Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / e849a1ae2d549c7b695f10f38691bd8b276bce15 / . / docs / SourceLevelDebugging.html
blob: f94547060b4a0b58b0cf6e5f81c6960844dc8199 [file] [log] [blame]
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>Source Level Debugging with LLVM</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
7</head>
8<body>
9
10<div class="doc_title">Source Level Debugging with LLVM</div>
11
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]12<table border="0" width="100%">
13<tr>
14<td valign="top">
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]15<ul>
16
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]17 <li><a href="#introduction">Introduction</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]18 <ol>
19 <li><a href="#phil">Philosophy behind LLVM debugging information</a></li>
20 <li><a href="#debugopt">Debugging optimized code</a></li>
21 <li><a href="#future">Future work</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]22 </ol></li>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]23 <li><a href="#llvm-db">Using the <tt>llvm-db</tt> tool</a>
24 <ol>
25 <li><a href="#limitations">Limitations of <tt>llvm-db</tt></a></li>
26 <li><a href="#sample">A sample <tt>llvm-db</tt> session</a></li>
27 <li><a href="#startup">Starting the debugger</a></li>
28 <li><a href="#commands">Commands recognized by the debugger</a></li>
29 </ol></li>
30
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]31 <li><a href="#architecture">Architecture of the LLVM debugger</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]32 <ol>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]33 <li><a href="#arch_debugger">The Debugger and InferiorProcess classes</a></li>
34 <li><a href="#arch_info">The RuntimeInfo, ProgramInfo, and SourceLanguage classes</a></li>
35 <li><a href="#arch_llvm-db">The <tt>llvm-db</tt> tool</a></li>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]36 <li><a href="#arch_todo">Short-term TODO list</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]37 </ol></li>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]38
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]39 <li><a href="#format">Debugging information format</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]40 <ol>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]41 <li><a href="#format_common_anchors">Anchors for global objects</a></li>
42 <li><a href="#format_common_stoppoint">Representing stopping points in the source program</a></li>
43 <li><a href="#format_common_lifetime">Object lifetimes and scoping</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]44 <li><a href="#format_common_descriptors">Object descriptor formats</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]45 <ul>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]46 <li><a href="#format_common_source_files">Representation of source files</a></li>
47 <li><a href="#format_common_program_objects">Representation of program objects</a></li>
48 <li><a href="#format_common_object_contexts">Program object contexts</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]49 </ul></li>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]50 <li><a href="#format_common_intrinsics">Debugger intrinsic functions</a></li>
51 <li><a href="#format_common_tags">Values for debugger tags</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]52 </ol></li>
53 <li><a href="#ccxx_frontend">C/C++ front-end specific debug information</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]54 <ol>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]55 <li><a href="#ccxx_pse">Program Scope Entries</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]56 <ul>
57 <li><a href="#ccxx_compilation_units">Compilation unit entries</a></li>
58 <li><a href="#ccxx_modules">Module, namespace, and importing entries</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]59 </ul></li>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]60 <li><a href="#ccxx_dataobjects">Data objects (program variables)</a></li>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]61 </ol></li>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]62</ul>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]63</td>
64<td align="right" valign="top">
Misha Brukmane849a1a2004-05-12 21:26:16 +0000[diff] [blame^]65<img src="img/venusflytrap.jpg" alt="A leafy and green bug eater" width="247"
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]66height="369">
67</td>
68</tr>
69</table>
70
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]71
72<!-- *********************************************************************** -->
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]73<div class="doc_section"><a name="introduction">Introduction</a></div> <!--
74*********************************************************************** -->
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]75
76<div class="doc_text">
77
78<p>This document is the central repository for all information pertaining to
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]79debug information in LLVM. It describes the <a href="#llvm-db">user
80interface</a> for the <a href="CommandGuide/llvm-db.html"><tt>llvm-db</tt>
81tool</a>, which provides a powerful <a href="#llvm-db">source-level debugger</a>
82to users of LLVM-based compilers. It then describes the <a
83href="#architecture">various components</a> that make up the debugger and the
84libraries which future clients may use. Finally, it describes the <a
85href="#format">actual format that the LLVM debug information</a> takes,
86which is useful for those interested in creating front-ends or dealing directly
87with the information.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]88
89</div>
90
91<!-- ======================================================================= -->
92<div class="doc_subsection">
93 <a name="phil">Philosophy behind LLVM debugging information</a>
94</div>
95
96<div class="doc_text">
97
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]98<p>The idea of the LLVM debugging information is to capture how the important
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]99pieces of the source-language's Abstract Syntax Tree map onto LLVM code.
100Several design aspects have shaped the solution that appears here. The
101important ones are:</p>
102
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]103<ul>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]104<li>Debugging information should have very little impact on the rest of the
105compiler. No transformations, analyses, or code generators should need to be
106modified because of debugging information.</li>
107
108<li>LLVM optimizations should interact in <a href="#debugopt">well-defined and
109easily described ways</a> with the debugging information.</li>
110
111<li>Because LLVM is designed to support arbitrary programming languages,
112LLVM-to-LLVM tools should not need to know anything about the semantics of the
113source-level-language.</li>
114
115<li>Source-level languages are often <b>widely</b> different from one another.
116LLVM should not put any restrictions of the flavor of the source-language, and
117the debugging information should work with any language.</li>
118
119<li>With code generator support, it should be possible to use an LLVM compiler
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]120to compile a program to native machine code and standard debugging formats.
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]121This allows compatibility with traditional machine-code level debuggers, like
122GDB or DBX.</li>
123
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]124</ul>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]125
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]126<p>The approach used by the LLVM implementation is to use a small set of <a
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]127href="#format_common_intrinsics">intrinsic functions</a> to define a mapping
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]128between LLVM program objects and the source-level objects. The description of
129the source-level program is maintained in LLVM global variables in an <a
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]130href="#ccxx_frontend">implementation-defined format</a> (the C/C++ front-end
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]131currently uses working draft 7 of the <a
132href="http://www.eagercon.com/dwarf/dwarf3std.htm">Dwarf 3 standard</a>).</p>
133
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]134<p>When a program is debugged, the debugger interacts with the user and turns
135the stored debug information into source-language specific information. As
136such, the debugger must be aware of the source-language, and is thus tied to a
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]137specific language of family of languages. The <a href="#llvm-db">LLVM
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]138debugger</a> is designed to be modular in its support for source-languages.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]139
140</div>
141
142
143<!-- ======================================================================= -->
144<div class="doc_subsection">
145 <a name="debugopt">Debugging optimized code</a>
146</div>
147
148<div class="doc_text">
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]149
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]150<p>An extremely high priority of LLVM debugging information is to make it
151interact well with optimizations and analysis. In particular, the LLVM debug
152information provides the following guarantees:</p>
153
154<ul>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]155
156<li>LLVM debug information <b>always provides information to accurately read the
157source-level state of the program</b>, regardless of which LLVM optimizations
158have been run, and without any modification to the optimizations themselves.
159However, some optimizations may impact the ability to modify the current state
160of the program with a debugger, such as setting program variables, or calling
161function that have been deleted.</li>
162
163<li>LLVM optimizations gracefully interact with debugging information. If they
164are not aware of debug information, they are automatically disabled as necessary
165in the cases that would invalidate the debug info. This retains the LLVM
166features making it easy to write new transformations.</li>
167
168<li>As desired, LLVM optimizations can be upgraded to be aware of the LLVM
169debugging information, allowing them to update the debugging information as they
170perform aggressive optimizations. This means that, with effort, the LLVM
171optimizers could optimize debug code just as well as non-debug code.</li>
172
173<li>LLVM debug information does not prevent many important optimizations from
174happening (for example inlining, basic block reordering/merging/cleanup, tail
175duplication, etc), further reducing the amount of the compiler that eventually
176is "aware" of debugging information.</li>
177
178<li>LLVM debug information is automatically optimized along with the rest of the
179program, using existing facilities. For example, duplicate information is
180automatically merged by the linker, and unused information is automatically
181removed.</li>
182
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]183</ul>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]184
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]185<p>Basically, the debug information allows you to compile a program with
186"<tt>-O0 -g</tt>" and get full debug information, allowing you to arbitrarily
187modify the program as it executes from the debugger. Compiling a program with
188"<tt>-O3 -g</tt>" gives you full debug information that is always available and
189accurate for reading (e.g., you get accurate stack traces despite tail call
190elimination and inlining), but you might lose the ability to modify the program
191and call functions where were optimized out of the program, or inlined away
192completely.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]193
194</div>
195
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]196<!-- ======================================================================= -->
197<div class="doc_subsection">
198 <a name="future">Future work</a>
199</div>
200
201<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]202<p>There are several important extensions that could be eventually added to the
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]203LLVM debugger. The most important extension would be to upgrade the LLVM code
204generators to support debugging information. This would also allow, for
205example, the X86 code generator to emit native objects that contain debugging
206information consumable by traditional source-level debuggers like GDB or
207DBX.</p>
208
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]209<p>Additionally, LLVM optimizations can be upgraded to incrementally update the
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]210debugging information, <a href="#commands">new commands</a> can be added to the
211debugger, and thread support could be added to the debugger.</p>
212
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]213<p>The "SourceLanguage" modules provided by <tt>llvm-db</tt> could be
214substantially improved to provide good support for C++ language features like
215namespaces and scoping rules.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]216
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]217<p>After working with the debugger for a while, perhaps the nicest improvement
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]218would be to add some sort of line editor, such as GNU readline (but one that is
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]219compatible with the LLVM license).</p>
220
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]221<p>For someone so inclined, it should be straight-forward to write different
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]222front-ends for the LLVM debugger, as the LLVM debugging engine is cleanly
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]223separated from the <tt>llvm-db</tt> front-end. A new LLVM GUI debugger or IDE
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]224would be nice. :)</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]225
226</div>
227
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]228<!-- *********************************************************************** -->
229<div class="doc_section">
230 <a name="llvm-db">Using the <tt>llvm-db</tt> tool</a>
231</div>
232<!-- *********************************************************************** -->
233
234<div class="doc_text">
235
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]236<p>The <tt>llvm-db</tt> tool provides a GDB-like interface for source-level
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]237debugging of programs. This tool provides many standard commands for inspecting
238and modifying the program as it executes, loading new programs, single stepping,
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]239placing breakpoints, etc. This section describes how to use the debugger.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]240
241<p><tt>llvm-db</tt> has been designed to be as similar to GDB in its user
242interface as possible. This should make it extremely easy to learn
243<tt>llvm-db</tt> if you already know <tt>GDB</tt>. In general, <tt>llvm-db</tt>
244provides the subset of GDB commands that are applicable to LLVM debugging users.
245If there is a command missing that make a reasonable amount of sense within the
246<a href="#limitations">limitations of <tt>llvm-db</tt></a>, please report it as
247a bug or, better yet, submit a patch to add it. :)</p>
248
249</div>
250
251<!-- ======================================================================= -->
252<div class="doc_subsection">
253 <a name="limitations">Limitations of <tt>llvm-db</tt></a>
254</div>
255
256<div class="doc_text">
257
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]258<p><tt>llvm-db</tt> is designed to be modular and easy to extend. This
259extensibility was key to getting the debugger up-and-running quickly, because we
260can start with simple-but-unsophisicated implementations of various components.
261Because of this, it is currently missing many features, though they should be
262easy to add over time (patches welcomed!). The biggest inherent limitations of
263<tt>llvm-db</tt> are currently due to extremely simple <a
264href="#arch_debugger">debugger backend</a> (implemented in
265"lib/Debugger/UnixLocalInferiorProcess.cpp") which is designed to work without
266any cooperation from the code generators. Because it is so simple, it suffers
267from the following inherent limitations:</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]268
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]269<ul>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]270
271<li>Running a program in <tt>llvm-db</tt> is a bit slower than running it with
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]272<tt>lli</tt> (i.e., in the JIT).</li>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]273
274<li>Inspection of the target hardware is not supported. This means that you
275cannot, for example, print the contents of X86 registers.</li>
276
277<li>Inspection of LLVM code is not supported. This means that you cannot print
278the contents of arbitrary LLVM values, or use commands such as <tt>stepi</tt>.
279This also means that you cannot debug code without debug information.</li>
280
281<li>Portions of the debugger run in the same address space as the program being
282debugged. This means that memory corruption by the program could trample on
283portions of the debugger.</li>
284
285<li>Attaching to existing processes and core files is not currently
286supported.</li>
287
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]288</ul>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]289
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]290<p>That said, the debugger is still quite useful, and all of these limitations
291can be eliminated by integrating support for the debugger into the code
292generators, and writing a new <a href="#arch_debugger">InferiorProcess</a>
293subclass to use it. See the <a href="#future">future work</a> section for ideas
294of how to extend the LLVM debugger despite these limitations.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]295
296</div>
297
298
299<!-- ======================================================================= -->
300<div class="doc_subsection">
301 <a name="sample">A sample <tt>llvm-db</tt> session</a>
302</div>
303
304<div class="doc_text">
305
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]306<p>TODO: this is obviously lame, when more is implemented, this can be much
307better.</p>
308
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]309<pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]310$ <b>llvm-db funccall</b>
311llvm-db: The LLVM source-level debugger
312Loading program... successfully loaded 'funccall.bc'!
313(llvm-db) <b>create</b>
314Starting program: funccall.bc
315main at funccall.c:9:2
3169 -> q = 0;
317(llvm-db) <b>list main</b>
3184 void foo() {
3195 int t = q;
3206 q = t + 1;
3217 }
3228 int main() {
3239 -> q = 0;
32410 foo();
32511 q = q - 1;
32612
32713 return q;
328(llvm-db) <b>list</b>
32914 }
330(llvm-db) <b>step</b>
33110 -> foo();
332(llvm-db) <b>s</b>
333foo at funccall.c:5:2
3345 -> int t = q;
335(llvm-db) <b>bt</b>
336#0 -> 0x85ffba0 in foo at funccall.c:5:2
337#1 0x85ffd98 in main at funccall.c:10:2
338(llvm-db) <b>finish</b>
339main at funccall.c:11:2
34011 -> q = q - 1;
341(llvm-db) <b>s</b>
34213 -> return q;
343(llvm-db) <b>s</b>
344The program stopped with exit code 0
345(llvm-db) <b>quit</b>
346$
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]347</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]348
349</div>
350
351
352
353<!-- ======================================================================= -->
354<div class="doc_subsection">
355 <a name="startup">Starting the debugger</a>
356</div>
357
358<div class="doc_text">
359
360<p>There are three ways to start up the <tt>llvm-db</tt> debugger:</p>
361
362<p>When run with no options, just <tt>llvm-db</tt>, the debugger starts up
363without a program loaded at all. You must use the <a
364href="#c_file"><tt>file</tt> command</a> to load a program, and the <a
365href="c_set_args"><tt>set args</tt></a> or <a href="#c_run"><tt>run</tt></a>
366commands to specify the arguments for the program.</p>
367
368<p>If you start the debugger with one argument, as <tt>llvm-db
369&lt;program&gt;</tt>, the debugger will start up and load in the specified
370program. You can then optionally specify arguments to the program with the <a
371href="c_set_args"><tt>set args</tt></a> or <a href="#c_run"><tt>run</tt></a>
372commands.</p>
373
374<p>The third way to start the program is with the <tt>--args</tt> option. This
375option allows you to specify the program to load and the arguments to start out
376with. <!-- No options to <tt>llvm-db</tt> may be specified after the
377<tt>-args</tt> option. --> Example use: <tt>llvm-db --args ls /home</tt></p>
378
379</div>
380
381<!-- ======================================================================= -->
382<div class="doc_subsection">
383 <a name="commands">Commands recognized by the debugger</a>
384</div>
385
386<div class="doc_text">
387
388<p>FIXME: this needs work obviously. See the <a
389href="http://sources.redhat.com/gdb/documentation/">GDB documentation</a> for
390information about what these do, or try '<tt>help [command]</tt>' within
391<tt>llvm-db</tt> to get information.</p>
392
393<p>
394<h2>General usage:</h2>
395<ul>
396<li>help [command]</li>
397<li>quit</li>
398<li><a name="c_file">file</a> [program]</li>
399</ul>
400
401<h2>Program inspection and interaction:</h2>
402<ul>
403<li>create (start the program, stopping it ASAP in <tt>main</tt>)</li>
404<li>kill</li>
405<li>run [args]</li>
406<li>step [num]</li>
407<li>next [num]</li>
408<li>cont</li>
409<li>finish</li>
410
411<li>list [start[, end]]</li>
412<li>info source</li>
413<li>info sources</li>
414<li>info functions</li>
415</ul>
416
417<h2>Call stack inspection:</h2>
418<ul>
419<li>backtrace</li>
420<li>up [n]</li>
421<li>down [n]</li>
422<li>frame [n]</li>
423</ul>
424
425
426<h2>Debugger inspection and interaction:</h2>
427<ul>
428<li>info target</li>
429<li>show prompt</li>
430<li>set prompt</li>
431<li>show listsize</li>
432<li>set listsize</li>
433<li>show language</li>
434<li>set language</li>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]435<li>show args</li>
436<li>set args [args]</li>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]437</ul>
438
439<h2>TODO:</h2>
440<ul>
441<li>info frame</li>
442<li>break</li>
443<li>print</li>
444<li>ptype</li>
445
446<li>info types</li>
447<li>info variables</li>
448<li>info program</li>
449
450<li>info args</li>
451<li>info locals</li>
452<li>info catch</li>
453<li>... many others</li>
454</ul>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]455
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]456</div>
457
458<!-- *********************************************************************** -->
459<div class="doc_section">
460 <a name="architecture">Architecture of the LLVM debugger</a>
461</div>
462<!-- *********************************************************************** -->
463
464<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]465<p>The LLVM debugger is built out of three distinct layers of software. These
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]466layers provide clients with different interface options depending on what pieces
467of they want to implement themselves, and it also promotes code modularity and
468good design. The three layers are the <a href="#arch_debugger">Debugger
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]469interface</a>, the <a href="#arch_info">"info" interfaces</a>, and the <a
470href="#arch_llvm-db"><tt>llvm-db</tt> tool</a> itself.</p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]471</div>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]472
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]473<!-- ======================================================================= -->
474<div class="doc_subsection">
475 <a name="arch_debugger">The Debugger and InferiorProcess classes</a>
476</div>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]477
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]478<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]479<p>The Debugger class (defined in the <tt>include/llvm/Debugger/</tt> directory)
480is a low-level class which is used to maintain information about the loaded
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]481program, as well as start and stop the program running as necessary. This class
482does not provide any high-level analysis or control over the program, only
483exposing simple interfaces like <tt>load/unloadProgram</tt>,
484<tt>create/killProgram</tt>, <tt>step/next/finish/contProgram</tt>, and
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]485low-level methods for installing breakpoints.</p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]486
487<p>
488The Debugger class is itself a wrapper around the lowest-level InferiorProcess
489class. This class is used to represent an instance of the program running under
490debugger control. The InferiorProcess class can be implemented in different
491ways for different targets and execution scenarios (e.g., remote debugging).
492The InferiorProcess class exposes a small and simple collection of interfaces
493which are useful for inspecting the current state of the program (such as
494collecting stack trace information, reading the memory image of the process,
495etc). The interfaces in this class are designed to be as low-level and simple
496as possible, to make it easy to create new instances of the class.
497</p>
498
499<p>
500The Debugger class exposes the currently active instance of InferiorProcess
501through the <tt>Debugger::getRunningProcess</tt> method, which returns a
502<tt>const</tt> reference to the class. This means that clients of the Debugger
503class can only <b>inspect</b> the running instance of the program directly. To
504change the executing process in some way, they must use the interces exposed by
505the Debugger class.
506</p>
507</div>
508
509<!-- ======================================================================= -->
510<div class="doc_subsection">
511 <a name="arch_info">The RuntimeInfo, ProgramInfo, and SourceLanguage classes</a>
512</div>
513
514<div class="doc_text">
515<p>
516The next-highest level of debugger abstraction is provided through the
517ProgramInfo, RuntimeInfo, SourceLanguage and related classes (also defined in
518the <tt>include/llvm/Debugger/</tt> directory). These classes efficiently
519decode the debugging information and low-level interfaces exposed by
520InferiorProcess into a higher-level representation, suitable for analysis by the
521debugger.
522</p>
523
524<p>
525The ProgramInfo class exposes a variety of different kinds of information about
526the program objects in the source-level-language. The SourceFileInfo class
527represents a source-file in the program (e.g. a .cpp or .h file). The
528SourceFileInfo class captures information such as which SourceLanguage was used
529to compile the file, where the debugger can get access to the actual file text
530(which is lazily loaded on demand), etc. The SourceFunctionInfo class
531represents a... <b>FIXME: finish</b>. The ProgramInfo class provides interfaces
532to lazily find and decode the information needed to create the Source*Info
533classes requested by the debugger.
534</p>
535
536<p>
537The RuntimeInfo class exposes information about the currently executed program,
538by decoding information from the InferiorProcess and ProgramInfo classes. It
539provides a StackFrame class which provides an easy-to-use interface for
540inspecting the current and suspended stack frames in the program.
541</p>
542
543<p>
544The SourceLanguage class is an abstract interface used by the debugger to
545perform all source-language-specific tasks. For example, this interface is used
546by the ProgramInfo class to decode language-specific types and functions and by
547the debugger front-end (such as <a href="#arch_llvm-db"><tt>llvm-db</tt></a> to
548evaluate source-langauge expressions typed into the debugger. This class uses
549the RuntimeInfo &amp; ProgramInfo classes to get information about the current
550execution context and the loaded program, respectively.
551</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]552
553</div>
554
555<!-- ======================================================================= -->
556<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]557 <a name="arch_llvm-db">The <tt>llvm-db</tt> tool</a>
558</div>
559
560<div class="doc_text">
561<p>
562The <tt>llvm-db</tt> is designed to be a debugger providing an interface as <a
563href="#llvm-db">similar to GDB</a> as reasonable, but no more so than that.
564Because the <a href="#arch_debugger">Debugger</a> and <a
565href="#arch_info">info</a> classes implement all of the heavy lifting and
566analysis, <tt>llvm-db</tt> (which lives in <tt>llvm/tools/llvm-db</tt>) consists
567mainly of of code to interact with the user and parse commands. The CLIDebugger
568constructor registers all of the builtin commands for the debugger, and each
569command is implemented as a CLIDebugger::[name]Command method.
570</p>
571</div>
572
573
574<!-- ======================================================================= -->
575<div class="doc_subsection">
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]576 <a name="arch_todo">Short-term TODO list</a>
577</div>
578
579<div class="doc_text">
580
581<p>
582FIXME: this section will eventually go away. These are notes to myself of
583things that should be implemented, but haven't yet.
584</p>
585
586<p>
587<b>Breakpoints:</b> Support is already implemented in the 'InferiorProcess'
588class, though it hasn't been tested yet. To finish breakpoint support, we need
589to implement breakCommand (which should reuse the linespec parser from the list
590command), and handle the fact that 'break foo' or 'break file.c:53' may insert
591multiple breakpoints. Also, if you say 'break file.c:53' and there is no
592stoppoint on line 53, the breakpoint should go on the next available line. My
593idea was to have the Debugger class provide a "Breakpoint" class which
594encapsulated this messiness, giving the debugger front-end a simple interface.
595The debugger front-end would have to map the really complex semantics of
596temporary breakpoints and 'conditional' breakpoints onto this intermediate
597level. Also, breakpoints should survive as much as possible across program
598reloads.
599</p>
600
601<p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]602<b>UnixLocalInferiorProcess.cpp speedup</b>: There is no reason for the debugged
603process to code gen the globals corresponding to debug information. The
604IntrinsicLowering object could instead change descriptors into constant expr
605casts of the constant address of the LLVM objects for the descriptors. This
606would also allow us to eliminate the mapping back and forth between physical
607addresses that must be done.</p>
608
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]609<p>
610<b>Process deaths</b>: The InferiorProcessDead exception should be extended to
611know "how" a process died, i.e., it was killed by a signal. This is easy to
612collect in the UnixLocalInferiorProcess, we just need to represent it.</p>
613
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]614</div>
615
616<!-- *********************************************************************** -->
617<div class="doc_section">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]618 <a name="format">Debugging information format</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]619</div>
620<!-- *********************************************************************** -->
621
622<div class="doc_text">
623
624<p>LLVM debugging information has been carefully designed to make it possible
625for the optimizer to optimize the program and debugging information without
626necessarily having to know anything about debugging information. In particular,
627the global constant merging pass automatically eliminates duplicated debugging
628information (often caused by header files), the global dead code elimination
629pass automatically deletes debugging information for a function if it decides to
630delete the function, and the linker eliminates debug information when it merges
631<tt>linkonce</tt> functions.</p>
632
633<p>To do this, most of the debugging information (descriptors for types,
634variables, functions, source files, etc) is inserted by the language front-end
635in the form of LLVM global variables. These LLVM global variables are no
636different from any other global variables, except that they have a web of LLVM
637intrinsic functions that point to them. If the last references to a particular
638piece of debugging information are deleted (for example, by the
639<tt>-globaldce</tt> pass), the extraneous debug information will automatically
640become dead and be removed by the optimizer.</p>
641
642<p>The debugger is designed to be agnostic about the contents of most of the
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]643debugging information. It uses a <a href="#arch_info">source-language-specific
644module</a> to decode the information that represents variables, types,
645functions, namespaces, etc: this allows for arbitrary source-language semantics
646and type-systems to be used, as long as there is a module written for the
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]647debugger to interpret the information.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]648
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]649<p>To provide basic functionality, the LLVM debugger does have to make some
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]650assumptions about the source-level language being debugged, though it keeps
651these to a minimum. The only common features that the LLVM debugger assumes
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]652exist are <a href="#format_common_source_files">source files</a>, and <a
653href="#format_program_objects">program objects</a>. These abstract objects are
654used by the debugger to form stack traces, show information about local
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]655variables, etc.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]656
657<p>This section of the documentation first describes the representation aspects
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]658common to any source-language. The <a href="#ccxx_frontend">next section</a>
659describes the data layout conventions used by the C and C++ front-ends.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]660
661</div>
662
663<!-- ======================================================================= -->
664<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]665 <a name="format_common_anchors">Anchors for global objects</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]666</div>
667
668<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]669<p>One important aspect of the LLVM debug representation is that it allows the
670LLVM debugger to efficiently index all of the global objects without having the
671scan the program. To do this, all of the global objects use "anchor" globals of
672type "<tt>{}</tt>", with designated names. These anchor objects obviously do
673not contain any content or meaning by themselves, but all of the global objects
674of a particular type (e.g., source file descriptors) contain a pointer to the
675anchor. This pointer allows the debugger to use def-use chains to find all
676global objects of that type.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]677
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]678<p>So far, the following names are recognized as anchors by the LLVM
679debugger:</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]680
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]681<pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]682 %<a href="#format_common_source_files">llvm.dbg.translation_units</a> = linkonce global {} {}
683 %<a href="#format_program_objects">llvm.dbg.globals</a> = linkonce global {} {}
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]684</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]685
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]686<p>Using anchors in this way (where the source file descriptor points to the
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]687anchors, as opposed to having a list of source file descriptors) allows for the
688standard dead global elimination and merging passes to automatically remove
689unused debugging information. If the globals were kept track of through lists,
690there would always be an object pointing to the descriptors, thus would never be
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]691deleted.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]692
693</div>
694
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]695<!-- ======================================================================= -->
696<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]697 <a name="format_common_stoppoint">
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]698 Representing stopping points in the source program
699 </a>
700</div>
701
702<div class="doc_text">
703
704<p>LLVM debugger "stop points" are a key part of the debugging representation
705that allows the LLVM to maintain simple semantics for <a
706href="#debugopt">debugging optimized code</a>. The basic idea is that the
707front-end inserts calls to the <tt>%llvm.dbg.stoppoint</tt> intrinsic function
708at every point in the program where the debugger should be able to inspect the
709program (these correspond to places the debugger stops when you "<tt>step</tt>"
710through it). The front-end can choose to place these as fine-grained as it
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]711would like (for example, before every subexpression evaluated), but it is
712recommended to only put them after every source statement that includes
713executable code.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]714
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]715<p>Using calls to this intrinsic function to demark legal points for the
716debugger to inspect the program automatically disables any optimizations that
717could potentially confuse debugging information. To non-debug-information-aware
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]718transformations, these calls simply look like calls to an external function,
719which they must assume to do anything (including reading or writing to any part
720of reachable memory). On the other hand, it does not impact many optimizations,
721such as code motion of non-trapping instructions, nor does it impact
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]722optimization of subexpressions, code duplication transformations, or basic-block
723reordering transformations.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]724
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]725<p>An important aspect of the calls to the <tt>%llvm.dbg.stoppoint</tt>
726intrinsic is that the function-local debugging information is woven together
727with use-def chains. This makes it easy for the debugger to, for example,
728locate the 'next' stop point. For a concrete example of stop points, see the
729example in <a href="#format_common_lifetime">the next section</a>.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]730
731</div>
732
733
734<!-- ======================================================================= -->
735<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]736 <a name="format_common_lifetime">Object lifetimes and scoping</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]737</div>
738
739<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]740<p>In many languages, the local variables in functions can have their lifetime
741or scope limited to a subset of a function. In the C family of languages, for
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]742example, variables are only live (readable and writable) within the source block
743that they are defined in. In functional languages, values are only readable
744after they have been defined. Though this is a very obvious concept, it is also
745non-trivial to model in LLVM, because it has no notion of scoping in this sense,
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]746and does not want to be tied to a language's scoping rules.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]747
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]748<p>In order to handle this, the LLVM debug format uses the notion of "regions"
749of a function, delineated by calls to intrinsic functions. These intrinsic
750functions define new regions of the program and indicate when the region
751lifetime expires. Consider the following C fragment, for example:</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]752
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]753<pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]7541. void foo() {
7552. int X = ...;
7563. int Y = ...;
7574. {
7585. int Z = ...;
7596. ...
7607. }
7618. ...
7629. }
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]763</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]764
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]765<p>Compiled to LLVM, this function would be represented like this (FIXME: CHECK
766AND UPDATE THIS):</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]767
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]768<pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]769void %foo() {
770 %X = alloca int
771 %Y = alloca int
772 %Z = alloca int
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]773 <a name="#icl_ex_D1">%D1</a> = call {}* %llvm.dbg.func.start(<a href="#format_program_objects">%lldb.global</a>* %d.foo)
774 %D2 = call {}* <a href="#format_common_stoppoint">%llvm.dbg.stoppoint</a>({}* %D1, uint 2, uint 2, <a href="#format_common_source_files">%lldb.compile_unit</a>* %file)
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]775
776 %D3 = call {}* %llvm.dbg.DEFINEVARIABLE({}* %D2, ...)
777 <i>;; Evaluate expression on line 2, assigning to X.</i>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]778 %D4 = call {}* <a href="#format_common_stoppoint">%llvm.dbg.stoppoint</a>({}* %D3, uint 3, uint 2, <a href="#format_common_source_files">%lldb.compile_unit</a>* %file)
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]779
780 %D5 = call {}* %llvm.dbg.DEFINEVARIABLE({}* %D4, ...)
781 <i>;; Evaluate expression on line 3, assigning to Y.</i>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]782 %D6 = call {}* <a href="#format_common_stoppoint">%llvm.dbg.stoppoint</a>({}* %D5, uint 5, uint 4, <a href="#format_common_source_files">%lldb.compile_unit</a>* %file)
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]783
784 <a name="#icl_ex_D1">%D7</a> = call {}* %llvm.region.start({}* %D6)
785 %D8 = call {}* %llvm.dbg.DEFINEVARIABLE({}* %D7, ...)
786 <i>;; Evaluate expression on line 5, assigning to Z.</i>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]787 %D9 = call {}* <a href="#format_common_stoppoint">%llvm.dbg.stoppoint</a>({}* %D8, uint 6, uint 4, <a href="#format_common_source_files">%lldb.compile_unit</a>* %file)
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]788
789 <i>;; Code for line 6.</i>
790 %D10 = call {}* %llvm.region.end({}* %D9)
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]791 %D11 = call {}* <a href="#format_common_stoppoint">%llvm.dbg.stoppoint</a>({}* %D10, uint 8, uint 2, <a href="#format_common_source_files">%lldb.compile_unit</a>* %file)
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]792
793 <i>;; Code for line 8.</i>
794 <a name="#icl_ex_D1">%D12</a> = call {}* %llvm.region.end({}* %D11)
795 ret void
796}
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]797</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]798
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]799<p>This example illustrates a few important details about the LLVM debugging
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]800information. In particular, it shows how the various intrinsics used are woven
801together with def-use and use-def chains, similar to how <a
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]802href="#format_common_anchors">anchors</a> are used with globals. This allows
803the debugger to analyze the relationship between statements, variable
804definitions, and the code used to implement the function.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]805
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]806<p>In this example, two explicit regions are defined, one with the <a
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]807href="#icl_ex_D1">definition of the <tt>%D1</tt> variable</a> and one with the
808<a href="#icl_ex_D7">definition of <tt>%D7</tt></a>. In the case of
809<tt>%D1</tt>, the debug information indicates that the function whose <a
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]810href="#format_program_objects">descriptor</a> is specified as an argument to the
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]811intrinsic. This defines a new stack frame whose lifetime ends when the region
812is ended by <a href="#icl_ex_D12">the <tt>%D12</tt> call</a>.</p>
813
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]814<p>Using regions to represent the boundaries of source-level functions allow
815LLVM interprocedural optimizations to arbitrarily modify LLVM functions without
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]816having to worry about breaking mapping information between the LLVM code and the
817and source-level program. In particular, the inliner requires no modification
818to support inlining with debugging information: there is no explicit correlation
819drawn between LLVM functions and their source-level counterparts (note however,
820that if the inliner inlines all instances of a non-strong-linkage function into
821its caller that it will not be possible for the user to manually invoke the
822inlined function from the debugger).</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]823
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]824<p>Once the function has been defined, the <a
825href="#format_common_stoppoint">stopping point</a> corresponding to line #2 of
826the function is encountered. At this point in the function, <b>no</b> local
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]827variables are live. As lines 2 and 3 of the example are executed, their
828variable definitions are automatically introduced into the program, without the
829need to specify a new region. These variables do not require new regions to be
830introduced because they go out of scope at the same point in the program: line
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]8319.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]832
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]833<p>In contrast, the <tt>Z</tt> variable goes out of scope at a different time,
834on line 7. For this reason, it is defined within <a href="#icl_ex_D7">the
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]835<tt>%D7</tt> region</a>, which kills the availability of <tt>Z</tt> before the
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]836code for line 8 is executed. In this way, regions can support arbitrary
837source-language scoping rules, as long as they can only be nested (ie, one scope
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]838cannot partially overlap with a part of another scope).</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]839
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]840<p>It is worth noting that this scoping mechanism is used to control scoping of
841all declarations, not just variable declarations. For example, the scope of a
842C++ using declaration is controlled with this, and the <tt>llvm-db</tt> C++
843support routines could use this to change how name lookup is performed (though
844this is not implemented yet).</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]845
846</div>
847
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]848<!-- ======================================================================= -->
849<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]850 <a name="format_common_descriptors">Object descriptor formats</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]851</div>
852
853<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]854<p>The LLVM debugger expects the descriptors for program objects to start in a
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]855canonical format, but the descriptors can include additional information
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]856appended at the end that is source-language specific. All LLVM debugging
857information is versioned, allowing backwards compatibility in the case that the
858core structures need to change in some way. Also, all debugging information
859objects start with a <a href="#format_common_tags">tag</a> to indicate what type
860of object it is. The source-language is allows to define its own objects, by
861using unreserved tag numbers.</p>
862
863<p>The lowest-level descriptor are those describing <a
864href="#format_common_source_files">the files containing the program source
865code</a>, as most other descriptors (sometimes indirectly) refer to them.
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]866</p>
867</div>
868
869
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]870<!-- ------------------------------------------------------------------------ ->
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]871<div class="doc_subsubsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]872 <a name="format_common_source_files">Representation of source files</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]873</div>
874
875<div class="doc_text">
876<p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]877Source file descriptors are patterned after the Dwarf "compile_unit" object.
878The descriptor currently is defined to have at least the following LLVM
879type entries:</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]880
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]881<pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]882%lldb.compile_unit = type {
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]883 uint, <i>;; Tag: <a href="#tag_compile_unit">LLVM_COMPILE_UNIT</a></i>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]884 ushort, <i>;; LLVM debug version number</i>
885 ushort, <i>;; Dwarf language identifier</i>
886 sbyte*, <i>;; Filename</i>
887 sbyte*, <i>;; Working directory when compiled</i>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]888 sbyte* <i>;; Producer of the debug information</i>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]889}
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]890</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]891
892<p>
893These descriptors contain the version number for the debug info, a source
894language ID for the file (we use the Dwarf 3.0 ID numbers, such as
895<tt>DW_LANG_C89</tt>, <tt>DW_LANG_C_plus_plus</tt>, <tt>DW_LANG_Cobol74</tt>,
896etc), three strings describing the filename, working directory of the compiler,
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]897and an identifier string for the compiler that produced it. Note that actual
898compile_unit declarations must also include an <a
899href="#format_common_anchors">anchor</a> to <tt>llvm.dbg.translation_units</tt>,
900but it is not specified where the anchor is to be located. Here is an example
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]901descriptor:
902</p>
903
904<p><pre>
905%arraytest_source_file = internal constant %lldb.compile_unit {
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]906 <a href="#tag_compile_unit">uint 17</a>, ; Tag value
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]907 ushort 0, ; Version #0
908 ushort 1, ; DW_LANG_C89
909 sbyte* getelementptr ([12 x sbyte]* %.str_1, long 0, long 0), ; filename
910 sbyte* getelementptr ([12 x sbyte]* %.str_2, long 0, long 0), ; working dir
911 sbyte* getelementptr ([12 x sbyte]* %.str_3, long 0, long 0), ; producer
912 {}* %llvm.dbg.translation_units ; Anchor
913}
914%.str_1 = internal constant [12 x sbyte] c"arraytest.c\00"
915%.str_2 = internal constant [12 x sbyte] c"/home/sabre\00"
916%.str_3 = internal constant [12 x sbyte] c"llvmgcc 3.4\00"
917</pre></p>
918
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]919<p>
920Note that the LLVM constant merging pass should eliminate duplicate copies of
921the strings that get emitted to each translation unit, such as the producer.
922</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]923
924</div>
925
926
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]927<!-- ----------------------------------------------------------------------- -->
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]928<div class="doc_subsubsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]929 <a name="format_program_objects">Representation of program objects</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]930</div>
931
932<div class="doc_text">
933<p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]934The LLVM debugger needs to know about some source-language program objects, in
935order to build stack traces, print information about local variables, and other
936related activities. The LLVM debugger differentiates between three different
937types of program objects: subprograms (functions, messages, methods, etc),
938variables (locals and globals), and others. Because source-languages have
939widely varying forms of these objects, the LLVM debugger expects only a few
940fields in the descriptor for each object:
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]941</p>
942
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]943<pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]944%lldb.object = type {
945 uint, <i>;; <a href="#format_common_tag">A tag</a></i>
946 <i>any</i>*, <i>;; The <a href="#format_common_object_contexts">context</a> for the object</i>
947 sbyte* <i>;; The object 'name'</i>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]948}
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]949</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]950
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]951<p>The first field contains a tag for the descriptor. The second field contains
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]952either a pointer to the descriptor for the containing <a
953href="#format_common_source_files">source file</a>, or it contains a pointer to
954another program object whose context pointer eventually reaches a source file.
955Through this <a href="#format_common_object_contexts">context</a> pointer, the
956LLVM debugger can establish the debug version number of the object.</p>
957
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]958<p>The third field contains a string that the debugger can use to identify the
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]959object if it does not contain explicit support for the source-language in use
960(ie, the 'unknown' source language handler uses this string). This should be
961some sort of unmangled string that corresponds to the object, but it is a
962quality of implementation issue what exactly it contains (it is legal, though
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]963not useful, for all of these strings to be null).</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]964
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]965<p>Note again that descriptors can be extended to include
966source-language-specific information in addition to the fields required by the
967LLVM debugger. See the <a href="#ccxx_descriptors">section on the C/C++
968front-end</a> for more information. Also remember that global objects
969(functions, selectors, global variables, etc) must contain an <a
970href="format_common_anchors">anchor</a> to the <tt>llvm.dbg.globals</tt>
971variable.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]972</div>
973
974
975<!-- ======================================================================= -->
976<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]977 <a name="format_common_object_contexts">Program object contexts</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]978</div>
979
980<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]981<pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]982Allow source-language specific contexts, use to identify namespaces etc
983Must end up in a source file descriptor.
984Debugger core ignores all unknown context objects.
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]985</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]986</div>
987
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]988<!-- ======================================================================= -->
989<div class="doc_subsection">
990 <a name="format_common_intrinsics">Debugger intrinsic functions</a>
991</div>
992
993<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]994<pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]995Define each intrinsics, as an extension of the language reference manual.
996
997llvm.dbg.stoppoint
998llvm.dbg.region.start
999llvm.dbg.region.end
1000llvm.dbg.function.start
1001llvm.dbg.declare
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1002</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1003</div>
1004
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1005<!-- ======================================================================= -->
1006<div class="doc_subsection">
1007 <a name="format_common_tags">Values for debugger tags</a>
1008</div>
1009
1010<div class="doc_text">
1011
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1012<p>Happen to be the same value as the similarly named Dwarf-3 tags, this may
1013change in the future.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1014
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1015<pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1016 <a name="tag_compile_unit">LLVM_COMPILE_UNIT</a> : 17
1017 <a name="tag_subprogram">LLVM_SUBPROGRAM</a> : 46
1018 <a name="tag_variable">LLVM_VARIABLE</a> : 52
1019<!-- <a name="tag_formal_parameter">LLVM_FORMAL_PARAMETER : 5-->
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1020</pre>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1021</div>
1022
1023
1024
1025<!-- *********************************************************************** -->
1026<div class="doc_section">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1027 <a name="ccxx_frontend">C/C++ front-end specific debug information</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1028</div>
1029
1030<div class="doc_text">
1031
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1032<p>The C and C++ front-ends represent information about the program in a format
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1033that is effectively identical to <a
1034href="http://www.eagercon.com/dwarf/dwarf3std.htm">Dwarf 3.0</a> in terms of
1035information content. This allows code generators to trivially support native
1036debuggers by generating standard dwarf information, and contains enough
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1037information for non-dwarf targets to translate it as needed.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1038
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1039<p>The basic debug information required by the debugger is (intentionally)
1040designed to be as minimal as possible. This basic information is so minimal
1041that it is unlikely that <b>any</b> source-language could be adequately
1042described by it. Because of this, the debugger format was designed for
1043extension to support source-language-specific information. The extended
1044descriptors are read and interpreted by the <a
1045href="#arch_info">language-specific</a> modules in the debugger if there is
1046support available, otherwise it is ignored.</p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1047
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1048<p>This section describes the extensions used to represent C and C++ programs.
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1049Other languages could pattern themselves after this (which itself is tuned to
1050representing programs in the same way that Dwarf 3 does), or they could choose
1051to provide completely different extensions if they don't fit into the Dwarf
1052model. As support for debugging information gets added to the various LLVM
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1053source-language front-ends, the information used should be documented here.</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1054
1055</div>
1056
1057<!-- ======================================================================= -->
1058<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1059 <a name="ccxx_pse">Program Scope Entries</a>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1060</div>
1061
1062<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1063<p>TODO</p>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1064</div>
1065
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1066<!-- -------------------------------------------------------------------------->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1067<div class="doc_subsubsection">
1068 <a name="ccxx_compilation_units">Compilation unit entries</a>
1069</div>
1070
1071<div class="doc_text">
1072<p>
1073Translation units do not add any information over the standard <a
1074href="#format_common_source_files">source file representation</a> already
1075expected by the debugger. As such, it uses descriptors of the type specified,
1076with a trailing <a href="#format_common_anchors">anchor</a>.
1077</p>
1078</div>
1079
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1080<!-- -------------------------------------------------------------------------->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1081<div class="doc_subsubsection">
1082 <a name="ccxx_modules">Module, namespace, and importing entries</a>
1083</div>
1084
1085<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1086<p>TODO</p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1087</div>
1088
1089<!-- ======================================================================= -->
1090<div class="doc_subsection">
1091 <a name="ccxx_dataobjects">Data objects (program variables)</a>
1092</div>
1093
1094<div class="doc_text">
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1095<p>TODO</p>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]1096</div>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1097
1098
1099<!-- *********************************************************************** -->
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1100
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1101<hr>
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1102<address>
1103 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1104 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
1105 <a href="http://validator.w3.org/check/referer"><img
1106 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
1107
1108 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
1109 <a href="http://llvm.cs.uiuc.edu">LLVM Compiler Infrastructure</a><br>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1110 Last modified: $Date$
Misha Brukman82873732004-05-12 19:21:57 +0000[diff] [blame]1111</address>
Chris Lattnerbdfb3392004-01-05 05:06:33 +0000[diff] [blame]1112
1113</body>
1114</html>