if (value.compare("vector-uint32") == 0)
reg_info.format = eFormatVectorOfUInt32;
else if (value.compare("vector-float32") == 0)
reg_info.format = eFormatVectorOfFloat32;
else if (value.compare("vector-uint128") == 0)
reg_info.format = eFormatVectorOfUInt128;
}
else if (name.compare("set") == 0)
{
set_name.SetCString(value.c_str());
}
else if (name.compare("gcc") == 0)
{
reg_info.kinds[eRegisterKindGCC] = Args::StringToUInt32(value.c_str(), LLDB_INVALID_REGNUM, 0);
}
else if (name.compare("dwarf") == 0)
{
reg_info.kinds[eRegisterKindDWARF] = Args::StringToUInt32(value.c_str(), LLDB_INVALID_REGNUM, 0);
}
else if (name.compare("generic") == 0)
{
if (value.compare("pc") == 0)
reg_info.kinds[eRegisterKindGeneric] = LLDB_REGNUM_GENERIC_PC;
else if (value.compare("sp") == 0)
reg_info.kinds[eRegisterKindGeneric] = LLDB_REGNUM_GENERIC_SP;
else if (value.compare("fp") == 0)
reg_info.kinds[eRegisterKindGeneric] = LLDB_REGNUM_GENERIC_FP;
else if (value.compare("ra") == 0)
reg_info.kinds[eRegisterKindGeneric] = LLDB_REGNUM_GENERIC_RA;
else if (value.compare("flags") == 0)
reg_info.kinds[eRegisterKindGeneric] = LLDB_REGNUM_GENERIC_FLAGS;
}
}
Applied PluginManager.cpp patch from Jean-Daniel Dupas. Fixed problem Jean-Daniel Dupas found in ProcessGDBRemote.cpp. llvm-svn: 105857
2010-06-11 23:44:18 +00:00
reg_info.byte_offset = reg_offset;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
assert (reg_info.byte_size != 0);
reg_offset += reg_info.byte_size;
m_register_info.AddRegister(reg_info, reg_name, alt_name, set_name);
}
}
else
{
packet_type = StringExtractorGDBRemote::eError;
}
}
if (reg_num == 0)
{
// We didn't get anything. See if we are debugging ARM and fill with
// a hard coded register set until we can get an updated debugserver
// down on the devices.
ArchSpec arm_arch ("arm");
if (GetTarget().GetArchitecture() == arm_arch)
m_register_info.HardcodeARMRegisters();
}
m_register_info.Finalize ();
}
Error
ProcessGDBRemote::WillLaunch (Module* module)
{
return WillLaunchOrAttach ();
}
Error
Fixed Process::Halt() as it was broken for "process halt" after recent changes to the DoHalt down in ProcessGDBRemote. I also moved the functionality that was in ProcessGDBRemote::DoHalt up into Process::Halt so not every class has to implement a tricky halt/resume on the internal state thread. The functionality is the same as it was before with two changes: - when we eat the event we now just reuse the event we consume when the private state thread is paused and set the interrupted bool on the event if needed - we also properly update the Process::m_public_state with the state of the event we consume. Prior to this, if you issued a "process halt" it would eat the event, not update the process state, and then produce a new event with the interrupted bit set and send it. Anyone listening to the event would get the stopped event with a process that whose state was set to "running". Fixed debugserver to not have to be spawned with the architecture of the inferior process. This worked fine for launching processes, but when attaching to processes by name or pid without a file in lldb, it would fail. Now debugserver can support multiple architectures for a native debug session on the current host. This currently means i386 and x86_64 are supported in the same binary and a x86_64 debugserver can attach to a i386 executable. This change involved a lot of changes to make sure we dynamically detect the correct registers for the inferior process. llvm-svn: 119680
2010-11-18 05:57:03 +00:00
ProcessGDBRemote::WillAttachToProcessWithID (lldb::pid_t pid)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
return WillLaunchOrAttach ();
}
Error
Fixed Process::Halt() as it was broken for "process halt" after recent changes to the DoHalt down in ProcessGDBRemote. I also moved the functionality that was in ProcessGDBRemote::DoHalt up into Process::Halt so not every class has to implement a tricky halt/resume on the internal state thread. The functionality is the same as it was before with two changes: - when we eat the event we now just reuse the event we consume when the private state thread is paused and set the interrupted bool on the event if needed - we also properly update the Process::m_public_state with the state of the event we consume. Prior to this, if you issued a "process halt" it would eat the event, not update the process state, and then produce a new event with the interrupted bit set and send it. Anyone listening to the event would get the stopped event with a process that whose state was set to "running". Fixed debugserver to not have to be spawned with the architecture of the inferior process. This worked fine for launching processes, but when attaching to processes by name or pid without a file in lldb, it would fail. Now debugserver can support multiple architectures for a native debug session on the current host. This currently means i386 and x86_64 are supported in the same binary and a x86_64 debugserver can attach to a i386 executable. This change involved a lot of changes to make sure we dynamically detect the correct registers for the inferior process. llvm-svn: 119680
2010-11-18 05:57:03 +00:00
ProcessGDBRemote::WillAttachToProcessWithName (const char *process_name, bool wait_for_launch)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
return WillLaunchOrAttach ();
}
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
Error
ProcessGDBRemote::DoConnectRemote (const char *remote_url)
{
Error error (WillLaunchOrAttach ());
if (error.Fail())
return error;
if (strncmp (remote_url, "connect://", strlen ("connect://")) == 0)
{
error = ConnectToDebugserver (remote_url);
}
else
{
error.SetErrorStringWithFormat ("unsupported remote url: %s", remote_url);
}
if (error.Fail())
return error;
StartAsyncThread ();
lldb::pid_t pid = m_gdb_comm.GetCurrentProcessID (m_packet_timeout);
if (pid == LLDB_INVALID_PROCESS_ID)
{
// We don't have a valid process ID, so note that we are connected
// and could now request to launch or attach, or get remote process
// listings...
SetPrivateState (eStateConnected);
}
else
{
// We have a valid process
SetID (pid);
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse("?", 1, response, m_packet_timeout, false))
{
const StateType state = SetThreadStopInfo (response);
if (state == eStateStopped)
{
SetPrivateState (state);
}
else
error.SetErrorStringWithFormat ("Process %i was reported after connecting to '%s', but state was not stopped: %s", pid, remote_url, StateAsCString (state));
}
else
error.SetErrorStringWithFormat ("Process %i was reported after connecting to '%s', but no stop reply packet was received", pid, remote_url);
}
return error;
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Error
ProcessGDBRemote::WillLaunchOrAttach ()
{
Error error;
// TODO: this is hardcoded for macosx right now. We need this to be more dynamic
m_dynamic_loader_ap.reset(DynamicLoader::FindPlugin(this, "dynamic-loader.macosx-dyld"));
if (m_dynamic_loader_ap.get() == NULL)
error.SetErrorString("unable to find the dynamic loader named 'dynamic-loader.macosx-dyld'");
m_stdio_communication.Clear ();
return error;
}
//----------------------------------------------------------------------
// Process Control
//----------------------------------------------------------------------
Error
ProcessGDBRemote::DoLaunch
(
Module* module,
char const *argv[],
char const *envp[],
Added the ability to disable ASLR (Address Space Layout Randomization). ASLR is disabled by default, and can be enabled using: (lldb) set disable-aslr 0 llvm-svn: 112616
2010-08-31 18:35:14 +00:00
uint32_t launch_flags,
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
const char *stdin_path,
const char *stdout_path,
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
const char *stderr_path,
const char *working_dir
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
)
{
Cleaned up a unused member variable in Debugger. Added the start of Host specific launch services, though it currently isn't hookup up to anything. We want to be able to launch a process and use the native launch services to launch an app like it would be launched by the user double clicking on the app. We also eventually want to be able to run a command line app in a newly spawned terminal to avoid terminal sharing. Fixed an issue with the new DWARF forward type declaration stuff. A crasher was found that was happening when trying to properly expand the forward declarations. llvm-svn: 115213
2010-09-30 21:49:03 +00:00
Error error;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// ::LogSetBitMask (GDBR_LOG_DEFAULT);
// ::LogSetOptions (LLDB_LOG_OPTION_THREADSAFE | LLDB_LOG_OPTION_PREPEND_TIMESTAMP | LLDB_LOG_OPTION_PREPEND_PROC_AND_THREAD);
// ::LogSetLogFile ("/dev/stdout");
ObjectFile * object_file = module->GetObjectFile();
if (object_file)
{
ArchSpec inferior_arch(module->GetArchitecture());
char host_port[128];
snprintf (host_port, sizeof(host_port), "localhost:%u", get_random_port ());
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
char connect_url[128];
snprintf (connect_url, sizeof(connect_url), "connect://%s", host_port);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
const bool launch_process = true;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
bool start_debugserver_with_inferior_args = false;
if (start_debugserver_with_inferior_args)
{
// We want to launch debugserver with the inferior program and its
// arguments on the command line. We should only do this if we
// the GDB server we are talking to doesn't support the 'A' packet.
error = StartDebugserverProcess (host_port,
argv,
envp,
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
stdin_path,
stdout_path,
stderr_path,
working_dir,
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
launch_process,
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
LLDB_INVALID_PROCESS_ID,
NULL, false,
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
launch_flags,
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
inferior_arch);
if (error.Fail())
return error;
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
error = ConnectToDebugserver (connect_url);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Success())
{
SetID (m_gdb_comm.GetCurrentProcessID (m_packet_timeout));
}
}
else
{
error = StartDebugserverProcess (host_port,
NULL,
NULL,
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
stdin_path,
stdout_path,
stderr_path,
working_dir,
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
launch_process,
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
LLDB_INVALID_PROCESS_ID,
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
NULL,
false,
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
launch_flags,
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
inferior_arch);
if (error.Fail())
return error;
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
error = ConnectToDebugserver (connect_url);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Success())
{
// Send the environment and the program + arguments after we connect
if (envp)
{
const char *env_entry;
for (int i=0; (env_entry = envp[i]); ++i)
{
if (m_gdb_comm.SendEnvironmentPacket(env_entry, m_packet_timeout) != 0)
break;
}
}
Added support for objective C built-in types: id, Class, and SEL. This involved watching for the objective C built-in types in DWARF and making sure when we convert the DWARF types into clang types that we use the appropriate ASTContext types. Added a way to find and dump types in lldb (something equivalent to gdb's "ptype" command): image lookup --type <TYPENAME> This only works for looking up types by name and won't work with variables. It also currently dumps out verbose internal information. I will modify it to dump more appropriate user level info in my next submission. Hookup up the "FindTypes()" functions in the SymbolFile and SymbolVendor so we can lookup types by name in one or more images. Fixed "image lookup --address <ADDRESS>" to be able to correctly show all symbol context information, but it will only show this extra information when the new "--verbose" flag is used. Updated to latest LLVM to get a few needed fixes. llvm-svn: 110089
2010-08-03 00:35:52 +00:00
// FIXME: convert this to use the new set/show variables when they are available
#if 0
if (::getenv ("LLDB_DEBUG_DEBUGSERVER"))
{
const uint32_t attach_debugserver_secs = 10;
::printf ("attach to debugserver (pid = %i)\n", m_debugserver_pid);
for (uint32_t i=0; i<attach_debugserver_secs; ++i)
{
printf ("%i\n", attach_debugserver_secs - i);
sleep (1);
}
}
#endif
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
const uint32_t arg_timeout_seconds = 10;
int arg_packet_err = m_gdb_comm.SendArgumentsPacket (argv, arg_timeout_seconds);
if (arg_packet_err == 0)
{
std::string error_str;
if (m_gdb_comm.GetLaunchSuccess (m_packet_timeout, error_str))
{
SetID (m_gdb_comm.GetCurrentProcessID (m_packet_timeout));
}
else
{
error.SetErrorString (error_str.c_str());
}
}
else
{
error.SetErrorStringWithFormat("'A' packet returned an error: %i.\n", arg_packet_err);
}
SetID (m_gdb_comm.GetCurrentProcessID (m_packet_timeout));
}
}
if (GetID() == LLDB_INVALID_PROCESS_ID)
{
KillDebugserverProcess ();
return error;
}
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse("?", 1, response, m_packet_timeout, false))
SetPrivateState (SetThreadStopInfo (response));
}
else
{
// Set our user ID to an invalid process ID.
SetID(LLDB_INVALID_PROCESS_ID);
error.SetErrorStringWithFormat("Failed to get object file from '%s' for arch %s.\n", module->GetFileSpec().GetFilename().AsCString(), module->GetArchitecture().AsCString());
}
return error;
Cleaned up a unused member variable in Debugger. Added the start of Host specific launch services, though it currently isn't hookup up to anything. We want to be able to launch a process and use the native launch services to launch an app like it would be launched by the user double clicking on the app. We also eventually want to be able to run a command line app in a newly spawned terminal to avoid terminal sharing. Fixed an issue with the new DWARF forward type declaration stuff. A crasher was found that was happening when trying to properly expand the forward declarations. llvm-svn: 115213
2010-09-30 21:49:03 +00:00
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
Error
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
ProcessGDBRemote::ConnectToDebugserver (const char *connect_url)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Error error;
// Sleep and wait a bit for debugserver to start to listen...
std::auto_ptr<ConnectionFileDescriptor> conn_ap(new ConnectionFileDescriptor());
if (conn_ap.get())
{
const uint32_t max_retry_count = 50;
uint32_t retry_count = 0;
while (!m_gdb_comm.IsConnected())
{
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
if (conn_ap->Connect(connect_url, &error) == eConnectionStatusSuccess)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
m_gdb_comm.SetConnection (conn_ap.release());
break;
}
retry_count++;
if (retry_count >= max_retry_count)
break;
usleep (100000);
}
}
if (!m_gdb_comm.IsConnected())
{
if (error.Success())
error.SetErrorString("not connected to remote gdb server");
return error;
}
if (m_gdb_comm.StartReadThread(&error))
{
// Send an initial ack
Sped up the shutdown time on MacOSX by quite a bit by making sure any threads that we spawn let us know when they are going away and that we don't timeout waiting for a message from threads that have gone away. We also now don't expect the "k" packet (kill) to send a response. This greatly speeds up debugger shutdown performance. The test suite now runs quite a bit faster. Added a fix to the variable display code that fixes the display of base classes. We were assuming the virtual or normal base class offsets were being given in bit sizes, but they were being given as character sizes, so we needed to multiply the offset by 8. This wasn't affecting the expression parser, but it was affecting the correct display of C++ class base classes and all of their children. llvm-svn: 124024
2011-01-22 07:12:45 +00:00
m_gdb_comm.SendAck();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (m_debugserver_pid != LLDB_INVALID_PROCESS_ID)
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
m_debugserver_thread = Host::StartMonitoringChildProcess (MonitorDebugserverProcess,
this,
m_debugserver_pid,
false);
Added the ability to detect which vCont packets (using the "vCont?") packet are supported by the remote GDB target. We can also now deal with the lack of vCont support and send packets that the remote GDB stub can use. We also error out of the continue if LLDB tries to do something too complex when vCont isn't supported. llvm-svn: 125433
2011-02-12 06:28:37 +00:00
m_gdb_comm.ResetDiscoverableSettings();
m_gdb_comm.GetSendAcks ();
m_gdb_comm.GetThreadSuffixSupported ();
m_gdb_comm.GetHostInfo ();
m_gdb_comm.GetVContSupported ('c');
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
return error;
}
void
ProcessGDBRemote::DidLaunchOrAttach ()
{
Added a quicker lookup in the SectionLoadList when looking things up by section by using a DenseMap. Fixed some logging calls to get the log shared pointer. llvm-svn: 124926
2011-02-05 02:25:06 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (log)
log->Printf ("ProcessGDBRemote::DidLaunch()");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (GetID() == LLDB_INVALID_PROCESS_ID)
{
m_dynamic_loader_ap.reset();
}
else
{
m_dispatch_queue_offsets_addr = LLDB_INVALID_ADDRESS;
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
BuildDynamicRegisterInfo (false);
Fixed Process::Halt() as it was broken for "process halt" after recent changes to the DoHalt down in ProcessGDBRemote. I also moved the functionality that was in ProcessGDBRemote::DoHalt up into Process::Halt so not every class has to implement a tricky halt/resume on the internal state thread. The functionality is the same as it was before with two changes: - when we eat the event we now just reuse the event we consume when the private state thread is paused and set the interrupted bool on the event if needed - we also properly update the Process::m_public_state with the state of the event we consume. Prior to this, if you issued a "process halt" it would eat the event, not update the process state, and then produce a new event with the interrupted bit set and send it. Anyone listening to the event would get the stopped event with a process that whose state was set to "running". Fixed debugserver to not have to be spawned with the architecture of the inferior process. This worked fine for launching processes, but when attaching to processes by name or pid without a file in lldb, it would fail. Now debugserver can support multiple architectures for a native debug session on the current host. This currently means i386 and x86_64 are supported in the same binary and a x86_64 debugserver can attach to a i386 executable. This change involved a lot of changes to make sure we dynamically detect the correct registers for the inferior process. llvm-svn: 119680
2010-11-18 05:57:03 +00:00
m_byte_order = m_gdb_comm.GetByteOrder();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
StreamString strm;
// See if the GDB server supports the qHostInfo information
const char *vendor = m_gdb_comm.GetVendorString().AsCString();
const char *os_type = m_gdb_comm.GetOSString().AsCString();
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
ArchSpec target_arch (GetTarget().GetArchitecture());
ArchSpec gdb_remote_arch (m_gdb_comm.GetHostArchitecture());
Fix comment text. llvm-svn: 125162
2011-02-09 03:12:09 +00:00
// If the remote host is ARM and we have apple as the vendor, then
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
// ARM executables and shared libraries can have mixed ARM architectures.
// You can have an armv6 executable, and if the host is armv7, then the
// system will load the best possible architecture for all shared libraries
// it has, so we really need to take the remote host architecture as our
// defacto architecture in this case.
if (gdb_remote_arch == ArchSpec ("arm") &&
vendor != NULL &&
strcmp(vendor, "apple") == 0)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
GetTarget().SetArchitecture (gdb_remote_arch);
target_arch = gdb_remote_arch;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
if (!target_arch.IsValid())
target_arch = gdb_remote_arch;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
if (target_arch.IsValid())
{
if (vendor == NULL)
vendor = Host::GetVendorString().AsCString("apple");
if (os_type == NULL)
os_type = Host::GetOSString().AsCString("darwin");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
strm.Printf ("%s-%s-%s", target_arch.AsCString(), vendor, os_type);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
<rdar://problem/8972204> Test failure: ./dotest.py -v -t -f UniversalTestCase.test_process_launch_for_universal Fix for bad architecture settings that were being used from the qHostInfo. llvm-svn: 125161
2011-02-09 03:09:55 +00:00
std::transform (strm.GetString().begin(),
strm.GetString().end(),
strm.GetString().begin(),
::tolower);
m_target_triple.SetCString(strm.GetString().c_str());
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
}
void
ProcessGDBRemote::DidLaunch ()
{
DidLaunchOrAttach ();
if (m_dynamic_loader_ap.get())
m_dynamic_loader_ap->DidLaunch();
}
Error
Merged Eli Friedman's linux build changes where he added Makefile files that enabled LLVM make style building and made this compile LLDB on Mac OS X. We can now iterate on this to make the build work on both linux and macosx. llvm-svn: 108009
2010-07-09 20:39:50 +00:00
ProcessGDBRemote::DoAttachToProcessWithID (lldb::pid_t attach_pid)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Error error;
// Clear out and clean up from any current state
Clear();
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
ArchSpec arch_spec = GetTarget().GetArchitecture();
Added a quicker lookup in the SectionLoadList when looking things up by section by using a DenseMap. Fixed some logging calls to get the log shared pointer. llvm-svn: 124926
2011-02-05 02:25:06 +00:00
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (attach_pid != LLDB_INVALID_PROCESS_ID)
{
char host_port[128];
snprintf (host_port, sizeof(host_port), "localhost:%u", get_random_port ());
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
char connect_url[128];
snprintf (connect_url, sizeof(connect_url), "connect://%s", host_port);
Added the ability to disable ASLR (Address Space Layout Randomization). ASLR is disabled by default, and can be enabled using: (lldb) set disable-aslr 0 llvm-svn: 112616
2010-08-31 18:35:14 +00:00
error = StartDebugserverProcess (host_port, // debugserver_url
NULL, // inferior_argv
NULL, // inferior_envp
NULL, // stdin_path
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
NULL, // stdout_path
NULL, // stderr_path
NULL, // working_dir
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
false, // launch_process == false (we are attaching)
LLDB_INVALID_PROCESS_ID, // Don't send any attach to pid options to debugserver
NULL, // Don't send any attach by process name option to debugserver
false, // Don't send any attach wait_for_launch flag as an option to debugserver
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
0, // launch_flags
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
arch_spec);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Fail())
{
const char *error_string = error.AsCString();
if (error_string == NULL)
error_string = "unable to launch " DEBUGSERVER_BASENAME;
SetExitStatus (-1, error_string);
}
else
{
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
error = ConnectToDebugserver (connect_url);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Success())
{
char packet[64];
const int packet_len = ::snprintf (packet, sizeof(packet), "vAttach;%x", attach_pid);
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
m_async_broadcaster.BroadcastEvent (eBroadcastBitAsyncContinue, new EventDataBytes (packet, packet_len));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
}
}
return error;
}
size_t
ProcessGDBRemote::AttachInputReaderCallback
(
void *baton,
InputReader *reader,
lldb::InputReaderAction notification,
const char *bytes,
size_t bytes_len
)
{
if (notification == eInputReaderGotToken)
{
ProcessGDBRemote *gdb_process = (ProcessGDBRemote *)baton;
if (gdb_process->m_waiting_for_attach)
gdb_process->m_waiting_for_attach = false;
reader->SetIsDone(true);
return 1;
}
return 0;
}
Error
Merged Eli Friedman's linux build changes where he added Makefile files that enabled LLVM make style building and made this compile LLDB on Mac OS X. We can now iterate on this to make the build work on both linux and macosx. llvm-svn: 108009
2010-07-09 20:39:50 +00:00
ProcessGDBRemote::DoAttachToProcessWithName (const char *process_name, bool wait_for_launch)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Error error;
// Clear out and clean up from any current state
Clear();
// HACK: require arch be set correctly at the target level until we can
// figure out a good way to determine the arch of what we are attaching to
if (process_name && process_name[0])
{
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
ArchSpec arch_spec = GetTarget().GetArchitecture();
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
char host_port[128];
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
snprintf (host_port, sizeof(host_port), "localhost:%u", get_random_port ());
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
char connect_url[128];
snprintf (connect_url, sizeof(connect_url), "connect://%s", host_port);
Added the ability to disable ASLR (Address Space Layout Randomization). ASLR is disabled by default, and can be enabled using: (lldb) set disable-aslr 0 llvm-svn: 112616
2010-08-31 18:35:14 +00:00
error = StartDebugserverProcess (host_port, // debugserver_url
NULL, // inferior_argv
NULL, // inferior_envp
NULL, // stdin_path
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
NULL, // stdout_path
NULL, // stderr_path
NULL, // working_dir
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
false, // launch_process == false (we are attaching)
LLDB_INVALID_PROCESS_ID, // Don't send any attach to pid options to debugserver
NULL, // Don't send any attach by process name option to debugserver
false, // Don't send any attach wait_for_launch flag as an option to debugserver
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
0, // launch_flags
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
arch_spec);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Fail())
{
const char *error_string = error.AsCString();
if (error_string == NULL)
error_string = "unable to launch " DEBUGSERVER_BASENAME;
SetExitStatus (-1, error_string);
}
else
{
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
error = ConnectToDebugserver (connect_url);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Success())
{
StreamString packet;
if (wait_for_launch)
Fixed debugserver to properly attach to a process by name with the "vAttachName;<PROCNAME>" packet, and wait for a new process by name to launch with the "vAttachWait;<PROCNAME>". Fixed a few issues with attaching where if DoAttach() returned no error, yet there was no valid process ID, we would deadlock waiting for an event that would never happen. Added a new "process launch" option "--tty" that will launch the process in a new terminal if the Host layer supports the "Host::LaunchInNewTerminal(...)" function. This currently works on MacOSX and will allow the debugging of terminal applications that do complex operations with the terminal. Cleaned up the output when the process resumes, stops and halts to be consistent with the output format. llvm-svn: 116693
2010-10-18 01:45:30 +00:00
packet.PutCString("vAttachWait");
else
packet.PutCString("vAttachName");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
packet.PutChar(';');
Endian patch from Kirk Beitz that allows better cross platform building. llvm-svn: 124643
2011-02-01 01:31:41 +00:00
packet.PutBytesAsRawHex8(process_name, strlen(process_name), lldb::endian::InlHostByteOrder(), lldb::endian::InlHostByteOrder());
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
m_async_broadcaster.BroadcastEvent (eBroadcastBitAsyncContinue, new EventDataBytes (packet.GetData(), packet.GetSize()));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
}
}
return error;
}
void
ProcessGDBRemote::DidAttach ()
{
Added support for attaching to a remote debug server with the new command: (lldb) process connect <remote-url> Currently when you specify a file with the file command it helps us to find a process plug-in that is suitable for debugging. If you specify a file you can rely upon this to find the correct debugger plug-in: % lldb a.out Current executable set to 'a.out' (x86_64). (lldb) process connect connect://localhost:2345 ... If you don't specify a file, you will need to specify the plug-in name that you wish to use: % lldb (lldb) process connect --plugin process.gdb-remote connect://localhost:2345 Other connection URL examples: (lldb) process connect connect://localhost:2345 (lldb) process connect tcp://127.0.0.1 (lldb) process connect file:///dev/ttyS1 We are currently treating the "connect://host:port" as a way to do raw socket connections. If there is a URL for this already, please let me know and we will adopt it. So now you can connect to a remote debug server with the ProcessGDBRemote plug-in. After connection, it will ask for the pid info using the "qC" packet and if it responds with a valid process ID, it will be equivalent to attaching. If it response with an error or invalid process ID, the LLDB process will be in a new state: eStateConnected. This allows us to then download a program or specify the program to run (using the 'A' packet), or specify a process to attach to (using the "vAttach" packets), or query info about the processes that might be available. llvm-svn: 124846
2011-02-04 01:58:07 +00:00
DidLaunchOrAttach ();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (m_dynamic_loader_ap.get())
m_dynamic_loader_ap->DidAttach();
}
Error
ProcessGDBRemote::WillResume ()
{
Added the ability to detect which vCont packets (using the "vCont?") packet are supported by the remote GDB target. We can also now deal with the lack of vCont support and send packets that the remote GDB stub can use. We also error out of the continue if LLDB tries to do something too complex when vCont isn't supported. llvm-svn: 125433
2011-02-12 06:28:37 +00:00
m_continue_c_tids.clear();
m_continue_C_tids.clear();
m_continue_s_tids.clear();
m_continue_S_tids.clear();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
return Error();
}
Error
ProcessGDBRemote::DoResume ()
{
Added an "Interrupted" bit to the ProcessEventData. Halt now generates an event with the Interrupted bit set. Process::HandlePrivateEvent ignores Interrupted events. DoHalt is changed to ensure that the stop even is processed, and an event with the Interrupted event is posted. Finally ClangFunction is rationalized to use this facility so the that Halt is handled more deterministically. llvm-svn: 119453
2010-11-17 02:32:00 +00:00
Error error;
Added a quicker lookup in the SectionLoadList when looking things up by section by using a DenseMap. Fixed some logging calls to get the log shared pointer. llvm-svn: 124926
2011-02-05 02:25:06 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (log)
log->Printf ("ProcessGDBRemote::Resume()");
Fixed a race condition that could cause ProcessGDBRemote::DoResume() to return an error saying the resume timed out. Previously the thread that was trying to resume the process would eventually call ProcessGDBRemote::DoResume() which would broadcast an event over to the async GDB remote thread which would sent the continue packet to the remote gdb server. Right after this was sent, it would set a predicate boolean value (protected by a mutex and condition) and then the thread that issued the ProcessGDBRemote::DoResume() would then wait for that condition variable to be set. If the async gdb thread was too quick though, the predicate boolean value could have been set to true and back to false by the time the thread that issued the ProcessGDBRemote::DoResume() checks the boolean value. So we can't use the predicate value as a handshake. I have changed the code over to using a Event by having the GDB remote communication object post an event: GDBRemoteCommunication::eBroadcastBitRunPacketSent This allows reliable handshaking between the two threads and avoids the erroneous ProcessGDBRemote::DoResume() errors. Added a host backtrace service to allow in process backtraces when trying to track down tricky issues. I need to see if LLVM has any backtracing abilities abstracted in it already, and if so, use that, but I needed something ASAP for the current issue I was working on. The static function is: void Host::Backtrace (Stream &strm, uint32_t max_frames); And it will backtrace at most "max_frames" frames for the current thread and can be used with any of the Stream subclasses for logging. llvm-svn: 120793
2010-12-03 06:02:24 +00:00
Listener listener ("gdb-remote.resume-packet-sent");
if (listener.StartListeningForEvents (&m_gdb_comm, GDBRemoteCommunication::eBroadcastBitRunPacketSent))
{
Added the ability to detect which vCont packets (using the "vCont?") packet are supported by the remote GDB target. We can also now deal with the lack of vCont support and send packets that the remote GDB stub can use. We also error out of the continue if LLDB tries to do something too complex when vCont isn't supported. llvm-svn: 125433
2011-02-12 06:28:37 +00:00
StreamString continue_packet;
bool continue_packet_error = false;
if (m_gdb_comm.HasAnyVContSupport ())
{
continue_packet.PutCString ("vCont");
if (!m_continue_c_tids.empty())
{
if (m_gdb_comm.GetVContSupported ('c'))
{
for (tid_collection::const_iterator t_pos = m_continue_c_tids.begin(), t_end = m_continue_c_tids.end(); t_pos != t_end; ++t_pos)
continue_packet.Printf(";c:%4.4x", *t_pos);
}
else
continue_packet_error = true;
}
if (!continue_packet_error && !m_continue_C_tids.empty())
{
if (m_gdb_comm.GetVContSupported ('C'))
{
for (tid_sig_collection::const_iterator s_pos = m_continue_C_tids.begin(), s_end = m_continue_C_tids.end(); s_pos != s_end; ++s_pos)
continue_packet.Printf(";C%2.2x:%4.4x", s_pos->second, s_pos->first);
}
else
continue_packet_error = true;
}
if (!continue_packet_error && !m_continue_s_tids.empty())
{
if (m_gdb_comm.GetVContSupported ('s'))
{
for (tid_collection::const_iterator t_pos = m_continue_s_tids.begin(), t_end = m_continue_s_tids.end(); t_pos != t_end; ++t_pos)
continue_packet.Printf(";s:%4.4x", *t_pos);
}
else
continue_packet_error = true;
}
if (!continue_packet_error && !m_continue_S_tids.empty())
{
if (m_gdb_comm.GetVContSupported ('S'))
{
for (tid_sig_collection::const_iterator s_pos = m_continue_S_tids.begin(), s_end = m_continue_S_tids.end(); s_pos != s_end; ++s_pos)
continue_packet.Printf(";S%2.2x:%4.4x", s_pos->second, s_pos->first);
}
else
continue_packet_error = true;
}
if (continue_packet_error)
continue_packet.GetString().clear();
}
else
continue_packet_error = true;
if (continue_packet_error)
{
continue_packet_error = false;
// Either no vCont support, or we tried to use part of the vCont
// packet that wasn't supported by the remote GDB server.
// We need to try and make a simple packet that can do our continue
const size_t num_threads = GetThreadList().GetSize();
const size_t num_continue_c_tids = m_continue_c_tids.size();
const size_t num_continue_C_tids = m_continue_C_tids.size();
const size_t num_continue_s_tids = m_continue_s_tids.size();
const size_t num_continue_S_tids = m_continue_S_tids.size();
if (num_continue_c_tids > 0)
{
if (num_continue_c_tids == num_threads)
{
// All threads are resuming...
SetCurrentGDBRemoteThreadForRun (-1);
continue_packet.PutChar ('c');
}
else if (num_continue_c_tids == 1 &&
num_continue_C_tids == 0 &&
num_continue_s_tids == 0 &&
num_continue_S_tids == 0 )
{
// Only one thread is continuing
SetCurrentGDBRemoteThreadForRun (m_continue_c_tids.front());
continue_packet.PutChar ('c');
}
else
{
// We can't represent this continue packet....
continue_packet_error = true;
}
}
if (!continue_packet_error && num_continue_C_tids > 0)
{
if (num_continue_C_tids == num_threads)
{
const int continue_signo = m_continue_C_tids.front().second;
if (num_continue_C_tids > 1)
{
for (size_t i=1; i<num_threads; ++i)
{
if (m_continue_C_tids[i].second != continue_signo)
continue_packet_error = true;
}
}
if (!continue_packet_error)
{
// Add threads continuing with the same signo...
SetCurrentGDBRemoteThreadForRun (-1);
continue_packet.Printf("C%2.2x", continue_signo);
}
}
else if (num_continue_c_tids == 0 &&
num_continue_C_tids == 1 &&
num_continue_s_tids == 0 &&
num_continue_S_tids == 0 )
{
// Only one thread is continuing with signal
SetCurrentGDBRemoteThreadForRun (m_continue_C_tids.front().first);
continue_packet.Printf("C%2.2x", m_continue_C_tids.front().second);
}
else
{
// We can't represent this continue packet....
continue_packet_error = true;
}
}
if (!continue_packet_error && num_continue_s_tids > 0)
{
if (num_continue_s_tids == num_threads)
{
// All threads are resuming...
SetCurrentGDBRemoteThreadForRun (-1);
continue_packet.PutChar ('s');
}
else if (num_continue_c_tids == 0 &&
num_continue_C_tids == 0 &&
num_continue_s_tids == 1 &&
num_continue_S_tids == 0 )
{
// Only one thread is stepping
SetCurrentGDBRemoteThreadForRun (m_continue_s_tids.front());
continue_packet.PutChar ('s');
}
else
{
// We can't represent this continue packet....
continue_packet_error = true;
}
}
if (!continue_packet_error && num_continue_S_tids > 0)
{
if (num_continue_S_tids == num_threads)
{
const int step_signo = m_continue_S_tids.front().second;
// Are all threads trying to step with the same signal?
if (num_continue_S_tids > 1)
{
for (size_t i=1; i<num_threads; ++i)
{
if (m_continue_S_tids[i].second != step_signo)
continue_packet_error = true;
}
}
if (!continue_packet_error)
{
// Add threads stepping with the same signo...
SetCurrentGDBRemoteThreadForRun (-1);
continue_packet.Printf("S%2.2x", step_signo);
}
}
else if (num_continue_c_tids == 0 &&
num_continue_C_tids == 0 &&
num_continue_s_tids == 0 &&
num_continue_S_tids == 1 )
{
// Only one thread is stepping with signal
SetCurrentGDBRemoteThreadForRun (m_continue_S_tids.front().first);
continue_packet.Printf("S%2.2x", m_continue_S_tids.front().second);
}
else
{
// We can't represent this continue packet....
continue_packet_error = true;
}
}
}
if (continue_packet_error)
{
error.SetErrorString ("can't make continue packet for this resume");
}
else
{
EventSP event_sp;
TimeValue timeout;
timeout = TimeValue::Now();
timeout.OffsetWithSeconds (5);
m_async_broadcaster.BroadcastEvent (eBroadcastBitAsyncContinue, new EventDataBytes (continue_packet.GetData(), continue_packet.GetSize()));
if (listener.WaitForEvent (&timeout, event_sp) == false)
error.SetErrorString("Resume timed out.");
}
Fixed a race condition that could cause ProcessGDBRemote::DoResume() to return an error saying the resume timed out. Previously the thread that was trying to resume the process would eventually call ProcessGDBRemote::DoResume() which would broadcast an event over to the async GDB remote thread which would sent the continue packet to the remote gdb server. Right after this was sent, it would set a predicate boolean value (protected by a mutex and condition) and then the thread that issued the ProcessGDBRemote::DoResume() would then wait for that condition variable to be set. If the async gdb thread was too quick though, the predicate boolean value could have been set to true and back to false by the time the thread that issued the ProcessGDBRemote::DoResume() checks the boolean value. So we can't use the predicate value as a handshake. I have changed the code over to using a Event by having the GDB remote communication object post an event: GDBRemoteCommunication::eBroadcastBitRunPacketSent This allows reliable handshaking between the two threads and avoids the erroneous ProcessGDBRemote::DoResume() errors. Added a host backtrace service to allow in process backtraces when trying to track down tricky issues. I need to see if LLVM has any backtracing abilities abstracted in it already, and if so, use that, but I needed something ASAP for the current issue I was working on. The static function is: void Host::Backtrace (Stream &strm, uint32_t max_frames); And it will backtrace at most "max_frames" frames for the current thread and can be used with any of the Stream subclasses for logging. llvm-svn: 120793
2010-12-03 06:02:24 +00:00
}
Added an "Interrupted" bit to the ProcessEventData. Halt now generates an event with the Interrupted bit set. Process::HandlePrivateEvent ignores Interrupted events. DoHalt is changed to ensure that the stop even is processed, and an event with the Interrupted event is posted. Finally ClangFunction is rationalized to use this facility so the that Halt is handled more deterministically. llvm-svn: 119453
2010-11-17 02:32:00 +00:00
return error;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
size_t
ProcessGDBRemote::GetSoftwareBreakpointTrapOpcode (BreakpointSite* bp_site)
{
const uint8_t *trap_opcode = NULL;
uint32_t trap_opcode_size = 0;
static const uint8_t g_arm_breakpoint_opcode[] = { 0xFE, 0xDE, 0xFF, 0xE7 };
//static const uint8_t g_thumb_breakpooint_opcode[] = { 0xFE, 0xDE };
static const uint8_t g_ppc_breakpoint_opcode[] = { 0x7F, 0xC0, 0x00, 0x08 };
static const uint8_t g_i386_breakpoint_opcode[] = { 0xCC };
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
ArchSpec::CPU arch_cpu = GetTarget().GetArchitecture().GetGenericCPUType();
Made lldb_private::ArchSpec more generic so that it can take a mach-o cpu type and sub-type, or an ELF e_machine value. Also added a generic CPU type to the arch spec class so we can have a single arch definition that the LLDB core code can use. Previously a lot of places in the code were using the mach-o definitions from a macosx header file. Switches over to using "llvm/Support/MachO.h" for the llvm::MachO::XXX for the CPU types and sub types for mach-o ArchSpecs. Added "llvm/Support/ELF.h" so we can use the "llvm::ELF::XXX" defines for the ELF ArchSpecs. Got rid of all CPU_TYPE_ and CPU_SUBTYPE_ defines that were previously being used in LLDB. llvm-svn: 105806
2010-06-11 03:25:34 +00:00
switch (arch_cpu)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Made lldb_private::ArchSpec more generic so that it can take a mach-o cpu type and sub-type, or an ELF e_machine value. Also added a generic CPU type to the arch spec class so we can have a single arch definition that the LLDB core code can use. Previously a lot of places in the code were using the mach-o definitions from a macosx header file. Switches over to using "llvm/Support/MachO.h" for the llvm::MachO::XXX for the CPU types and sub types for mach-o ArchSpecs. Added "llvm/Support/ELF.h" so we can use the "llvm::ELF::XXX" defines for the ELF ArchSpecs. Got rid of all CPU_TYPE_ and CPU_SUBTYPE_ defines that were previously being used in LLDB. llvm-svn: 105806
2010-06-11 03:25:34 +00:00
case ArchSpec::eCPU_i386:
case ArchSpec::eCPU_x86_64:
trap_opcode = g_i386_breakpoint_opcode;
trap_opcode_size = sizeof(g_i386_breakpoint_opcode);
break;
case ArchSpec::eCPU_arm:
// TODO: fill this in for ARM. We need to dig up the symbol for
// the address in the breakpoint locaiton and figure out if it is
// an ARM or Thumb breakpoint.
trap_opcode = g_arm_breakpoint_opcode;
trap_opcode_size = sizeof(g_arm_breakpoint_opcode);
break;
case ArchSpec::eCPU_ppc:
case ArchSpec::eCPU_ppc64:
trap_opcode = g_ppc_breakpoint_opcode;
trap_opcode_size = sizeof(g_ppc_breakpoint_opcode);
break;
default:
assert(!"Unhandled architecture in ProcessMacOSX::GetSoftwareBreakpointTrapOpcode()");
break;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
if (trap_opcode && trap_opcode_size)
{
if (bp_site->SetTrapOpcode(trap_opcode, trap_opcode_size))
return trap_opcode_size;
}
return 0;
}
uint32_t
ProcessGDBRemote::UpdateThreadListIfNeeded ()
{
// locker will keep a mutex locked until it goes out of scope
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_THREAD));
Updated the lldb_private::Flags class to have better method names and made all of the calls inlined in the header file for better performance. Fixed the summary for C string types (array of chars (with any combo if modifiers), and pointers to chars) work in all cases. Fixed an issue where a forward declaration to a clang type could cause itself to resolve itself more than once if, during the resolving of the type itself it caused something to try and resolve itself again. We now remove the clang type from the forward declaration map in the DWARF parser when we start to resolve it and avoid this additional call. This should stop any duplicate members from appearing and throwing all the alignment of structs, unions and classes. llvm-svn: 117437
2010-10-27 03:32:59 +00:00
if (log && log->GetMask().Test(GDBR_LOG_VERBOSE))
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
log->Printf ("ProcessGDBRemote::%s (pid = %i)", __FUNCTION__, GetID());
Fixed the StackFrame to correctly resolve the StackID's SymbolContextScope. Added extra logging for stepping. Fixed an issue where cached stack frame data could be lost between runs when the thread plans read a stack frame. llvm-svn: 112973
2010-09-03 17:10:42 +00:00
Mutex::Locker locker (m_thread_list.GetMutex ());
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
const uint32_t stop_id = GetStopID();
if (m_thread_list.GetSize(false) == 0 || stop_id != m_thread_list.GetStopID())
{
// Update the thread list's stop id immediately so we don't recurse into this function.
ThreadList curr_thread_list (this);
curr_thread_list.SetStopID(stop_id);
Error err;
StringExtractorGDBRemote response;
for (m_gdb_comm.SendPacketAndWaitForResponse("qfThreadInfo", response, 1, false);
response.IsNormalPacket();
m_gdb_comm.SendPacketAndWaitForResponse("qsThreadInfo", response, 1, false))
{
char ch = response.GetChar();
if (ch == 'l')
break;
if (ch == 'm')
{
do
{
tid_t tid = response.GetHexMaxU32(false, LLDB_INVALID_THREAD_ID);
if (tid != LLDB_INVALID_THREAD_ID)
{
ThreadSP thread_sp (GetThreadList().FindThreadByID (tid, false));
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-09 21:07:35 +00:00
if (!thread_sp)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
thread_sp.reset (new ThreadGDBRemote (*this, tid));
curr_thread_list.AddThread(thread_sp);
}
ch = response.GetChar();
} while (ch == ',');
}
}
m_thread_list = curr_thread_list;
SetThreadStopInfo (m_last_stop_packet);
}
return GetThreadList().GetSize(false);
}
StateType
ProcessGDBRemote::SetThreadStopInfo (StringExtractor& stop_packet)
{
const char stop_type = stop_packet.GetChar();
switch (stop_type)
{
case 'T':
case 'S':
{
All UnwindPlan objects are now passed around as shared pointers. ArchDefaultUnwindPlan plug-in interfaces are now cached per architecture instead of being leaked for every frame. Split the ArchDefaultUnwindPlan_x86 into ArchDefaultUnwindPlan_x86_64 and ArchDefaultUnwindPlan_i386 interfaces. There were sporadic crashes that were due to something leaking or being destroyed when doing stack crawls. This patch should clear up these issues. llvm-svn: 125541
2011-02-15 00:19:15 +00:00
if (GetStopID() == 0)
{
// Our first stop, make sure we have a process ID, and also make
// sure we know about our registers
if (GetID() == LLDB_INVALID_PROCESS_ID)
{
lldb::pid_t pid = m_gdb_comm.GetCurrentProcessID (1);
if (pid != LLDB_INVALID_PROCESS_ID)
SetID (pid);
}
BuildDynamicRegisterInfo (true);
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// Stop with signal and thread info
const uint8_t signo = stop_packet.GetHexU8();
std::string name;
std::string value;
std::string thread_name;
uint32_t exc_type = 0;
Centralized the Mach exception stop info code by adding it as a first class citizen on the StopInfo class. llvm-svn: 109235
2010-07-23 16:45:51 +00:00
std::vector<addr_t> exc_data;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
uint32_t tid = LLDB_INVALID_THREAD_ID;
addr_t thread_dispatch_qaddr = LLDB_INVALID_ADDRESS;
uint32_t exc_data_count = 0;
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-09 21:07:35 +00:00
ThreadSP thread_sp;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
while (stop_packet.GetNameColonValue(name, value))
{
if (name.compare("metype") == 0)
{
// exception type in big endian hex
exc_type = Args::StringToUInt32 (value.c_str(), 0, 16);
}
else if (name.compare("mecount") == 0)
{
// exception count in big endian hex
exc_data_count = Args::StringToUInt32 (value.c_str(), 0, 16);
}
else if (name.compare("medata") == 0)
{
// exception data in big endian hex
exc_data.push_back(Args::StringToUInt64 (value.c_str(), 0, 16));
}
else if (name.compare("thread") == 0)
{
// thread in big endian hex
tid = Args::StringToUInt32 (value.c_str(), 0, 16);
All UnwindPlan objects are now passed around as shared pointers. ArchDefaultUnwindPlan plug-in interfaces are now cached per architecture instead of being leaked for every frame. Split the ArchDefaultUnwindPlan_x86 into ArchDefaultUnwindPlan_x86_64 and ArchDefaultUnwindPlan_i386 interfaces. There were sporadic crashes that were due to something leaking or being destroyed when doing stack crawls. This patch should clear up these issues. llvm-svn: 125541
2011-02-15 00:19:15 +00:00
Mutex::Locker locker (m_thread_list.GetMutex ());
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-09 21:07:35 +00:00
thread_sp = m_thread_list.FindThreadByID(tid, false);
All UnwindPlan objects are now passed around as shared pointers. ArchDefaultUnwindPlan plug-in interfaces are now cached per architecture instead of being leaked for every frame. Split the ArchDefaultUnwindPlan_x86 into ArchDefaultUnwindPlan_x86_64 and ArchDefaultUnwindPlan_i386 interfaces. There were sporadic crashes that were due to something leaking or being destroyed when doing stack crawls. This patch should clear up these issues. llvm-svn: 125541
2011-02-15 00:19:15 +00:00
if (!thread_sp)
{
// Create the thread if we need to
thread_sp.reset (new ThreadGDBRemote (*this, tid));
m_thread_list.AddThread(thread_sp);
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
Modified the stop reply packet to be able to send the thread name using the new "hexname" key for the "key:value;" duple that is part of the packet. This allows for thread names to contain special characters such as $ # : ; + - Debugserver now detects if the thread name contains special characters and sends the chars in hex format if needed. llvm-svn: 123053
2011-01-08 03:17:57 +00:00
else if (name.compare("hexname") == 0)
{
StringExtractor name_extractor;
// Swap "value" over into "name_extractor"
name_extractor.GetStringRef().swap(value);
// Now convert the HEX bytes into a string value
name_extractor.GetHexByteString (value);
thread_name.swap (value);
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
else if (name.compare("name") == 0)
{
thread_name.swap (value);
}
Got the ARM version of debugserver up to date. Renamed the "dispatchqaddr" setting that was coming back for stop reply packets to be named "qaddr" so that gdb doesn't thing it is a register number. gdb was checking the first character and assuming "di" was a hex register number because 'd' is a hex digit. It has been shortened so gdb can safely ignore it. llvm-svn: 113475
2010-09-09 06:32:46 +00:00
else if (name.compare("qaddr") == 0)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
thread_dispatch_qaddr = Args::StringToUInt64 (value.c_str(), 0, 16);
}
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-09 21:07:35 +00:00
else if (name.size() == 2 && ::isxdigit(name[0]) && ::isxdigit(name[1]))
{
// We have a register number that contains an expedited
// register value. Lets supply this register to our thread
// so it won't have to go and read it.
if (thread_sp)
{
uint32_t reg = Args::StringToUInt32 (name.c_str(), UINT32_MAX, 16);
if (reg != UINT32_MAX)
{
StringExtractor reg_value_extractor;
// Swap "value" over into "reg_value_extractor"
reg_value_extractor.GetStringRef().swap(value);
All UnwindPlan objects are now passed around as shared pointers. ArchDefaultUnwindPlan plug-in interfaces are now cached per architecture instead of being leaked for every frame. Split the ArchDefaultUnwindPlan_x86 into ArchDefaultUnwindPlan_x86_64 and ArchDefaultUnwindPlan_i386 interfaces. There were sporadic crashes that were due to something leaking or being destroyed when doing stack crawls. This patch should clear up these issues. llvm-svn: 125541
2011-02-15 00:19:15 +00:00
if (!static_cast<ThreadGDBRemote *> (thread_sp.get())->PrivateSetRegisterValue (reg, reg_value_extractor))
{
Host::SetCrashDescriptionWithFormat("Setting thread register '%s' (decoded to %u (0x%x)) with value '%s' for stop packet: '%s'",
name.c_str(),
reg,
reg,
reg_value_extractor.GetStringRef().c_str(),
stop_packet.GetStringRef().c_str());
}
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-09 21:07:35 +00:00
}
}
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
if (thread_sp)
{
ThreadGDBRemote *gdb_thread = static_cast<ThreadGDBRemote *> (thread_sp.get());
gdb_thread->SetThreadDispatchQAddr (thread_dispatch_qaddr);
Fix a little thinko in sending down the thread name to SetName. llvm-svn: 124464
2011-01-28 02:23:12 +00:00
gdb_thread->SetName (thread_name.empty() ? NULL : thread_name.c_str());
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (exc_type != 0)
{
Enabled extra warnings and fixed a bunch of small issues. llvm-svn: 124250
2011-01-25 23:55:37 +00:00
const size_t exc_data_size = exc_data.size();
Abtracted the old "lldb_private::Thread::StopInfo" into an abtract class. This will allow debugger plug-ins to make any instance of "lldb_private::StopInfo" that can completely describe any stop reason. It also provides a framework for doing intelligent things with the stop info at important times in the lifetime of the inferior. Examples include the signal stop info in StopInfoUnixSignal. It will check with the process to see that the current action is for the signal. These actions include wether to stop for the signal, wether the notify that the signal was hit, and wether to pass the signal along to the inferior process. The StopInfoUnixSignal class overrides the "ShouldStop()" method of StopInfo and this allows the stop info to determine if it should stop at the signal or continue the process. StopInfo subclasses must override the following functions: virtual lldb::StopReason GetStopReason () const = 0; virtual const char * GetDescription () = 0; StopInfo subclasses can override the following functions: // If the subclass returns "false", the inferior will resume. The default // version of this function returns "true" which means the default stop // info will stop the process. The breakpoint subclass will check if // the breakpoint wants us to stop by calling any installed callback on // the breakpoint, and also checking if the breakpoint is for the current // thread. Signals will check if they should stop based off of the // UnixSignal settings in the process. virtual bool ShouldStop (Event *event_ptr); // Sublasses can state if they want to notify the debugger when "ShouldStop" // returns false. This would be handy for breakpoints where you want to // log information and continue and is also used by the signal stop info // to notify that a signal was received (after it checks with the process // signal settings). virtual bool ShouldNotify (Event *event_ptr) { return false; } // Allow subclasses to do something intelligent right before we resume. // The signal class will figure out if the signal should be propagated // to the inferior process and pass that along to the debugger plug-ins. virtual void WillResume (lldb::StateType resume_state) { // By default, don't do anything } The support the Mach exceptions was moved into the lldb/source/Plugins/Process/Utility folder and now doesn't polute the lldb_private::Thread class with platform specific code. llvm-svn: 110184
2010-08-04 01:40:35 +00:00
gdb_thread->SetStopInfo (StopInfoMachException::CreateStopReasonWithMachException (*thread_sp,
exc_type,
Enabled extra warnings and fixed a bunch of small issues. llvm-svn: 124250
2011-01-25 23:55:37 +00:00
exc_data_size,
exc_data_size >= 1 ? exc_data[0] : 0,
exc_data_size >= 2 ? exc_data[1] : 0));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
else if (signo)
{
Abtracted the old "lldb_private::Thread::StopInfo" into an abtract class. This will allow debugger plug-ins to make any instance of "lldb_private::StopInfo" that can completely describe any stop reason. It also provides a framework for doing intelligent things with the stop info at important times in the lifetime of the inferior. Examples include the signal stop info in StopInfoUnixSignal. It will check with the process to see that the current action is for the signal. These actions include wether to stop for the signal, wether the notify that the signal was hit, and wether to pass the signal along to the inferior process. The StopInfoUnixSignal class overrides the "ShouldStop()" method of StopInfo and this allows the stop info to determine if it should stop at the signal or continue the process. StopInfo subclasses must override the following functions: virtual lldb::StopReason GetStopReason () const = 0; virtual const char * GetDescription () = 0; StopInfo subclasses can override the following functions: // If the subclass returns "false", the inferior will resume. The default // version of this function returns "true" which means the default stop // info will stop the process. The breakpoint subclass will check if // the breakpoint wants us to stop by calling any installed callback on // the breakpoint, and also checking if the breakpoint is for the current // thread. Signals will check if they should stop based off of the // UnixSignal settings in the process. virtual bool ShouldStop (Event *event_ptr); // Sublasses can state if they want to notify the debugger when "ShouldStop" // returns false. This would be handy for breakpoints where you want to // log information and continue and is also used by the signal stop info // to notify that a signal was received (after it checks with the process // signal settings). virtual bool ShouldNotify (Event *event_ptr) { return false; } // Allow subclasses to do something intelligent right before we resume. // The signal class will figure out if the signal should be propagated // to the inferior process and pass that along to the debugger plug-ins. virtual void WillResume (lldb::StateType resume_state) { // By default, don't do anything } The support the Mach exceptions was moved into the lldb/source/Plugins/Process/Utility folder and now doesn't polute the lldb_private::Thread class with platform specific code. llvm-svn: 110184
2010-08-04 01:40:35 +00:00
gdb_thread->SetStopInfo (StopInfo::CreateStopReasonWithSignal (*thread_sp, signo));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
else
{
Abtracted the old "lldb_private::Thread::StopInfo" into an abtract class. This will allow debugger plug-ins to make any instance of "lldb_private::StopInfo" that can completely describe any stop reason. It also provides a framework for doing intelligent things with the stop info at important times in the lifetime of the inferior. Examples include the signal stop info in StopInfoUnixSignal. It will check with the process to see that the current action is for the signal. These actions include wether to stop for the signal, wether the notify that the signal was hit, and wether to pass the signal along to the inferior process. The StopInfoUnixSignal class overrides the "ShouldStop()" method of StopInfo and this allows the stop info to determine if it should stop at the signal or continue the process. StopInfo subclasses must override the following functions: virtual lldb::StopReason GetStopReason () const = 0; virtual const char * GetDescription () = 0; StopInfo subclasses can override the following functions: // If the subclass returns "false", the inferior will resume. The default // version of this function returns "true" which means the default stop // info will stop the process. The breakpoint subclass will check if // the breakpoint wants us to stop by calling any installed callback on // the breakpoint, and also checking if the breakpoint is for the current // thread. Signals will check if they should stop based off of the // UnixSignal settings in the process. virtual bool ShouldStop (Event *event_ptr); // Sublasses can state if they want to notify the debugger when "ShouldStop" // returns false. This would be handy for breakpoints where you want to // log information and continue and is also used by the signal stop info // to notify that a signal was received (after it checks with the process // signal settings). virtual bool ShouldNotify (Event *event_ptr) { return false; } // Allow subclasses to do something intelligent right before we resume. // The signal class will figure out if the signal should be propagated // to the inferior process and pass that along to the debugger plug-ins. virtual void WillResume (lldb::StateType resume_state) { // By default, don't do anything } The support the Mach exceptions was moved into the lldb/source/Plugins/Process/Utility folder and now doesn't polute the lldb_private::Thread class with platform specific code. llvm-svn: 110184
2010-08-04 01:40:35 +00:00
StopInfoSP invalid_stop_info_sp;
gdb_thread->SetStopInfo (invalid_stop_info_sp);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
}
return eStateStopped;
}
break;
case 'W':
// process exited
return eStateExited;
default:
break;
}
return eStateInvalid;
}
void
ProcessGDBRemote::RefreshStateAfterStop ()
{
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
// FIXME - add a variable to tell that we're in the middle of attaching if we
// need to know that.
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// We must be attaching if we don't already have a valid architecture
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
// if (!GetTarget().GetArchitecture().IsValid())
// {
// Module *exe_module = GetTarget().GetExecutableModule().get();
// if (exe_module)
// m_arch_spec = exe_module->GetArchitecture();
// }
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// Let all threads recover from stopping and do any clean up based
// on the previous thread state (if any).
m_thread_list.RefreshStateAfterStop();
// Discover new threads:
UpdateThreadListIfNeeded ();
}
Error
Added an "Interrupted" bit to the ProcessEventData. Halt now generates an event with the Interrupted bit set. Process::HandlePrivateEvent ignores Interrupted events. DoHalt is changed to ensure that the stop even is processed, and an event with the Interrupted event is posted. Finally ClangFunction is rationalized to use this facility so the that Halt is handled more deterministically. llvm-svn: 119453
2010-11-17 02:32:00 +00:00
ProcessGDBRemote::DoHalt (bool &caused_stop)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Error error;
Added an "Interrupted" bit to the ProcessEventData. Halt now generates an event with the Interrupted bit set. Process::HandlePrivateEvent ignores Interrupted events. DoHalt is changed to ensure that the stop even is processed, and an event with the Interrupted event is posted. Finally ClangFunction is rationalized to use this facility so the that Halt is handled more deterministically. llvm-svn: 119453
2010-11-17 02:32:00 +00:00
Sped up the shutdown time on MacOSX by quite a bit by making sure any threads that we spawn let us know when they are going away and that we don't timeout waiting for a message from threads that have gone away. We also now don't expect the "k" packet (kill) to send a response. This greatly speeds up debugger shutdown performance. The test suite now runs quite a bit faster. Added a fix to the variable display code that fixes the display of base classes. We were assuming the virtual or normal base class offsets were being given in bit sizes, but they were being given as character sizes, so we needed to multiply the offset by 8. This wasn't affecting the expression parser, but it was affecting the correct display of C++ class base classes and all of their children. llvm-svn: 124024
2011-01-22 07:12:45 +00:00
bool timed_out = false;
Mutex::Locker locker;
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
if (m_public_state.GetValue() == eStateAttaching)
Fixed Process::Halt() as it was broken for "process halt" after recent changes to the DoHalt down in ProcessGDBRemote. I also moved the functionality that was in ProcessGDBRemote::DoHalt up into Process::Halt so not every class has to implement a tricky halt/resume on the internal state thread. The functionality is the same as it was before with two changes: - when we eat the event we now just reuse the event we consume when the private state thread is paused and set the interrupted bool on the event if needed - we also properly update the Process::m_public_state with the state of the event we consume. Prior to this, if you issued a "process halt" it would eat the event, not update the process state, and then produce a new event with the interrupted bit set and send it. Anyone listening to the event would get the stopped event with a process that whose state was set to "running". Fixed debugserver to not have to be spawned with the architecture of the inferior process. This worked fine for launching processes, but when attaching to processes by name or pid without a file in lldb, it would fail. Now debugserver can support multiple architectures for a native debug session on the current host. This currently means i386 and x86_64 are supported in the same binary and a x86_64 debugserver can attach to a i386 executable. This change involved a lot of changes to make sure we dynamically detect the correct registers for the inferior process. llvm-svn: 119680
2010-11-18 05:57:03 +00:00
{
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
// We are being asked to halt during an attach. We need to just close
// our file handle and debugserver will go away, and we can be done...
m_gdb_comm.Disconnect();
}
else
{
if (!m_gdb_comm.SendInterrupt (locker, 2, caused_stop, timed_out))
{
if (timed_out)
error.SetErrorString("timed out sending interrupt packet");
else
error.SetErrorString("unknown error sending interrupt packet");
}
Fixed Process::Halt() as it was broken for "process halt" after recent changes to the DoHalt down in ProcessGDBRemote. I also moved the functionality that was in ProcessGDBRemote::DoHalt up into Process::Halt so not every class has to implement a tricky halt/resume on the internal state thread. The functionality is the same as it was before with two changes: - when we eat the event we now just reuse the event we consume when the private state thread is paused and set the interrupted bool on the event if needed - we also properly update the Process::m_public_state with the state of the event we consume. Prior to this, if you issued a "process halt" it would eat the event, not update the process state, and then produce a new event with the interrupted bit set and send it. Anyone listening to the event would get the stopped event with a process that whose state was set to "running". Fixed debugserver to not have to be spawned with the architecture of the inferior process. This worked fine for launching processes, but when attaching to processes by name or pid without a file in lldb, it would fail. Now debugserver can support multiple architectures for a native debug session on the current host. This currently means i386 and x86_64 are supported in the same binary and a x86_64 debugserver can attach to a i386 executable. This change involved a lot of changes to make sure we dynamically detect the correct registers for the inferior process. llvm-svn: 119680
2010-11-18 05:57:03 +00:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
return error;
}
Error
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
ProcessGDBRemote::InterruptIfRunning
(
bool discard_thread_plans,
bool catch_stop_event,
EventSP &stop_event_sp
)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Error error;
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
bool paused_private_state_thread = false;
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
const bool is_running = m_gdb_comm.IsRunning();
if (log)
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
log->Printf ("ProcessGDBRemote::InterruptIfRunning(discard_thread_plans=%i, catch_stop_event=%i) is_running=%i",
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
discard_thread_plans,
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
catch_stop_event,
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
is_running);
if (discard_thread_plans)
{
if (log)
log->Printf ("ProcessGDBRemote::InterruptIfRunning() discarding all thread plans");
m_thread_list.DiscardThreadPlans();
}
if (is_running)
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
{
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
if (catch_stop_event)
{
if (log)
log->Printf ("ProcessGDBRemote::InterruptIfRunning() pausing private state thread");
PausePrivateStateThread();
paused_private_state_thread = true;
}
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
bool timed_out = false;
Sped up the shutdown time on MacOSX by quite a bit by making sure any threads that we spawn let us know when they are going away and that we don't timeout waiting for a message from threads that have gone away. We also now don't expect the "k" packet (kill) to send a response. This greatly speeds up debugger shutdown performance. The test suite now runs quite a bit faster. Added a fix to the variable display code that fixes the display of base classes. We were assuming the virtual or normal base class offsets were being given in bit sizes, but they were being given as character sizes, so we needed to multiply the offset by 8. This wasn't affecting the expression parser, but it was affecting the correct display of C++ class base classes and all of their children. llvm-svn: 124024
2011-01-22 07:12:45 +00:00
bool sent_interrupt = false;
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
Mutex::Locker locker;
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
//m_debugserver_pid = LLDB_INVALID_PROCESS_ID;
if (!m_gdb_comm.SendInterrupt (locker, 1, sent_interrupt, timed_out))
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
{
if (timed_out)
error.SetErrorString("timed out sending interrupt packet");
else
error.SetErrorString("unknown error sending interrupt packet");
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
if (paused_private_state_thread)
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
ResumePrivateStateThread();
return error;
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
}
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
if (catch_stop_event)
{
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
// LISTEN HERE
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
TimeValue timeout_time;
timeout_time = TimeValue::Now();
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
timeout_time.OffsetWithSeconds(5);
StateType state = WaitForStateChangedEventsPrivate (&timeout_time, stop_event_sp);
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
Enabled extra warnings and fixed a bunch of small issues. llvm-svn: 124250
2011-01-25 23:55:37 +00:00
timed_out = state == eStateInvalid;
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
if (log)
log->Printf ("ProcessGDBRemote::InterruptIfRunning() catch stop event: state = %s, timed-out=%i", StateAsCString(state), timed_out);
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
if (timed_out)
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
error.SetErrorString("unable to verify target stopped");
}
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
if (paused_private_state_thread)
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
{
if (log)
log->Printf ("ProcessGDBRemote::InterruptIfRunning() resuming private state thread");
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
ResumePrivateStateThread();
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
}
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
return error;
}
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
Error
ProcessGDBRemote::WillDetach ()
{
Improved process logging for both lldb_private::Process and ProcessGDBRemote. llvm-svn: 124080
2011-01-23 19:58:49 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
if (log)
log->Printf ("ProcessGDBRemote::WillDetach()");
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
bool discard_thread_plans = true;
bool catch_stop_event = true;
EventSP event_sp;
Fixed ProcessGDBRemote to kill the process correctly when it is either running or stopped. Added support for sections to be able to state if they are encrypted or not. llvm-svn: 124171
2011-01-25 02:58:48 +00:00
return InterruptIfRunning (discard_thread_plans, catch_stop_event, event_sp);
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
}
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
Error
ProcessGDBRemote::DoDetach()
{
Error error;
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
if (log)
log->Printf ("ProcessGDBRemote::DoDetach()");
DisableAllBreakpointSites ();
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
m_thread_list.DiscardThreadPlans();
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
size_t response_size = m_gdb_comm.SendPacket ("D", 1);
if (log)
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
{
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
if (response_size)
log->PutCString ("ProcessGDBRemote::DoDetach() detach packet sent successfully");
else
log->PutCString ("ProcessGDBRemote::DoDetach() detach packet send failed");
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
}
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
// Sleep for one second to let the process get all detached...
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
StopAsyncThread ();
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
m_gdb_comm.StopReadThread();
m_gdb_comm.Disconnect(); // Disconnect from the debug server.
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
SetPrivateState (eStateDetached);
ResumePrivateStateThread();
//KillDebugserverProcess ();
Hooked up detach for ProcessGDBRemote. Remove the GetUserData()/SetUserData() from the DWARFDebugInfoEntry class. We now track everything with dense maps. llvm-svn: 114876
2010-09-27 21:07:38 +00:00
return error;
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Error
ProcessGDBRemote::DoDestroy ()
{
Error error;
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::DoDestroy()");
// Interrupt if our inferior is running...
Sped up the shutdown time on MacOSX by quite a bit by making sure any threads that we spawn let us know when they are going away and that we don't timeout waiting for a message from threads that have gone away. We also now don't expect the "k" packet (kill) to send a response. This greatly speeds up debugger shutdown performance. The test suite now runs quite a bit faster. Added a fix to the variable display code that fixes the display of base classes. We were assuming the virtual or normal base class offsets were being given in bit sizes, but they were being given as character sizes, so we needed to multiply the offset by 8. This wasn't affecting the expression parser, but it was affecting the correct display of C++ class base classes and all of their children. llvm-svn: 124024
2011-01-22 07:12:45 +00:00
if (m_gdb_comm.IsConnected())
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
if (m_public_state.GetValue() == eStateAttaching)
One more fix to ProcessGDBRemote::DoDestroy(). This one will make sure we parse the exit status in case we send this packet while stopped. llvm-svn: 124177
2011-01-25 04:57:42 +00:00
{
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
// We are being asked to halt during an attach. We need to just close
// our file handle and debugserver will go away, and we can be done...
m_gdb_comm.Disconnect();
One more fix to ProcessGDBRemote::DoDestroy(). This one will make sure we parse the exit status in case we send this packet while stopped. llvm-svn: 124177
2011-01-25 04:57:42 +00:00
}
else
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
{
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
StringExtractorGDBRemote response;
bool send_async = true;
if (m_gdb_comm.SendPacketAndWaitForResponse("k", 1, response, 2, send_async))
{
char packet_cmd = response.GetChar(0);
if (packet_cmd == 'W' || packet_cmd == 'X')
{
m_last_stop_packet = response;
SetExitStatus(response.GetHexU8(), NULL);
}
}
else
{
SetExitStatus(SIGABRT, NULL);
//error.SetErrorString("kill packet failed");
}
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
StopAsyncThread ();
m_gdb_comm.StopReadThread();
KillDebugserverProcess ();
Fixed an lldb infrastructure bug, where the debugger should reaaly update its execution context only when the process is still alive. When running the test suite, the debugger is launching and killing processes constantly. This might be the cause of the test hang as reported in rdar://problem/8377854, where the debugger was looping infinitely trying to update a supposedly stale thread list. llvm-svn: 113022
2010-09-03 22:35:47 +00:00
m_gdb_comm.Disconnect(); // Disconnect from the debug server.
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
return error;
}
//------------------------------------------------------------------
// Process Queries
//------------------------------------------------------------------
bool
ProcessGDBRemote::IsAlive ()
{
Added the ability to dump sections to a certain depth (for when sections have children sections). Modified SectionLoadList to do it's own multi-threaded protected on its map. The ThreadSafeSTLMap class was difficult to deal with and wasn't providing much utility, it was only getting in the way. Make sure when the communication read thread is about to exit, it clears the thread in the main class. Fixed the ModuleList to correctly ignore architectures and UUIDs if they aren't valid when searching for a matching module. If we specified a file with no arch, and then modified the file and loaded it again, it would not match on subsequent searches if the arch was invalid since it would compare an invalid architecture to the one that was found or selected within the shared library or executable. This was causing stale modules to stay around in the global module list when they should have been removed. Removed deprecated functions from the DynamicLoaderMacOSXDYLD class. Modified "ProcessGDBRemote::IsAlive" to check if we are connected to a gdb server and also make sure our process hasn't exited. llvm-svn: 121236
2010-12-08 05:08:21 +00:00
return m_gdb_comm.IsConnected() && m_private_state.GetValue() != eStateExited;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
addr_t
ProcessGDBRemote::GetImageInfoAddress()
{
if (!m_gdb_comm.IsRunning())
{
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse("qShlibInfoAddr", ::strlen ("qShlibInfoAddr"), response, 2, false))
{
if (response.IsNormalPacket())
return response.GetHexMaxU64(false, LLDB_INVALID_ADDRESS);
}
}
return LLDB_INVALID_ADDRESS;
}
DynamicLoader *
ProcessGDBRemote::GetDynamicLoader()
{
return m_dynamic_loader_ap.get();
}
//------------------------------------------------------------------
// Process Memory
//------------------------------------------------------------------
size_t
ProcessGDBRemote::DoReadMemory (addr_t addr, void *buf, size_t size, Error &error)
{
if (size > m_max_memory_size)
{
// Keep memory read sizes down to a sane limit. This function will be
// called multiple times in order to complete the task by
// lldb_private::Process so it is ok to do this.
size = m_max_memory_size;
}
char packet[64];
const int packet_len = ::snprintf (packet, sizeof(packet), "m%llx,%zx", (uint64_t)addr, size);
assert (packet_len + 1 < sizeof(packet));
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse(packet, packet_len, response, 2, true))
{
if (response.IsNormalPacket())
{
error.Clear();
return response.GetHexBytes(buf, size, '\xdd');
}
else if (response.IsErrorPacket())
error.SetErrorStringWithFormat("gdb remote returned an error: %s", response.GetStringRef().c_str());
else if (response.IsUnsupportedPacket())
error.SetErrorStringWithFormat("'%s' packet unsupported", packet);
else
error.SetErrorStringWithFormat("unexpected response to '%s': '%s'", packet, response.GetStringRef().c_str());
}
else
{
error.SetErrorStringWithFormat("failed to sent packet: '%s'", packet);
}
return 0;
}
size_t
ProcessGDBRemote::DoWriteMemory (addr_t addr, const void *buf, size_t size, Error &error)
{
StreamString packet;
packet.Printf("M%llx,%zx:", addr, size);
Endian patch from Kirk Beitz that allows better cross platform building. llvm-svn: 124643
2011-02-01 01:31:41 +00:00
packet.PutBytesAsRawHex8(buf, size, lldb::endian::InlHostByteOrder(), lldb::endian::InlHostByteOrder());
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse(packet.GetData(), packet.GetSize(), response, 2, true))
{
if (response.IsOKPacket())
{
error.Clear();
return size;
}
else if (response.IsErrorPacket())
error.SetErrorStringWithFormat("gdb remote returned an error: %s", response.GetStringRef().c_str());
else if (response.IsUnsupportedPacket())
error.SetErrorStringWithFormat("'%s' packet unsupported", packet.GetString().c_str());
else
error.SetErrorStringWithFormat("unexpected response to '%s': '%s'", packet.GetString().c_str(), response.GetStringRef().c_str());
}
else
{
error.SetErrorStringWithFormat("failed to sent packet: '%s'", packet.GetString().c_str());
}
return 0;
}
lldb::addr_t
ProcessGDBRemote::DoAllocateMemory (size_t size, uint32_t permissions, Error &error)
{
addr_t allocated_addr = m_gdb_comm.AllocateMemory (size, permissions, m_packet_timeout);
if (allocated_addr == LLDB_INVALID_ADDRESS)
error.SetErrorStringWithFormat("unable to allocate %zu bytes of memory with permissions %u", size, permissions);
else
error.Clear();
return allocated_addr;
}
Error
ProcessGDBRemote::DoDeallocateMemory (lldb::addr_t addr)
{
Error error;
if (!m_gdb_comm.DeallocateMemory (addr, m_packet_timeout))
error.SetErrorStringWithFormat("unable to deallocate memory at 0x%llx", addr);
return error;
}
//------------------------------------------------------------------
// Process STDIO
//------------------------------------------------------------------
size_t
ProcessGDBRemote::GetSTDOUT (char *buf, size_t buf_size, Error &error)
{
Mutex::Locker locker(m_stdio_mutex);
size_t bytes_available = m_stdout_data.size();
if (bytes_available > 0)
{
Added a quicker lookup in the SectionLoadList when looking things up by section by using a DenseMap. Fixed some logging calls to get the log shared pointer. llvm-svn: 124926
2011-02-05 02:25:06 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (log)
log->Printf ("ProcessGDBRemote::%s (&%p[%u]) ...", __FUNCTION__, buf, buf_size);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (bytes_available > buf_size)
{
Remove use of STL collection class use of the "data()" method since it isn't part of C++'98. Most of these were "std::vector<T>::data()" and "std::string::data()". llvm-svn: 108957
2010-07-20 22:52:08 +00:00
memcpy(buf, m_stdout_data.c_str(), buf_size);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
m_stdout_data.erase(0, buf_size);
bytes_available = buf_size;
}
else
{
Remove use of STL collection class use of the "data()" method since it isn't part of C++'98. Most of these were "std::vector<T>::data()" and "std::string::data()". llvm-svn: 108957
2010-07-20 22:52:08 +00:00
memcpy(buf, m_stdout_data.c_str(), bytes_available);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
m_stdout_data.clear();
//ResetEventBits(eBroadcastBitSTDOUT);
}
}
return bytes_available;
}
size_t
ProcessGDBRemote::GetSTDERR (char *buf, size_t buf_size, Error &error)
{
// Can we get STDERR through the remote protocol?
return 0;
}
size_t
ProcessGDBRemote::PutSTDIN (const char *src, size_t src_len, Error &error)
{
if (m_stdio_communication.IsConnected())
{
ConnectionStatus status;
m_stdio_communication.Write(src, src_len, status, NULL);
}
return 0;
}
Error
ProcessGDBRemote::EnableBreakpoint (BreakpointSite *bp_site)
{
Error error;
assert (bp_site != NULL);
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_BREAKPOINTS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
user_id_t site_id = bp_site->GetID();
const addr_t addr = bp_site->GetLoadAddress();
if (log)
log->Printf ("ProcessGDBRemote::EnableBreakpoint (size_id = %d) address = 0x%llx", site_id, (uint64_t)addr);
if (bp_site->IsEnabled())
{
if (log)
log->Printf ("ProcessGDBRemote::EnableBreakpoint (size_id = %d) address = 0x%llx -- SUCCESS (already enabled)", site_id, (uint64_t)addr);
return error;
}
else
{
const size_t bp_op_size = GetSoftwareBreakpointTrapOpcode (bp_site);
if (bp_site->HardwarePreferred())
{
// Try and set hardware breakpoint, and if that fails, fall through
// and set a software breakpoint?
}
if (m_z0_supported)
{
char packet[64];
const int packet_len = ::snprintf (packet, sizeof(packet), "Z0,%llx,%zx", addr, bp_op_size);
assert (packet_len + 1 < sizeof(packet));
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse(packet, packet_len, response, 2, true))
{
if (response.IsUnsupportedPacket())
{
// Disable z packet support and try again
m_z0_supported = 0;
return EnableBreakpoint (bp_site);
}
else if (response.IsOKPacket())
{
bp_site->SetEnabled(true);
bp_site->SetType (BreakpointSite::eExternal);
return error;
}
else
{
uint8_t error_byte = response.GetError();
if (error_byte)
error.SetErrorStringWithFormat("%x packet failed with error: %i (0x%2.2x).\n", packet, error_byte, error_byte);
}
}
}
else
{
return EnableSoftwareBreakpoint (bp_site);
}
}
if (log)
{
const char *err_string = error.AsCString();
log->Printf ("ProcessGDBRemote::EnableBreakpoint() error for breakpoint at 0x%8.8llx: %s",
bp_site->GetLoadAddress(),
err_string ? err_string : "NULL");
}
// We shouldn't reach here on a successful breakpoint enable...
if (error.Success())
error.SetErrorToGenericError();
return error;
}
Error
ProcessGDBRemote::DisableBreakpoint (BreakpointSite *bp_site)
{
Error error;
assert (bp_site != NULL);
addr_t addr = bp_site->GetLoadAddress();
user_id_t site_id = bp_site->GetID();
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_BREAKPOINTS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::DisableBreakpoint (site_id = %d) addr = 0x%8.8llx", site_id, (uint64_t)addr);
if (bp_site->IsEnabled())
{
const size_t bp_op_size = GetSoftwareBreakpointTrapOpcode (bp_site);
if (bp_site->IsHardware())
{
// TODO: disable hardware breakpoint...
}
else
{
if (m_z0_supported)
{
char packet[64];
const int packet_len = ::snprintf (packet, sizeof(packet), "z0,%llx,%zx", addr, bp_op_size);
assert (packet_len + 1 < sizeof(packet));
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse(packet, packet_len, response, 2, true))
{
if (response.IsUnsupportedPacket())
{
error.SetErrorString("Breakpoint site was set with Z packet, yet remote debugserver states z packets are not supported.");
}
else if (response.IsOKPacket())
{
if (log)
log->Printf ("ProcessGDBRemote::DisableBreakpoint (site_id = %d) addr = 0x%8.8llx -- SUCCESS", site_id, (uint64_t)addr);
bp_site->SetEnabled(false);
return error;
}
else
{
uint8_t error_byte = response.GetError();
if (error_byte)
error.SetErrorStringWithFormat("%x packet failed with error: %i (0x%2.2x).\n", packet, error_byte, error_byte);
}
}
}
else
{
return DisableSoftwareBreakpoint (bp_site);
}
}
}
else
{
if (log)
log->Printf ("ProcessGDBRemote::DisableBreakpoint (site_id = %d) addr = 0x%8.8llx -- SUCCESS (already disabled)", site_id, (uint64_t)addr);
return error;
}
if (error.Success())
error.SetErrorToGenericError();
return error;
}
Error
ProcessGDBRemote::EnableWatchpoint (WatchpointLocation *wp)
{
Error error;
if (wp)
{
user_id_t watchID = wp->GetID();
addr_t addr = wp->GetLoadAddress();
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_WATCHPOINTS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::EnableWatchpoint(watchID = %d)", watchID);
if (wp->IsEnabled())
{
if (log)
log->Printf("ProcessGDBRemote::EnableWatchpoint(watchID = %d) addr = 0x%8.8llx: watchpoint already enabled.", watchID, (uint64_t)addr);
return error;
}
else
{
// Pass down an appropriate z/Z packet...
error.SetErrorString("watchpoints not supported");
}
}
else
{
error.SetErrorString("Watchpoint location argument was NULL.");
}
if (error.Success())
error.SetErrorToGenericError();
return error;
}
Error
ProcessGDBRemote::DisableWatchpoint (WatchpointLocation *wp)
{
Error error;
if (wp)
{
user_id_t watchID = wp->GetID();
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_WATCHPOINTS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
addr_t addr = wp->GetLoadAddress();
if (log)
log->Printf ("ProcessGDBRemote::DisableWatchpoint (watchID = %d) addr = 0x%8.8llx", watchID, (uint64_t)addr);
if (wp->IsHardware())
{
// Pass down an appropriate z/Z packet...
error.SetErrorString("watchpoints not supported");
}
// TODO: clear software watchpoints if we implement them
}
else
{
error.SetErrorString("Watchpoint location argument was NULL.");
}
if (error.Success())
error.SetErrorToGenericError();
return error;
}
void
ProcessGDBRemote::Clear()
{
m_flags = 0;
m_thread_list.Clear();
{
Mutex::Locker locker(m_stdio_mutex);
m_stdout_data.clear();
}
}
Error
ProcessGDBRemote::DoSignal (int signo)
{
Error error;
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::DoSignal (signal = %d)", signo);
if (!m_gdb_comm.SendAsyncSignal (signo))
error.SetErrorStringWithFormat("failed to send signal %i", signo);
return error;
}
Error
ProcessGDBRemote::StartDebugserverProcess
(
const char *debugserver_url, // The connection string to use in the spawned debugserver ("localhost:1234" or "/dev/tty...")
char const *inferior_argv[], // Arguments for the inferior program including the path to the inferior itself as the first argument
char const *inferior_envp[], // Environment to pass along to the inferior program
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
const char *stdin_path,
const char *stdout_path,
const char *stderr_path,
const char *working_dir,
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
bool launch_process, // Set to true if we are going to be launching a the process
lldb::pid_t attach_pid, // If inferior inferior_argv == NULL, and attach_pid != LLDB_INVALID_PROCESS_ID send this pid as an argument to debugserver
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
const char *attach_name, // Wait for the next process to launch whose basename matches "attach_name"
bool wait_for_launch, // Wait for the process named "attach_name" to launch
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
uint32_t launch_flags, // Launch flags
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
ArchSpec& inferior_arch // The arch of the inferior that we will launch
)
{
Error error;
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
bool disable_aslr = (launch_flags & eLaunchFlagDisableASLR) != 0;
bool no_stdio = (launch_flags & eLaunchFlagDisableSTDIO) != 0;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (m_debugserver_pid == LLDB_INVALID_PROCESS_ID)
{
// If we locate debugserver, keep that located version around
static FileSpec g_debugserver_file_spec;
FileSpec debugserver_file_spec;
char debugserver_path[PATH_MAX];
// Always check to see if we have an environment override for the path
// to the debugserver to use and use it if we do.
const char *env_debugserver_path = getenv("LLDB_DEBUGSERVER_PATH");
if (env_debugserver_path)
Fixed an issue where we were resolving paths when we should have been. So the issue here was that we have lldb_private::FileSpec that by default was always resolving a path when using the: FileSpec::FileSpec (const char *path); and in the: void FileSpec::SetFile(const char *pathname, bool resolve = true); This isn't what we want in many many cases. One example is you have "/tmp" on your file system which is really "/private/tmp". You compile code in that directory and end up with debug info that mentions "/tmp/file.c". Then you type: (lldb) breakpoint set --file file.c --line 5 If your current working directory is "/tmp", then "file.c" would be turned into "/private/tmp/file.c" which won't match anything in the debug info. Also, it should have been just a FileSpec with no directory and a filename of "file.c" which could (and should) potentially match any instances of "file.c" in the debug info. So I removed the constructor that just takes a path: FileSpec::FileSpec (const char *path); // REMOVED You must now use the other constructor that has a "bool resolve" parameter that you must always supply: FileSpec::FileSpec (const char *path, bool resolve); I also removed the default parameter to SetFile(): void FileSpec::SetFile(const char *pathname, bool resolve); And fixed all of the code to use the right settings. llvm-svn: 116944
2010-10-20 20:54:39 +00:00
debugserver_file_spec.SetFile (env_debugserver_path, false);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
else
debugserver_file_spec = g_debugserver_file_spec;
bool debugserver_exists = debugserver_file_spec.Exists();
if (!debugserver_exists)
{
// The debugserver binary is in the LLDB.framework/Resources
// directory.
Added a new Host call to find LLDB related paths: static bool Host::GetLLDBPath (lldb::PathType path_type, FileSpec &file_spec); This will fill in "file_spec" with an appropriate path that is appropriate for the current Host OS. MacOSX will return paths within the LLDB.framework, and other unixes will return the paths they want. The current PathType enums are: typedef enum PathType { ePathTypeLLDBShlibDir, // The directory where the lldb.so (unix) or LLDB mach-o file in LLDB.framework (MacOSX) exists ePathTypeSupportExecutableDir, // Find LLDB support executable directory (debugserver, etc) ePathTypeHeaderDir, // Find LLDB header file directory ePathTypePythonDir // Find Python modules (PYTHONPATH) directory } PathType; All places that were finding executables are and python paths are now updated to use this Host call. Added another new host call to launch the inferior in a terminal. This ability will be very host specific and doesn't need to be supported on all systems. MacOSX currently will create a new .command file and tell Terminal.app to open the .command file. It also uses the new "darwin-debug" app which is a small app that uses posix to exec (no fork) and stop at the entry point of the program. The GDB remote plug-in is almost able launch a process and attach to it, it currently will spawn the process, but it won't attach to it just yet. This will let LLDB not have to share the terminal with another process and a new terminal window will pop up when you launch. This won't get hooked up until we work out all of the kinks. The new Host function is: static lldb::pid_t Host::LaunchInNewTerminal ( const char **argv, // argv[0] is executable const char **envp, const ArchSpec *arch_spec, bool stop_at_entry, bool disable_aslr); Cleaned up FileSpec::GetPath to not use strncpy() as it was always zero filling the entire path buffer. Fixed an issue with the dynamic checker function where I missed a '$' prefix that should have been added. llvm-svn: 116690
2010-10-17 22:03:32 +00:00
if (Host::GetLLDBPath (ePathTypeSupportExecutableDir, debugserver_file_spec))
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Added a new Host call to find LLDB related paths: static bool Host::GetLLDBPath (lldb::PathType path_type, FileSpec &file_spec); This will fill in "file_spec" with an appropriate path that is appropriate for the current Host OS. MacOSX will return paths within the LLDB.framework, and other unixes will return the paths they want. The current PathType enums are: typedef enum PathType { ePathTypeLLDBShlibDir, // The directory where the lldb.so (unix) or LLDB mach-o file in LLDB.framework (MacOSX) exists ePathTypeSupportExecutableDir, // Find LLDB support executable directory (debugserver, etc) ePathTypeHeaderDir, // Find LLDB header file directory ePathTypePythonDir // Find Python modules (PYTHONPATH) directory } PathType; All places that were finding executables are and python paths are now updated to use this Host call. Added another new host call to launch the inferior in a terminal. This ability will be very host specific and doesn't need to be supported on all systems. MacOSX currently will create a new .command file and tell Terminal.app to open the .command file. It also uses the new "darwin-debug" app which is a small app that uses posix to exec (no fork) and stop at the entry point of the program. The GDB remote plug-in is almost able launch a process and attach to it, it currently will spawn the process, but it won't attach to it just yet. This will let LLDB not have to share the terminal with another process and a new terminal window will pop up when you launch. This won't get hooked up until we work out all of the kinks. The new Host function is: static lldb::pid_t Host::LaunchInNewTerminal ( const char **argv, // argv[0] is executable const char **envp, const ArchSpec *arch_spec, bool stop_at_entry, bool disable_aslr); Cleaned up FileSpec::GetPath to not use strncpy() as it was always zero filling the entire path buffer. Fixed an issue with the dynamic checker function where I missed a '$' prefix that should have been added. llvm-svn: 116690
2010-10-17 22:03:32 +00:00
debugserver_file_spec.GetFilename().SetCString(DEBUGSERVER_BASENAME);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
debugserver_exists = debugserver_file_spec.Exists();
Added a new Host call to find LLDB related paths: static bool Host::GetLLDBPath (lldb::PathType path_type, FileSpec &file_spec); This will fill in "file_spec" with an appropriate path that is appropriate for the current Host OS. MacOSX will return paths within the LLDB.framework, and other unixes will return the paths they want. The current PathType enums are: typedef enum PathType { ePathTypeLLDBShlibDir, // The directory where the lldb.so (unix) or LLDB mach-o file in LLDB.framework (MacOSX) exists ePathTypeSupportExecutableDir, // Find LLDB support executable directory (debugserver, etc) ePathTypeHeaderDir, // Find LLDB header file directory ePathTypePythonDir // Find Python modules (PYTHONPATH) directory } PathType; All places that were finding executables are and python paths are now updated to use this Host call. Added another new host call to launch the inferior in a terminal. This ability will be very host specific and doesn't need to be supported on all systems. MacOSX currently will create a new .command file and tell Terminal.app to open the .command file. It also uses the new "darwin-debug" app which is a small app that uses posix to exec (no fork) and stop at the entry point of the program. The GDB remote plug-in is almost able launch a process and attach to it, it currently will spawn the process, but it won't attach to it just yet. This will let LLDB not have to share the terminal with another process and a new terminal window will pop up when you launch. This won't get hooked up until we work out all of the kinks. The new Host function is: static lldb::pid_t Host::LaunchInNewTerminal ( const char **argv, // argv[0] is executable const char **envp, const ArchSpec *arch_spec, bool stop_at_entry, bool disable_aslr); Cleaned up FileSpec::GetPath to not use strncpy() as it was always zero filling the entire path buffer. Fixed an issue with the dynamic checker function where I missed a '$' prefix that should have been added. llvm-svn: 116690
2010-10-17 22:03:32 +00:00
if (debugserver_exists)
{
g_debugserver_file_spec = debugserver_file_spec;
}
else
{
g_debugserver_file_spec.Clear();
debugserver_file_spec.Clear();
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
}
if (debugserver_exists)
{
debugserver_file_spec.GetPath (debugserver_path, sizeof(debugserver_path));
m_stdio_communication.Clear();
posix_spawnattr_t attr;
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Error local_err; // Errors that don't affect the spawning.
if (log)
log->Printf ("%s ( path='%s', argv=%p, envp=%p, arch=%s )", __FUNCTION__, debugserver_path, inferior_argv, inferior_envp, inferior_arch.AsCString());
error.SetError( ::posix_spawnattr_init (&attr), eErrorTypePOSIX);
if (error.Fail() || log)
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
error.PutToLog(log.get(), "::posix_spawnattr_init ( &attr )");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (error.Fail())
return error;;
#if !defined (__arm__)
Added a new Host call to find LLDB related paths: static bool Host::GetLLDBPath (lldb::PathType path_type, FileSpec &file_spec); This will fill in "file_spec" with an appropriate path that is appropriate for the current Host OS. MacOSX will return paths within the LLDB.framework, and other unixes will return the paths they want. The current PathType enums are: typedef enum PathType { ePathTypeLLDBShlibDir, // The directory where the lldb.so (unix) or LLDB mach-o file in LLDB.framework (MacOSX) exists ePathTypeSupportExecutableDir, // Find LLDB support executable directory (debugserver, etc) ePathTypeHeaderDir, // Find LLDB header file directory ePathTypePythonDir // Find Python modules (PYTHONPATH) directory } PathType; All places that were finding executables are and python paths are now updated to use this Host call. Added another new host call to launch the inferior in a terminal. This ability will be very host specific and doesn't need to be supported on all systems. MacOSX currently will create a new .command file and tell Terminal.app to open the .command file. It also uses the new "darwin-debug" app which is a small app that uses posix to exec (no fork) and stop at the entry point of the program. The GDB remote plug-in is almost able launch a process and attach to it, it currently will spawn the process, but it won't attach to it just yet. This will let LLDB not have to share the terminal with another process and a new terminal window will pop up when you launch. This won't get hooked up until we work out all of the kinks. The new Host function is: static lldb::pid_t Host::LaunchInNewTerminal ( const char **argv, // argv[0] is executable const char **envp, const ArchSpec *arch_spec, bool stop_at_entry, bool disable_aslr); Cleaned up FileSpec::GetPath to not use strncpy() as it was always zero filling the entire path buffer. Fixed an issue with the dynamic checker function where I missed a '$' prefix that should have been added. llvm-svn: 116690
2010-10-17 22:03:32 +00:00
// We don't need to do this for ARM, and we really shouldn't now
// that we have multiple CPU subtypes and no posix_spawnattr call
// that allows us to set which CPU subtype to launch...
Made lldb_private::ArchSpec more generic so that it can take a mach-o cpu type and sub-type, or an ELF e_machine value. Also added a generic CPU type to the arch spec class so we can have a single arch definition that the LLDB core code can use. Previously a lot of places in the code were using the mach-o definitions from a macosx header file. Switches over to using "llvm/Support/MachO.h" for the llvm::MachO::XXX for the CPU types and sub types for mach-o ArchSpecs. Added "llvm/Support/ELF.h" so we can use the "llvm::ELF::XXX" defines for the ELF ArchSpecs. Got rid of all CPU_TYPE_ and CPU_SUBTYPE_ defines that were previously being used in LLDB. llvm-svn: 105806
2010-06-11 03:25:34 +00:00
if (inferior_arch.GetType() == eArchTypeMachO)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Made lldb_private::ArchSpec more generic so that it can take a mach-o cpu type and sub-type, or an ELF e_machine value. Also added a generic CPU type to the arch spec class so we can have a single arch definition that the LLDB core code can use. Previously a lot of places in the code were using the mach-o definitions from a macosx header file. Switches over to using "llvm/Support/MachO.h" for the llvm::MachO::XXX for the CPU types and sub types for mach-o ArchSpecs. Added "llvm/Support/ELF.h" so we can use the "llvm::ELF::XXX" defines for the ELF ArchSpecs. Got rid of all CPU_TYPE_ and CPU_SUBTYPE_ defines that were previously being used in LLDB. llvm-svn: 105806
2010-06-11 03:25:34 +00:00
cpu_type_t cpu = inferior_arch.GetCPUType();
if (cpu != 0 && cpu != UINT32_MAX && cpu != LLDB_INVALID_CPUTYPE)
{
size_t ocount = 0;
error.SetError( ::posix_spawnattr_setbinpref_np (&attr, 1, &cpu, &ocount), eErrorTypePOSIX);
if (error.Fail() || log)
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
error.PutToLog(log.get(), "::posix_spawnattr_setbinpref_np ( &attr, 1, cpu_type = 0x%8.8x, count => %zu )", cpu, ocount);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Made lldb_private::ArchSpec more generic so that it can take a mach-o cpu type and sub-type, or an ELF e_machine value. Also added a generic CPU type to the arch spec class so we can have a single arch definition that the LLDB core code can use. Previously a lot of places in the code were using the mach-o definitions from a macosx header file. Switches over to using "llvm/Support/MachO.h" for the llvm::MachO::XXX for the CPU types and sub types for mach-o ArchSpecs. Added "llvm/Support/ELF.h" so we can use the "llvm::ELF::XXX" defines for the ELF ArchSpecs. Got rid of all CPU_TYPE_ and CPU_SUBTYPE_ defines that were previously being used in LLDB. llvm-svn: 105806
2010-06-11 03:25:34 +00:00
if (error.Fail() != 0 || ocount != 1)
return error;
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
#endif
Args debugserver_args;
char arg_cstr[PATH_MAX];
lldb_utility::PseudoTerminal pty;
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
const char *stdio_path = NULL;
if (launch_process &&
If the user specfies one of stdin, stdout or stderr, don't automatically set the non-specified ones to /dev/null. llvm-svn: 124459
2011-01-28 00:19:58 +00:00
(stdin_path == NULL || stdout_path == NULL || stderr_path == NULL) &&
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
m_local_debugserver &&
no_stdio == false)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
if (pty.OpenFirstAvailableMaster(O_RDWR|O_NOCTTY, NULL, 0))
If the user specfies one of stdin, stdout or stderr, don't automatically set the non-specified ones to /dev/null. llvm-svn: 124459
2011-01-28 00:19:58 +00:00
{
const char *slave_name = pty.GetSlaveName (NULL, 0);
if (stdin_path == NULL
&& stdout_path == NULL
&& stderr_path == NULL)
stdio_path = slave_name;
else
{
if (stdin_path == NULL)
stdin_path = slave_name;
if (stdout_path == NULL)
stdout_path = slave_name;
if (stderr_path == NULL)
stderr_path = slave_name;
}
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
// Start args with "debugserver /file/path -r --"
debugserver_args.AppendArgument(debugserver_path);
debugserver_args.AppendArgument(debugserver_url);
Added a new Host call to find LLDB related paths: static bool Host::GetLLDBPath (lldb::PathType path_type, FileSpec &file_spec); This will fill in "file_spec" with an appropriate path that is appropriate for the current Host OS. MacOSX will return paths within the LLDB.framework, and other unixes will return the paths they want. The current PathType enums are: typedef enum PathType { ePathTypeLLDBShlibDir, // The directory where the lldb.so (unix) or LLDB mach-o file in LLDB.framework (MacOSX) exists ePathTypeSupportExecutableDir, // Find LLDB support executable directory (debugserver, etc) ePathTypeHeaderDir, // Find LLDB header file directory ePathTypePythonDir // Find Python modules (PYTHONPATH) directory } PathType; All places that were finding executables are and python paths are now updated to use this Host call. Added another new host call to launch the inferior in a terminal. This ability will be very host specific and doesn't need to be supported on all systems. MacOSX currently will create a new .command file and tell Terminal.app to open the .command file. It also uses the new "darwin-debug" app which is a small app that uses posix to exec (no fork) and stop at the entry point of the program. The GDB remote plug-in is almost able launch a process and attach to it, it currently will spawn the process, but it won't attach to it just yet. This will let LLDB not have to share the terminal with another process and a new terminal window will pop up when you launch. This won't get hooked up until we work out all of the kinks. The new Host function is: static lldb::pid_t Host::LaunchInNewTerminal ( const char **argv, // argv[0] is executable const char **envp, const ArchSpec *arch_spec, bool stop_at_entry, bool disable_aslr); Cleaned up FileSpec::GetPath to not use strncpy() as it was always zero filling the entire path buffer. Fixed an issue with the dynamic checker function where I missed a '$' prefix that should have been added. llvm-svn: 116690
2010-10-17 22:03:32 +00:00
// use native registers, not the GDB registers
debugserver_args.AppendArgument("--native-regs");
// make debugserver run in its own session so signals generated by
// special terminal key sequences (^C) don't affect debugserver
debugserver_args.AppendArgument("--setsid");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Added the ability to disable ASLR (Address Space Layout Randomization). ASLR is disabled by default, and can be enabled using: (lldb) set disable-aslr 0 llvm-svn: 112616
2010-08-31 18:35:14 +00:00
if (disable_aslr)
debugserver_args.AppendArguments("--disable-aslr");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// Only set the inferior
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
if (launch_process)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
if (no_stdio)
debugserver_args.AppendArgument("--no-stdio");
else
{
if (stdin_path && stdout_path && stderr_path &&
strcmp(stdin_path, stdout_path) == 0 &&
strcmp(stdin_path, stderr_path) == 0)
{
stdio_path = stdin_path;
stdin_path = stdout_path = stderr_path = NULL;
}
if (stdio_path)
{
// All file handles to stdin, stdout, stderr are the same...
debugserver_args.AppendArgument("--stdio-path");
debugserver_args.AppendArgument(stdio_path);
}
else
{
if (stdin_path == NULL && (stdout_path || stderr_path))
stdin_path = "/dev/null";
if (stdout_path == NULL && (stdin_path || stderr_path))
stdout_path = "/dev/null";
if (stderr_path == NULL && (stdin_path || stdout_path))
stderr_path = "/dev/null";
if (stdin_path)
{
debugserver_args.AppendArgument("--stdin-path");
debugserver_args.AppendArgument(stdin_path);
}
if (stdout_path)
{
debugserver_args.AppendArgument("--stdout-path");
debugserver_args.AppendArgument(stdout_path);
}
if (stderr_path)
{
debugserver_args.AppendArgument("--stderr-path");
debugserver_args.AppendArgument(stderr_path);
}
}
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
if (working_dir)
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
{
Added a new variant of SBTarget::Launch() that deprectates the old one that takes separate file handles for stdin, stdout, and stder and also allows for the working directory to be specified. Added support to "process launch" to a new option: --working-dir=PATH. We can now set the working directory. If this is not set, it defaults to that of the process that has LLDB loaded. Added the working directory to the host LaunchInNewTerminal function to allows the current working directory to be set in processes that are spawned in their own terminal. Also hooked this up to the lldb_private::Process and all mac plug-ins. The linux plug-in had its API changed, but nothing is making use of it yet. Modfied "debugserver" and "darwin-debug" to also handle the current working directory options and modified the code in LLDB that spawns these tools to pass the info along. Fixed ProcessGDBRemote to properly pass along all file handles for stdin, stdout and stderr. After clearing the default values for the stdin/out/err file handles for process to be NULL, we had a crasher in UserSettingsController::UpdateStringVariable which is now fixed. Also fixed the setting of boolean values to be able to be set as "true", "yes", "on", "1" for true (case insensitive) and "false", "no", "off", or "0" for false. Fixed debugserver to properly handle files for STDIN, STDOUT and STDERR that are not already opened. Previous to this fix debugserver would only correctly open and dupe file handles for the slave side of a pseudo terminal. It now correctly handles getting STDIN for the inferior from a file, and spitting STDOUT and STDERR out to files. Also made sure the file handles were correctly opened with the NOCTTY flag for terminals. llvm-svn: 124060
2011-01-23 05:56:20 +00:00
debugserver_args.AppendArgument("--working-dir");
debugserver_args.AppendArgument(working_dir);
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
const char *env_debugserver_log_file = getenv("LLDB_DEBUGSERVER_LOG_FILE");
if (env_debugserver_log_file)
{
::snprintf (arg_cstr, sizeof(arg_cstr), "--log-file=%s", env_debugserver_log_file);
debugserver_args.AppendArgument(arg_cstr);
}
const char *env_debugserver_log_flags = getenv("LLDB_DEBUGSERVER_LOG_FLAGS");
if (env_debugserver_log_flags)
{
::snprintf (arg_cstr, sizeof(arg_cstr), "--log-flags=%s", env_debugserver_log_flags);
debugserver_args.AppendArgument(arg_cstr);
}
Reverting recent thread resume changes as it was causing testing issues. We will need to try again soon, but this change was causing instability. llvm-svn: 124180
2011-01-25 06:55:13 +00:00
// debugserver_args.AppendArgument("--log-file=/tmp/debugserver.txt");
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
// debugserver_args.AppendArgument("--log-flags=0x802e0e");
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
// Now append the program arguments
if (launch_process)
{
if (inferior_argv)
{
// Terminate the debugserver args so we can now append the inferior args
debugserver_args.AppendArgument("--");
for (int i = 0; inferior_argv[i] != NULL; ++i)
debugserver_args.AppendArgument (inferior_argv[i]);
}
else
{
// Will send environment entries with the 'QEnvironment:' packet
// Will send arguments with the 'A' packet
}
}
else if (attach_pid != LLDB_INVALID_PROCESS_ID)
{
::snprintf (arg_cstr, sizeof(arg_cstr), "--attach=%u", attach_pid);
debugserver_args.AppendArgument (arg_cstr);
}
else if (attach_name && attach_name[0])
{
if (wait_for_launch)
debugserver_args.AppendArgument ("--waitfor");
else
debugserver_args.AppendArgument ("--attach");
debugserver_args.AppendArgument (attach_name);
}
Error file_actions_err;
posix_spawn_file_actions_t file_actions;
#if DONT_CLOSE_DEBUGSERVER_STDIO
file_actions_err.SetErrorString ("Remove this after uncommenting the code block below.");
#else
file_actions_err.SetError( ::posix_spawn_file_actions_init (&file_actions), eErrorTypePOSIX);
if (file_actions_err.Success())
{
::posix_spawn_file_actions_addclose (&file_actions, STDIN_FILENO);
::posix_spawn_file_actions_addclose (&file_actions, STDOUT_FILENO);
::posix_spawn_file_actions_addclose (&file_actions, STDERR_FILENO);
}
#endif
if (log)
{
StreamString strm;
debugserver_args.Dump (&strm);
log->Printf("%s arguments:\n%s", debugserver_args.GetArgumentAtIndex(0), strm.GetData());
}
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
error.SetError (::posix_spawnp (&m_debugserver_pid,
debugserver_path,
file_actions_err.Success() ? &file_actions : NULL,
&attr,
debugserver_args.GetArgumentVector(),
(char * const*)inferior_envp),
eErrorTypePOSIX);
More leaks detection: - fixed 3 posix spawn attributes leaks - fixed us always leaking CXXBaseSpecifier objects when we create class base classes. Clang apparently copies the base classes we pass in. Fixed some code formatting in ClangASTContext.cpp. llvm-svn: 107459
2010-07-02 01:29:13 +00:00
::posix_spawnattr_destroy (&attr);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (file_actions_err.Success())
::posix_spawn_file_actions_destroy (&file_actions);
// We have seen some cases where posix_spawnp was returning a valid
// looking pid even when an error was returned, so clear it out
if (error.Fail())
m_debugserver_pid = LLDB_INVALID_PROCESS_ID;
if (error.Fail() || log)
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
error.PutToLog(log.get(), "::posix_spawnp ( pid => %i, path = '%s', file_actions = %p, attr = %p, argv = %p, envp = %p )", m_debugserver_pid, debugserver_path, NULL, &attr, inferior_argv, inferior_envp);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
Add '-no-stdio' option to 'process launch' command, which causes the inferior to be launched without setting up terminal stdin/stdout for it (leaving the lldb command line accessible while the program is executing). Also add a user settings variable, 'target.process.disable-stdio' to allow the user to set this globally rather than having to use the command option each time the process is launched. llvm-svn: 120825
2010-12-03 18:46:09 +00:00
if (m_debugserver_pid != LLDB_INVALID_PROCESS_ID && !no_stdio)
If debugserver is running on the local machine, pass it a pseudoterminal to pass to the inferior for the inferior's I/O (to allow direct writing, rather than passing all the I/O around via packets). llvm-svn: 118308
2010-11-05 22:37:44 +00:00
{
Cleaned up the pseudo terminal code in ProcessGDBRemote as it was spawning a pseudo terminal even when the process being attached to. Fixed a possible crasher in the in: bool ClangASTContext::IsAggregateType (clang_type_t clang_type); It seems that if you pass in a record decl, enum decl, or objc class decl and ask it if it is an aggregate type, clang will crash. llvm-svn: 118404
2010-11-08 04:29:11 +00:00
if (pty.GetMasterFileDescriptor() != lldb_utility::PseudoTerminal::invalid_fd)
Make processes use InputReaders for their input. Move the process ReadThread stuff into the main Process class (out of the Process Plugins). This has the (intended) side effect of disabling the command line tool from reading input/commands while the process is running (the input is directed to the running process rather than to the command interpreter). llvm-svn: 119329
2010-11-16 05:07:41 +00:00
SetUpProcessInputReader (pty.ReleaseMasterFileDescriptor());
If debugserver is running on the local machine, pass it a pseudoterminal to pass to the inferior for the inferior's I/O (to allow direct writing, rather than passing all the I/O around via packets). llvm-svn: 118308
2010-11-05 22:37:44 +00:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
else
{
error.SetErrorStringWithFormat ("Unable to locate " DEBUGSERVER_BASENAME ".\n");
}
if (m_debugserver_pid != LLDB_INVALID_PROCESS_ID)
StartAsyncThread ();
}
return error;
}
bool
ProcessGDBRemote::MonitorDebugserverProcess
(
void *callback_baton,
lldb::pid_t debugserver_pid,
int signo, // Zero for no signal
int exit_status // Exit value of process if signal is zero
)
{
// We pass in the ProcessGDBRemote inferior process it and name it
// "gdb_remote_pid". The process ID is passed in the "callback_baton"
// pointer value itself, thus we need the double cast...
// "debugserver_pid" argument passed in is the process ID for
// debugserver that we are tracking...
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
ProcessGDBRemote *process = (ProcessGDBRemote *)callback_baton;
Fixed an issue in "SBError SBProcess::Destroy ()" where it wasn't properly checking the validity of the shared pointer prior to using it. Fixed the GDB remote plug-in to once again watch for a reply from the "k" packet, and fixed the logic to make sure the thread requesting the kill and the async thread play nice (and very quickly) by synchronizing the packet sending and reply. I also tweaked some of the shut down packet ("k" kill, "D" detach, and the halt packet) to make sure they do the right thing. Fixed "StateType Process::WaitForProcessStopPrivate (...)" to correctly pass the timeout along to WaitForStateChangedEventsPrivate() and made the function behave correctly with respect to timing out. Added separate STDIN, STDOUT, and STDERR support to debugserver. Also added the start of being able to set the working directory for the inferior process. llvm-svn: 124049
2011-01-22 23:43:18 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
if (log)
log->Printf ("ProcessGDBRemote::MonitorDebugserverProcess (baton=%p, pid=%i, signo=%i (0x%x), exit_status=%i)", callback_baton, debugserver_pid, signo, signo, exit_status);
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
if (process)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
// Sleep for a half a second to make sure our inferior process has
// time to set its exit status before we set it incorrectly when
// both the debugserver and the inferior process shut down.
usleep (500000);
// If our process hasn't yet exited, debugserver might have died.
// If the process did exit, the we are reaping it.
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
const StateType state = process->GetState();
if (process->m_debugserver_pid != LLDB_INVALID_PROCESS_ID &&
state != eStateInvalid &&
state != eStateUnloaded &&
state != eStateExited &&
state != eStateDetached)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
char error_str[1024];
if (signo)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
const char *signal_cstr = process->GetUnixSignals().GetSignalAsCString (signo);
if (signal_cstr)
::snprintf (error_str, sizeof (error_str), DEBUGSERVER_BASENAME " died with signal %s", signal_cstr);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
else
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
::snprintf (error_str, sizeof (error_str), DEBUGSERVER_BASENAME " died with signal %i", signo);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
else
{
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
::snprintf (error_str, sizeof (error_str), DEBUGSERVER_BASENAME " died with an exit status of 0x%8.8x", exit_status);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
Modified the host process monitor callback function Host::StartMonitoringChildProcess to spawn a thread for each process that is being monitored. Previously LLDB would spawn a single thread that would wait for any child process which isn't ok to do as a shared library (LLDB.framework on Mac OSX, or lldb.so on linux). The old single thread used to call wait4() with a pid of -1 which could cause it to reap child processes that it shouldn't have. Re-wrote the way Function blocks are handles. Previously I attempted to keep all blocks in a single memory allocation (in a std::vector). This made the code somewhat efficient, but hard to work with. I got rid of the old BlockList class, and went to a straight parent with children relationship. This new approach will allow for partial parsing of the blocks within a function. llvm-svn: 111706
2010-08-21 02:22:51 +00:00
process->SetExitStatus (-1, error_str);
}
Still trying to get detach to work with debugserver. Got a bit closer, but something is still killing our inferior. Fixed an issue with darwin-debug where it wasn't passing all needed arguments to the inferior. Fixed a race condition with the attach to named process code. llvm-svn: 116697
2010-10-18 04:14:23 +00:00
// Debugserver has exited we need to let our ProcessGDBRemote
// know that it no longer has a debugserver instance
process->m_debugserver_pid = LLDB_INVALID_PROCESS_ID;
// We are returning true to this function below, so we can
// forget about the monitor handle.
process->m_debugserver_thread = LLDB_INVALID_HOST_THREAD;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
return true;
}
void
ProcessGDBRemote::KillDebugserverProcess ()
{
if (m_debugserver_pid != LLDB_INVALID_PROCESS_ID)
{
::kill (m_debugserver_pid, SIGINT);
m_debugserver_pid = LLDB_INVALID_PROCESS_ID;
}
}
void
ProcessGDBRemote::Initialize()
{
static bool g_initialized = false;
if (g_initialized == false)
{
g_initialized = true;
PluginManager::RegisterPlugin (GetPluginNameStatic(),
GetPluginDescriptionStatic(),
CreateInstance);
Log::Callbacks log_callbacks = {
ProcessGDBRemoteLog::DisableLog,
ProcessGDBRemoteLog::EnableLog,
ProcessGDBRemoteLog::ListLogCategories
};
Log::RegisterLogChannel (ProcessGDBRemote::GetPluginNameStatic(), log_callbacks);
}
}
bool
ProcessGDBRemote::SetCurrentGDBRemoteThread (int tid)
{
if (m_curr_tid == tid)
return true;
char packet[32];
Added the ability to detect which vCont packets (using the "vCont?") packet are supported by the remote GDB target. We can also now deal with the lack of vCont support and send packets that the remote GDB stub can use. We also error out of the continue if LLDB tries to do something too complex when vCont isn't supported. llvm-svn: 125433
2011-02-12 06:28:37 +00:00
int packet_len;
if (tid <= 0)
packet_len = ::snprintf (packet, sizeof(packet), "Hg%i", tid);
else
packet_len = ::snprintf (packet, sizeof(packet), "Hg%x", tid);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
assert (packet_len + 1 < sizeof(packet));
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse(packet, packet_len, response, 2, false))
{
if (response.IsOKPacket())
{
m_curr_tid = tid;
return true;
}
}
return false;
}
bool
ProcessGDBRemote::SetCurrentGDBRemoteThreadForRun (int tid)
{
if (m_curr_tid_run == tid)
return true;
char packet[32];
Added the ability to detect which vCont packets (using the "vCont?") packet are supported by the remote GDB target. We can also now deal with the lack of vCont support and send packets that the remote GDB stub can use. We also error out of the continue if LLDB tries to do something too complex when vCont isn't supported. llvm-svn: 125433
2011-02-12 06:28:37 +00:00
int packet_len;
if (tid <= 0)
packet_len = ::snprintf (packet, sizeof(packet), "Hc%i", tid);
else
packet_len = ::snprintf (packet, sizeof(packet), "Hc%x", tid);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
assert (packet_len + 1 < sizeof(packet));
StringExtractorGDBRemote response;
if (m_gdb_comm.SendPacketAndWaitForResponse(packet, packet_len, response, 2, false))
{
if (response.IsOKPacket())
{
m_curr_tid_run = tid;
return true;
}
}
return false;
}
void
ProcessGDBRemote::ResetGDBRemoteState ()
{
// Reset and GDB remote state
m_curr_tid = LLDB_INVALID_THREAD_ID;
m_curr_tid_run = LLDB_INVALID_THREAD_ID;
m_z0_supported = 1;
}
bool
ProcessGDBRemote::StartAsyncThread ()
{
ResetGDBRemoteState ();
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::%s ()", __FUNCTION__);
// Create a thread that watches our internal state and controls which
// events make it to clients (into the DCProcess event queue).
m_async_thread = Host::ThreadCreate ("<lldb.process.gdb-remote.async>", ProcessGDBRemote::AsyncThread, this, NULL);
Patch that allows for thread_t to be something more complex than an integer. Modified patch from Kirk Beitz. llvm-svn: 125067
2011-02-08 01:34:25 +00:00
return IS_VALID_LLDB_HOST_THREAD(m_async_thread);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
}
void
ProcessGDBRemote::StopAsyncThread ()
{
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet(GDBR_LOG_PROCESS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::%s ()", __FUNCTION__);
m_async_broadcaster.BroadcastEvent (eBroadcastBitAsyncThreadShouldExit);
// Stop the stdio thread
Patch that allows for thread_t to be something more complex than an integer. Modified patch from Kirk Beitz. llvm-svn: 125067
2011-02-08 01:34:25 +00:00
if (IS_VALID_LLDB_HOST_THREAD(m_async_thread))
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
{
Host::ThreadJoin (m_async_thread, NULL, NULL);
}
}
void *
ProcessGDBRemote::AsyncThread (void *arg)
{
ProcessGDBRemote *process = (ProcessGDBRemote*) arg;
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 01:53:30 +00:00
LogSP log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) thread starting...", __FUNCTION__, arg, process->GetID());
Listener listener ("ProcessGDBRemote::AsyncThread");
EventSP event_sp;
const uint32_t desired_event_mask = eBroadcastBitAsyncContinue |
eBroadcastBitAsyncThreadShouldExit;
if (listener.StartListeningForEvents (&process->m_async_broadcaster, desired_event_mask) == desired_event_mask)
{
bool done = false;
while (!done)
{
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) listener.WaitForEvent (NULL, event_sp)...", __FUNCTION__, arg, process->GetID());
if (listener.WaitForEvent (NULL, event_sp))
{
const uint32_t event_type = event_sp->GetType();
Added an "Interrupted" bit to the ProcessEventData. Halt now generates an event with the Interrupted bit set. Process::HandlePrivateEvent ignores Interrupted events. DoHalt is changed to ensure that the stop even is processed, and an event with the Interrupted event is posted. Finally ClangFunction is rationalized to use this facility so the that Halt is handled more deterministically. llvm-svn: 119453
2010-11-17 02:32:00 +00:00
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) Got an event of type: %d...", __FUNCTION__, arg, process->GetID(), event_type);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
switch (event_type)
{
case eBroadcastBitAsyncContinue:
{
const EventDataBytes *continue_packet = EventDataBytes::GetEventDataFromEvent(event_sp.get());
if (continue_packet)
{
const char *continue_cstr = (const char *)continue_packet->GetBytes ();
const size_t continue_cstr_len = continue_packet->GetByteSize ();
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) got eBroadcastBitAsyncContinue: %s", __FUNCTION__, arg, process->GetID(), continue_cstr);
Finished up the async attach support. This allows us to request to attach by name or by pid (with or without waiting for a process to launch) and catch the response asynchronously. llvm-svn: 124530
2011-01-29 07:10:55 +00:00
if (::strstr (continue_cstr, "vAttach") == NULL)
process->SetPrivateState(eStateRunning);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
StringExtractorGDBRemote response;
StateType stop_state = process->GetGDBRemote().SendContinuePacketAndWaitForResponse (process, continue_cstr, continue_cstr_len, response);
switch (stop_state)
{
case eStateStopped:
case eStateCrashed:
case eStateSuspended:
process->m_last_stop_packet = response;
process->m_last_stop_packet.SetFilePos (0);
process->SetPrivateState (stop_state);
break;
case eStateExited:
process->m_last_stop_packet = response;
process->m_last_stop_packet.SetFilePos (0);
response.SetFilePos(1);
process->SetExitStatus(response.GetHexU8(), NULL);
done = true;
break;
case eStateInvalid:
break;
default:
process->SetPrivateState (stop_state);
break;
}
}
}
break;
case eBroadcastBitAsyncThreadShouldExit:
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) got eBroadcastBitAsyncThreadShouldExit...", __FUNCTION__, arg, process->GetID());
done = true;
break;
default:
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) got unknown event 0x%8.8x", __FUNCTION__, arg, process->GetID(), event_type);
done = true;
break;
}
}
else
{
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) listener.WaitForEvent (NULL, event_sp) => false", __FUNCTION__, arg, process->GetID());
done = true;
}
}
}
if (log)
log->Printf ("ProcessGDBRemote::%s (arg = %p, pid = %i) thread exiting...", __FUNCTION__, arg, process->GetID());
process->m_async_thread = LLDB_INVALID_HOST_THREAD;
return NULL;
}
const char *
ProcessGDBRemote::GetDispatchQueueNameForThread
(
addr_t thread_dispatch_qaddr,
std::string &dispatch_queue_name
)
{
dispatch_queue_name.clear();
if (thread_dispatch_qaddr != 0 && thread_dispatch_qaddr != LLDB_INVALID_ADDRESS)
{
// Cache the dispatch_queue_offsets_addr value so we don't always have
// to look it up
if (m_dispatch_queue_offsets_addr == LLDB_INVALID_ADDRESS)
{
Fixed the dispatch queue name retrieval for threads by looking in an extra shlib. llvm-svn: 116315
2010-10-12 17:33:06 +00:00
static ConstString g_dispatch_queue_offsets_symbol_name ("dispatch_queue_offsets");
const Symbol *dispatch_queue_offsets_symbol = NULL;
Fixed an issue where we were resolving paths when we should have been. So the issue here was that we have lldb_private::FileSpec that by default was always resolving a path when using the: FileSpec::FileSpec (const char *path); and in the: void FileSpec::SetFile(const char *pathname, bool resolve = true); This isn't what we want in many many cases. One example is you have "/tmp" on your file system which is really "/private/tmp". You compile code in that directory and end up with debug info that mentions "/tmp/file.c". Then you type: (lldb) breakpoint set --file file.c --line 5 If your current working directory is "/tmp", then "file.c" would be turned into "/private/tmp/file.c" which won't match anything in the debug info. Also, it should have been just a FileSpec with no directory and a filename of "file.c" which could (and should) potentially match any instances of "file.c" in the debug info. So I removed the constructor that just takes a path: FileSpec::FileSpec (const char *path); // REMOVED You must now use the other constructor that has a "bool resolve" parameter that you must always supply: FileSpec::FileSpec (const char *path, bool resolve); I also removed the default parameter to SetFile(): void FileSpec::SetFile(const char *pathname, bool resolve); And fixed all of the code to use the right settings. llvm-svn: 116944
2010-10-20 20:54:39 +00:00
ModuleSP module_sp(GetTarget().GetImages().FindFirstModuleForFileSpec (FileSpec("libSystem.B.dylib", false)));
Fixed the dispatch queue name retrieval for threads by looking in an extra shlib. llvm-svn: 116315
2010-10-12 17:33:06 +00:00
if (module_sp)
dispatch_queue_offsets_symbol = module_sp->FindFirstSymbolWithNameAndType (g_dispatch_queue_offsets_symbol_name, eSymbolTypeData);
if (dispatch_queue_offsets_symbol == NULL)
{
Fixed an issue where we were resolving paths when we should have been. So the issue here was that we have lldb_private::FileSpec that by default was always resolving a path when using the: FileSpec::FileSpec (const char *path); and in the: void FileSpec::SetFile(const char *pathname, bool resolve = true); This isn't what we want in many many cases. One example is you have "/tmp" on your file system which is really "/private/tmp". You compile code in that directory and end up with debug info that mentions "/tmp/file.c". Then you type: (lldb) breakpoint set --file file.c --line 5 If your current working directory is "/tmp", then "file.c" would be turned into "/private/tmp/file.c" which won't match anything in the debug info. Also, it should have been just a FileSpec with no directory and a filename of "file.c" which could (and should) potentially match any instances of "file.c" in the debug info. So I removed the constructor that just takes a path: FileSpec::FileSpec (const char *path); // REMOVED You must now use the other constructor that has a "bool resolve" parameter that you must always supply: FileSpec::FileSpec (const char *path, bool resolve); I also removed the default parameter to SetFile(): void FileSpec::SetFile(const char *pathname, bool resolve); And fixed all of the code to use the right settings. llvm-svn: 116944
2010-10-20 20:54:39 +00:00
module_sp = GetTarget().GetImages().FindFirstModuleForFileSpec (FileSpec("libdispatch.dylib", false));
Fixed the dispatch queue name retrieval for threads by looking in an extra shlib. llvm-svn: 116315
2010-10-12 17:33:06 +00:00
if (module_sp)
dispatch_queue_offsets_symbol = module_sp->FindFirstSymbolWithNameAndType (g_dispatch_queue_offsets_symbol_name, eSymbolTypeData);
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (dispatch_queue_offsets_symbol)
Moved the section load list up into the target so we can use the target to symbolicate things without the need for a valid process subclass. llvm-svn: 113895
2010-09-14 23:36:40 +00:00
m_dispatch_queue_offsets_addr = dispatch_queue_offsets_symbol->GetValue().GetLoadAddress(&m_target);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-08 16:52:24 +00:00
if (m_dispatch_queue_offsets_addr == LLDB_INVALID_ADDRESS)
return NULL;
}
uint8_t memory_buffer[8];
DataExtractor data(memory_buffer, sizeof(memory_buffer), GetByteOrder(), GetAddressByteSize());
// Excerpt from src/queue_private.h
struct dispatch_queue_offsets_s
{
uint16_t dqo_version;
uint16_t dqo_label;
uint16_t dqo_label_size;
} dispatch_queue_offsets;
Error error;
if (ReadMemory (m_dispatch_queue_offsets_addr, memory_buffer, sizeof(dispatch_queue_offsets), error) == sizeof(dispatch_queue_offsets))
{
uint32_t data_offset = 0;
if (data.GetU16(&data_offset, &dispatch_queue_offsets.dqo_version, sizeof(dispatch_queue_offsets)/sizeof(uint16_t)))
{
if (ReadMemory (thread_dispatch_qaddr, &memory_buffer, data.GetAddressByteSize(), error) == data.GetAddressByteSize())
{
data_offset = 0;
lldb::addr_t queue_addr = data.GetAddress(&data_offset);
lldb::addr_t label_addr = queue_addr + dispatch_queue_offsets.dqo_label;
dispatch_queue_name.resize(dispatch_queue_offsets.dqo_label_size, '\0');
size_t bytes_read = ReadMemory (label_addr, &dispatch_queue_name[0], dispatch_queue_offsets.dqo_label_size, error);
if (bytes_read < dispatch_queue_offsets.dqo_label_size)
dispatch_queue_name.erase (bytes_read);
}
}
}
}
if (dispatch_queue_name.empty())
return NULL;
return dispatch_queue_name.c_str();
}
Change Target & Process so they can really be initialized with an invalid architecture. Arrange that this then gets properly set on attach, or when a "file" is set. Add a completer for "process attach -n". Caveats: there isn't currently a way to handle multiple processes with the same name. That will have to wait on a way to pass annotations along with the completion strings. llvm-svn: 110624
2010-08-09 23:31:02 +00:00
uint32_t
ProcessGDBRemote::ListProcessesMatchingName (const char *name, StringList &matches, std::vector<lldb::pid_t> &pids)
{
// If we are planning to launch the debugserver remotely, then we need to fire up a debugserver
// process and ask it for the list of processes. But if we are local, we can let the Host do it.
if (m_local_debugserver)
{
return Host::ListProcessesMatchingName (name, matches, pids);
}
else
{
// FIXME: Implement talking to the remote debugserver.
return 0;
}
}
Added an interface for noticing new thread creation. At this point, I only turn it on when we are requesting a single thread to run. May seem like a silly thing to do, but the kernel on MacOS X will inject new threads into a program willy-nilly, and I would like to keep them from running if I can. llvm-svn: 124018
2011-01-22 01:33:44 +00:00
bool
ProcessGDBRemote::NewThreadNotifyBreakpointHit (void *baton,
lldb_private::StoppointCallbackContext *context,
lldb::user_id_t break_id,
lldb::user_id_t break_loc_id)
{
// I don't think I have to do anything here, just make sure I notice the new thread when it starts to
// run so I can stop it if that's what I want to do.
LogSP log (lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_STEP));
if (log)
log->Printf("Hit New Thread Notification breakpoint.");
return false;
}
bool
ProcessGDBRemote::StartNoticingNewThreads()
{
static const char *bp_names[] =
{
"start_wqthread",
Add _pthread_wqthread to the list of thread stop points, and change the logging a bit. llvm-svn: 125081
2011-02-08 05:19:01 +00:00
"_pthread_wqthread",
Added an interface for noticing new thread creation. At this point, I only turn it on when we are requesting a single thread to run. May seem like a silly thing to do, but the kernel on MacOS X will inject new threads into a program willy-nilly, and I would like to keep them from running if I can. llvm-svn: 124018
2011-01-22 01:33:44 +00:00
"_pthread_start",
NULL
};
LogSP log (lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_STEP));
size_t num_bps = m_thread_observation_bps.size();
if (num_bps != 0)
{
for (int i = 0; i < num_bps; i++)
{
lldb::BreakpointSP break_sp = m_target.GetBreakpointByID(m_thread_observation_bps[i]);
if (break_sp)
{
if (log)
log->Printf("Enabled noticing new thread breakpoint.");
break_sp->SetEnabled(true);
}
}
}
else
{
for (int i = 0; bp_names[i] != NULL; i++)
{
Breakpoint *breakpoint = m_target.CreateBreakpoint (NULL, bp_names[i], eFunctionNameTypeFull, true).get();
if (breakpoint)
{
if (log)
log->Printf("Successfully created new thread notification breakpoint at \"%s\".", bp_names[i]);
m_thread_observation_bps.push_back(breakpoint->GetID());
breakpoint->SetCallback (ProcessGDBRemote::NewThreadNotifyBreakpointHit, this, true);
}
else
{
if (log)
log->Printf("Failed to create new thread notification breakpoint.");
return false;
}
}
}
return true;
}
bool
ProcessGDBRemote::StopNoticingNewThreads()
{
Add _pthread_wqthread to the list of thread stop points, and change the logging a bit. llvm-svn: 125081
2011-02-08 05:19:01 +00:00
LogSP log (lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_STEP));
if (log)
log->Printf ("Disabling new thread notification breakpoint.");
Added an interface for noticing new thread creation. At this point, I only turn it on when we are requesting a single thread to run. May seem like a silly thing to do, but the kernel on MacOS X will inject new threads into a program willy-nilly, and I would like to keep them from running if I can. llvm-svn: 124018
2011-01-22 01:33:44 +00:00
size_t num_bps = m_thread_observation_bps.size();
if (num_bps != 0)
{
for (int i = 0; i < num_bps; i++)
{
lldb::BreakpointSP break_sp = m_target.GetBreakpointByID(m_thread_observation_bps[i]);
if (break_sp)
{
break_sp->SetEnabled(false);
}
}
}
return true;
}
Reference in New Issue Copy Permalink