llvm-project/lldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationClien...

3361 lines
122 KiB
C++
Raw Normal View History

//===-- GDBRemoteCommunicationClient.cpp ------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "GDBRemoteCommunicationClient.h"
// C Includes
#include <sys/stat.h>
// C++ Includes
#include <sstream>
// Other libraries and framework includes
#include "llvm/ADT/Triple.h"
#include "lldb/Interpreter/Args.h"
#include "lldb/Core/ConnectionFileDescriptor.h"
#include "lldb/Core/Log.h"
#include "lldb/Core/State.h"
#include "lldb/Core/StreamGDBRemote.h"
#include "lldb/Core/StreamString.h"
#include "lldb/Host/Endian.h"
#include "lldb/Host/Host.h"
#include "lldb/Host/TimeValue.h"
// Project includes
#include "Utility/StringExtractorGDBRemote.h"
#include "ProcessGDBRemote.h"
#include "ProcessGDBRemoteLog.h"
#include "lldb/Host/Config.h"
using namespace lldb;
using namespace lldb_private;
#ifdef LLDB_DISABLE_POSIX
#define SIGSTOP 17
#endif
//----------------------------------------------------------------------
// GDBRemoteCommunicationClient constructor
//----------------------------------------------------------------------
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
GDBRemoteCommunicationClient::GDBRemoteCommunicationClient(bool is_platform) :
GDBRemoteCommunication("gdb-remote.client", "gdb-remote.client.rx_packet", is_platform),
m_supports_not_sending_acks (eLazyBoolCalculate),
m_supports_thread_suffix (eLazyBoolCalculate),
m_supports_threads_in_stop_reply (eLazyBoolCalculate),
m_supports_vCont_all (eLazyBoolCalculate),
m_supports_vCont_any (eLazyBoolCalculate),
m_supports_vCont_c (eLazyBoolCalculate),
m_supports_vCont_C (eLazyBoolCalculate),
m_supports_vCont_s (eLazyBoolCalculate),
m_supports_vCont_S (eLazyBoolCalculate),
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
m_qHostInfo_is_valid (eLazyBoolCalculate),
m_qProcessInfo_is_valid (eLazyBoolCalculate),
Added the ability to get the return value from a ThreadPlanCallFunction thread plan. In order to get the return value, you can call: void ThreadPlanCallFunction::RequestReturnValue (lldb::ValueSP &return_value_sp); This registers a shared pointer to a return value that will get filled in if everything goes well. After the thread plan is run the return value will be extracted for you. Added an ifdef to be able to switch between the LLVM MCJIT and the standand JIT. We currently have the standard JIT selected because we have some work to do to get the MCJIT fuctioning properly. Added the ability to call functions with 6 argument in the x86_64 ABI. Added the ability for GDBRemoteCommunicationClient to detect if the allocate and deallocate memory packets are supported and to not call allocate memory ("_M") or deallocate ("_m") if we find they aren't supported. Modified the ProcessGDBRemote::DoAllocateMemory(...) and ProcessGDBRemote::DoDeallocateMemory(...) to be able to deal with the allocate and deallocate memory packets not being supported. If they are not supported, ProcessGDBRemote will switch to calling "mmap" and "munmap" to allocate and deallocate memory instead using our trivial function call support. Modified the "void ProcessGDBRemote::DidLaunchOrAttach()" to correctly ignore the qHostInfo triple information if any was specified in the target. Currently if the target only specifies an architecture when creating the target: (lldb) target create --arch i386 a.out Then the vendor, os and environemnt will be adopted by the target. If the target was created with any triple that specifies more than the arch: (lldb) target create --arch i386-unknown-unknown a.out Then the target will maintain its triple and not adopt any new values. This can be used to help force bare board debugging where the dynamic loader for static files will get used and users can then use "target modules load ..." to set addressses for any files that are desired. Added back some convenience functions to the lldb_private::RegisterContext class for writing registers with unsigned values. Also made all RegisterContext constructors explicit to make sure we know when an integer is being converted to a RegisterValue. llvm-svn: 131370
2011-05-15 09:25:55 +08:00
m_supports_alloc_dealloc_memory (eLazyBoolCalculate),
m_supports_memory_region_info (eLazyBoolCalculate),
m_supports_watchpoint_support_info (eLazyBoolCalculate),
m_supports_detach_stay_stopped (eLazyBoolCalculate),
m_watchpoints_trigger_after_instruction(eLazyBoolCalculate),
m_attach_or_wait_reply(eLazyBoolCalculate),
m_prepare_for_reg_writing_reply (eLazyBoolCalculate),
m_supports_p (eLazyBoolCalculate),
m_supports_QSaveRegisterState (eLazyBoolCalculate),
m_supports_qXfer_auxv_read (eLazyBoolCalculate),
m_supports_qXfer_libraries_read (eLazyBoolCalculate),
m_supports_qXfer_libraries_svr4_read (eLazyBoolCalculate),
m_supports_augmented_libraries_svr4_read (eLazyBoolCalculate),
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
m_supports_qProcessInfoPID (true),
m_supports_qfProcessInfo (true),
m_supports_qUserName (true),
m_supports_qGroupName (true),
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
m_supports_qThreadStopInfo (true),
m_supports_z0 (true),
m_supports_z1 (true),
m_supports_z2 (true),
m_supports_z3 (true),
m_supports_z4 (true),
m_supports_QEnvironment (true),
m_supports_QEnvironmentHexEncoded (true),
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
m_curr_tid (LLDB_INVALID_THREAD_ID),
m_curr_tid_run (LLDB_INVALID_THREAD_ID),
m_num_supported_hardware_watchpoints (0),
m_async_mutex (Mutex::eMutexTypeRecursive),
m_async_packet_predicate (false),
m_async_packet (),
m_async_result (PacketResult::Success),
m_async_response (),
m_async_signal (-1),
m_thread_id_to_used_usec_map (),
m_host_arch(),
m_process_arch(),
m_os_version_major (UINT32_MAX),
m_os_version_minor (UINT32_MAX),
m_os_version_update (UINT32_MAX),
m_os_build (),
m_os_kernel (),
m_hostname (),
m_default_packet_timeout (0),
m_max_packet_size (0)
{
}
//----------------------------------------------------------------------
// Destructor
//----------------------------------------------------------------------
GDBRemoteCommunicationClient::~GDBRemoteCommunicationClient()
{
if (IsConnected())
Disconnect();
}
bool
GDBRemoteCommunicationClient::HandshakeWithServer (Error *error_ptr)
{
ResetDiscoverableSettings();
// Start the read thread after we send the handshake ack since if we
// fail to send the handshake ack, there is no reason to continue...
if (SendAck())
{
// Wait for any responses that might have been queued up in the remote
// GDB server and flush them all
StringExtractorGDBRemote response;
PacketResult packet_result = PacketResult::Success;
const uint32_t timeout_usec = 10 * 1000; // Wait for 10 ms for a response
while (packet_result == PacketResult::Success)
packet_result = WaitForPacketWithTimeoutMicroSecondsNoLock (response, timeout_usec);
// The return value from QueryNoAckModeSupported() is true if the packet
// was sent and _any_ response (including UNIMPLEMENTED) was received),
// or false if no response was received. This quickly tells us if we have
// a live connection to a remote GDB server...
if (QueryNoAckModeSupported())
{
#if 0
// Set above line to "#if 1" to test packet speed if remote GDB server
// supports the qSpeedTest packet...
TestPacketSpeed(10000);
#endif
return true;
}
else
{
if (error_ptr)
error_ptr->SetErrorString("failed to get reply to handshake packet");
}
}
else
{
if (error_ptr)
error_ptr->SetErrorString("failed to send the handshake ack");
}
return false;
}
bool
GDBRemoteCommunicationClient::GetAugmentedLibrariesSVR4ReadSupported ()
{
if (m_supports_augmented_libraries_svr4_read == eLazyBoolCalculate)
{
GetRemoteQSupported();
}
return (m_supports_augmented_libraries_svr4_read == eLazyBoolYes);
}
bool
GDBRemoteCommunicationClient::GetQXferLibrariesSVR4ReadSupported ()
{
if (m_supports_qXfer_libraries_svr4_read == eLazyBoolCalculate)
{
GetRemoteQSupported();
}
return (m_supports_qXfer_libraries_svr4_read == eLazyBoolYes);
}
bool
GDBRemoteCommunicationClient::GetQXferLibrariesReadSupported ()
{
if (m_supports_qXfer_libraries_read == eLazyBoolCalculate)
{
GetRemoteQSupported();
}
return (m_supports_qXfer_libraries_read == eLazyBoolYes);
}
bool
GDBRemoteCommunicationClient::GetQXferAuxvReadSupported ()
{
if (m_supports_qXfer_auxv_read == eLazyBoolCalculate)
{
GetRemoteQSupported();
}
return (m_supports_qXfer_auxv_read == eLazyBoolYes);
}
uint64_t
GDBRemoteCommunicationClient::GetRemoteMaxPacketSize()
{
if (m_max_packet_size == 0)
{
GetRemoteQSupported();
}
return m_max_packet_size;
}
bool
GDBRemoteCommunicationClient::QueryNoAckModeSupported ()
{
if (m_supports_not_sending_acks == eLazyBoolCalculate)
{
m_send_acks = true;
m_supports_not_sending_acks = eLazyBoolNo;
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("QStartNoAckMode", response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
{
m_send_acks = false;
m_supports_not_sending_acks = eLazyBoolYes;
}
return true;
}
}
return false;
}
void
GDBRemoteCommunicationClient::GetListThreadsInStopReplySupported ()
{
if (m_supports_threads_in_stop_reply == eLazyBoolCalculate)
{
m_supports_threads_in_stop_reply = eLazyBoolNo;
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("QListThreadsInStopReply", response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
m_supports_threads_in_stop_reply = eLazyBoolYes;
}
}
}
bool
GDBRemoteCommunicationClient::GetVAttachOrWaitSupported ()
{
if (m_attach_or_wait_reply == eLazyBoolCalculate)
{
m_attach_or_wait_reply = eLazyBoolNo;
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("qVAttachOrWaitSupported", response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
m_attach_or_wait_reply = eLazyBoolYes;
}
}
if (m_attach_or_wait_reply == eLazyBoolYes)
return true;
else
return false;
}
bool
GDBRemoteCommunicationClient::GetSyncThreadStateSupported ()
{
if (m_prepare_for_reg_writing_reply == eLazyBoolCalculate)
{
m_prepare_for_reg_writing_reply = eLazyBoolNo;
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("qSyncThreadStateSupported", response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
m_prepare_for_reg_writing_reply = eLazyBoolYes;
}
}
if (m_prepare_for_reg_writing_reply == eLazyBoolYes)
return true;
else
return false;
}
void
GDBRemoteCommunicationClient::ResetDiscoverableSettings()
{
m_supports_not_sending_acks = eLazyBoolCalculate;
m_supports_thread_suffix = eLazyBoolCalculate;
m_supports_threads_in_stop_reply = eLazyBoolCalculate;
m_supports_vCont_c = eLazyBoolCalculate;
m_supports_vCont_C = eLazyBoolCalculate;
m_supports_vCont_s = eLazyBoolCalculate;
m_supports_vCont_S = eLazyBoolCalculate;
m_supports_p = eLazyBoolCalculate;
m_supports_QSaveRegisterState = eLazyBoolCalculate;
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
m_qHostInfo_is_valid = eLazyBoolCalculate;
m_qProcessInfo_is_valid = eLazyBoolCalculate;
Added the ability to get the return value from a ThreadPlanCallFunction thread plan. In order to get the return value, you can call: void ThreadPlanCallFunction::RequestReturnValue (lldb::ValueSP &return_value_sp); This registers a shared pointer to a return value that will get filled in if everything goes well. After the thread plan is run the return value will be extracted for you. Added an ifdef to be able to switch between the LLVM MCJIT and the standand JIT. We currently have the standard JIT selected because we have some work to do to get the MCJIT fuctioning properly. Added the ability to call functions with 6 argument in the x86_64 ABI. Added the ability for GDBRemoteCommunicationClient to detect if the allocate and deallocate memory packets are supported and to not call allocate memory ("_M") or deallocate ("_m") if we find they aren't supported. Modified the ProcessGDBRemote::DoAllocateMemory(...) and ProcessGDBRemote::DoDeallocateMemory(...) to be able to deal with the allocate and deallocate memory packets not being supported. If they are not supported, ProcessGDBRemote will switch to calling "mmap" and "munmap" to allocate and deallocate memory instead using our trivial function call support. Modified the "void ProcessGDBRemote::DidLaunchOrAttach()" to correctly ignore the qHostInfo triple information if any was specified in the target. Currently if the target only specifies an architecture when creating the target: (lldb) target create --arch i386 a.out Then the vendor, os and environemnt will be adopted by the target. If the target was created with any triple that specifies more than the arch: (lldb) target create --arch i386-unknown-unknown a.out Then the target will maintain its triple and not adopt any new values. This can be used to help force bare board debugging where the dynamic loader for static files will get used and users can then use "target modules load ..." to set addressses for any files that are desired. Added back some convenience functions to the lldb_private::RegisterContext class for writing registers with unsigned values. Also made all RegisterContext constructors explicit to make sure we know when an integer is being converted to a RegisterValue. llvm-svn: 131370
2011-05-15 09:25:55 +08:00
m_supports_alloc_dealloc_memory = eLazyBoolCalculate;
m_supports_memory_region_info = eLazyBoolCalculate;
m_prepare_for_reg_writing_reply = eLazyBoolCalculate;
m_attach_or_wait_reply = eLazyBoolCalculate;
m_supports_qXfer_auxv_read = eLazyBoolCalculate;
m_supports_qXfer_libraries_read = eLazyBoolCalculate;
m_supports_qXfer_libraries_svr4_read = eLazyBoolCalculate;
m_supports_augmented_libraries_svr4_read = eLazyBoolCalculate;
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
m_supports_qProcessInfoPID = true;
m_supports_qfProcessInfo = true;
m_supports_qUserName = true;
m_supports_qGroupName = true;
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
m_supports_qThreadStopInfo = true;
m_supports_z0 = true;
m_supports_z1 = true;
m_supports_z2 = true;
m_supports_z3 = true;
m_supports_z4 = true;
m_supports_QEnvironment = true;
m_supports_QEnvironmentHexEncoded = true;
m_host_arch.Clear();
m_process_arch.Clear();
m_max_packet_size = 0;
}
void
GDBRemoteCommunicationClient::GetRemoteQSupported ()
{
// Clear out any capabilities we expect to see in the qSupported response
m_supports_qXfer_auxv_read = eLazyBoolNo;
m_supports_qXfer_libraries_read = eLazyBoolNo;
m_supports_qXfer_libraries_svr4_read = eLazyBoolNo;
m_supports_augmented_libraries_svr4_read = eLazyBoolNo;
m_max_packet_size = UINT64_MAX; // It's supposed to always be there, but if not, we assume no limit
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("qSupported",
response,
/*send_async=*/false) == PacketResult::Success)
{
const char *response_cstr = response.GetStringRef().c_str();
if (::strstr (response_cstr, "qXfer:auxv:read+"))
m_supports_qXfer_auxv_read = eLazyBoolYes;
if (::strstr (response_cstr, "qXfer:libraries-svr4:read+"))
m_supports_qXfer_libraries_svr4_read = eLazyBoolYes;
if (::strstr (response_cstr, "augmented-libraries-svr4-read"))
{
m_supports_qXfer_libraries_svr4_read = eLazyBoolYes; // implied
m_supports_augmented_libraries_svr4_read = eLazyBoolYes;
}
if (::strstr (response_cstr, "qXfer:libraries:read+"))
m_supports_qXfer_libraries_read = eLazyBoolYes;
const char *packet_size_str = ::strstr (response_cstr, "PacketSize=");
if (packet_size_str)
{
StringExtractorGDBRemote packet_response(packet_size_str + strlen("PacketSize="));
m_max_packet_size = packet_response.GetHexMaxU64(/*little_endian=*/false, UINT64_MAX);
if (m_max_packet_size == 0)
{
m_max_packet_size = UINT64_MAX; // Must have been a garbled response
Log *log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (log)
log->Printf ("Garbled PacketSize spec in qSupported response");
}
}
}
}
bool
GDBRemoteCommunicationClient::GetThreadSuffixSupported ()
{
if (m_supports_thread_suffix == eLazyBoolCalculate)
{
StringExtractorGDBRemote response;
m_supports_thread_suffix = eLazyBoolNo;
if (SendPacketAndWaitForResponse("QThreadSuffixSupported", response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
m_supports_thread_suffix = eLazyBoolYes;
}
}
return m_supports_thread_suffix;
}
bool
GDBRemoteCommunicationClient::GetVContSupported (char flavor)
{
if (m_supports_vCont_c == eLazyBoolCalculate)
{
StringExtractorGDBRemote response;
m_supports_vCont_any = eLazyBoolNo;
m_supports_vCont_all = eLazyBoolNo;
m_supports_vCont_c = eLazyBoolNo;
m_supports_vCont_C = eLazyBoolNo;
m_supports_vCont_s = eLazyBoolNo;
m_supports_vCont_S = eLazyBoolNo;
if (SendPacketAndWaitForResponse("vCont?", response, false) == PacketResult::Success)
{
const char *response_cstr = response.GetStringRef().c_str();
if (::strstr (response_cstr, ";c"))
m_supports_vCont_c = eLazyBoolYes;
if (::strstr (response_cstr, ";C"))
m_supports_vCont_C = eLazyBoolYes;
if (::strstr (response_cstr, ";s"))
m_supports_vCont_s = eLazyBoolYes;
if (::strstr (response_cstr, ";S"))
m_supports_vCont_S = eLazyBoolYes;
if (m_supports_vCont_c == eLazyBoolYes &&
m_supports_vCont_C == eLazyBoolYes &&
m_supports_vCont_s == eLazyBoolYes &&
m_supports_vCont_S == eLazyBoolYes)
{
m_supports_vCont_all = eLazyBoolYes;
}
if (m_supports_vCont_c == eLazyBoolYes ||
m_supports_vCont_C == eLazyBoolYes ||
m_supports_vCont_s == eLazyBoolYes ||
m_supports_vCont_S == eLazyBoolYes)
{
m_supports_vCont_any = eLazyBoolYes;
}
}
}
switch (flavor)
{
case 'a': return m_supports_vCont_any;
case 'A': return m_supports_vCont_all;
case 'c': return m_supports_vCont_c;
case 'C': return m_supports_vCont_C;
case 's': return m_supports_vCont_s;
case 'S': return m_supports_vCont_S;
default: break;
}
return false;
}
// Check if the target supports 'p' packet. It sends out a 'p'
// packet and checks the response. A normal packet will tell us
// that support is available.
//
// Takes a valid thread ID because p needs to apply to a thread.
bool
GDBRemoteCommunicationClient::GetpPacketSupported (lldb::tid_t tid)
{
if (m_supports_p == eLazyBoolCalculate)
{
StringExtractorGDBRemote response;
m_supports_p = eLazyBoolNo;
char packet[256];
if (GetThreadSuffixSupported())
snprintf(packet, sizeof(packet), "p0;thread:%" PRIx64 ";", tid);
else
snprintf(packet, sizeof(packet), "p0");
if (SendPacketAndWaitForResponse(packet, response, false) == PacketResult::Success)
{
if (response.IsNormalResponse())
m_supports_p = eLazyBoolYes;
}
}
return m_supports_p;
}
GDBRemoteCommunicationClient::PacketResult
GDBRemoteCommunicationClient::SendPacketsAndConcatenateResponses
(
const char *payload_prefix,
std::string &response_string
)
{
Mutex::Locker locker;
if (!GetSequenceMutex(locker,
"ProcessGDBRemote::SendPacketsAndConcatenateResponses() failed due to not getting the sequence mutex"))
{
Log *log (ProcessGDBRemoteLog::GetLogIfAnyCategoryIsSet (GDBR_LOG_PROCESS | GDBR_LOG_PACKETS));
if (log)
log->Printf("error: failed to get packet sequence mutex, not sending packets with prefix '%s'",
payload_prefix);
return PacketResult::ErrorNoSequenceLock;
}
response_string = "";
std::string payload_prefix_str(payload_prefix);
unsigned int response_size = 0x1000;
if (response_size > GetRemoteMaxPacketSize()) { // May send qSupported packet
response_size = GetRemoteMaxPacketSize();
}
for (unsigned int offset = 0; true; offset += response_size)
{
StringExtractorGDBRemote this_response;
// Construct payload
char sizeDescriptor[128];
snprintf(sizeDescriptor, sizeof(sizeDescriptor), "%x,%x", offset, response_size);
PacketResult result = SendPacketAndWaitForResponse((payload_prefix_str + sizeDescriptor).c_str(),
this_response,
/*send_async=*/false);
if (result != PacketResult::Success)
return result;
const std::string &this_string = this_response.GetStringRef();
// Check for m or l as first character; l seems to mean this is the last chunk
char first_char = *this_string.c_str();
if (first_char != 'm' && first_char != 'l')
{
return PacketResult::ErrorReplyInvalid;
}
// Concatenate the result so far (skipping 'm' or 'l')
response_string.append(this_string, 1, std::string::npos);
if (first_char == 'l')
// We're done
return PacketResult::Success;
}
}
GDBRemoteCommunicationClient::PacketResult
GDBRemoteCommunicationClient::SendPacketAndWaitForResponse
(
const char *payload,
StringExtractorGDBRemote &response,
bool send_async
)
{
return SendPacketAndWaitForResponse (payload,
::strlen (payload),
response,
send_async);
}
GDBRemoteCommunicationClient::PacketResult
GDBRemoteCommunicationClient::SendPacketAndWaitForResponseNoLock (const char *payload,
size_t payload_length,
StringExtractorGDBRemote &response)
{
PacketResult packet_result = SendPacketNoLock (payload, payload_length);
if (packet_result == PacketResult::Success)
packet_result = WaitForPacketWithTimeoutMicroSecondsNoLock (response, GetPacketTimeoutInMicroSeconds ());
return packet_result;
}
GDBRemoteCommunicationClient::PacketResult
GDBRemoteCommunicationClient::SendPacketAndWaitForResponse
(
const char *payload,
size_t payload_length,
StringExtractorGDBRemote &response,
bool send_async
)
{
PacketResult packet_result = PacketResult::ErrorSendFailed;
Mutex::Locker locker;
Log *log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (GetSequenceMutex (locker))
{
packet_result = SendPacketAndWaitForResponseNoLock (payload, payload_length, response);
}
else
{
if (send_async)
{
if (IsRunning())
{
Mutex::Locker async_locker (m_async_mutex);
m_async_packet.assign(payload, payload_length);
m_async_packet_predicate.SetValue (true, eBroadcastNever);
if (log)
log->Printf ("async: async packet = %s", m_async_packet.c_str());
bool timed_out = false;
if (SendInterrupt(locker, 2, timed_out))
{
if (m_interrupt_sent)
{
m_interrupt_sent = false;
TimeValue timeout_time;
timeout_time = TimeValue::Now();
timeout_time.OffsetWithSeconds (m_packet_timeout);
if (log)
log->Printf ("async: sent interrupt");
if (m_async_packet_predicate.WaitForValueEqualTo (false, &timeout_time, &timed_out))
{
if (log)
log->Printf ("async: got response");
// Swap the response buffer to avoid malloc and string copy
response.GetStringRef().swap (m_async_response.GetStringRef());
packet_result = m_async_result;
}
else
{
if (log)
log->Printf ("async: timed out waiting for response");
}
// Make sure we wait until the continue packet has been sent again...
if (m_private_is_running.WaitForValueEqualTo (true, &timeout_time, &timed_out))
{
if (log)
{
if (timed_out)
log->Printf ("async: timed out waiting for process to resume, but process was resumed");
else
log->Printf ("async: async packet sent");
}
}
else
{
if (log)
log->Printf ("async: timed out waiting for process to resume");
}
}
else
{
// We had a racy condition where we went to send the interrupt
// yet we were able to get the lock, so the process must have
// just stopped?
if (log)
log->Printf ("async: got lock without sending interrupt");
// Send the packet normally since we got the lock
packet_result = SendPacketAndWaitForResponseNoLock (payload, payload_length, response);
}
}
else
{
if (log)
log->Printf ("async: failed to interrupt");
}
}
else
{
if (log)
log->Printf ("async: not running, async is ignored");
}
}
else
{
if (log)
log->Printf("error: failed to get packet sequence mutex, not sending packet '%*s'", (int) payload_length, payload);
}
}
return packet_result;
}
static const char *end_delimiter = "--end--;";
static const int end_delimiter_len = 8;
std::string
GDBRemoteCommunicationClient::HarmonizeThreadIdsForProfileData
( ProcessGDBRemote *process,
StringExtractorGDBRemote& profileDataExtractor
)
{
std::map<uint64_t, uint32_t> new_thread_id_to_used_usec_map;
std::stringstream final_output;
std::string name, value;
// Going to assuming thread_used_usec comes first, else bail out.
while (profileDataExtractor.GetNameColonValue(name, value))
{
if (name.compare("thread_used_id") == 0)
{
StringExtractor threadIDHexExtractor(value.c_str());
uint64_t thread_id = threadIDHexExtractor.GetHexMaxU64(false, 0);
bool has_used_usec = false;
uint32_t curr_used_usec = 0;
std::string usec_name, usec_value;
uint32_t input_file_pos = profileDataExtractor.GetFilePos();
if (profileDataExtractor.GetNameColonValue(usec_name, usec_value))
{
if (usec_name.compare("thread_used_usec") == 0)
{
has_used_usec = true;
curr_used_usec = strtoull(usec_value.c_str(), NULL, 0);
}
else
{
// We didn't find what we want, it is probably
// an older version. Bail out.
profileDataExtractor.SetFilePos(input_file_pos);
}
}
if (has_used_usec)
{
uint32_t prev_used_usec = 0;
std::map<uint64_t, uint32_t>::iterator iterator = m_thread_id_to_used_usec_map.find(thread_id);
if (iterator != m_thread_id_to_used_usec_map.end())
{
prev_used_usec = m_thread_id_to_used_usec_map[thread_id];
}
uint32_t real_used_usec = curr_used_usec - prev_used_usec;
// A good first time record is one that runs for at least 0.25 sec
bool good_first_time = (prev_used_usec == 0) && (real_used_usec > 250000);
bool good_subsequent_time = (prev_used_usec > 0) &&
((real_used_usec > 0) || (process->HasAssignedIndexIDToThread(thread_id)));
if (good_first_time || good_subsequent_time)
{
// We try to avoid doing too many index id reservation,
// resulting in fast increase of index ids.
final_output << name << ":";
int32_t index_id = process->AssignIndexIDToThread(thread_id);
final_output << index_id << ";";
final_output << usec_name << ":" << usec_value << ";";
}
else
{
// Skip past 'thread_used_name'.
std::string local_name, local_value;
profileDataExtractor.GetNameColonValue(local_name, local_value);
}
// Store current time as previous time so that they can be compared later.
new_thread_id_to_used_usec_map[thread_id] = curr_used_usec;
}
else
{
// Bail out and use old string.
final_output << name << ":" << value << ";";
}
}
else
{
final_output << name << ":" << value << ";";
}
}
final_output << end_delimiter;
m_thread_id_to_used_usec_map = new_thread_id_to_used_usec_map;
return final_output.str();
}
StateType
GDBRemoteCommunicationClient::SendContinuePacketAndWaitForResponse
(
ProcessGDBRemote *process,
const char *payload,
size_t packet_length,
StringExtractorGDBRemote &response
)
{
m_curr_tid = LLDB_INVALID_THREAD_ID;
Log *log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PROCESS));
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s ()", __FUNCTION__);
Mutex::Locker locker(m_sequence_mutex);
StateType state = eStateRunning;
BroadcastEvent(eBroadcastBitRunPacketSent, NULL);
m_public_is_running.SetValue (true, eBroadcastNever);
// Set the starting continue packet into "continue_packet". This packet
// may change if we are interrupted and we continue after an async packet...
std::string continue_packet(payload, packet_length);
bool got_async_packet = false;
while (state == eStateRunning)
{
if (!got_async_packet)
{
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s () sending continue packet: %s", __FUNCTION__, continue_packet.c_str());
if (SendPacketNoLock(continue_packet.c_str(), continue_packet.size()) != PacketResult::Success)
state = eStateInvalid;
m_private_is_running.SetValue (true, eBroadcastAlways);
}
got_async_packet = false;
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s () WaitForPacket(%s)", __FUNCTION__, continue_packet.c_str());
if (WaitForPacketWithTimeoutMicroSecondsNoLock(response, UINT32_MAX) == PacketResult::Success)
{
if (response.Empty())
state = eStateInvalid;
else
{
const char stop_type = response.GetChar();
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s () got packet: %s", __FUNCTION__, response.GetStringRef().c_str());
switch (stop_type)
{
case 'T':
case 'S':
{
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
if (process->GetStopID() == 0)
{
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
if (process->GetID() == LLDB_INVALID_PROCESS_ID)
{
lldb::pid_t pid = GetCurrentProcessID ();
if (pid != LLDB_INVALID_PROCESS_ID)
process->SetID (pid);
}
process->BuildDynamicRegisterInfo (true);
}
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
// Privately notify any internal threads that we have stopped
// in case we wanted to interrupt our process, yet we might
// send a packet and continue without returning control to the
// user.
m_private_is_running.SetValue (false, eBroadcastAlways);
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
const uint8_t signo = response.GetHexU8 (UINT8_MAX);
bool continue_after_async = m_async_signal != -1 || m_async_packet_predicate.GetValue();
if (continue_after_async || m_interrupt_sent)
{
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
// We sent an interrupt packet to stop the inferior process
// for an async signal or to send an async packet while running
// but we might have been single stepping and received the
// stop packet for the step instead of for the interrupt packet.
// Typically when an interrupt is sent a SIGINT or SIGSTOP
// is used, so if we get anything else, we need to try and
// get another stop reply packet that may have been sent
// due to sending the interrupt when the target is stopped
// which will just re-send a copy of the last stop reply
// packet. If we don't do this, then the reply for our
// async packet will be the repeat stop reply packet and cause
// a lot of trouble for us!
if (signo != SIGINT && signo != SIGSTOP)
{
continue_after_async = false;
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
// We didn't get a a SIGINT or SIGSTOP, so try for a
// very brief time (1 ms) to get another stop reply
// packet to make sure it doesn't get in the way
StringExtractorGDBRemote extra_stop_reply_packet;
uint32_t timeout_usec = 1000;
if (WaitForPacketWithTimeoutMicroSecondsNoLock (extra_stop_reply_packet, timeout_usec) == PacketResult::Success)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
{
switch (extra_stop_reply_packet.GetChar())
{
case 'T':
case 'S':
// We did get an extra stop reply, which means
// our interrupt didn't stop the target so we
// shouldn't continue after the async signal
// or packet is sent...
continue_after_async = false;
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
break;
}
}
}
}
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
if (m_async_signal != -1)
{
if (log)
log->Printf ("async: send signo = %s", Host::GetSignalAsCString (m_async_signal));
// Save off the async signal we are supposed to send
const int async_signal = m_async_signal;
// Clear the async signal member so we don't end up
// sending the signal multiple times...
m_async_signal = -1;
// Check which signal we stopped with
if (signo == async_signal)
{
if (log)
log->Printf ("async: stopped with signal %s, we are done running", Host::GetSignalAsCString (signo));
// We already stopped with a signal that we wanted
// to stop with, so we are done
}
else
{
// We stopped with a different signal that the one
// we wanted to stop with, so now we must resume
// with the signal we want
char signal_packet[32];
int signal_packet_len = 0;
signal_packet_len = ::snprintf (signal_packet,
sizeof (signal_packet),
"C%2.2x",
async_signal);
if (log)
log->Printf ("async: stopped with signal %s, resume with %s",
Host::GetSignalAsCString (signo),
Host::GetSignalAsCString (async_signal));
// Set the continue packet to resume even if the
// interrupt didn't cause our stop (ignore continue_after_async)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
continue_packet.assign(signal_packet, signal_packet_len);
continue;
}
}
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
else if (m_async_packet_predicate.GetValue())
{
Log * packet_log (ProcessGDBRemoteLog::GetLogIfAllCategoriesSet (GDBR_LOG_PACKETS));
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
// We are supposed to send an asynchronous packet while
// we are running.
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
m_async_response.Clear();
if (m_async_packet.empty())
{
m_async_result = PacketResult::ErrorSendFailed;
if (packet_log)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
packet_log->Printf ("async: error: empty async packet");
}
else
{
if (packet_log)
packet_log->Printf ("async: sending packet");
m_async_result = SendPacketAndWaitForResponse (&m_async_packet[0],
m_async_packet.size(),
m_async_response,
false);
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
}
// Let the other thread that was trying to send the async
// packet know that the packet has been sent and response is
// ready...
m_async_packet_predicate.SetValue(false, eBroadcastAlways);
if (packet_log)
packet_log->Printf ("async: sent packet, continue_after_async = %i", continue_after_async);
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
// Set the continue packet to resume if our interrupt
// for the async packet did cause the stop
if (continue_after_async)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
{
// Reverting this for now as it is causing deadlocks
// in programs (<rdar://problem/11529853>). In the future
// we should check our thread list and "do the right thing"
// for new threads that show up while we stop and run async
// packets. Setting the packet to 'c' to continue all threads
// is the right thing to do 99.99% of the time because if a
// thread was single stepping, and we sent an interrupt, we
// will notice above that we didn't stop due to an interrupt
// but stopped due to stepping and we would _not_ continue.
continue_packet.assign (1, 'c');
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
continue;
}
}
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
// Stop with signal and thread info
state = eStateStopped;
}
break;
case 'W':
case 'X':
// process exited
state = eStateExited;
break;
case 'O':
// STDOUT
{
got_async_packet = true;
std::string inferior_stdout;
inferior_stdout.reserve(response.GetBytesLeft () / 2);
char ch;
while ((ch = response.GetHexU8()) != '\0')
inferior_stdout.append(1, ch);
process->AppendSTDOUT (inferior_stdout.c_str(), inferior_stdout.size());
}
break;
case 'A':
// Async miscellaneous reply. Right now, only profile data is coming through this channel.
{
got_async_packet = true;
std::string input = response.GetStringRef().substr(1); // '1' to move beyond 'A'
if (m_partial_profile_data.length() > 0)
{
m_partial_profile_data.append(input);
input = m_partial_profile_data;
m_partial_profile_data.clear();
}
size_t found, pos = 0, len = input.length();
while ((found = input.find(end_delimiter, pos)) != std::string::npos)
{
StringExtractorGDBRemote profileDataExtractor(input.substr(pos, found).c_str());
std::string profile_data = HarmonizeThreadIdsForProfileData(process, profileDataExtractor);
process->BroadcastAsyncProfileData (profile_data);
pos = found + end_delimiter_len;
}
if (pos < len)
{
// Last incomplete chunk.
m_partial_profile_data = input.substr(pos);
}
}
break;
case 'E':
// ERROR
state = eStateInvalid;
break;
default:
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s () unrecognized async packet", __FUNCTION__);
state = eStateInvalid;
break;
}
}
}
else
{
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s () WaitForPacket(...) => false", __FUNCTION__);
state = eStateInvalid;
}
}
if (log)
log->Printf ("GDBRemoteCommunicationClient::%s () => %s", __FUNCTION__, StateAsCString(state));
response.SetFilePos(0);
m_private_is_running.SetValue (false, eBroadcastAlways);
m_public_is_running.SetValue (false, eBroadcastAlways);
return state;
}
bool
GDBRemoteCommunicationClient::SendAsyncSignal (int signo)
{
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
Mutex::Locker async_locker (m_async_mutex);
m_async_signal = signo;
bool timed_out = false;
Mutex::Locker locker;
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
if (SendInterrupt (locker, 1, timed_out))
return true;
m_async_signal = -1;
return false;
}
// This function takes a mutex locker as a parameter in case the GetSequenceMutex
// actually succeeds. If it doesn't succeed in acquiring the sequence mutex
// (the expected result), then it will send the halt packet. If it does succeed
// then the caller that requested the interrupt will want to keep the sequence
// locked down so that no one else can send packets while the caller has control.
// This function usually gets called when we are running and need to stop the
// target. It can also be used when we are running and and we need to do something
// else (like read/write memory), so we need to interrupt the running process
// (gdb remote protocol requires this), and do what we need to do, then resume.
bool
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
GDBRemoteCommunicationClient::SendInterrupt
(
Mutex::Locker& locker,
uint32_t seconds_to_wait_for_stop,
bool &timed_out
)
{
timed_out = false;
Log *log (ProcessGDBRemoteLog::GetLogIfAnyCategoryIsSet (GDBR_LOG_PROCESS | GDBR_LOG_PACKETS));
if (IsRunning())
{
// Only send an interrupt if our debugserver is running...
if (GetSequenceMutex (locker))
{
if (log)
log->Printf ("SendInterrupt () - got sequence mutex without having to interrupt");
}
else
{
// Someone has the mutex locked waiting for a response or for the
// inferior to stop, so send the interrupt on the down low...
char ctrl_c = '\x03';
ConnectionStatus status = eConnectionStatusSuccess;
size_t bytes_written = Write (&ctrl_c, 1, status, NULL);
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
if (log)
log->PutCString("send packet: \\x03");
if (bytes_written > 0)
{
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
m_interrupt_sent = true;
if (seconds_to_wait_for_stop)
{
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
TimeValue timeout;
if (seconds_to_wait_for_stop)
{
timeout = TimeValue::Now();
timeout.OffsetWithSeconds (seconds_to_wait_for_stop);
}
if (m_private_is_running.WaitForValueEqualTo (false, &timeout, &timed_out))
{
if (log)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
log->PutCString ("SendInterrupt () - sent interrupt, private state stopped");
return true;
}
else
{
if (log)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
log->Printf ("SendInterrupt () - sent interrupt, timed out wating for async thread resume");
}
}
else
{
if (log)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
log->Printf ("SendInterrupt () - sent interrupt, not waiting for stop...");
return true;
}
}
else
{
if (log)
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
log->Printf ("SendInterrupt () - failed to write interrupt");
}
return false;
}
}
<rdar://problem/11052174> <rdar://problem/11051056> Found a race condition when sending async packets in the ProcessGDBRemote. A little background: GDB remote clients can only send one packet at a time. You must send a packet and wait for a response. So when we continue, we obviously can't hold up the calling thread waiting for the process to stop again, so we have an async thread in the ProcessGDBRemote whose only job is to run packets that control the inferior process. When you send a continue packet, the only packet you can send is an interrupt packet (which consists of sending a CTRL+C (or a '\x03' byte)). This then stops the inferior and we can send the async packet, and then resume the target. There was a race condition that often happened during stepping where we are doing a source level single step which consists of many instruction steps and a few runs here and there when we step into a function. So the flow looks like: inst single step inst single step inst single step inst single step inst single step step BP and run inst single step inst single step inst single step Now if we got an async packet while the program is running we get something like: send --> continue send --> interrupt recv <-- interrupt stop reply packet send --> async packet recv <-- async response send --> continue again and wait for actual stop Problems arise when this was happening when single stepping a thread where we would get: send --> step thread 123 send --> interrupt send --> stop reply for thread 123 (from the step) Now we _might_ have an extra stop reply packet from the "interrupt" which we weren't checking for and we could end up with: send --> async packet (like memory read!) recv <-- async response (which is the interrupt stop reply packet) Now we have the read memroy reply sitting in our buffer and waiting to be used as the reply for the next packet... To further complicate things, the single step should have exited the async thread since the run control is finished, but now it will continue if it was interrupted. The fixes I checked in to two major things: - watch for the extra stop reply if we need to - make sure we exit from the async thread run loop when the previous run control (like the instruction level single step) is finished. Needless to say this makes very fast stepping in Xcode much more reliable. llvm-svn: 153629
2012-03-29 09:55:41 +08:00
else
{
if (log)
log->Printf ("SendInterrupt () - not running");
}
return true;
}
lldb::pid_t
GDBRemoteCommunicationClient::GetCurrentProcessID ()
{
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("qC", strlen("qC"), response, false) == PacketResult::Success)
{
if (response.GetChar() == 'Q')
if (response.GetChar() == 'C')
return response.GetHexMaxU32 (false, LLDB_INVALID_PROCESS_ID);
}
return LLDB_INVALID_PROCESS_ID;
}
bool
GDBRemoteCommunicationClient::GetLaunchSuccess (std::string &error_str)
{
error_str.clear();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("qLaunchSuccess", strlen("qLaunchSuccess"), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return true;
if (response.GetChar() == 'E')
{
// A string the describes what failed when launching...
error_str = response.GetStringRef().substr(1);
}
else
{
error_str.assign ("unknown error occurred launching process");
}
}
else
{
error_str.assign ("timed out waiting for app to launch");
}
return false;
}
int
GDBRemoteCommunicationClient::SendArgumentsPacket (const ProcessLaunchInfo &launch_info)
{
// Since we don't get the send argv0 separate from the executable path, we need to
// make sure to use the actual exectuable path found in the launch_info...
std::vector<const char *> argv;
FileSpec exe_file = launch_info.GetExecutableFile();
std::string exe_path;
const char *arg = NULL;
const Args &launch_args = launch_info.GetArguments();
if (exe_file)
exe_path = exe_file.GetPath();
else
{
arg = launch_args.GetArgumentAtIndex(0);
if (arg)
exe_path = arg;
}
if (!exe_path.empty())
{
argv.push_back(exe_path.c_str());
for (uint32_t i=1; (arg = launch_args.GetArgumentAtIndex(i)) != NULL; ++i)
{
if (arg)
argv.push_back(arg);
}
}
if (!argv.empty())
{
StreamString packet;
packet.PutChar('A');
for (size_t i = 0, n = argv.size(); i < n; ++i)
{
arg = argv[i];
const int arg_len = strlen(arg);
if (i > 0)
packet.PutChar(',');
packet.Printf("%i,%i,", arg_len * 2, (int)i);
packet.PutBytesAsRawHex8 (arg, arg_len);
}
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
}
return -1;
}
int
GDBRemoteCommunicationClient::SendEnvironmentPacket (char const *name_equal_value)
{
if (name_equal_value && name_equal_value[0])
{
StreamString packet;
bool send_hex_encoding = false;
for (const char *p = name_equal_value; *p != '\0' && send_hex_encoding == false; ++p)
{
if (isprint(*p))
{
switch (*p)
{
case '$':
case '#':
send_hex_encoding = true;
break;
default:
break;
}
}
else
{
// We have non printable characters, lets hex encode this...
send_hex_encoding = true;
}
}
StringExtractorGDBRemote response;
if (send_hex_encoding)
{
if (m_supports_QEnvironmentHexEncoded)
{
packet.PutCString("QEnvironmentHexEncoded:");
packet.PutBytesAsRawHex8 (name_equal_value, strlen(name_equal_value));
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
if (response.IsUnsupportedResponse())
m_supports_QEnvironmentHexEncoded = false;
}
}
}
else if (m_supports_QEnvironment)
{
packet.Printf("QEnvironment:%s", name_equal_value);
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
if (response.IsUnsupportedResponse())
m_supports_QEnvironment = false;
}
}
}
return -1;
}
int
GDBRemoteCommunicationClient::SendLaunchArchPacket (char const *arch)
{
if (arch && arch[0])
{
StreamString packet;
packet.Printf("QLaunchArch:%s", arch);
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
}
return -1;
}
bool
GDBRemoteCommunicationClient::GetOSVersion (uint32_t &major,
uint32_t &minor,
uint32_t &update)
{
if (GetHostInfo ())
{
if (m_os_version_major != UINT32_MAX)
{
major = m_os_version_major;
minor = m_os_version_minor;
update = m_os_version_update;
return true;
}
}
return false;
}
bool
GDBRemoteCommunicationClient::GetOSBuildString (std::string &s)
{
if (GetHostInfo ())
{
if (!m_os_build.empty())
{
s = m_os_build;
return true;
}
}
s.clear();
return false;
}
bool
GDBRemoteCommunicationClient::GetOSKernelDescription (std::string &s)
{
if (GetHostInfo ())
{
if (!m_os_kernel.empty())
{
s = m_os_kernel;
return true;
}
}
s.clear();
return false;
}
bool
GDBRemoteCommunicationClient::GetHostname (std::string &s)
{
if (GetHostInfo ())
{
if (!m_hostname.empty())
{
s = m_hostname;
return true;
}
}
s.clear();
return false;
}
ArchSpec
GDBRemoteCommunicationClient::GetSystemArchitecture ()
{
if (GetHostInfo ())
return m_host_arch;
return ArchSpec();
}
const lldb_private::ArchSpec &
GDBRemoteCommunicationClient::GetProcessArchitecture ()
{
if (m_qProcessInfo_is_valid == eLazyBoolCalculate)
GetCurrentProcessInfo ();
return m_process_arch;
}
bool
GDBRemoteCommunicationClient::GetHostInfo (bool force)
{
if (force || m_qHostInfo_is_valid == eLazyBoolCalculate)
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
m_qHostInfo_is_valid = eLazyBoolNo;
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse ("qHostInfo", response, false) == PacketResult::Success)
{
if (response.IsNormalResponse())
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
std::string name;
std::string value;
uint32_t cpu = LLDB_INVALID_CPUTYPE;
uint32_t sub = 0;
std::string arch_name;
std::string os_name;
std::string vendor_name;
std::string triple;
std::string distribution_id;
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
uint32_t pointer_byte_size = 0;
StringExtractor extractor;
ByteOrder byte_order = eByteOrderInvalid;
uint32_t num_keys_decoded = 0;
while (response.GetNameColonValue(name, value))
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (name.compare("cputype") == 0)
{
// exception type in big endian hex
cpu = Args::StringToUInt32 (value.c_str(), LLDB_INVALID_CPUTYPE, 0);
if (cpu != LLDB_INVALID_CPUTYPE)
++num_keys_decoded;
}
else if (name.compare("cpusubtype") == 0)
{
// exception count in big endian hex
sub = Args::StringToUInt32 (value.c_str(), 0, 0);
if (sub != 0)
++num_keys_decoded;
}
else if (name.compare("arch") == 0)
{
arch_name.swap (value);
++num_keys_decoded;
}
else if (name.compare("triple") == 0)
{
// The triple comes as ASCII hex bytes since it contains '-' chars
extractor.GetStringRef().swap(value);
extractor.SetFilePos(0);
extractor.GetHexByteString (triple);
++num_keys_decoded;
}
else if (name.compare ("distribution_id") == 0)
{
extractor.GetStringRef ().swap (value);
extractor.SetFilePos (0);
extractor.GetHexByteString (distribution_id);
++num_keys_decoded;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
else if (name.compare("os_build") == 0)
{
extractor.GetStringRef().swap(value);
extractor.SetFilePos(0);
extractor.GetHexByteString (m_os_build);
++num_keys_decoded;
}
else if (name.compare("hostname") == 0)
{
extractor.GetStringRef().swap(value);
extractor.SetFilePos(0);
extractor.GetHexByteString (m_hostname);
++num_keys_decoded;
}
else if (name.compare("os_kernel") == 0)
{
extractor.GetStringRef().swap(value);
extractor.SetFilePos(0);
extractor.GetHexByteString (m_os_kernel);
++num_keys_decoded;
}
else if (name.compare("ostype") == 0)
{
os_name.swap (value);
++num_keys_decoded;
}
else if (name.compare("vendor") == 0)
{
vendor_name.swap(value);
++num_keys_decoded;
}
else if (name.compare("endian") == 0)
{
++num_keys_decoded;
if (value.compare("little") == 0)
byte_order = eByteOrderLittle;
else if (value.compare("big") == 0)
byte_order = eByteOrderBig;
else if (value.compare("pdp") == 0)
byte_order = eByteOrderPDP;
else
--num_keys_decoded;
}
else if (name.compare("ptrsize") == 0)
{
pointer_byte_size = Args::StringToUInt32 (value.c_str(), 0, 0);
if (pointer_byte_size != 0)
++num_keys_decoded;
}
else if (name.compare("os_version") == 0)
{
Args::StringToVersion (value.c_str(),
m_os_version_major,
m_os_version_minor,
m_os_version_update);
if (m_os_version_major != UINT32_MAX)
++num_keys_decoded;
}
else if (name.compare("watchpoint_exceptions_received") == 0)
{
++num_keys_decoded;
if (strcmp(value.c_str(),"before") == 0)
m_watchpoints_trigger_after_instruction = eLazyBoolNo;
else if (strcmp(value.c_str(),"after") == 0)
m_watchpoints_trigger_after_instruction = eLazyBoolYes;
else
--num_keys_decoded;
}
else if (name.compare("default_packet_timeout") == 0)
{
m_default_packet_timeout = Args::StringToUInt32(value.c_str(), 0);
if (m_default_packet_timeout > 0)
{
SetPacketTimeout(m_default_packet_timeout);
++num_keys_decoded;
}
}
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (num_keys_decoded > 0)
m_qHostInfo_is_valid = eLazyBoolYes;
if (triple.empty())
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (arch_name.empty())
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (cpu != LLDB_INVALID_CPUTYPE)
{
m_host_arch.SetArchitecture (eArchTypeMachO, cpu, sub);
if (pointer_byte_size)
{
assert (pointer_byte_size == m_host_arch.GetAddressByteSize());
}
if (byte_order != eByteOrderInvalid)
{
assert (byte_order == m_host_arch.GetByteOrder());
}
if (!os_name.empty() && vendor_name.compare("apple") == 0 && os_name.find("darwin") == 0)
{
switch (m_host_arch.GetMachine())
{
case llvm::Triple::arm:
case llvm::Triple::thumb:
os_name = "ios";
break;
default:
os_name = "macosx";
break;
}
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (!vendor_name.empty())
m_host_arch.GetTriple().setVendorName (llvm::StringRef (vendor_name));
if (!os_name.empty())
m_host_arch.GetTriple().setOSName (llvm::StringRef (os_name));
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
}
else
{
std::string triple;
triple += arch_name;
if (!vendor_name.empty() || !os_name.empty())
{
triple += '-';
if (vendor_name.empty())
triple += "unknown";
else
triple += vendor_name;
triple += '-';
if (os_name.empty())
triple += "unknown";
else
triple += os_name;
}
m_host_arch.SetTriple (triple.c_str());
llvm::Triple &host_triple = m_host_arch.GetTriple();
if (host_triple.getVendor() == llvm::Triple::Apple && host_triple.getOS() == llvm::Triple::Darwin)
{
switch (m_host_arch.GetMachine())
{
case llvm::Triple::arm:
case llvm::Triple::thumb:
host_triple.setOS(llvm::Triple::IOS);
break;
default:
host_triple.setOS(llvm::Triple::MacOSX);
break;
}
}
if (pointer_byte_size)
{
assert (pointer_byte_size == m_host_arch.GetAddressByteSize());
}
if (byte_order != eByteOrderInvalid)
{
assert (byte_order == m_host_arch.GetByteOrder());
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
}
else
{
m_host_arch.SetTriple (triple.c_str());
if (pointer_byte_size)
{
assert (pointer_byte_size == m_host_arch.GetAddressByteSize());
}
if (byte_order != eByteOrderInvalid)
{
assert (byte_order == m_host_arch.GetByteOrder());
}
}
if (!distribution_id.empty ())
m_host_arch.SetDistributionId (distribution_id.c_str ());
}
}
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
return m_qHostInfo_is_valid == eLazyBoolYes;
}
int
GDBRemoteCommunicationClient::SendAttach
(
lldb::pid_t pid,
StringExtractorGDBRemote& response
)
{
if (pid != LLDB_INVALID_PROCESS_ID)
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
char packet[64];
const int packet_len = ::snprintf (packet, sizeof(packet), "vAttach;%" PRIx64, pid);
assert (packet_len < (int)sizeof(packet));
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
if (response.IsErrorResponse())
return response.GetError();
return 0;
}
}
return -1;
}
const lldb_private::ArchSpec &
GDBRemoteCommunicationClient::GetHostArchitecture ()
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (m_qHostInfo_is_valid == eLazyBoolCalculate)
GetHostInfo ();
return m_host_arch;
}
uint32_t
GDBRemoteCommunicationClient::GetHostDefaultPacketTimeout ()
{
if (m_qHostInfo_is_valid == eLazyBoolCalculate)
GetHostInfo ();
return m_default_packet_timeout;
}
addr_t
GDBRemoteCommunicationClient::AllocateMemory (size_t size, uint32_t permissions)
{
Added the ability to get the return value from a ThreadPlanCallFunction thread plan. In order to get the return value, you can call: void ThreadPlanCallFunction::RequestReturnValue (lldb::ValueSP &return_value_sp); This registers a shared pointer to a return value that will get filled in if everything goes well. After the thread plan is run the return value will be extracted for you. Added an ifdef to be able to switch between the LLVM MCJIT and the standand JIT. We currently have the standard JIT selected because we have some work to do to get the MCJIT fuctioning properly. Added the ability to call functions with 6 argument in the x86_64 ABI. Added the ability for GDBRemoteCommunicationClient to detect if the allocate and deallocate memory packets are supported and to not call allocate memory ("_M") or deallocate ("_m") if we find they aren't supported. Modified the ProcessGDBRemote::DoAllocateMemory(...) and ProcessGDBRemote::DoDeallocateMemory(...) to be able to deal with the allocate and deallocate memory packets not being supported. If they are not supported, ProcessGDBRemote will switch to calling "mmap" and "munmap" to allocate and deallocate memory instead using our trivial function call support. Modified the "void ProcessGDBRemote::DidLaunchOrAttach()" to correctly ignore the qHostInfo triple information if any was specified in the target. Currently if the target only specifies an architecture when creating the target: (lldb) target create --arch i386 a.out Then the vendor, os and environemnt will be adopted by the target. If the target was created with any triple that specifies more than the arch: (lldb) target create --arch i386-unknown-unknown a.out Then the target will maintain its triple and not adopt any new values. This can be used to help force bare board debugging where the dynamic loader for static files will get used and users can then use "target modules load ..." to set addressses for any files that are desired. Added back some convenience functions to the lldb_private::RegisterContext class for writing registers with unsigned values. Also made all RegisterContext constructors explicit to make sure we know when an integer is being converted to a RegisterValue. llvm-svn: 131370
2011-05-15 09:25:55 +08:00
if (m_supports_alloc_dealloc_memory != eLazyBoolNo)
{
Added the ability to get the return value from a ThreadPlanCallFunction thread plan. In order to get the return value, you can call: void ThreadPlanCallFunction::RequestReturnValue (lldb::ValueSP &return_value_sp); This registers a shared pointer to a return value that will get filled in if everything goes well. After the thread plan is run the return value will be extracted for you. Added an ifdef to be able to switch between the LLVM MCJIT and the standand JIT. We currently have the standard JIT selected because we have some work to do to get the MCJIT fuctioning properly. Added the ability to call functions with 6 argument in the x86_64 ABI. Added the ability for GDBRemoteCommunicationClient to detect if the allocate and deallocate memory packets are supported and to not call allocate memory ("_M") or deallocate ("_m") if we find they aren't supported. Modified the ProcessGDBRemote::DoAllocateMemory(...) and ProcessGDBRemote::DoDeallocateMemory(...) to be able to deal with the allocate and deallocate memory packets not being supported. If they are not supported, ProcessGDBRemote will switch to calling "mmap" and "munmap" to allocate and deallocate memory instead using our trivial function call support. Modified the "void ProcessGDBRemote::DidLaunchOrAttach()" to correctly ignore the qHostInfo triple information if any was specified in the target. Currently if the target only specifies an architecture when creating the target: (lldb) target create --arch i386 a.out Then the vendor, os and environemnt will be adopted by the target. If the target was created with any triple that specifies more than the arch: (lldb) target create --arch i386-unknown-unknown a.out Then the target will maintain its triple and not adopt any new values. This can be used to help force bare board debugging where the dynamic loader for static files will get used and users can then use "target modules load ..." to set addressses for any files that are desired. Added back some convenience functions to the lldb_private::RegisterContext class for writing registers with unsigned values. Also made all RegisterContext constructors explicit to make sure we know when an integer is being converted to a RegisterValue. llvm-svn: 131370
2011-05-15 09:25:55 +08:00
m_supports_alloc_dealloc_memory = eLazyBoolYes;
char packet[64];
const int packet_len = ::snprintf (packet, sizeof(packet), "_M%" PRIx64 ",%s%s%s",
(uint64_t)size,
permissions & lldb::ePermissionsReadable ? "r" : "",
permissions & lldb::ePermissionsWritable ? "w" : "",
permissions & lldb::ePermissionsExecutable ? "x" : "");
assert (packet_len < (int)sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
if (!response.IsErrorResponse())
return response.GetHexMaxU64(false, LLDB_INVALID_ADDRESS);
}
else
{
m_supports_alloc_dealloc_memory = eLazyBoolNo;
}
}
return LLDB_INVALID_ADDRESS;
}
bool
GDBRemoteCommunicationClient::DeallocateMemory (addr_t addr)
{
Added the ability to get the return value from a ThreadPlanCallFunction thread plan. In order to get the return value, you can call: void ThreadPlanCallFunction::RequestReturnValue (lldb::ValueSP &return_value_sp); This registers a shared pointer to a return value that will get filled in if everything goes well. After the thread plan is run the return value will be extracted for you. Added an ifdef to be able to switch between the LLVM MCJIT and the standand JIT. We currently have the standard JIT selected because we have some work to do to get the MCJIT fuctioning properly. Added the ability to call functions with 6 argument in the x86_64 ABI. Added the ability for GDBRemoteCommunicationClient to detect if the allocate and deallocate memory packets are supported and to not call allocate memory ("_M") or deallocate ("_m") if we find they aren't supported. Modified the ProcessGDBRemote::DoAllocateMemory(...) and ProcessGDBRemote::DoDeallocateMemory(...) to be able to deal with the allocate and deallocate memory packets not being supported. If they are not supported, ProcessGDBRemote will switch to calling "mmap" and "munmap" to allocate and deallocate memory instead using our trivial function call support. Modified the "void ProcessGDBRemote::DidLaunchOrAttach()" to correctly ignore the qHostInfo triple information if any was specified in the target. Currently if the target only specifies an architecture when creating the target: (lldb) target create --arch i386 a.out Then the vendor, os and environemnt will be adopted by the target. If the target was created with any triple that specifies more than the arch: (lldb) target create --arch i386-unknown-unknown a.out Then the target will maintain its triple and not adopt any new values. This can be used to help force bare board debugging where the dynamic loader for static files will get used and users can then use "target modules load ..." to set addressses for any files that are desired. Added back some convenience functions to the lldb_private::RegisterContext class for writing registers with unsigned values. Also made all RegisterContext constructors explicit to make sure we know when an integer is being converted to a RegisterValue. llvm-svn: 131370
2011-05-15 09:25:55 +08:00
if (m_supports_alloc_dealloc_memory != eLazyBoolNo)
{
Added the ability to get the return value from a ThreadPlanCallFunction thread plan. In order to get the return value, you can call: void ThreadPlanCallFunction::RequestReturnValue (lldb::ValueSP &return_value_sp); This registers a shared pointer to a return value that will get filled in if everything goes well. After the thread plan is run the return value will be extracted for you. Added an ifdef to be able to switch between the LLVM MCJIT and the standand JIT. We currently have the standard JIT selected because we have some work to do to get the MCJIT fuctioning properly. Added the ability to call functions with 6 argument in the x86_64 ABI. Added the ability for GDBRemoteCommunicationClient to detect if the allocate and deallocate memory packets are supported and to not call allocate memory ("_M") or deallocate ("_m") if we find they aren't supported. Modified the ProcessGDBRemote::DoAllocateMemory(...) and ProcessGDBRemote::DoDeallocateMemory(...) to be able to deal with the allocate and deallocate memory packets not being supported. If they are not supported, ProcessGDBRemote will switch to calling "mmap" and "munmap" to allocate and deallocate memory instead using our trivial function call support. Modified the "void ProcessGDBRemote::DidLaunchOrAttach()" to correctly ignore the qHostInfo triple information if any was specified in the target. Currently if the target only specifies an architecture when creating the target: (lldb) target create --arch i386 a.out Then the vendor, os and environemnt will be adopted by the target. If the target was created with any triple that specifies more than the arch: (lldb) target create --arch i386-unknown-unknown a.out Then the target will maintain its triple and not adopt any new values. This can be used to help force bare board debugging where the dynamic loader for static files will get used and users can then use "target modules load ..." to set addressses for any files that are desired. Added back some convenience functions to the lldb_private::RegisterContext class for writing registers with unsigned values. Also made all RegisterContext constructors explicit to make sure we know when an integer is being converted to a RegisterValue. llvm-svn: 131370
2011-05-15 09:25:55 +08:00
m_supports_alloc_dealloc_memory = eLazyBoolYes;
char packet[64];
const int packet_len = ::snprintf(packet, sizeof(packet), "_m%" PRIx64, (uint64_t)addr);
assert (packet_len < (int)sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return true;
}
else
{
m_supports_alloc_dealloc_memory = eLazyBoolNo;
}
}
return false;
}
Error
GDBRemoteCommunicationClient::Detach (bool keep_stopped)
{
Error error;
if (keep_stopped)
{
if (m_supports_detach_stay_stopped == eLazyBoolCalculate)
{
char packet[64];
const int packet_len = ::snprintf(packet, sizeof(packet), "qSupportsDetachAndStayStopped:");
assert (packet_len < (int)sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
m_supports_detach_stay_stopped = eLazyBoolYes;
}
else
{
m_supports_detach_stay_stopped = eLazyBoolNo;
}
}
if (m_supports_detach_stay_stopped == eLazyBoolNo)
{
error.SetErrorString("Stays stopped not supported by this target.");
return error;
}
else
{
StringExtractorGDBRemote response;
PacketResult packet_result = SendPacketAndWaitForResponse ("D1", 1, response, false);
if (packet_result != PacketResult::Success)
error.SetErrorString ("Sending extended disconnect packet failed.");
}
}
else
{
StringExtractorGDBRemote response;
PacketResult packet_result = SendPacketAndWaitForResponse ("D", 1, response, false);
if (packet_result != PacketResult::Success)
error.SetErrorString ("Sending disconnect packet failed.");
}
return error;
}
Error
GDBRemoteCommunicationClient::GetMemoryRegionInfo (lldb::addr_t addr,
lldb_private::MemoryRegionInfo &region_info)
{
Error error;
region_info.Clear();
if (m_supports_memory_region_info != eLazyBoolNo)
{
m_supports_memory_region_info = eLazyBoolYes;
char packet[64];
const int packet_len = ::snprintf(packet, sizeof(packet), "qMemoryRegionInfo:%" PRIx64, (uint64_t)addr);
assert (packet_len < (int)sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
std::string name;
std::string value;
addr_t addr_value;
bool success = true;
bool saw_permissions = false;
while (success && response.GetNameColonValue(name, value))
{
if (name.compare ("start") == 0)
{
addr_value = Args::StringToUInt64(value.c_str(), LLDB_INVALID_ADDRESS, 16, &success);
if (success)
region_info.GetRange().SetRangeBase(addr_value);
}
else if (name.compare ("size") == 0)
{
addr_value = Args::StringToUInt64(value.c_str(), 0, 16, &success);
if (success)
region_info.GetRange().SetByteSize (addr_value);
}
else if (name.compare ("permissions") == 0 && region_info.GetRange().IsValid())
{
saw_permissions = true;
if (region_info.GetRange().Contains (addr))
{
if (value.find('r') != std::string::npos)
region_info.SetReadable (MemoryRegionInfo::eYes);
else
region_info.SetReadable (MemoryRegionInfo::eNo);
if (value.find('w') != std::string::npos)
region_info.SetWritable (MemoryRegionInfo::eYes);
else
region_info.SetWritable (MemoryRegionInfo::eNo);
if (value.find('x') != std::string::npos)
region_info.SetExecutable (MemoryRegionInfo::eYes);
else
region_info.SetExecutable (MemoryRegionInfo::eNo);
}
else
{
// The reported region does not contain this address -- we're looking at an unmapped page
region_info.SetReadable (MemoryRegionInfo::eNo);
region_info.SetWritable (MemoryRegionInfo::eNo);
region_info.SetExecutable (MemoryRegionInfo::eNo);
}
}
else if (name.compare ("error") == 0)
{
StringExtractorGDBRemote 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);
error.SetErrorString(value.c_str());
}
}
// We got a valid address range back but no permissions -- which means this is an unmapped page
if (region_info.GetRange().IsValid() && saw_permissions == false)
{
region_info.SetReadable (MemoryRegionInfo::eNo);
region_info.SetWritable (MemoryRegionInfo::eNo);
region_info.SetExecutable (MemoryRegionInfo::eNo);
}
}
else
{
m_supports_memory_region_info = eLazyBoolNo;
}
}
if (m_supports_memory_region_info == eLazyBoolNo)
{
error.SetErrorString("qMemoryRegionInfo is not supported");
}
if (error.Fail())
region_info.Clear();
return error;
}
Error
GDBRemoteCommunicationClient::GetWatchpointSupportInfo (uint32_t &num)
{
Error error;
if (m_supports_watchpoint_support_info == eLazyBoolYes)
{
num = m_num_supported_hardware_watchpoints;
return error;
}
// Set num to 0 first.
num = 0;
if (m_supports_watchpoint_support_info != eLazyBoolNo)
{
char packet[64];
const int packet_len = ::snprintf(packet, sizeof(packet), "qWatchpointSupportInfo:");
assert (packet_len < (int)sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
m_supports_watchpoint_support_info = eLazyBoolYes;
std::string name;
std::string value;
while (response.GetNameColonValue(name, value))
{
if (name.compare ("num") == 0)
{
num = Args::StringToUInt32(value.c_str(), 0, 0);
m_num_supported_hardware_watchpoints = num;
}
}
}
else
{
m_supports_watchpoint_support_info = eLazyBoolNo;
}
}
if (m_supports_watchpoint_support_info == eLazyBoolNo)
{
error.SetErrorString("qWatchpointSupportInfo is not supported");
}
return error;
}
lldb_private::Error
GDBRemoteCommunicationClient::GetWatchpointSupportInfo (uint32_t &num, bool& after)
{
Error error(GetWatchpointSupportInfo(num));
if (error.Success())
error = GetWatchpointsTriggerAfterInstruction(after);
return error;
}
lldb_private::Error
GDBRemoteCommunicationClient::GetWatchpointsTriggerAfterInstruction (bool &after)
{
Error error;
// we assume watchpoints will happen after running the relevant opcode
// and we only want to override this behavior if we have explicitly
// received a qHostInfo telling us otherwise
if (m_qHostInfo_is_valid != eLazyBoolYes)
after = true;
else
after = (m_watchpoints_trigger_after_instruction != eLazyBoolNo);
return error;
}
int
GDBRemoteCommunicationClient::SetSTDIN (char const *path)
{
if (path && path[0])
{
StreamString packet;
packet.PutCString("QSetSTDIN:");
packet.PutBytesAsRawHex8(path, strlen(path));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
}
return -1;
}
int
GDBRemoteCommunicationClient::SetSTDOUT (char const *path)
{
if (path && path[0])
{
StreamString packet;
packet.PutCString("QSetSTDOUT:");
packet.PutBytesAsRawHex8(path, strlen(path));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
}
return -1;
}
int
GDBRemoteCommunicationClient::SetSTDERR (char const *path)
{
if (path && path[0])
{
StreamString packet;
packet.PutCString("QSetSTDERR:");
packet.PutBytesAsRawHex8(path, strlen(path));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
}
return -1;
}
bool
GDBRemoteCommunicationClient::GetWorkingDir (std::string &cwd)
{
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse ("qGetWorkingDir", response, false) == PacketResult::Success)
{
if (response.IsUnsupportedResponse())
return false;
if (response.IsErrorResponse())
return false;
response.GetHexByteString (cwd);
return !cwd.empty();
}
return false;
}
int
GDBRemoteCommunicationClient::SetWorkingDir (char const *path)
{
if (path && path[0])
{
StreamString packet;
packet.PutCString("QSetWorkingDir:");
packet.PutBytesAsRawHex8(path, strlen(path));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
}
return -1;
}
int
GDBRemoteCommunicationClient::SetDisableASLR (bool enable)
{
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
char packet[32];
const int packet_len = ::snprintf (packet, sizeof (packet), "QSetDisableASLR:%i", enable ? 1 : 0);
assert (packet_len < (int)sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return 0;
uint8_t error = response.GetError();
if (error)
return error;
}
return -1;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
bool
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
GDBRemoteCommunicationClient::DecodeProcessInfoResponse (StringExtractorGDBRemote &response, ProcessInstanceInfo &process_info)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
if (response.IsNormalResponse())
{
std::string name;
std::string value;
StringExtractor extractor;
uint32_t cpu = LLDB_INVALID_CPUTYPE;
uint32_t sub = 0;
std::string vendor;
std::string os_type;
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
while (response.GetNameColonValue(name, value))
{
if (name.compare("pid") == 0)
{
process_info.SetProcessID (Args::StringToUInt32 (value.c_str(), LLDB_INVALID_PROCESS_ID, 0));
}
else if (name.compare("ppid") == 0)
{
process_info.SetParentProcessID (Args::StringToUInt32 (value.c_str(), LLDB_INVALID_PROCESS_ID, 0));
}
else if (name.compare("uid") == 0)
{
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
process_info.SetUserID (Args::StringToUInt32 (value.c_str(), UINT32_MAX, 0));
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
else if (name.compare("euid") == 0)
{
process_info.SetEffectiveUserID (Args::StringToUInt32 (value.c_str(), UINT32_MAX, 0));
}
else if (name.compare("gid") == 0)
{
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
process_info.SetGroupID (Args::StringToUInt32 (value.c_str(), UINT32_MAX, 0));
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
else if (name.compare("egid") == 0)
{
process_info.SetEffectiveGroupID (Args::StringToUInt32 (value.c_str(), UINT32_MAX, 0));
}
else if (name.compare("triple") == 0)
{
// The triple comes as ASCII hex bytes since it contains '-' chars
extractor.GetStringRef().swap(value);
extractor.SetFilePos(0);
extractor.GetHexByteString (value);
process_info.GetArchitecture ().SetTriple (value.c_str());
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
else if (name.compare("name") == 0)
{
StringExtractor extractor;
2012-05-07 17:30:51 +08:00
// The process name from ASCII hex bytes since we can't
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
// control the characters in a process name
extractor.GetStringRef().swap(value);
extractor.SetFilePos(0);
extractor.GetHexByteString (value);
process_info.GetExecutableFile().SetFile (value.c_str(), false);
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
else if (name.compare("cputype") == 0)
{
cpu = Args::StringToUInt32 (value.c_str(), LLDB_INVALID_CPUTYPE, 16);
}
else if (name.compare("cpusubtype") == 0)
{
sub = Args::StringToUInt32 (value.c_str(), 0, 16);
}
else if (name.compare("vendor") == 0)
{
vendor = value;
}
else if (name.compare("ostype") == 0)
{
os_type = value;
}
}
if (cpu != LLDB_INVALID_CPUTYPE && !vendor.empty() && !os_type.empty())
{
if (vendor == "apple")
{
process_info.GetArchitecture().SetArchitecture (eArchTypeMachO, cpu, sub);
process_info.GetArchitecture().GetTriple().setVendorName (llvm::StringRef (vendor));
process_info.GetArchitecture().GetTriple().setOSName (llvm::StringRef (os_type));
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
if (process_info.GetProcessID() != LLDB_INVALID_PROCESS_ID)
return true;
}
return false;
}
bool
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
GDBRemoteCommunicationClient::GetProcessInfo (lldb::pid_t pid, ProcessInstanceInfo &process_info)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
process_info.Clear();
if (m_supports_qProcessInfoPID)
{
char packet[32];
const int packet_len = ::snprintf (packet, sizeof (packet), "qProcessInfoPID:%" PRIu64, pid);
assert (packet_len < (int)sizeof(packet));
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
return DecodeProcessInfoResponse (response, process_info);
}
else
{
m_supports_qProcessInfoPID = false;
return false;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
return false;
}
bool
GDBRemoteCommunicationClient::GetCurrentProcessInfo ()
{
if (m_qProcessInfo_is_valid == eLazyBoolYes)
return true;
if (m_qProcessInfo_is_valid == eLazyBoolNo)
return false;
GetHostInfo ();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse ("qProcessInfo", response, false) == PacketResult::Success)
{
if (response.IsNormalResponse())
{
std::string name;
std::string value;
uint32_t cpu = LLDB_INVALID_CPUTYPE;
uint32_t sub = 0;
std::string arch_name;
std::string os_name;
std::string vendor_name;
std::string triple;
uint32_t pointer_byte_size = 0;
StringExtractor extractor;
ByteOrder byte_order = eByteOrderInvalid;
uint32_t num_keys_decoded = 0;
while (response.GetNameColonValue(name, value))
{
if (name.compare("cputype") == 0)
{
cpu = Args::StringToUInt32 (value.c_str(), LLDB_INVALID_CPUTYPE, 16);
if (cpu != LLDB_INVALID_CPUTYPE)
++num_keys_decoded;
}
else if (name.compare("cpusubtype") == 0)
{
sub = Args::StringToUInt32 (value.c_str(), 0, 16);
if (sub != 0)
++num_keys_decoded;
}
else if (name.compare("ostype") == 0)
{
os_name.swap (value);
++num_keys_decoded;
}
else if (name.compare("vendor") == 0)
{
vendor_name.swap(value);
++num_keys_decoded;
}
else if (name.compare("endian") == 0)
{
++num_keys_decoded;
if (value.compare("little") == 0)
byte_order = eByteOrderLittle;
else if (value.compare("big") == 0)
byte_order = eByteOrderBig;
else if (value.compare("pdp") == 0)
byte_order = eByteOrderPDP;
else
--num_keys_decoded;
}
else if (name.compare("ptrsize") == 0)
{
pointer_byte_size = Args::StringToUInt32 (value.c_str(), 0, 16);
if (pointer_byte_size != 0)
++num_keys_decoded;
}
}
if (num_keys_decoded > 0)
m_qProcessInfo_is_valid = eLazyBoolYes;
if (cpu != LLDB_INVALID_CPUTYPE && !os_name.empty() && !vendor_name.empty())
{
m_process_arch.SetArchitecture (eArchTypeMachO, cpu, sub);
if (pointer_byte_size)
{
assert (pointer_byte_size == m_process_arch.GetAddressByteSize());
}
m_host_arch.GetTriple().setVendorName (llvm::StringRef (vendor_name));
m_host_arch.GetTriple().setOSName (llvm::StringRef (os_name));
return true;
}
}
}
else
{
m_qProcessInfo_is_valid = eLazyBoolNo;
}
return false;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
uint32_t
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
GDBRemoteCommunicationClient::FindProcesses (const ProcessInstanceInfoMatch &match_info,
ProcessInstanceInfoList &process_infos)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
process_infos.Clear();
if (m_supports_qfProcessInfo)
{
StreamString packet;
packet.PutCString ("qfProcessInfo");
if (!match_info.MatchAllProcesses())
{
packet.PutChar (':');
const char *name = match_info.GetProcessInfo().GetName();
bool has_name_match = false;
if (name && name[0])
{
has_name_match = true;
NameMatchType name_match_type = match_info.GetNameMatchType();
switch (name_match_type)
{
case eNameMatchIgnore:
has_name_match = false;
break;
case eNameMatchEquals:
packet.PutCString ("name_match:equals;");
break;
case eNameMatchContains:
packet.PutCString ("name_match:contains;");
break;
case eNameMatchStartsWith:
packet.PutCString ("name_match:starts_with;");
break;
case eNameMatchEndsWith:
packet.PutCString ("name_match:ends_with;");
break;
case eNameMatchRegularExpression:
packet.PutCString ("name_match:regex;");
break;
}
if (has_name_match)
{
packet.PutCString ("name:");
packet.PutBytesAsRawHex8(name, ::strlen(name));
packet.PutChar (';');
}
}
if (match_info.GetProcessInfo().ProcessIDIsValid())
packet.Printf("pid:%" PRIu64 ";",match_info.GetProcessInfo().GetProcessID());
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (match_info.GetProcessInfo().ParentProcessIDIsValid())
packet.Printf("parent_pid:%" PRIu64 ";",match_info.GetProcessInfo().GetParentProcessID());
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
if (match_info.GetProcessInfo().UserIDIsValid())
packet.Printf("uid:%u;",match_info.GetProcessInfo().GetUserID());
if (match_info.GetProcessInfo().GroupIDIsValid())
packet.Printf("gid:%u;",match_info.GetProcessInfo().GetGroupID());
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (match_info.GetProcessInfo().EffectiveUserIDIsValid())
packet.Printf("euid:%u;",match_info.GetProcessInfo().GetEffectiveUserID());
if (match_info.GetProcessInfo().EffectiveGroupIDIsValid())
packet.Printf("egid:%u;",match_info.GetProcessInfo().GetEffectiveGroupID());
if (match_info.GetProcessInfo().EffectiveGroupIDIsValid())
packet.Printf("all_users:%u;",match_info.GetMatchAllUsers() ? 1 : 0);
if (match_info.GetProcessInfo().GetArchitecture().IsValid())
{
const ArchSpec &match_arch = match_info.GetProcessInfo().GetArchitecture();
const llvm::Triple &triple = match_arch.GetTriple();
packet.PutCString("triple:");
packet.PutCStringAsRawHex8(triple.getTriple().c_str());
packet.PutChar (';');
}
}
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
do
{
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
ProcessInstanceInfo process_info;
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
if (!DecodeProcessInfoResponse (response, process_info))
break;
process_infos.Append(process_info);
response.GetStringRef().clear();
response.SetFilePos(0);
} while (SendPacketAndWaitForResponse ("qsProcessInfo", strlen ("qsProcessInfo"), response, false) == PacketResult::Success);
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
else
{
m_supports_qfProcessInfo = false;
return 0;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
return process_infos.GetSize();
}
bool
GDBRemoteCommunicationClient::GetUserName (uint32_t uid, std::string &name)
{
if (m_supports_qUserName)
{
char packet[32];
const int packet_len = ::snprintf (packet, sizeof (packet), "qUserName:%i", uid);
assert (packet_len < (int)sizeof(packet));
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
if (response.IsNormalResponse())
{
// Make sure we parsed the right number of characters. The response is
// the hex encoded user name and should make up the entire packet.
// If there are any non-hex ASCII bytes, the length won't match below..
if (response.GetHexByteString (name) * 2 == response.GetStringRef().size())
return true;
}
}
else
{
m_supports_qUserName = false;
return false;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
return false;
}
bool
GDBRemoteCommunicationClient::GetGroupName (uint32_t gid, std::string &name)
{
if (m_supports_qGroupName)
{
char packet[32];
const int packet_len = ::snprintf (packet, sizeof (packet), "qGroupName:%i", gid);
assert (packet_len < (int)sizeof(packet));
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse (packet, packet_len, response, false) == PacketResult::Success)
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
{
if (response.IsNormalResponse())
{
// Make sure we parsed the right number of characters. The response is
// the hex encoded group name and should make up the entire packet.
// If there are any non-hex ASCII bytes, the length won't match below..
if (response.GetHexByteString (name) * 2 == response.GetStringRef().size())
return true;
}
}
else
{
m_supports_qGroupName = false;
return false;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
return false;
}
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
void
GDBRemoteCommunicationClient::TestPacketSpeed (const uint32_t num_packets)
{
uint32_t i;
TimeValue start_time, end_time;
uint64_t total_time_nsec;
if (SendSpeedTestPacket (0, 0))
{
static uint32_t g_send_sizes[] = { 0, 64, 128, 512, 1024 };
static uint32_t g_recv_sizes[] = { 0, 64, 128, 512, 1024 }; //, 4*1024, 8*1024, 16*1024, 32*1024, 48*1024, 64*1024, 96*1024, 128*1024 };
const size_t k_num_send_sizes = sizeof(g_send_sizes)/sizeof(uint32_t);
const size_t k_num_recv_sizes = sizeof(g_recv_sizes)/sizeof(uint32_t);
const uint64_t k_recv_amount = 4*1024*1024; // Receive 4MB
for (uint32_t send_idx = 0; send_idx < k_num_send_sizes; ++send_idx)
{
const uint32_t send_size = g_send_sizes[send_idx];
for (uint32_t recv_idx = 0; recv_idx < k_num_recv_sizes; ++recv_idx)
{
const uint32_t recv_size = g_recv_sizes[recv_idx];
StreamString packet;
packet.Printf ("qSpeedTest:response_size:%i;data:", recv_size);
uint32_t bytes_left = send_size;
while (bytes_left > 0)
{
if (bytes_left >= 26)
{
packet.PutCString("abcdefghijklmnopqrstuvwxyz");
bytes_left -= 26;
}
else
{
packet.Printf ("%*.*s;", bytes_left, bytes_left, "abcdefghijklmnopqrstuvwxyz");
bytes_left = 0;
}
}
start_time = TimeValue::Now();
if (recv_size == 0)
{
for (i=0; i<num_packets; ++i)
{
StringExtractorGDBRemote response;
SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false);
}
}
else
{
uint32_t bytes_read = 0;
while (bytes_read < k_recv_amount)
{
StringExtractorGDBRemote response;
SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false);
bytes_read += recv_size;
}
}
end_time = TimeValue::Now();
total_time_nsec = end_time.GetAsNanoSecondsSinceJan1_1970() - start_time.GetAsNanoSecondsSinceJan1_1970();
if (recv_size == 0)
{
float packets_per_second = (((float)num_packets)/(float)total_time_nsec) * (float)TimeValue::NanoSecPerSec;
printf ("%u qSpeedTest(send=%-7u, recv=%-7u) in %" PRIu64 ".%9.9" PRIu64 " sec for %f packets/sec.\n",
num_packets,
send_size,
recv_size,
total_time_nsec / TimeValue::NanoSecPerSec,
total_time_nsec % TimeValue::NanoSecPerSec,
packets_per_second);
}
else
{
float mb_second = ((((float)k_recv_amount)/(float)total_time_nsec) * (float)TimeValue::NanoSecPerSec) / (1024.0*1024.0);
printf ("%u qSpeedTest(send=%-7u, recv=%-7u) sent 4MB in %" PRIu64 ".%9.9" PRIu64 " sec for %f MB/sec.\n",
num_packets,
send_size,
recv_size,
total_time_nsec / TimeValue::NanoSecPerSec,
total_time_nsec % TimeValue::NanoSecPerSec,
mb_second);
}
}
}
}
}
bool
GDBRemoteCommunicationClient::SendSpeedTestPacket (uint32_t send_size, uint32_t recv_size)
{
StreamString packet;
packet.Printf ("qSpeedTest:response_size:%i;data:", recv_size);
uint32_t bytes_left = send_size;
while (bytes_left > 0)
{
if (bytes_left >= 26)
{
packet.PutCString("abcdefghijklmnopqrstuvwxyz");
bytes_left -= 26;
}
else
{
packet.Printf ("%*.*s;", bytes_left, bytes_left, "abcdefghijklmnopqrstuvwxyz");
bytes_left = 0;
}
}
StringExtractorGDBRemote response;
return SendPacketAndWaitForResponse (packet.GetData(), packet.GetSize(), response, false) == PacketResult::Success;
Many improvements to the Platform base class and subclasses. The base Platform class now implements the Host functionality for a lot of things that make sense by default so that subclasses can check: int PlatformSubclass::Foo () { if (IsHost()) return Platform::Foo (); // Let the platform base class do the host specific stuff // Platform subclass specific code... int result = ... return result; } Added new functions to the platform: virtual const char *Platform::GetUserName (uint32_t uid); virtual const char *Platform::GetGroupName (uint32_t gid); The user and group names are cached locally so that remote platforms can avoid sending packets multiple times to resolve this information. Added the parent process ID to the ProcessInfo class. Added a new ProcessInfoMatch class which helps us to match processes up and changed the Host layer over to using this new class. The new class allows us to search for processs: 1 - by name (equal to, starts with, ends with, contains, and regex) 2 - by pid 3 - And further check for parent pid == value, uid == value, gid == value, euid == value, egid == value, arch == value, parent == value. This is all hookup up to the "platform process list" command which required adding dumping routines to dump process information. If the Host class implements the process lookup routines, you can now lists processes on your local machine: machine1.foo.com % lldb (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari 94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode 92742 92710 username usergroup username usergroup i386-apple-darwin debugserver This of course also works remotely with the lldb-platform: machine1.foo.com % lldb-platform --listen 1234 machine2.foo.com % lldb (lldb) platform create remote-macosx Platform: remote-macosx Connected: no (lldb) platform connect connect://localhost:1444 Platform: remote-macosx Triple: x86_64-apple-darwin OS Version: 10.6.7 (10J869) Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386 Hostname: machine1.foo.com Connected: yes (lldb) platform process list PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME ====== ====== ========== ========== ========== ========== ======================== ============================ 99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation 99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb 99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge 94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker 94852 244 username usergroup username usergroup x86_64-apple-darwin Safari The lldb-platform implements everything with the Host:: layer, so this should "just work" for linux. I will probably be adding more stuff to the Host layer for launching processes and attaching to processes so that this support should eventually just work as well. Modified the target to be able to be created with an architecture that differs from the main executable. This is needed for iOS debugging since we can have an "armv6" binary which can run on an "armv7" machine, so we want to be able to do: % lldb (lldb) platform create remote-ios (lldb) file --arch armv7 a.out Where "a.out" is an armv6 executable. The platform then can correctly decide to open all "armv7" images for all dependent shared libraries. Modified the disassembly to show the current PC value. Example output: (lldb) disassemble --frame a.out`main: 0x1eb7: pushl %ebp 0x1eb8: movl %esp, %ebp 0x1eba: pushl %ebx 0x1ebb: subl $20, %esp 0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18 0x1ec3: popl %ebx -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf 0x1edb: leal 213(%ebx), %eax 0x1ee1: movl %eax, (%esp) 0x1ee4: calll 0x1f1e ; puts 0x1ee9: calll 0x1f0c ; getchar 0x1eee: movl $20, (%esp) 0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6 0x1efa: movl $12, %eax 0x1eff: addl $20, %esp 0x1f02: popl %ebx 0x1f03: leave 0x1f04: ret This can be handy when dealing with the new --line options that was recently added: (lldb) disassemble --line a.out`main + 13 at test.c:19 18 { -> 19 printf("Process: %i\n\n", getpid()); 20 puts("Press any key to continue..."); getchar(); -> 0x1ec4: calll 0x1f12 ; getpid 0x1ec9: movl %eax, 4(%esp) 0x1ecd: leal 199(%ebx), %eax 0x1ed3: movl %eax, (%esp) 0x1ed6: calll 0x1f18 ; printf Modified the ModuleList to have a lookup based solely on a UUID. Since the UUID is typically the MD5 checksum of a binary image, there is no need to give the path and architecture when searching for a pre-existing image in an image list. Now that we support remote debugging a bit better, our lldb_private::Module needs to be able to track what the original path for file was as the platform knows it, as well as where the file is locally. The module has the two following functions to retrieve both paths: const FileSpec &Module::GetFileSpec () const; const FileSpec &Module::GetPlatformFileSpec () const; llvm-svn: 128563
2011-03-31 02:16:51 +08:00
}
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
uint16_t
GDBRemoteCommunicationClient::LaunchGDBserverAndGetPort (lldb::pid_t &pid, const char *remote_accept_hostname)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
pid = LLDB_INVALID_PROCESS_ID;
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
StringExtractorGDBRemote response;
StreamString stream;
stream.PutCString("qLaunchGDBServer;");
std::string hostname;
if (remote_accept_hostname && remote_accept_hostname[0])
hostname = remote_accept_hostname;
else
{
if (Host::GetHostname (hostname))
{
// Make the GDB server we launch only accept connections from this host
stream.Printf("host:%s;", hostname.c_str());
}
else
{
// Make the GDB server we launch accept connections from any host since we can't figure out the hostname
stream.Printf("host:*;");
}
}
const char *packet = stream.GetData();
int packet_len = stream.GetSize();
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
std::string name;
std::string value;
uint16_t port = 0;
while (response.GetNameColonValue(name, value))
{
if (name.compare("port") == 0)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
port = Args::StringToUInt32(value.c_str(), 0, 0);
else if (name.compare("pid") == 0)
pid = Args::StringToUInt64(value.c_str(), LLDB_INVALID_PROCESS_ID, 0);
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
}
return port;
}
return 0;
}
bool
GDBRemoteCommunicationClient::KillSpawnedProcess (lldb::pid_t pid)
{
StreamString stream;
stream.Printf ("qKillSpawnedProcess:%" PRId64 , pid);
const char *packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
return true;
}
return false;
}
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
bool
GDBRemoteCommunicationClient::SetCurrentThread (uint64_t tid)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
if (m_curr_tid == tid)
return true;
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
char packet[32];
int packet_len;
if (tid == UINT64_MAX)
packet_len = ::snprintf (packet, sizeof(packet), "Hg-1");
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
else
packet_len = ::snprintf (packet, sizeof(packet), "Hg%" PRIx64, tid);
assert (packet_len + 1 < (int)sizeof(packet));
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
if (response.IsOKResponse())
{
m_curr_tid = tid;
return true;
}
}
return false;
}
bool
GDBRemoteCommunicationClient::SetCurrentThreadForRun (uint64_t tid)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
if (m_curr_tid_run == tid)
return true;
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
char packet[32];
int packet_len;
if (tid == UINT64_MAX)
packet_len = ::snprintf (packet, sizeof(packet), "Hc-1");
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
else
packet_len = ::snprintf (packet, sizeof(packet), "Hc%" PRIx64, tid);
assert (packet_len + 1 < (int)sizeof(packet));
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
if (response.IsOKResponse())
{
m_curr_tid_run = tid;
return true;
}
}
return false;
}
bool
GDBRemoteCommunicationClient::GetStopReply (StringExtractorGDBRemote &response)
{
if (SendPacketAndWaitForResponse("?", 1, response, false) == PacketResult::Success)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
return response.IsNormalResponse();
return false;
}
bool
GDBRemoteCommunicationClient::GetThreadStopInfo (lldb::tid_t tid, StringExtractorGDBRemote &response)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
if (m_supports_qThreadStopInfo)
{
char packet[256];
int packet_len = ::snprintf(packet, sizeof(packet), "qThreadStopInfo%" PRIx64, tid);
assert (packet_len < (int)sizeof(packet));
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
<rdar://problem/14972424> When debugging with the GDB remote in LLDB, LLDB uses special packets to discover the registers on the remote server. When those packets aren't supported, LLDB doesn't know what the registers look like. This checkin implements a setting that can be used to specify a python file that contains the registers definitions. The setting is: (lldb) settings set plugin.process.gdb-remote.target-definition-file /path/to/module.py Inside module there should be a function: def get_dynamic_setting(target, setting_name): This dynamic setting function is handed the "target" which is a SBTarget, and the "setting_name", which is the name of the dynamic setting to retrieve. For the GDB remote target definition the setting name is 'gdb-server-target-definition'. The return value is a dictionary that follows the same format as the OperatingSystem plugins follow. I have checked in an example file that implements the x86_64 GDB register set for people to see: examples/python/x86_64_target_definition.py This allows LLDB to debug to any archticture that is support and allows users to define the registers contexts when the discovery packets (qRegisterInfo, qHostInfo) are not supported by the remote GDB server. A few benefits of doing this in Python: 1 - The dynamic register context was already supported in the OperatingSystem plug-in 2 - Register contexts can use all of the LLDB enumerations and definitions for things like lldb::Format, lldb::Encoding, generic register numbers, invalid registers numbers, etc. 3 - The code that generates the register context can use the program to calculate the register context contents (like offsets, register numbers, and more) 4 - True dynamic detection could be used where variables and types could be read from the target program itself in order to determine which registers are available since the target is passed into the python function. This is designed to be used instead of XML since it is more dynamic and code flow and functions can be used to make the dictionary. llvm-svn: 192646
2013-10-15 08:14:28 +08:00
if (response.IsUnsupportedResponse())
m_supports_qThreadStopInfo = false;
else if (response.IsNormalResponse())
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
return true;
else
return false;
}
else
{
m_supports_qThreadStopInfo = false;
}
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
}
return false;
}
uint8_t
GDBRemoteCommunicationClient::SendGDBStoppointTypePacket (GDBStoppointType type, bool insert, addr_t addr, uint32_t length)
{
// Check if the stub is known not to support this breakpoint type
if (!SupportsGDBStoppointPacket(type))
return UINT8_MAX;
// Construct the breakpoint packet
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
char packet[64];
const int packet_len = ::snprintf (packet,
sizeof(packet),
"%c%i,%" PRIx64 ",%x",
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
insert ? 'Z' : 'z',
type,
addr,
length);
// Check we havent overwritten the end of the packet buffer
assert (packet_len + 1 < (int)sizeof(packet));
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
StringExtractorGDBRemote response;
// Try to send the breakpoint packet, and check that it was correctly sent
if (SendPacketAndWaitForResponse(packet, packet_len, response, true) == PacketResult::Success)
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
{
// Receive and OK packet when the breakpoint successfully placed
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
if (response.IsOKResponse())
return 0;
// Error while setting breakpoint, send back specific error
if (response.IsErrorResponse())
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
return response.GetError();
// Empty packet informs us that breakpoint is not supported
if (response.IsUnsupportedResponse())
{
// Disable this breakpoint type since it is unsupported
switch (type)
{
case eBreakpointSoftware: m_supports_z0 = false; break;
case eBreakpointHardware: m_supports_z1 = false; break;
case eWatchpointWrite: m_supports_z2 = false; break;
case eWatchpointRead: m_supports_z3 = false; break;
case eWatchpointReadWrite: m_supports_z4 = false; break;
}
}
}
// Signal generic faliure
Moved the execution context that was in the Debugger into the CommandInterpreter where it was always being used. Make sure that Modules can track their object file offsets correctly to allow opening of sub object files (like the "__commpage" on darwin). Modified the Platforms to be able to launch processes. The first part of this move is the platform soon will become the entity that launches your program and when it does, it uses a new ProcessLaunchInfo class which encapsulates all process launching settings. This simplifies the internal APIs needed for launching. I want to slowly phase out process launching from the process classes, so for now we can still launch just as we used to, but eventually the platform is the object that should do the launching. Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able to launch processes with all of the new eLaunchFlag settings. Modified any code that was manually launching processes to use the Host::LaunchProcess functions. Fixed an issue where lldb_private::Args had implicitly defined copy constructors that could do the wrong thing. This has now been fixed by adding an appropriate copy constructor and assignment operator. Make sure we don't add empty ModuleSP entries to a module list. Fixed the commpage module creation on MacOSX, but we still need to train the MacOSX dynamic loader to not get rid of it when it doesn't have an entry in the all image infos. Abstracted many more calls from in ProcessGDBRemote down into the GDBRemoteCommunicationClient subclass to make the classes cleaner and more efficient. Fixed the default iOS ARM register context to be correct and also added support for targets that don't support the qThreadStopInfo packet by selecting the current thread (only if needed) and then sending a stop reply packet. Debugserver can now start up with a --unix-socket (-u for short) and can then bind to port zero and send the port it bound to to a listening process on the other end. This allows the GDB remote platform to spawn new GDB server instances (debugserver) to allow platform debugging. llvm-svn: 129351
2011-04-12 13:54:46 +08:00
return UINT8_MAX;
}
size_t
GDBRemoteCommunicationClient::GetCurrentThreadIDs (std::vector<lldb::tid_t> &thread_ids,
bool &sequence_mutex_unavailable)
{
Mutex::Locker locker;
thread_ids.clear();
if (GetSequenceMutex (locker, "ProcessGDBRemote::UpdateThreadList() failed due to not getting the sequence mutex"))
{
sequence_mutex_unavailable = false;
StringExtractorGDBRemote response;
PacketResult packet_result;
for (packet_result = SendPacketAndWaitForResponseNoLock ("qfThreadInfo", strlen("qfThreadInfo"), response);
packet_result == PacketResult::Success && response.IsNormalResponse();
packet_result = SendPacketAndWaitForResponseNoLock ("qsThreadInfo", strlen("qsThreadInfo"), response))
{
char ch = response.GetChar();
if (ch == 'l')
break;
if (ch == 'm')
{
do
{
tid_t tid = response.GetHexMaxU64(false, LLDB_INVALID_THREAD_ID);
if (tid != LLDB_INVALID_THREAD_ID)
{
thread_ids.push_back (tid);
}
ch = response.GetChar(); // Skip the command separator
} while (ch == ','); // Make sure we got a comma separator
}
}
}
else
{
#if defined (LLDB_CONFIGURATION_DEBUG)
// assert(!"ProcessGDBRemote::UpdateThreadList() failed due to not getting the sequence mutex");
#else
Log *log (ProcessGDBRemoteLog::GetLogIfAnyCategoryIsSet (GDBR_LOG_PROCESS | GDBR_LOG_PACKETS));
if (log)
log->Printf("error: failed to get packet sequence mutex, not sending packet 'qfThreadInfo'");
#endif
sequence_mutex_unavailable = true;
}
return thread_ids.size();
}
lldb::addr_t
GDBRemoteCommunicationClient::GetShlibInfoAddr()
{
if (!IsRunning())
{
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse("qShlibInfoAddr", ::strlen ("qShlibInfoAddr"), response, false) == PacketResult::Success)
{
if (response.IsNormalResponse())
return response.GetHexMaxU64(false, LLDB_INVALID_ADDRESS);
}
}
return LLDB_INVALID_ADDRESS;
}
lldb_private::Error
GDBRemoteCommunicationClient::RunShellCommand (const char *command, // Shouldn't be NULL
const char *working_dir, // Pass NULL to use the current working directory
int *status_ptr, // Pass NULL if you don't want the process exit status
int *signo_ptr, // Pass NULL if you don't want the signal that caused the process to exit
std::string *command_output, // Pass NULL if you don't want the command output
uint32_t timeout_sec) // Timeout in seconds to wait for shell program to finish
{
lldb_private::StreamString stream;
stream.PutCString("qPlatform_shell:");
stream.PutBytesAsRawHex8(command, strlen(command));
stream.PutChar(',');
stream.PutHex32(timeout_sec);
if (working_dir && *working_dir)
{
stream.PutChar(',');
stream.PutBytesAsRawHex8(working_dir, strlen(working_dir));
}
const char *packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
return Error("malformed reply");
if (response.GetChar() != ',')
return Error("malformed reply");
uint32_t exitcode = response.GetHexMaxU32(false, UINT32_MAX);
if (exitcode == UINT32_MAX)
return Error("unable to run remote process");
else if (status_ptr)
*status_ptr = exitcode;
if (response.GetChar() != ',')
return Error("malformed reply");
uint32_t signo = response.GetHexMaxU32(false, UINT32_MAX);
if (signo_ptr)
*signo_ptr = signo;
if (response.GetChar() != ',')
return Error("malformed reply");
std::string output;
response.GetEscapedBinaryData(output);
if (command_output)
command_output->assign(output);
return Error();
}
return Error("unable to send packet");
}
Error
GDBRemoteCommunicationClient::MakeDirectory (const char *path,
uint32_t file_permissions)
{
lldb_private::StreamString stream;
stream.PutCString("qPlatform_mkdir:");
stream.PutHex32(file_permissions);
stream.PutChar(',');
stream.PutBytesAsRawHex8(path, strlen(path));
const char *packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
return Error(response.GetHexMaxU32(false, UINT32_MAX), eErrorTypePOSIX);
}
return Error();
}
Error
GDBRemoteCommunicationClient::SetFilePermissions (const char *path,
uint32_t file_permissions)
{
lldb_private::StreamString stream;
stream.PutCString("qPlatform_chmod:");
stream.PutHex32(file_permissions);
stream.PutChar(',');
stream.PutBytesAsRawHex8(path, strlen(path));
const char *packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
return Error(response.GetHexMaxU32(false, UINT32_MAX), eErrorTypePOSIX);
}
return Error();
}
static uint64_t
ParseHostIOPacketResponse (StringExtractorGDBRemote &response,
uint64_t fail_result,
Error &error)
{
response.SetFilePos(0);
if (response.GetChar() != 'F')
return fail_result;
int32_t result = response.GetS32 (-2);
if (result == -2)
return fail_result;
if (response.GetChar() == ',')
{
int result_errno = response.GetS32 (-2);
if (result_errno != -2)
error.SetError(result_errno, eErrorTypePOSIX);
else
error.SetError(-1, eErrorTypeGeneric);
}
else
error.Clear();
return result;
}
lldb::user_id_t
GDBRemoteCommunicationClient::OpenFile (const lldb_private::FileSpec& file_spec,
uint32_t flags,
mode_t mode,
Error &error)
{
lldb_private::StreamString stream;
stream.PutCString("vFile:open:");
std::string path (file_spec.GetPath());
if (path.empty())
return UINT64_MAX;
stream.PutCStringAsRawHex8(path.c_str());
stream.PutChar(',');
const uint32_t posix_open_flags = File::ConvertOpenOptionsForPOSIXOpen(flags);
stream.PutHex32(posix_open_flags);
stream.PutChar(',');
stream.PutHex32(mode);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
return ParseHostIOPacketResponse (response, UINT64_MAX, error);
}
return UINT64_MAX;
}
bool
GDBRemoteCommunicationClient::CloseFile (lldb::user_id_t fd,
Error &error)
{
lldb_private::StreamString stream;
stream.Printf("vFile:close:%i", (int)fd);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
return ParseHostIOPacketResponse (response, -1, error) == 0;
}
return false;
}
// Extension of host I/O packets to get the file size.
lldb::user_id_t
GDBRemoteCommunicationClient::GetFileSize (const lldb_private::FileSpec& file_spec)
{
lldb_private::StreamString stream;
stream.PutCString("vFile:size:");
std::string path (file_spec.GetPath());
stream.PutCStringAsRawHex8(path.c_str());
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
return UINT64_MAX;
uint32_t retcode = response.GetHexMaxU64(false, UINT64_MAX);
return retcode;
}
return UINT64_MAX;
}
Error
GDBRemoteCommunicationClient::GetFilePermissions(const char *path, uint32_t &file_permissions)
{
Error error;
lldb_private::StreamString stream;
stream.PutCString("vFile:mode:");
stream.PutCStringAsRawHex8(path);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
{
error.SetErrorStringWithFormat ("invalid response to '%s' packet", packet);
}
else
{
const uint32_t mode = response.GetS32(-1);
if (mode == -1)
{
if (response.GetChar() == ',')
{
int response_errno = response.GetS32(-1);
if (response_errno > 0)
error.SetError(response_errno, lldb::eErrorTypePOSIX);
else
error.SetErrorToGenericError();
}
else
error.SetErrorToGenericError();
}
else
{
file_permissions = mode & (S_IRWXU|S_IRWXG|S_IRWXO);
}
}
}
else
{
error.SetErrorStringWithFormat ("failed to send '%s' packet", packet);
}
return error;
}
uint64_t
GDBRemoteCommunicationClient::ReadFile (lldb::user_id_t fd,
uint64_t offset,
void *dst,
uint64_t dst_len,
Error &error)
{
lldb_private::StreamString stream;
stream.Printf("vFile:pread:%i,%" PRId64 ",%" PRId64, (int)fd, dst_len, offset);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
return 0;
uint32_t retcode = response.GetHexMaxU32(false, UINT32_MAX);
if (retcode == UINT32_MAX)
return retcode;
const char next = (response.Peek() ? *response.Peek() : 0);
if (next == ',')
return 0;
if (next == ';')
{
response.GetChar(); // skip the semicolon
std::string buffer;
if (response.GetEscapedBinaryData(buffer))
{
const uint64_t data_to_write = std::min<uint64_t>(dst_len, buffer.size());
if (data_to_write > 0)
memcpy(dst, &buffer[0], data_to_write);
return data_to_write;
}
}
}
return 0;
}
uint64_t
GDBRemoteCommunicationClient::WriteFile (lldb::user_id_t fd,
uint64_t offset,
const void* src,
uint64_t src_len,
Error &error)
{
lldb_private::StreamGDBRemote stream;
stream.Printf("vFile:pwrite:%i,%" PRId64 ",", (int)fd, offset);
stream.PutEscapedBytes(src, src_len);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
{
error.SetErrorStringWithFormat("write file failed");
return 0;
}
uint64_t bytes_written = response.GetU64(UINT64_MAX);
if (bytes_written == UINT64_MAX)
{
error.SetErrorToGenericError();
if (response.GetChar() == ',')
{
int response_errno = response.GetS32(-1);
if (response_errno > 0)
error.SetError(response_errno, lldb::eErrorTypePOSIX);
}
return 0;
}
return bytes_written;
}
else
{
error.SetErrorString ("failed to send vFile:pwrite packet");
}
return 0;
}
Error
GDBRemoteCommunicationClient::CreateSymlink (const char *src, const char *dst)
{
Error error;
lldb_private::StreamGDBRemote stream;
stream.PutCString("vFile:symlink:");
// the unix symlink() command reverses its parameters where the dst if first,
// so we follow suit here
stream.PutCStringAsRawHex8(dst);
stream.PutChar(',');
stream.PutCStringAsRawHex8(src);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() == 'F')
{
uint32_t result = response.GetU32(UINT32_MAX);
if (result != 0)
{
error.SetErrorToGenericError();
if (response.GetChar() == ',')
{
int response_errno = response.GetS32(-1);
if (response_errno > 0)
error.SetError(response_errno, lldb::eErrorTypePOSIX);
}
}
}
else
{
// Should have returned with 'F<result>[,<errno>]'
error.SetErrorStringWithFormat("symlink failed");
}
}
else
{
error.SetErrorString ("failed to send vFile:symlink packet");
}
return error;
}
Error
GDBRemoteCommunicationClient::Unlink (const char *path)
{
Error error;
lldb_private::StreamGDBRemote stream;
stream.PutCString("vFile:unlink:");
// the unix symlink() command reverses its parameters where the dst if first,
// so we follow suit here
stream.PutCStringAsRawHex8(path);
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() == 'F')
{
uint32_t result = response.GetU32(UINT32_MAX);
if (result != 0)
{
error.SetErrorToGenericError();
if (response.GetChar() == ',')
{
int response_errno = response.GetS32(-1);
if (response_errno > 0)
error.SetError(response_errno, lldb::eErrorTypePOSIX);
}
}
}
else
{
// Should have returned with 'F<result>[,<errno>]'
error.SetErrorStringWithFormat("unlink failed");
}
}
else
{
error.SetErrorString ("failed to send vFile:unlink packet");
}
return error;
}
// Extension of host I/O packets to get whether a file exists.
bool
GDBRemoteCommunicationClient::GetFileExists (const lldb_private::FileSpec& file_spec)
{
lldb_private::StreamString stream;
stream.PutCString("vFile:exists:");
std::string path (file_spec.GetPath());
stream.PutCStringAsRawHex8(path.c_str());
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
return false;
if (response.GetChar() != ',')
return false;
bool retcode = (response.GetChar() != '0');
return retcode;
}
return false;
}
bool
GDBRemoteCommunicationClient::CalculateMD5 (const lldb_private::FileSpec& file_spec,
uint64_t &high,
uint64_t &low)
{
lldb_private::StreamString stream;
stream.PutCString("vFile:MD5:");
std::string path (file_spec.GetPath());
stream.PutCStringAsRawHex8(path.c_str());
const char* packet = stream.GetData();
int packet_len = stream.GetSize();
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, packet_len, response, false) == PacketResult::Success)
{
if (response.GetChar() != 'F')
return false;
if (response.GetChar() != ',')
return false;
if (response.Peek() && *response.Peek() == 'x')
return false;
low = response.GetHexMaxU64(false, UINT64_MAX);
high = response.GetHexMaxU64(false, UINT64_MAX);
return true;
}
return false;
}
bool
GDBRemoteCommunicationClient::ReadRegister(lldb::tid_t tid, uint32_t reg, StringExtractorGDBRemote &response)
{
Mutex::Locker locker;
if (GetSequenceMutex (locker, "Didn't get sequence mutex for p packet."))
{
const bool thread_suffix_supported = GetThreadSuffixSupported();
if (thread_suffix_supported || SetCurrentThread(tid))
{
char packet[64];
int packet_len = 0;
if (thread_suffix_supported)
packet_len = ::snprintf (packet, sizeof(packet), "p%x;thread:%4.4" PRIx64 ";", reg, tid);
else
packet_len = ::snprintf (packet, sizeof(packet), "p%x", reg);
assert (packet_len < ((int)sizeof(packet) - 1));
return SendPacketAndWaitForResponse(packet, response, false) == PacketResult::Success;
}
}
return false;
}
bool
GDBRemoteCommunicationClient::ReadAllRegisters (lldb::tid_t tid, StringExtractorGDBRemote &response)
{
Mutex::Locker locker;
if (GetSequenceMutex (locker, "Didn't get sequence mutex for g packet."))
{
const bool thread_suffix_supported = GetThreadSuffixSupported();
if (thread_suffix_supported || SetCurrentThread(tid))
{
char packet[64];
int packet_len = 0;
// Get all registers in one packet
if (thread_suffix_supported)
packet_len = ::snprintf (packet, sizeof(packet), "g;thread:%4.4" PRIx64 ";", tid);
else
packet_len = ::snprintf (packet, sizeof(packet), "g");
assert (packet_len < ((int)sizeof(packet) - 1));
return SendPacketAndWaitForResponse(packet, response, false) == PacketResult::Success;
}
}
return false;
}
bool
GDBRemoteCommunicationClient::SaveRegisterState (lldb::tid_t tid, uint32_t &save_id)
{
save_id = 0; // Set to invalid save ID
if (m_supports_QSaveRegisterState == eLazyBoolNo)
return false;
m_supports_QSaveRegisterState = eLazyBoolYes;
Mutex::Locker locker;
if (GetSequenceMutex (locker, "Didn't get sequence mutex for QSaveRegisterState."))
{
const bool thread_suffix_supported = GetThreadSuffixSupported();
if (thread_suffix_supported || SetCurrentThread(tid))
{
char packet[256];
if (thread_suffix_supported)
::snprintf (packet, sizeof(packet), "QSaveRegisterState;thread:%4.4" PRIx64 ";", tid);
else
::strncpy (packet, "QSaveRegisterState", sizeof(packet));
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, response, false) == PacketResult::Success)
{
if (response.IsUnsupportedResponse())
{
// This packet isn't supported, don't try calling it again
m_supports_QSaveRegisterState = eLazyBoolNo;
}
const uint32_t response_save_id = response.GetU32(0);
if (response_save_id != 0)
{
save_id = response_save_id;
return true;
}
}
}
}
return false;
}
bool
GDBRemoteCommunicationClient::RestoreRegisterState (lldb::tid_t tid, uint32_t save_id)
{
// We use the "m_supports_QSaveRegisterState" variable here becuase the
// QSaveRegisterState and QRestoreRegisterState packets must both be supported in
// order to be useful
if (m_supports_QSaveRegisterState == eLazyBoolNo)
return false;
Mutex::Locker locker;
if (GetSequenceMutex (locker, "Didn't get sequence mutex for QRestoreRegisterState."))
{
const bool thread_suffix_supported = GetThreadSuffixSupported();
if (thread_suffix_supported || SetCurrentThread(tid))
{
char packet[256];
if (thread_suffix_supported)
::snprintf (packet, sizeof(packet), "QRestoreRegisterState:%u;thread:%4.4" PRIx64 ";", save_id, tid);
else
::snprintf (packet, sizeof(packet), "QRestoreRegisterState:%u" PRIx64 ";", save_id);
StringExtractorGDBRemote response;
if (SendPacketAndWaitForResponse(packet, response, false) == PacketResult::Success)
{
if (response.IsOKResponse())
{
return true;
}
else if (response.IsUnsupportedResponse())
{
// This packet isn't supported, don't try calling this packet or
// QSaveRegisterState again...
m_supports_QSaveRegisterState = eLazyBoolNo;
}
}
}
}
return false;
}