1Python Reference 2================ 3 4The entire LLDB API is available as Python functions through a script bridging 5interface. This means the LLDB API's can be used directly from python either 6interactively or to build python apps that provide debugger features. 7 8Additionally, Python can be used as a programmatic interface within the lldb 9command interpreter (we refer to this for brevity as the embedded interpreter). 10Of course, in this context it has full access to the LLDB API - with some 11additional conveniences we will call out in the FAQ. 12 13.. contents:: 14 :local: 15 16Documentation 17-------------- 18 19The LLDB API is contained in a python module named lldb. A useful resource when 20writing Python extensions is the lldb Python classes reference guide. 21 22The documentation is also accessible in an interactive debugger session with 23the following command: 24 25:: 26 27 (lldb) script help(lldb) 28 Help on package lldb: 29 30 NAME 31 lldb - The lldb module contains the public APIs for Python binding. 32 33 FILE 34 /System/Library/PrivateFrameworks/LLDB.framework/Versions/A/Resources/Python/lldb/__init__.py 35 36 DESCRIPTION 37 ... 38 39You can also get help using a module class name. The full API that is exposed 40for that class will be displayed in a man page style window. Below we want to 41get help on the lldb.SBFrame class: 42 43:: 44 45 (lldb) script help(lldb.SBFrame) 46 Help on class SBFrame in module lldb: 47 48 class SBFrame(__builtin__.object) 49 | Represents one of the stack frames associated with a thread. 50 | SBThread contains SBFrame(s). For example (from test/lldbutil.py), 51 | 52 | def print_stacktrace(thread, string_buffer = False): 53 | '''Prints a simple stack trace of this thread.''' 54 | 55 ... 56 57Or you can get help using any python object, here we use the lldb.process 58object which is a global variable in the lldb module which represents the 59currently selected process: 60 61:: 62 63 (lldb) script help(lldb.process) 64 Help on SBProcess in module lldb object: 65 66 class SBProcess(__builtin__.object) 67 | Represents the process associated with the target program. 68 | 69 | SBProcess supports thread iteration. For example (from test/lldbutil.py), 70 | 71 | # ================================================== 72 | # Utility functions related to Threads and Processes 73 | # ================================================== 74 | 75 ... 76 77Embedded Python Interpreter 78--------------------------- 79 80The embedded python interpreter can be accessed in a variety of ways from 81within LLDB. The easiest way is to use the lldb command script with no 82arguments at the lldb command prompt: 83 84:: 85 86 (lldb) script 87 Python Interactive Interpreter. To exit, type 'quit()', 'exit()' or Ctrl-D. 88 >>> 2+3 89 5 90 >>> hex(12345) 91 '0x3039' 92 >>> 93 94This drops you into the embedded python interpreter. When running under the 95script command, lldb sets some convenience variables that give you quick access 96to the currently selected entities that characterize the program and debugger 97state. In each case, if there is no currently selected entity of the 98appropriate type, the variable's IsValid method will return false. These 99variables are: 100 101+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 102| Variable | Type | Equivalent | Description | 103+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 104| ``lldb.debugger`` | `lldb.SBDebugger` | `SBTarget.GetDebugger` | Contains the debugger object whose ``script`` command was invoked. | 105| | | | The `lldb.SBDebugger` object owns the command interpreter | 106| | | | and all the targets in your debug session. There will always be a | 107| | | | Debugger in the embedded interpreter. | 108+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 109| ``lldb.target`` | `lldb.SBTarget` | `SBDebugger.GetSelectedTarget` | Contains the currently selected target - for instance the one made with the | 110| | | | ``file`` or selected by the ``target select <target-index>`` command. | 111| | | `SBProcess.GetTarget` | The `lldb.SBTarget` manages one running process, and all the executable | 112| | | | and debug files for the process. | 113+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 114| ``lldb.process`` | `lldb.SBProcess` | `SBTarget.GetProcess` | Contains the process of the currently selected target. | 115| | | | The `lldb.SBProcess` object manages the threads and allows access to | 116| | | `SBThread.GetProcess` | memory for the process. | 117+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 118| ``lldb.thread`` | `lldb.SBThread` | `SBProcess.GetSelectedThread` | Contains the currently selected thread. | 119| | | | The `lldb.SBThread` object manages the stack frames in that thread. | 120| | | `SBFrame.GetThread` | A thread is always selected in the command interpreter when a target stops. | 121| | | | The ``thread select <thread-index>`` command can be used to change the | 122| | | | currently selected thread. So as long as you have a stopped process, there will be | 123| | | | some selected thread. | 124+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 125| ``lldb.frame`` | `lldb.SBFrame` | `SBThread.GetSelectedFrame` | Contains the currently selected stack frame. | 126| | | | The `lldb.SBFrame` object manage the stack locals and the register set for | 127| | | | that stack. | 128| | | | A stack frame is always selected in the command interpreter when a target stops. | 129| | | | The ``frame select <frame-index>`` command can be used to change the | 130| | | | currently selected frame. So as long as you have a stopped process, there will | 131| | | | be some selected frame. | 132+-------------------+---------------------+-------------------------------------+-------------------------------------------------------------------------------------+ 133 134While extremely convenient, these variables have a couple caveats that you 135should be aware of. First of all, they hold the values of the selected objects 136on entry to the embedded interpreter. They do not update as you use the LLDB 137API's to change, for example, the currently selected stack frame or thread. 138 139Moreover, they are only defined and meaningful while in the interactive Python 140interpreter. There is no guarantee on their value in any other situation, hence 141you should not use them when defining Python formatters, breakpoint scripts and 142commands (or any other Python extension point that LLDB provides). For the 143latter you'll be passed an `SBDebugger`, `SBTarget`, `SBProcess`, `SBThread` or 144`SBFrame` instance and you can use the functions from the "Equivalent" column 145to navigate between them. 146 147As a rationale for such behavior, consider that lldb can run in a multithreaded 148environment, and another thread might call the "script" command, changing the 149value out from under you. 150 151To get started with these objects and LLDB scripting, please note that almost 152all of the lldb Python objects are able to briefly describe themselves when you 153pass them to the Python print function: 154 155:: 156 157 (lldb) script 158 Python Interactive Interpreter. To exit, type 'quit()', 'exit()' or Ctrl-D. 159 >>> print lldb.debugger 160 Debugger (instance: "debugger_1", id: 1) 161 >>> print lldb.target 162 a.out 163 >>> print lldb.process 164 SBProcess: pid = 59289, state = stopped, threads = 1, executable = a.out 165 >>> print lldb.thread 166 SBThread: tid = 0x1f03 167 >>> print lldb.frame 168 frame #0: 0x0000000100000bb6 a.out main + 54 at main.c:16 169 170 171Running a python script when a breakpoint gets hit 172-------------------------------------------------- 173 174One very powerful use of the lldb Python API is to have a python script run 175when a breakpoint gets hit. Adding python scripts to breakpoints provides a way 176to create complex breakpoint conditions and also allows for smart logging and 177data gathering. 178 179When your process hits a breakpoint to which you have attached some python 180code, the code is executed as the body of a function which takes three 181arguments: 182 183:: 184 185 def breakpoint_function_wrapper(frame, bp_loc, internal_dict): 186 # Your code goes here 187 188or: 189 190:: 191 192 def breakpoint_function_wrapper(frame, bp_loc, extra_args, internal_dict): 193 # Your code goes here 194 195 196+-------------------+-------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 197| Argument | Type | Description | 198+-------------------+-------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 199| ``frame`` | `lldb.SBFrame` | The current stack frame where the breakpoint got hit. | 200| | | The object will always be valid. | 201| | | This ``frame`` argument might *not* match the currently selected stack frame found in the `lldb` module global variable ``lldb.frame``. | 202+-------------------+-------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 203| ``bp_loc`` | `lldb.SBBreakpointLocation` | The breakpoint location that just got hit. Breakpoints are represented by `lldb.SBBreakpoint` | 204| | | objects. These breakpoint objects can have one or more locations. These locations | 205| | | are represented by `lldb.SBBreakpointLocation` objects. | 206+-------------------+-------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 207| ``extra_args`` | `lldb.SBStructuredData` | ``Optional`` If your breakpoint callback function takes this extra parameter, then when the callback gets added to a breakpoint, its | 208| | | contents can parametrize this use of the callback. For instance, instead of writing a callback that stops when the caller is "Foo", | 209| | | you could take the function name from a field in the ``extra_args``, making the callback more general. The ``-k`` and ``-v`` options | 210| | | to ``breakpoint command add`` will be passed as a Dictionary in the ``extra_args`` parameter, or you can provide it with the SB API's. | 211+-------------------+-------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 212| ``internal_dict`` | ``dict`` | The python session dictionary as a standard python dictionary object. | 213+-------------------+-------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------+ 214 215Optionally, a Python breakpoint command can return a value. Returning False 216tells LLDB that you do not want to stop at the breakpoint. Any other return 217value (including None or leaving out the return statement altogether) is akin 218to telling LLDB to actually stop at the breakpoint. This can be useful in 219situations where a breakpoint only needs to stop the process when certain 220conditions are met, and you do not want to inspect the program state manually 221at every stop and then continue. 222 223An example will show how simple it is to write some python code and attach it 224to a breakpoint. The following example will allow you to track the order in 225which the functions in a given shared library are first executed during one run 226of your program. This is a simple method to gather an order file which can be 227used to optimize function placement within a binary for execution locality. 228 229We do this by setting a regular expression breakpoint that will match every 230function in the shared library. The regular expression '.' will match any 231string that has at least one character in it, so we will use that. This will 232result in one lldb.SBBreakpoint object that contains an 233lldb.SBBreakpointLocation object for each function. As the breakpoint gets hit, 234we use a counter to track the order in which the function at this particular 235breakpoint location got hit. Since our code is passed the location that was 236hit, we can get the name of the function from the location, disable the 237location so we won't count this function again; then log some info and continue 238the process. 239 240Note we also have to initialize our counter, which we do with the simple 241one-line version of the script command. 242 243Here is the code: 244 245:: 246 247 (lldb) breakpoint set --func-regex=. --shlib=libfoo.dylib 248 Breakpoint created: 1: regex = '.', module = libfoo.dylib, locations = 223 249 (lldb) script counter = 0 250 (lldb) breakpoint command add --script-type python 1 251 Enter your Python command(s). Type 'DONE' to end. 252 > # Increment our counter. Since we are in a function, this must be a global python variable 253 > global counter 254 > counter += 1 255 > # Get the name of the function 256 > name = frame.GetFunctionName() 257 > # Print the order and the function name 258 > print '[%i] %s' % (counter, name) 259 > # Disable the current breakpoint location so it doesn't get hit again 260 > bp_loc.SetEnabled(False) 261 > # No need to stop here 262 > return False 263 > DONE 264 265The breakpoint command add command above attaches a python script to breakpoint 1. To remove the breakpoint command: 266 267:: 268 269 (lldb) breakpoint command delete 1 270 271 272Using the python api's to create custom breakpoints 273--------------------------------------------------- 274 275 276Another use of the Python API's in lldb is to create a custom breakpoint 277resolver. This facility was added in r342259. 278 279It allows you to provide the algorithm which will be used in the breakpoint's 280search of the space of the code in a given Target to determine where to set the 281breakpoint locations - the actual places where the breakpoint will trigger. To 282understand how this works you need to know a little about how lldb handles 283breakpoints. 284 285In lldb, a breakpoint is composed of three parts: the Searcher, the Resolver, 286and the Stop Options. The Searcher and Resolver cooperate to determine how 287breakpoint locations are set and differ between each breakpoint type. Stop 288options determine what happens when a location triggers and includes the 289commands, conditions, ignore counts, etc. Stop options are common between all 290breakpoint types, so for our purposes only the Searcher and Resolver are 291relevant. 292 293The Searcher's job is to traverse in a structured way the code in the current 294target. It proceeds from the Target, to search all the Modules in the Target, 295in each Module it can recurse into the Compile Units in that module, and within 296each Compile Unit it can recurse over the Functions it contains. 297 298The Searcher can be provided with a SearchFilter that it will use to restrict 299this search. For instance, if the SearchFilter specifies a list of Modules, the 300Searcher will not recurse into Modules that aren't on the list. When you pass 301the -s modulename flag to break set you are creating a Module-based search 302filter. When you pass -f filename.c to break set -n you are creating a file 303based search filter. If neither of these is specified, the breakpoint will have 304a no-op search filter, so all parts of the program are searched and all 305locations accepted. 306 307The Resolver has two functions. The most important one is the callback it 308provides. This will get called at the appropriate time in the course of the 309search. The callback is where the job of adding locations to the breakpoint 310gets done. 311 312The other function is specifying to the Searcher at what depth in the above 313described recursion it wants to be called. Setting a search depth also provides 314a stop for the recursion. For instance, if you request a Module depth search, 315then the callback will be called for each Module as it gets added to the 316Target, but the searcher will not recurse into the Compile Units in the module. 317 318One other slight subtlety is that the depth at which you get called back is not 319necessarily the depth at which the SearchFilter is specified. For instance, 320if you are doing symbol searches, it is convenient to use the Module depth for 321the search, since symbols are stored in the module. But the SearchFilter might 322specify some subset of CompileUnits, so not all the symbols you might find in 323each module will pass the search. You don't need to handle this situation 324yourself, since SBBreakpoint::AddLocation will only add locations that pass the 325Search Filter. This API returns an SBError to inform you whether your location 326was added. 327 328When the breakpoint is originally created, its Searcher will process all the 329currently loaded modules. The Searcher will also visit any new modules as they 330are added to the target. This happens, for instance, when a new shared library 331gets added to the target in the course of running, or on rerunning if any of 332the currently loaded modules have been changed. Note, in the latter case, all 333the locations set in the old module will get deleted and you will be asked to 334recreate them in the new version of the module when your callback gets called 335with that module. For this reason, you shouldn't try to manage the locations 336you add to the breakpoint yourself. Note that the Breakpoint takes care of 337deduplicating equal addresses in AddLocation, so you shouldn't need to worry 338about that anyway. 339 340At present, when adding a scripted Breakpoint type, you can only provide a 341custom Resolver, not a custom SearchFilter. 342 343The custom Resolver is provided as a Python class with the following methods: 344 345+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 346| Name | Arguments | Description | 347+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 348| ``__init__`` | ``bkpt``:`lldb.SBBreakpoint` | This is the constructor for the new Resolver. | 349| | ``extra_args``:`lldb.SBStructuredData`| | 350| | | | 351| | | ``bkpt`` is the breakpoint owning this Resolver. | 352| | | | 353| | | | 354| | | ``extra_args`` is an `SBStructuredData` object that the user can pass in when creating instances of this | 355| | | breakpoint. It is not required, but is quite handy. For instance if you were implementing a breakpoint on some | 356| | | symbol name, you could write a generic symbol name based Resolver, and then allow the user to pass | 357| | | in the particular symbol in the extra_args | 358+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 359| ``__callback__`` | ``sym_ctx``:`lldb.SBSymbolContext` | This is the Resolver callback. | 360| | | The ``sym_ctx`` argument will be filled with the current stage | 361| | | of the search. | 362| | | | 363| | | | 364| | | For instance, if you asked for a search depth of lldb.eSearchDepthCompUnit, then the | 365| | | target, module and compile_unit fields of the sym_ctx will be filled. The callback should look just in the | 366| | | context passed in ``sym_ctx`` for new locations. If the callback finds an address of interest, it | 367| | | can add it to the breakpoint with the `SBBreakpoint.AddLocation` method, using the breakpoint passed | 368| | | in to the ``__init__`` method. | 369+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 370| ``__get_depth__`` | ``None`` | Specify the depth at which you wish your callback to get called. The currently supported options are: | 371| | | | 372| | | `lldb.eSearchDepthModule` | 373| | | `lldb.eSearchDepthCompUnit` | 374| | | `lldb.eSearchDepthFunction` | 375| | | | 376| | | For instance, if you are looking | 377| | | up symbols, which are stored at the Module level, you will want to get called back module by module. | 378| | | So you would want to return `lldb.eSearchDepthModule`. This method is optional. If not provided the search | 379| | | will be done at Module depth. | 380+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 381| ``get_short_help`` | ``None`` | This is an optional method. If provided, the returned string will be printed at the beginning of | 382| | | the description for this breakpoint. | 383+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 384 385To define a new breakpoint command defined by this class from the lldb command 386line, use the command: 387 388:: 389 390 (lldb) breakpoint set -P MyModule.MyResolverClass 391 392You can also populate the extra_args SBStructuredData with a dictionary of 393key/value pairs with: 394 395:: 396 397 (lldb) breakpoint set -P MyModule.MyResolverClass -k key_1 -v value_1 -k key_2 -v value_2 398 399Although you can't write a scripted SearchFilter, both the command line and the 400SB API's for adding a scripted resolver allow you to specify a SearchFilter 401restricted to certain modules or certain compile units. When using the command 402line to create the resolver, you can specify a Module specific SearchFilter by 403passing the -s ModuleName option - which can be specified multiple times. You 404can also specify a SearchFilter restricted to certain compile units by passing 405in the -f CompUnitName option. This can also be specified more than once. And 406you can mix the two to specify "this comp unit in this module". So, for 407instance, 408 409:: 410 411 (lldb) breakpoint set -P MyModule.MyResolverClass -s a.out 412 413will use your resolver, but will only recurse into or accept new locations in 414the module a.out. 415 416Another option for creating scripted breakpoints is to use the 417SBTarget.CreateBreakpointFromScript API. This one has the advantage that you 418can pass in an arbitrary SBStructuredData object, so you can create more 419complex parametrizations. SBStructuredData has a handy SetFromJSON method which 420you can use for this purpose. Your __init__ function gets passed this 421SBStructuredData object. This API also allows you to directly provide the list 422of Modules and the list of CompileUnits that will make up the SearchFilter. If 423you pass in empty lists, the breakpoint will use the default "search 424everywhere,accept everything" filter. 425 426Using the python API' to create custom stepping logic 427----------------------------------------------------- 428 429A slightly esoteric use of the Python API's is to construct custom stepping 430types. LLDB's stepping is driven by a stack of "thread plans" and a fairly 431simple state machine that runs the plans. You can create a Python class that 432works as a thread plan, and responds to the requests the state machine makes to 433run its operations. 434 435There is a longer discussion of scripted thread plans and the state machine, 436and several interesting examples of their use in: 437 438https://github.com/llvm/llvm-project/blob/main/lldb/examples/python/scripted_step.py 439 440And for a MUCH fuller discussion of the whole state machine, see: 441 442https://github.com/llvm/llvm-project/blob/main/lldb/include/lldb/Target/ThreadPlan.h 443 444If you are reading those comments it is useful to know that scripted thread 445plans are set to be "ControllingPlans", and not "OkayToDiscard". 446 447To implement a scripted step, you define a python class that has the following 448methods: 449 450+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 451| Name | Arguments | Description | 452+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 453| ``__init__`` | ``thread_plan``:`lldb.SBThreadPlan`| This is the underlying `SBThreadPlan` that is pushed onto the plan stack. | 454| | | You will want to store this away in an ivar. Also, if you are going to | 455| | | use one of the canned thread plans, you can queue it at this point. | 456+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 457| ``explains_stop`` | ``event``: `lldb.SBEvent` | Return True if this stop is part of your thread plans logic, false otherwise. | 458+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 459| ``is_stale`` | ``None`` | If your plan is no longer relevant (for instance, you were | 460| | | stepping in a particular stack frame, but some other operation | 461| | | pushed that frame off the stack) return True and your plan will | 462| | | get popped. | 463+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 464| ``should_step`` | ``None`` | Return ``True`` if you want lldb to instruction step one instruction, | 465| | | or False to continue till the next breakpoint is hit. | 466+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 467| ``should_stop`` | ``event``: `lldb.SBEvent` | If your plan wants to stop and return control to the user at this point, return True. | 468| | | If your plan is done at this point, call SetPlanComplete on your | 469| | | thread plan instance. | 470| | | Also, do any work you need here to set up the next stage of stepping. | 471+-------------------+------------------------------------+---------------------------------------------------------------------------------------+ 472 473To use this class to implement a step, use the command: 474 475:: 476 477 (lldb) thread step-scripted -C MyModule.MyStepPlanClass 478 479Or use the SBThread.StepUsingScriptedThreadPlan API. The SBThreadPlan passed 480into your __init__ function can also push several common plans (step 481in/out/over and run-to-address) in front of itself on the stack, which can be 482used to compose more complex stepping operations. When you use subsidiary plans 483your explains_stop and should_stop methods won't get called until the 484subsidiary plan is done, or the process stops for an event the subsidiary plan 485doesn't explain. For instance, step over plans don't explain a breakpoint hit 486while performing the step-over. 487 488 489Create a new lldb command using a Python function 490------------------------------------------------- 491 492Python functions can be used to create new LLDB command interpreter commands, 493which will work like all the natively defined lldb commands. This provides a 494very flexible and easy way to extend LLDB to meet your debugging requirements. 495 496To write a python function that implements a new LLDB command define the 497function to take four arguments as follows: 498 499:: 500 501 def command_function(debugger, command, result, internal_dict): 502 # Your code goes here 503 504Optionally, you can also provide a Python docstring, and LLDB will use it when providing help for your command, as in: 505 506:: 507 508 def command_function(debugger, command, result, internal_dict): 509 """This command takes a lot of options and does many fancy things""" 510 # Your code goes here 511 512Since lldb 3.5.2, LLDB Python commands can also take an SBExecutionContext as an 513argument. This is useful in cases where the command's notion of where to act is 514independent of the currently-selected entities in the debugger. 515 516This feature is enabled if the command-implementing function can be recognized 517as taking 5 arguments, or a variable number of arguments, and it alters the 518signature as such: 519 520:: 521 522 def command_function(debugger, command, exe_ctx, result, internal_dict): 523 # Your code goes here 524 525+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 526| Argument | Type | Description | 527+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 528| ``debugger`` | `lldb.SBDebugger` | The current debugger object. | 529+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 530| ``command`` | ``python string`` | A python string containing all arguments for your command. If you need to chop up the arguments | 531| | | try using the ``shlex`` module's ``shlex.split(command)`` to properly extract the | 532| | | arguments. | 533+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 534| ``exe_ctx`` | `lldb.SBExecutionContext` | An execution context object carrying around information on the inferior process' context in which the command is expected to act | 535| | | | 536| | | *Optional since lldb 3.5.2, unavailable before* | 537+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 538| ``result`` | `lldb.SBCommandReturnObject` | A return object which encapsulates success/failure information for the command and output text | 539| | | that needs to be printed as a result of the command. The plain Python "print" command also works but | 540| | | text won't go in the result by default (it is useful as a temporary logging facility). | 541+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 542| ``internal_dict`` | ``python dict object`` | The dictionary for the current embedded script session which contains all variables | 543| | | and functions. | 544+-------------------+--------------------------------+----------------------------------------------------------------------------------------------------------------------------------+ 545 546Since lldb 3.7, Python commands can also be implemented by means of a class 547which should implement the following interface: 548 549:: 550 551 class CommandObjectType: 552 def __init__(self, debugger, internal_dict): 553 this call should initialize the command with respect to the command interpreter for the passed-in debugger 554 def __call__(self, debugger, command, exe_ctx, result): 555 this is the actual bulk of the command, akin to Python command functions 556 def get_short_help(self): 557 this call should return the short help text for this command[1] 558 def get_long_help(self): 559 this call should return the long help text for this command[1] 560 561[1] This method is optional. 562 563As a convenience, you can treat the result object as a Python file object, and 564say 565 566:: 567 568 print >>result, "my command does lots of cool stuff" 569 570SBCommandReturnObject and SBStream both support this file-like behavior by 571providing write() and flush() calls at the Python layer. 572 573One other handy convenience when defining lldb command-line commands is the 574command command script import which will import a module specified by file 575path, so you don't have to change your PYTHONPATH for temporary scripts. It 576also has another convenience that if your new script module has a function of 577the form: 578 579:: 580 581 def __lldb_init_module(debugger, internal_dict): 582 # Command Initialization code goes here 583 584where debugger and internal_dict are as above, that function will get run when 585the module is loaded allowing you to add whatever commands you want into the 586current debugger. Note that this function will only be run when using the LLDB 587command ``command script import``, it will not get run if anyone imports your 588module from another module. 589 590The standard test for ``__main__``, like many python modules do, is useful for 591creating scripts that can be run from the command line. However, for command 592line scripts, the debugger instance must be created manually. Sample code would 593look like: 594 595:: 596 597 if __name__ == '__main__': 598 # Initialize the debugger before making any API calls. 599 lldb.SBDebugger.Initialize() 600 # Create a new debugger instance in your module if your module 601 # can be run from the command line. When we run a script from 602 # the command line, we won't have any debugger object in 603 # lldb.debugger, so we can just create it if it will be needed 604 debugger = lldb.SBDebugger.Create() 605 606 # Next, do whatever work this module should do when run as a command. 607 # ... 608 609 # Finally, dispose of the debugger you just made. 610 lldb.SBDebugger.Destroy(debugger) 611 # Terminate the debug session 612 lldb.SBDebugger.Terminate() 613 614 615Now we can create a module called ls.py in the file ~/ls.py that will implement 616a function that can be used by LLDB's python command code: 617 618:: 619 620 #!/usr/bin/env python 621 622 import lldb 623 import commands 624 import optparse 625 import shlex 626 627 def ls(debugger, command, result, internal_dict): 628 print >>result, (commands.getoutput('/bin/ls %s' % command)) 629 630 # And the initialization code to add your commands 631 def __lldb_init_module(debugger, internal_dict): 632 debugger.HandleCommand('command script add -f ls.ls ls') 633 print 'The "ls" python command has been installed and is ready for use.' 634 635Now we can load the module into LLDB and use it 636 637:: 638 639 $ lldb 640 (lldb) command script import ~/ls.py 641 The "ls" python command has been installed and is ready for use. 642 (lldb) ls -l /tmp/ 643 total 365848 644 -rw-r--r--@ 1 someuser wheel 6148 Jan 19 17:27 .DS_Store 645 -rw------- 1 someuser wheel 7331 Jan 19 15:37 crash.log 646 647You can also make "container" commands to organize the commands you are adding to 648lldb. Most of the lldb built-in commands structure themselves this way, and using 649a tree structure has the benefit of leaving the one-word command space free for user 650aliases. It can also make it easier to find commands if you are adding more than 651a few of them. Here's a trivial example of adding two "utility" commands into a 652"my-utilities" container: 653 654:: 655 656 #!/usr/bin/env python 657 658 import lldb 659 660 def first_utility(debugger, command, result, internal_dict): 661 print("I am the first utility") 662 663 def second_utility(debugger, command, result, internal_dict): 664 print("I am the second utility") 665 666 # And the initialization code to add your commands 667 def __lldb_init_module(debugger, internal_dict): 668 debugger.HandleCommand('command container add -h "A container for my utilities" my-utilities') 669 debugger.HandleCommand('command script add -f my_utilities.first_utility -h "My first utility" my-utilities first') 670 debugger.HandleCommand('command script add -f my_utilities.second_utility -h "My second utility" my-utilities second') 671 print('The "my-utilities" python command has been installed and its subcommands are ready for use.') 672 673Then your new commands are available under the my-utilities node: 674 675:: 676 677 (lldb) help my-utilities 678 A container for my utilities 679 680 Syntax: my-utilities 681 682 The following subcommands are supported: 683 684 first -- My first utility Expects 'raw' input (see 'help raw-input'.) 685 second -- My second utility Expects 'raw' input (see 'help raw-input'.) 686 687 For more help on any particular subcommand, type 'help <command> <subcommand>'. 688 (lldb) my-utilities first 689 I am the first utility 690 691 692A more interesting template has been created in the source repository that can 693help you to create lldb command quickly: 694 695https://github.com/llvm/llvm-project/blob/main/lldb/examples/python/cmdtemplate.py 696 697A commonly required facility is being able to create a command that does some 698token substitution, and then runs a different debugger command (usually, it 699po'es the result of an expression evaluated on its argument). For instance, 700given the following program: 701 702:: 703 704 #import <Foundation/Foundation.h> 705 NSString* 706 ModifyString(NSString* src) 707 { 708 return [src stringByAppendingString:@"foobar"]; 709 } 710 711 int main() 712 { 713 NSString* aString = @"Hello world"; 714 NSString* anotherString = @"Let's be friends"; 715 return 1; 716 } 717 718you may want a pofoo X command, that equates po [ModifyString(X) 719capitalizedString]. The following debugger interaction shows how to achieve 720that goal: 721 722:: 723 724 (lldb) script 725 Python Interactive Interpreter. To exit, type 'quit()', 'exit()' or Ctrl-D. 726 >>> def pofoo_funct(debugger, command, result, internal_dict): 727 ... cmd = "po [ModifyString(" + command + ") capitalizedString]" 728 ... debugger.HandleCommand(cmd) 729 ... 730 >>> ^D 731 (lldb) command script add pofoo -f pofoo_funct 732 (lldb) pofoo aString 733 $1 = 0x000000010010aa00 Hello Worldfoobar 734 (lldb) pofoo anotherString 735 $2 = 0x000000010010aba0 Let's Be Friendsfoobar 736 737Using the lldb.py module in Python 738---------------------------------- 739 740LLDB has all of its core code build into a shared library which gets used by 741the `lldb` command line application. On macOS this shared library is a 742framework: LLDB.framework and on other unix variants the program is a shared 743library: lldb.so. LLDB also provides an lldb.py module that contains the 744bindings from LLDB into Python. To use the LLDB.framework to create your own 745stand-alone python programs, you will need to tell python where to look in 746order to find this module. This is done by setting the PYTHONPATH environment 747variable, adding a path to the directory that contains the lldb.py python 748module. The lldb driver program has an option to report the path to the lldb 749module. You can use that to point to correct lldb.py: 750 751For csh and tcsh: 752 753:: 754 755 % setenv PYTHONPATH `lldb -P` 756 757For sh and bash: 758 759:: 760 761 $ export PYTHONPATH=`lldb -P` 762 763Alternately, you can append the LLDB Python directory to the sys.path list 764directly in your Python code before importing the lldb module. 765 766Now your python scripts are ready to import the lldb module. Below is a python 767script that will launch a program from the current working directory called 768"a.out", set a breakpoint at "main", and then run and hit the breakpoint, and 769print the process, thread and frame objects if the process stopped: 770 771:: 772 773 #!/usr/bin/env python 774 775 import lldb 776 import os 777 778 def disassemble_instructions(insts): 779 for i in insts: 780 print i 781 782 # Set the path to the executable to debug 783 exe = "./a.out" 784 785 # Create a new debugger instance 786 debugger = lldb.SBDebugger.Create() 787 788 # When we step or continue, don't return from the function until the process 789 # stops. Otherwise we would have to handle the process events ourselves which, while doable is 790 #a little tricky. We do this by setting the async mode to false. 791 debugger.SetAsync (False) 792 793 # Create a target from a file and arch 794 print "Creating a target for '%s'" % exe 795 796 target = debugger.CreateTargetWithFileAndArch (exe, lldb.LLDB_ARCH_DEFAULT) 797 798 if target: 799 # If the target is valid set a breakpoint at main 800 main_bp = target.BreakpointCreateByName ("main", target.GetExecutable().GetFilename()); 801 802 print main_bp 803 804 # Launch the process. Since we specified synchronous mode, we won't return 805 # from this function until we hit the breakpoint at main 806 process = target.LaunchSimple (None, None, os.getcwd()) 807 808 # Make sure the launch went ok 809 if process: 810 # Print some simple process info 811 state = process.GetState () 812 print process 813 if state == lldb.eStateStopped: 814 # Get the first thread 815 thread = process.GetThreadAtIndex (0) 816 if thread: 817 # Print some simple thread info 818 print thread 819 # Get the first frame 820 frame = thread.GetFrameAtIndex (0) 821 if frame: 822 # Print some simple frame info 823 print frame 824 function = frame.GetFunction() 825 # See if we have debug info (a function) 826 if function: 827 # We do have a function, print some info for the function 828 print function 829 # Now get all instructions for this function and print them 830 insts = function.GetInstructions(target) 831 disassemble_instructions (insts) 832 else: 833 # See if we have a symbol in the symbol table for where we stopped 834 symbol = frame.GetSymbol(); 835 if symbol: 836 # We do have a symbol, print some info for the symbol 837 print symbol 838 839Writing lldb frame recognizers in Python 840---------------------------------------- 841 842Frame recognizers allow for retrieving information about special frames based 843on ABI, arguments or other special properties of that frame, even without 844source code or debug info. Currently, one use case is to extract function 845arguments that would otherwise be inaccessible, or augment existing arguments. 846 847Adding a custom frame recognizer is done by implementing a Python class and 848using the 'frame recognizer add' command. The Python class should have a 849'get_recognized_arguments' method and it will receive an argument of type 850lldb.SBFrame representing the current frame that we are trying to recognize. 851The method should return a (possibly empty) list of lldb.SBValue objects that 852represent the recognized arguments. 853 854An example of a recognizer that retrieves the file descriptor values from libc 855functions 'read', 'write' and 'close' follows: 856 857:: 858 859 class LibcFdRecognizer(object): 860 def get_recognized_arguments(self, frame): 861 if frame.name in ["read", "write", "close"]: 862 fd = frame.EvaluateExpression("$arg1").unsigned 863 target = frame.thread.process.target 864 value = target.CreateValueFromExpression("fd", "(int)%d" % fd) 865 return [value] 866 return [] 867 868The file containing this implementation can be imported via ``command script import`` 869and then we can register this recognizer with ``frame recognizer add``. 870It's important to restrict the recognizer to the libc library (which is 871libsystem_kernel.dylib on macOS) to avoid matching functions with the same name 872in other modules: 873 874:: 875 876 (lldb) command script import .../fd_recognizer.py 877 (lldb) frame recognizer add -l fd_recognizer.LibcFdRecognizer -n read -s libsystem_kernel.dylib 878 879When the program is stopped at the beginning of the 'read' function in libc, we can view the recognizer arguments in 'frame variable': 880 881:: 882 883 (lldb) b read 884 (lldb) r 885 Process 1234 stopped 886 * thread #1, queue = 'com.apple.main-thread', stop reason = breakpoint 1.3 887 frame #0: 0x00007fff06013ca0 libsystem_kernel.dylib`read 888 (lldb) frame variable 889 (int) fd = 3 890 891Writing Target Stop-Hooks in Python: 892------------------------------------ 893 894Stop hooks fire whenever the process stops just before control is returned to the 895user. Stop hooks can either be a set of lldb command-line commands, or can 896be implemented by a suitably defined Python class. The Python based stop-hooks 897can also be passed as set of -key -value pairs when they are added, and those 898will get packaged up into a SBStructuredData Dictionary and passed to the 899constructor of the Python object managing the stop hook. This allows for 900parametrization of the stop hooks. 901 902To add a Python-based stop hook, first define a class with the following methods: 903 904+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 905| Name | Arguments | Description | 906+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 907| ``__init__`` | ``target: lldb.SBTarget`` | This is the constructor for the new stop-hook. | 908| | ``extra_args: lldb.SBStructuredData`` | | 909| | | | 910| | | ``target`` is the SBTarget to which the stop hook is added. | 911| | | | 912| | | ``extra_args`` is an SBStructuredData object that the user can pass in when creating instances of this | 913| | | breakpoint. It is not required, but allows for reuse of stop-hook classes. | 914+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 915| ``handle_stop`` | ``exe_ctx: lldb.SBExecutionContext`` | This is the called when the target stops. | 916| | ``stream: lldb.SBStream`` | | 917| | | ``exe_ctx`` argument will be filled with the current stop point for which the stop hook is | 918| | | being evaluated. | 919| | | | 920| | | ``stream`` an lldb.SBStream, anything written to this stream will be written to the debugger console. | 921| | | | 922| | | The return value is a "Should Stop" vote from this thread. If the method returns either True or no return | 923| | | this thread votes to stop. If it returns False, then the thread votes to continue after all the stop-hooks | 924| | | are evaluated. | 925| | | Note, the --auto-continue flag to 'target stop-hook add' overrides a True return value from the method. | 926+--------------------+---------------------------------------+------------------------------------------------------------------------------------------------------------------+ 927 928To use this class in lldb, run the command: 929 930:: 931 932 (lldb) command script import MyModule.py 933 (lldb) target stop-hook add -P MyModule.MyStopHook -k first -v 1 -k second -v 2 934 935where MyModule.py is the file containing the class definition MyStopHook. 936