Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 8dabb50287ce0064be87d0c38f773d982e889773 / . / docs / BytecodeFormat.html
blob: 9902ecdf5edacc820c308cead08d1885408b05c9 [file] [log] [blame]
Reid Spencer50026612004-05-22 02:28:36 +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>LLVM Bytecode File Format</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
Reid Spencer6f1d6992004-05-23 17:12:45 +0000[diff] [blame]7 <style type="css">
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]8 table, tr, td { border: 2px solid gray }
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]9 th { border: 2px solid gray; font-weight: bold; }
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]10 table { border-collapse: collapse; margin-top: 1em margin-bottom: 1em }
11 </style>
12</head>
13<body>
14 <div class="doc_title"> LLVM Bytecode File Format </div>
15<ol>
16 <li><a href="#abstract">Abstract</a></li>
Reid Spencer6f1d6992004-05-23 17:12:45 +0000[diff] [blame]17 <li><a href="#general">General Concepts</a>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]18 <ol>
19 <li><a href="#blocks">Blocks</a></li>
20 <li><a href="#lists">Lists</a></li>
21 <li><a href="#fields">Fields</a></li>
Reid Spencer7aa940d2004-05-25 15:47:57 +0000[diff] [blame]22 <li><a href="#slots">Slots</a></li>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]23 <li><a href="#encoding">Encoding Rules</a></li>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]24 <li><a href="#align">Alignment</a></li>
25 </ol>
Reid Spencer6f1d6992004-05-23 17:12:45 +0000[diff] [blame]26 </li>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]27 <li><a href="#details">Detailed Layout</a>
28 <ol>
29 <li><a href="#notation">Notation</a></li>
30 <li><a href="#blocktypes">Blocks Types</a></li>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]31 <li><a href="#signature">Signature Block</a></li>
32 <li><a href="#module">Module Block</a></li>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]33 <li><a href="#typeool">Global Type Pool</a></li>
34 <li><a href="#modinfo">Module Info Block</a></li>
35 <li><a href="#constants">Global Constant Pool</a></li>
Chris Lattner2ca1fd12004-05-24 04:55:32 +0000[diff] [blame]36 <li><a href="#functions">Function Blocks</a></li>
37 <li><a href="#symtab">Module Symbol Table</a></li>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]38 </ol>
39 </li>
40</ol>
Chris Lattner8dabb502004-05-25 17:44:58 +0000[diff] [blame^]41<div class="doc_author">
42<p>Written by <a href="mailto:rspencer@x10sys.com">Reid Spencer</a>
43</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]44</div>
45<!-- *********************************************************************** -->
46<div class="doc_section"> <a name="abstract">Abstract </a></div>
47<!-- *********************************************************************** -->
48<div class="doc_text">
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]49<p>This document describes the LLVM bytecode
50file format. It specifies the binary encoding rules of the bytecode file format
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]51so that equivalent systems can encode bytecode files correctly. The LLVM
52bytecode representation is used to store the intermediate representation on
53disk in compacted form.
54</p>
55</div>
56<!-- *********************************************************************** -->
57<div class="doc_section"> <a name="general">General Concepts</a> </div>
58<!-- *********************************************************************** -->
59<div class="doc_text">
60<p>This section describes the general concepts of the bytecode file format
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]61without getting into bit and byte level specifics. Note that the LLVM bytecode
62format may change in the future, but will always be backwards compatible with
63older formats. This document only describes the most current version of the
64bytecode format.</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]65</div>
66<!-- _______________________________________________________________________ -->
67<div class="doc_subsection"><a name="blocks">Blocks</a> </div>
68<div class="doc_text">
69<p>LLVM bytecode files consist simply of a sequence of blocks of bytes.
70Each block begins with an identification value that determines the type of
71the next block. The possible types of blocks are described below in the section
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]72<a href="#blocktypes">Block Types</a>. The block identifier is used because
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]73it is possible for entire blocks to be omitted from the file if they are
74empty. The block identifier helps the reader determine which kind of block is
75next in the file.</p>
Reid Spencer939290f2004-05-22 05:56:41 +0000[diff] [blame]76<p>The following block identifiers are currently in use
77(from llvm/Bytecode/Format.h):</p>
78<ol>
79 <li><b>Module (0x01)</b>.</li>
80 <li><b>Function (0x11)</b>.</li>
81 <li><b>ConstantPool (0x12)</b>.</li>
82 <li><b>SymbolTable (0x13)</b>.</li>
83 <li><b>ModuleGlobalInfo (0x14)</b>.</li>
84 <li><b>GlobalTypePlane (0x15)</b>.</li>
85 <li><b>BasicBlock (0x31)</b>.</li>
86 <li><b>InstructionList (0x32)</b>.</li>
87 <li><b>CompactionTable (0x33)</b>.</li>
88</ol>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]89<p> All blocks are variable length, and the block header specifies the size of
90the block. All blocks are rounded aligned to even 32-bit boundaries, so they
91always start and end of this boundary. Each block begins with an integer
92identifier and the length of the block, which does not include the padding
93bytes needed for alignment.</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]94</div>
95<!-- _______________________________________________________________________ -->
96<div class="doc_subsection"><a name="lists">Lists</a> </div>
97<div class="doc_text">
98<p>Most blocks are constructed of lists of information. Lists can be constructed
99of other lists, etc. This decomposition of information follows the containment
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]100hierarchy of the LLVM Intermediate Representation. For example, a function
101contains a list of instructions (the terminator instructions implicitly define
102the end of the basic blocks).</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]103<p>A list is encoded into the file simply by encoding the number of entries as
104an integer followed by each of the entries. The reader knows when the list is
105done because it will have filled the list with the required numbe of entries.
106</p>
107</div>
108<!-- _______________________________________________________________________ -->
109<div class="doc_subsection"><a name="fields">Fields</a> </div>
110<div class="doc_text">
111<p>Fields are units of information that LLVM knows how to write atomically.
112Most fields have a uniform length or some kind of length indication built into
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]113their encoding. For example, a constant string (array of bytes) is
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]114written simply as the length followed by the characters. Although this is
115similar to a list, constant strings are treated atomically and are thus
116fields.</p>
117<p>Fields use a condensed bit format specific to the type of information
118they must contain. As few bits as possible are written for each field. The
119sections that follow will provide the details on how these fields are
120written and how the bits are to be interpreted.</p>
121</div>
122<!-- _______________________________________________________________________ -->
Reid Spencer7aa940d2004-05-25 15:47:57 +0000[diff] [blame]123<div class="doc_subsection"><a name="slots">Slots</a> </div>
124<div class="doc_text">
125<p>The bytecode format uses the notion of a "slot" to reference Types and
126Values. Since the bytecode file is a <em>direct</em> representation of LLVM's
127intermediate representation, there is a need to represent pointers in the file.
128Slots are used for this purpose. For example, if one has the following assembly:
129</p>
Chris Lattner8dabb502004-05-25 17:44:58 +0000[diff] [blame^]130
131<div class="doc_code">
132 %MyType = type { int, sbyte }<br>
133 %MyVar = external global %MyType
134</div>
135
136<p>there are two definitions. The definition of <tt>%MyVar</tt> uses
137<tt>%MyType</tt>. In the C++ IR this linkage between <tt>%MyVar</tt> and
138<tt>%MyType</tt> is
139explicit through the use of C++ pointers. In bytecode, however, there's no
Reid Spencer7aa940d2004-05-25 15:47:57 +0000[diff] [blame]140ability to store memory addresses. Instead, we compute and write out slot
141numbers for every type and Value written to the file.</p>
142<p>A slot number is simply an unsigned 32-bit integer encoded in the variable
143bit rate scheme (see <a href="#encoding">encoding</a> below). This ensures that
144low slot numbers are encoded in one byte. Through various bits of magic LLVM
145attempts to always keep the slot numbers low. The first attempt is to associate
146slot numbers with their "type plane". That is, Values of the same type are
147written to the bytecode file in a list (sequentially). Their order in that list
148determines their slot number. This means that slot #1 doesn't mean anything
149unless you also specify for which type you want slot #1. Types are handled
150specially and are always written to the file first (in the Global Type Pool) and
Chris Lattner8dabb502004-05-25 17:44:58 +0000[diff] [blame^]151in such a way that both forward and backward references of the types can often be
Reid Spencer7aa940d2004-05-25 15:47:57 +0000[diff] [blame]152resolved with a single pass through the type pool. </p>
153<p>Slot numbers are also kept small by rearranging their order. Because of the
154structure of LLVM, certain values are much more likely to be used frequently
155in the body of a function. For this reason, a compaction table is provided in
156the body of a function if its use would make the function body smaller.
157Suppose you have a function body that uses just the types "int*" and "{double}"
158but uses them thousands of time. Its worthwhile to ensure that the slot number
159for these types are low so they can be encoded in a single byte (via vbr).
160This is exactly what the compaction table does.</p>
161</div>
162<!-- _______________________________________________________________________ -->
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]163<div class="doc_subsection"><a name="encoding">Encoding Primitives</a> </div>
164<div class="doc_text">
165<p>Each field that can be put out is encoded into the file using a small set
166of primitives. The rules for these primitives are described below.</p>
167<h3>Variable Bit Rate Encoding</h3>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]168<p>Most of the values written to LLVM bytecode files are small integers. To
169minimize the number of bytes written for these quantities, an encoding
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]170scheme similar to UTF-8 is used to write integer data. The scheme is known as
171variable bit rate (vbr) encoding. In this encoding, the high bit of each
172byte is used to indicate if more bytes follow. If (byte &amp; 0x80) is non-zero
173in any given byte, it means there is another byte immediately following that
174also contributes to the value. For the final byte (byte &amp; 0x80) is false
175(the high bit is not set). In each byte only the low seven bits contribute to
176the value. Consequently 32-bit quantities can take from one to <em>five</em>
177bytes to encode. In general, smaller quantities will encode in fewer bytes,
178as follows:</p>
179<table class="doc_table_nw">
180 <tr>
181 <th>Byte #</th>
182 <th>Significant Bits</th>
183 <th>Maximum Value</th>
184 </tr>
185 <tr><td>1</td><td>0-6</td><td>127</td></tr>
186 <tr><td>2</td><td>7-13</td><td>16,383</td></tr>
187 <tr><td>3</td><td>14-20</td><td>2,097,151</td></tr>
188 <tr><td>4</td><td>21-27</td><td>268,435,455</td></tr>
189 <tr><td>5</td><td>28-34</td><td>34,359,738,367</td></tr>
190 <tr><td>6</td><td>35-41</td><td>4,398,046,511,103</td></tr>
191 <tr><td>7</td><td>42-48</td><td>562,949,953,421,311</td></tr>
192 <tr><td>8</td><td>49-55</td><td>72,057,594,037,927,935</td></tr>
193 <tr><td>9</td><td>56-62</td><td>9,223,372,036,854,775,807</td></tr>
194 <tr><td>10</td><td>63-69</td><td>1,180,591,620,717,411,303,423</td></tr>
195</table>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]196<p>Note that in practice, the tenth byte could only encode bit 63
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]197since the maximum quantity to use this encoding is a 64-bit integer.</p>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]198
199<p><em>Signed</em> VBR values are encoded with the standard vbr encoding, but
200with the sign bit as the low order bit instead of the high order bit. This
201allows small negative quantities to be encoded efficiently. For example, -3
202is encoded as "((3 &lt;&lt; 1) | 1)" and 3 is encoded as "(3 &lt;&lt; 1) |
2030)", emitted with the standard vbr encoding above.</p>
204
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]205<p>The table below defines the encoding rules for type names used in the
206descriptions of blocks and fields in the next section. Any type name with
207the suffix <em>_vbr</em> indicate a quantity that is encoded using
208variable bit rate encoding as described above.</p>
209<table class="doc_table" >
210 <tr>
211 <th><b>Type</b></th>
212 <th align="left"><b>Rule</b></th>
213 </tr>
214 <tr>
215 <td>unsigned</td>
216 <td align="left">A 32-bit unsigned integer that always occupies four
217 consecutive bytes. The unsigned integer is encoded using LSB first
218 ordering. That is bits 2<sup>0</sup> through 2<sup>7</sup> are in the
219 byte with the lowest file offset (little endian).</td>
220 </tr><tr>
221 <td>uint_vbr</td>
222 <td align="left">A 32-bit unsigned integer that occupies from one to five
223 bytes using variable bit rate encoding.</td>
224 </tr><tr>
225 <td>uint64_vbr</td>
226 <td align="left">A 64-bit unsigned integer that occupies from one to ten
227 bytes using variable bit rate encoding.</td>
228 </tr><tr>
229 <td>int64_vbr</td>
230 <td align="left">A 64-bit signed integer that occupies from one to ten
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]231 bytes using the signed variable bit rate encoding.</td>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]232 </tr><tr>
233 <td>char</td>
234 <td align="left">A single unsigned character encoded into one byte</td>
235 </tr><tr>
236 <td>bit</td>
237 <td align="left">A single bit within a byte.</td>
238 </tr><tr>
239 <td>string</td>
240 <td align="left">A uint_vbr indicating the length of the character string
241 immediately followed by the characters of the string. There is no
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]242 terminating null byte in the string.</td>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]243 </tr><tr>
244 <td>data</td>
245 <td align="left">An arbitrarily long segment of data to which no
246 interpretation is implied. This is used for float, double, and constant
247 initializers.</td>
248 </tr>
249</table>
250</div>
251<!-- _______________________________________________________________________ -->
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]252<div class="doc_subsection"><a name="align">Alignment</a> </div>
253<div class="doc_text">
254<p>To support cross-platform differences, the bytecode file is aligned on
255certain boundaries. This means that a small amount of padding (at most 3 bytes)
256will be added to ensure that the next entry is aligned to a 32-bit boundary.
257</p>
258</div>
259<!-- *********************************************************************** -->
260<div class="doc_section"> <a name="details">Detailed Layout</a> </div>
261<!-- *********************************************************************** -->
262<div class="doc_text">
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]263<p>This section provides the detailed layout of the LLVM bytecode file format.
264bit and byte level specifics.</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]265</div>
266<!-- _______________________________________________________________________ -->
267<div class="doc_subsection"><a name="notation">Notation</a></div>
268<div class="doc_text">
269 <p>The descriptions of the bytecode format that follow describe the bit
270 fields in detail. These descriptions are provided in tabular form. Each table
271 has four columns that specify:</p>
272 <ol>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]273 <li><b>Byte(s)</b>: The offset in bytes of the field from the start of
Reid Spencer6f1d6992004-05-23 17:12:45 +0000[diff] [blame]274 its container (block, list, other field).</li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]275 <li><b>Bit(s)</b>: The offset in bits of the field from the start of
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]276 the byte field. Bits are always little endian. That is, bit addresses with
Reid Spencer6f1d6992004-05-23 17:12:45 +0000[diff] [blame]277 smaller values have smaller address (i.e. 2<sup>0</sup> is at bit 0,
278 2<sup>1</sup> at 1, etc.)
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]279 </li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]280 <li><b>Align?</b>: Indicates if this field is aligned to 32 bits or not.
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]281 This indicates where the <em>next</em> field starts, always on a 32 bit
282 boundary.</li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]283 <li><b>Type</b>: The basic type of information contained in the field.</li>
284 <li><b>Description</b>: Describes the contents of the field.</li>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]285 </ol>
286</div>
287<!-- _______________________________________________________________________ -->
288<div class="doc_subsection"><a name="blocktypes">Block Types</a></div>
289<div class="doc_text">
290 <p>The bytecode format encodes the intermediate representation into groups
291 of bytes known as blocks. The blocks are written sequentially to the file in
292 the following order:</p>
293<ol>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]294 <li><a href="#signature">Signature</a>: This contains the file signature
295 (magic number) that identifies the file as LLVM bytecode and the bytecode
296 version number.</li>
297 <li><a href="#module">Module Block</a>: This is the top level block in a
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]298 bytecode file. It contains all the other blocks.</li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]299 <li><a href="#gtypepool">Global Type Pool</a>: This block contains all the
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]300 global (module) level types.</li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]301 <li><a href="#modinfo">Module Info</a>: This block contains the types of the
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]302 global variables and functions in the module as well as the constant
303 initializers for the global variables</li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]304 <li><a href="#constants">Constants</a>: This block contains all the global
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]305 constants except function arguments, global values and constant strings.</li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]306 <li><a href="#functions">Functions</a>: One function block is written for
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]307 each function in the module. </li>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]308 <li><a href="$symtab">Symbol Table</a>: The module level symbol table that
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]309 provides names for the various other entries in the file is the final block
310 written.</li>
311</ol>
312</div>
313<!-- _______________________________________________________________________ -->
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]314<div class="doc_subsection"><a name="signature">Signature Block</a> </div>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]315<div class="doc_text">
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]316<p>The signature occurs in every LLVM bytecode file and is always first.
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]317It simply provides a few bytes of data to identify the file as being an LLVM
318bytecode file. This block is always four bytes in length and differs from the
319other blocks because there is no identifier and no block length at the start
320of the block. Essentially, this block is just the "magic number" for the file.
321<table class="doc_table_nw" >
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]322 <tr>
323 <th><b>Byte(s)</b></th>
324 <th><b>Bit(s)</b></th>
325 <th><b>Align?</b></th>
Reid Spencer939290f2004-05-22 05:56:41 +0000[diff] [blame]326 <th><b>Type</b></th>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]327 <th align="left"><b>Field Description</b></th>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]328 </tr><tr>
329 <td>00</td><td>-</td><td>No</td><td>char</td>
Reid Spencer939290f2004-05-22 05:56:41 +0000[diff] [blame]330 <td align="left">Constant "l" (0x6C)</td>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]331 </tr><tr>
332 <td>01</td><td>-</td><td>No</td><td>char</td>
Reid Spencer939290f2004-05-22 05:56:41 +0000[diff] [blame]333 <td align="left">Constant "l" (0x6C)</td>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]334 </tr><tr>
335 <td>02</td><td>-</td><td>No</td><td>char</td>
Reid Spencer939290f2004-05-22 05:56:41 +0000[diff] [blame]336 <td align="left">Constant "v" (0x76)</td>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]337 </tr><tr>
338 <td>03</td><td>-</td><td>No</td><td>char</td>
Reid Spencer939290f2004-05-22 05:56:41 +0000[diff] [blame]339 <td align="left">Constant "m" (0x6D)</td>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]340 </tr>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]341</table>
342</div>
343<!-- _______________________________________________________________________ -->
344<div class="doc_subsection"><a name="module">Module Block</a> </div>
345<div class="doc_text">
346<p>The module block contains a small pre-amble and all the other blocks in
347the file. Of particular note, the bytecode format number is simply a 28-bit
348monotonically increase integer that identifiers the version of the bytecode
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]349format (which is not directly related to the LLVM release number). The
350bytecode versions defined so far are (note that this document only describes
351the latest version): </p>
352
353<ul>
354<li>#0: LLVM 1.0 &amp; 1.1</li>
355<li>#1: LLVM 1.2</li>
356<li>#2: LLVM 1.3</li>
357</ul>
358
359<p>The table below shows the format of the module block header. It is defined
360by blocks described in other sections.</p>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]361<table class="doc_table_nw" >
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]362 <tr>
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]363 <th><b>Byte(s)</b></th>
364 <th><b>Bit(s)</b></th>
365 <th><b>Align?</b></th>
366 <th><b>Type</b></th>
367 <th align="left"><b>Field Description</b></th>
368 </tr><tr>
369 <td>04-07</td><td>-</td><td>No</td><td>unsigned</td>
370 <td align="left">Module Identifier (0x01)</td>
371 </tr><tr>
372 <td>08-11</td><td>-</td><td>No</td><td>unsigned</td>
373 <td align="left">Size of the module block in bytes</td>
374 </tr><tr>
375 <td>12-15</td><td>00</td><td>Yes</td><td>uint32_vbr</td>
376 <td align="left">Format Information</td>
377 </tr><tr>
378 <td>''</td><td>0</td><td>-</td><td>bit</td>
379 <td align="left">Big Endian?</td>
380 </tr><tr>
381 <td>''</td><td>1</td><td>-</td><td>bit</td>
382 <td align="left">Pointers Are 64-bit?</td>
383 </tr><tr>
384 <td>''</td><td>2</td><td>-</td><td>bit</td>
385 <td align="left">Has No Endianess?</td>
386 </tr><tr>
387 <td>''</td><td>3</td><td>-</td><td>bit</td>
388 <td align="left">Has No Pointer Size?</td>
389 </tr><tr>
390 <td>''</td><td>4-31</td><td>-</td><td>bit</td>
391 <td align="left">Bytecode Format Version</td>
392 </tr><tr>
393 <td>16-end</td><td>-</td><td>-</td><td>blocks</td>
394 <td align="left">The remaining bytes in the block consist
395 solely of other block types in sequence.</td>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]396 </tr>
397</table>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]398
399<p>Note that we plan to eventually expand the target description capabilities
400of bytecode files to <a href="http://llvm.cs.uiuc.edu/PR263">target
401triples</a>.</p>
402
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]403</div>
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]404
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]405<!-- _______________________________________________________________________ -->
406<div class="doc_subsection"><a name="gtypepool">Global Type Pool</a> </div>
407<div class="doc_text">
Chris Lattner2b905652004-05-24 05:35:17 +0000[diff] [blame]408<p>The global type pool consists of type definitions. Their order of appearance
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]409in the file determines their slot number (0 based). Slot numbers are used to
410replace pointers in the intermediate representation. Each slot number uniquely
411identifies one entry in a type plane (a collection of values of the same type).
412Since all values have types and are associated with the order in which the type
413pool is written, the global type pool <em>must</em> be written as the first
414block of a module. If it is not, attempts to read the file will fail because
415both forward and backward type resolution will not be possible.</p>
416<p>The type pool is simply a list of types definitions, as shown in the table
417below.</p>
418<table class="doc_table_nw" >
419 <tr>
420 <th><b>Byte(s)</b></th>
421 <th><b>Bit(s)</b></th>
422 <th><b>Align?</b></th>
423 <th><b>Type</b></th>
424 <th align="left"><b>Field Description</b></th>
425 </tr><tr>
426 <td>00-03</td><td>-</td><td>No</td><td>unsigned</td>
427 <td align="left">Type Pool Identifier (0x13)</td>
428 </tr><tr>
429 <td>04-07</td><td>-</td><td>No</td><td>unsigned</td>
430 <td align="left">Size in bytes of the symbol table block.</td>
431 </tr><tr>
432 <td>08-11<sup>1</sup></td><td>-</td><td>No</td><td>uint32_vbr</td>
433 <td align="left">Number of entries in type plane</td>
434 </tr><tr>
435 <td>12-15<sup>1</sup></td><td>-</td><td>No</td><td>uint32_vbr</td>
436 <td align="left">Type plane index for following entries</td>
437 </tr><tr>
438 <td>16-end<sup>1,2</sup></td><td>-</td><td>No</td><td>type</td>
439 <td align="left">Each of the type definitions.</td>
440 </tr><tr>
441 <td align="left" colspan="5"><sup>1</sup>Maximum length shown,
442 may be smaller<br><sup>2</sup>Repeated field.
443 </tr>
444</table>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]445</div>
446<!-- _______________________________________________________________________ -->
447<div class="doc_subsection"><a name="modinfo">Module Info</a> </div>
448<div class="doc_text">
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]449 <p>To be determined.</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]450</div>
451<!-- _______________________________________________________________________ -->
452<div class="doc_subsection"><a name="constants">Constants</a> </div>
453<div class="doc_text">
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]454 <p>To be determined.</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]455</div>
456<!-- _______________________________________________________________________ -->
457<div class="doc_subsection"><a name="functions">Functions</a> </div>
458<div class="doc_text">
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]459 <p>To be determined.</p>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]460</div>
461<!-- _______________________________________________________________________ -->
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]462<div class="doc_subsection"><a name="symtab">Symbol Table</a> </div>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]463<div class="doc_text">
Reid Spencerb39021b2004-05-23 17:05:09 +0000[diff] [blame]464<p>A symbol table can be put out in conjunction with a module or a function.
465A symbol table is a list of type planes. Each type plane starts with the number
466of entries in the plane and the type plane's slot number (so the type can be
467looked up in the global type pool). For each entry in a type plane, the slot
468number of the value and the name associated with that value are written. The
469format is given in the table below. </p>
470<table class="doc_table_nw" >
471 <tr>
472 <th><b>Byte(s)</b></th>
473 <th><b>Bit(s)</b></th>
474 <th><b>Align?</b></th>
475 <th><b>Type</b></th>
476 <th align="left"><b>Field Description</b></th>
477 </tr><tr>
478 <td>00-03</td><td>-</td><td>No</td><td>unsigned</td>
479 <td align="left">Symbol Table Identifier (0x13)</td>
480 </tr><tr>
481 <td>04-07</td><td>-</td><td>No</td><td>unsigned</td>
482 <td align="left">Size in bytes of the symbol table block.</td>
483 </tr><tr>
484 <td>08-11<sup>1</sup></td><td>-</td><td>No</td><td>uint32_vbr</td>
485 <td align="left">Number of entries in type plane</td>
486 </tr><tr>
487 <td>12-15<sup>1</sup></td><td>-</td><td>No</td><td>uint32_vbr</td>
488 <td align="left">Type plane index for following entries</td>
489 </tr><tr>
490 <td>16-19<sup>1,2</sup></td><td>-</td><td>No</td><td>uint32_vbr</td>
491 <td align="left">Slot number of a value.</td>
492 </tr><tr>
493 <td>variable<sup>1,2</sup></td><td>-</td><td>No</td><td>string</td>
494 <td align="left">Name of the value in the symbol table.</td>
495 </tr>
496 <tr>
497 <td align="left" colspan="5"><sup>1</sup>Maximum length shown,
498 may be smaller<br><sup>2</sup>Repeated field.
499 </tr>
500</table>
Reid Spencer50026612004-05-22 02:28:36 +0000[diff] [blame]501</div>
502
503<!-- *********************************************************************** -->
504<hr>
505<address>
506 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
507 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
508 <a href="http://validator.w3.org/check/referer"><img
509 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
510
511 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
512 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
513 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
514 Last modified: $Date$
515</address>
516</body>
517</html>
518<!-- vim: sw=2
519-->