diff options
author | The Android Open Source Project <initial-contribution@android.com> | 2009-03-03 19:28:47 -0800 |
---|---|---|
committer | The Android Open Source Project <initial-contribution@android.com> | 2009-03-03 19:28:47 -0800 |
commit | f6c387128427e121477c1b32ad35cdcaa5101ba3 (patch) | |
tree | 2aa25fa8c8c3a9caeecf98fd8ac4cd9b12717997 /docs | |
parent | f72d5de56a522ac3be03873bdde26f23a5eeeb3c (diff) | |
download | android_dalvik-f6c387128427e121477c1b32ad35cdcaa5101ba3.tar.gz android_dalvik-f6c387128427e121477c1b32ad35cdcaa5101ba3.tar.bz2 android_dalvik-f6c387128427e121477c1b32ad35cdcaa5101ba3.zip |
auto import from //depot/cupcake/@135843
Diffstat (limited to 'docs')
68 files changed, 16388 insertions, 0 deletions
diff --git a/docs/dalvik-bytecode.css b/docs/dalvik-bytecode.css new file mode 100644 index 000000000..e4a5caa3c --- /dev/null +++ b/docs/dalvik-bytecode.css @@ -0,0 +1,165 @@ +h1 { + font-family: serif; + color: #222266; +} + +h2 { + font-family: serif; + border-top-style: solid; + border-top-width: 2px; + border-color: #ccccdd; + padding-top: 12px; + margin-top: 48px; + margin-bottom: 2px; + color: #222266; +} + +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + + +/* general for all tables */ + +table { + border-collapse: collapse; + margin-top: 12px; +} + +table th { + font-family: sans-serif; + background: #aabbff; +} + +table td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + border-color: #aaaaff; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 4px; + padding-right: 6px; + background: #eeeeff; +} + +table td p { + margin-top: 4pt; + margin-bottom: 0pt; +} + + + +/* opcodes table */ + +table.instruc { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.instruc td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 2px; + padding-right: 2px; +} + +table.instruc td:first-child { + font-family: monospace; + font-size: 90%; + vertical-align: top; + width: 12%; +} + +table.instruc td:first-child + td { + font-family: monospace; + font-size: 90%; + vertical-align: top; + width: 23%; +} + +table.instruc td:first-child + td i { + font-family: sans-serif; + font-size: 90%; +} + +table.instruc td:first-child + td + td { + vertical-align: top; + width: 28%; +} + +table.instruc td:first-child + td + td + td { + vertical-align: top; + width: 37%; +} + + +/* supplemental opcode format table */ + +table.supplement { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.supplement td:first-child { + font-family: monospace; + vertical-align: top; + width: 20%; +} + +table.supplement td:first-child + td { + font-family: monospace; + vertical-align: top; + width: 20%; +} + +table.supplement td:first-child + td + td { + font-family: sans-serif; + vertical-align: top; + width: 60%; +} + + +/* math details table */ + +table.math { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.math td:first-child { + font-family: monospace; + vertical-align: top; + width: 10%; +} + +table.math td:first-child + td { + font-family: monospace; + vertical-align: top; + width: 30%; +} + +table.math td:first-child + td + td { + font-family: sans-serif; + vertical-align: top; + width: 60%; +} diff --git a/docs/dalvik-bytecode.html b/docs/dalvik-bytecode.html new file mode 100644 index 000000000..4945d60f4 --- /dev/null +++ b/docs/dalvik-bytecode.html @@ -0,0 +1,1500 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>Bytecode for the Dalvik VM</title> +<link rel=stylesheet href="dalvik-bytecode.css"> +</head> + +<body> + +<h1>Bytecode for the Dalvik VM</h1> +<p>Copyright © 2007 The Android Open Source Project + +<h2>General Design</h2> + +<ul> +<li>The machine model and calling conventions are meant to approximately + imitate common real architectures and C-style calling conventions: + <ul> + <li>The VM is register-based, and frames are fixed in size upon creation. + Each frame consists of a particular number of registers (specified by + the method) as well as any adjunct data needed to execute the method, + such as (but not limited to) the program counter and a reference to the + <code>.dex</code> file that contains the method. + </li> + <li>Registers are 32 bits wide. Adjacent register pairs are used for 64-bit + values. + </li> + <li>In terms of bitwise representation, <code>(Object) null == (int) + 0</code>. + </li> + <li>The <i>N</i> arguments to a method land in the last <i>N</i> registers + of the method's invocation frame, in order. Wide arguments consume + two registers. Instance methods are passed a <code>this</code> reference + as their first argument. + </li> + </ul> +<li>The storage unit in the instruction stream is a 16-bit unsigned quantity. + Some bits in some instructions are ignored / must-be-zero. +</li> +<li>Instructions aren't gratuitously limited to a particular type. For + example, instructions that move 32-bit register values without interpretation + don't have to specify whether they are moving ints or floats. +</li> +<li>There are separately enumerated and indexed constant pools for + references to strings, types, fields, and methods. +</li> +<li>Bitwise literal data is represented in-line in the instruction stream.</li> +<li>Because, in practice, it is uncommon for a method to need more than + 16 registers, and because needing more than eight registers <i>is</i> + reasonably common, many instructions may only address the first 16 + registers. When reasonably possible, instructions allow references to + up to the first 256 registers. In cases where an instruction variant isn't + available to address a desired register, it is expected that the register + contents get moved from the original register to a low register (before the + operation) and/or moved from a low result register to a high register + (after the operation). +</li> +<li>There are several "pseudo-instructions" that are used to hold + variable-length data referred to by regular instructions (for example, + <code>fill-array-data</code>). Such instructions must never be + encountered during the normal flow of execution. In addition, the + instructions must be located on even-numbered bytecode offsets (that is, + 4-byte aligned). In order to meet this requirement, dex generation tools + should emit an extra <code>nop</code> instruction as a spacer if such an + instruction would otherwise be unaligned. Finally, though not required, + it is expected that most tools will choose to emit these instructions at + the ends of methods, since otherwise it would likely be the case that + additional instructions would be needed to branch around them. +</li> +<li>When installed on a running system, some instructions may be altered, + changing their format, as an install-time static linking optimization. + This is to allow for faster execution once linkage is known. + See the associated + <a href="instruction-formats.html">instruction formats document</a> + for the suggested variants. The word "suggested" is used advisedly; + it is not mandatory to implement these. +</li> +<li>Human-syntax and mnemonics: + <ul> + <li>Dest-then-source ordering for arguments.</li> + <li>Some opcodes have a disambiguating suffix with respect to the type(s) + they operate on: Type-general 64-bit opcodes + are suffixed with <code>-wide</code>. + Type-specific opcodes are suffixed with their type (or a + straightforward abbreviation), one of: <code>-boolean</code> + <code>-byte</code> <code>-char</code> <code>-short</code> + <code>-int</code> <code>-long</code> <code>-float</code> + <code>-double</code> <code>-object</code> <code>-string</code> + <code>-class</code> <code>-void</code>. Type-general 32-bit opcodes + are unmarked. + </li> + <li>Some opcodes have a disambiguating suffix to distinguish + otherwise-identical operations that have different instruction layouts + or options. These suffixes are separated from the main names with a slash + ("<code>/</code>") and mainly exist at all to make there be a one-to-one + mapping with static constants in the code that generates and interprets + executables (that is, to reduce ambiguity for humans). + </li> + </ul> +</li> +<li>See the <a href="instruction-formats.html">instruction formats + document</a> for more details about the various instruction formats + (listed under "Op & Format") as well as details about the opcode + syntax. +</li> +</ul> + +<h2>Summary of Instruction Set</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>00 10x</td> + <td>nop</td> + <td> </td> + <td>Waste cycles.</td> +</tr> +<tr> + <td>01 12x</td> + <td>move vA, vB</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> source register (4 bits)</td> + <td>Move the contents of one non-object register to another.</td> +</tr> +<tr> + <td>02 22x</td> + <td>move/from16 vAA, vBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> source register (16 bits)</td> + <td>Move the contents of one non-object register to another.</td> +</tr> +<tr> + <td>03 32x</td> + <td>move/16 vAAAA, vBBBB</td> + <td><code>A:</code> destination register (16 bits)<br/> + <code>B:</code> source register (16 bits)</td> + <td>Move the contents of one non-object register to another.</td> +</tr> +<tr> + <td>04 12x</td> + <td>move-wide vA, vB</td> + <td><code>A:</code> destination register pair (4 bits)<br/> + <code>B:</code> source register pair (4 bits)</td> + <td>Move the contents of one register-pair to another. + <p><b>Note:</b> + It is legal to move from <code>v<i>N</i></code> to either + <code>v<i>N-1</i></code> or <code>v<i>N+1</i></code>, so implementations + must arrange for both halves of a register pair to be read before + anything is written.</p> + </td> +</tr> +<tr> + <td>05 22x</td> + <td>move-wide/from16 vAA, vBBBB</td> + <td><code>A:</code> destination register pair (8 bits)<br/> + <code>B:</code> source register pair (16 bits)</td> + <td>Move the contents of one register-pair to another. + <p><b>Note:</b> + Implementation considerations are the same as <code>move-wide</code>, + above.</p> + </td> +</tr> +<tr> + <td>06 32x</td> + <td>move-wide/16 vAAAA, vBBBB</td> + <td><code>A:</code> destination register pair (16 bits)<br/> + <code>B:</code> source register pair (16 bits)</td> + <td>Move the contents of one register-pair to another. + <p><b>Note:</b> + Implementation considerations are the same as <code>move-wide</code>, + above.</p> + </td> +</tr> +<tr> + <td>07 12x</td> + <td>move-object vA, vB</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> source register (4 bits)</td> + <td>Move the contents of one object-bearing register to another.</td> +</tr> +<tr> + <td>08 22x</td> + <td>move-object/from16 vAA, vBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> source register (16 bits)</td> + <td>Move the contents of one object-bearing register to another.</td> +</tr> +<tr> + <td>09 32x</td> + <td>move-object/16 vAAAA, vBBBB</td> + <td><code>A:</code> destination register (16 bits)<br/> + <code>B:</code> source register (16 bits)</td> + <td>Move the contents of one object-bearing register to another.</td> +</tr> +<tr> + <td>0a 11x</td> + <td>move-result vAA</td> + <td><code>A:</code> destination register (8 bits)</td> + <td>Move the single-word non-object result of the most recent + <code>invoke-<i>kind</i></code> into the indicated register. + This must be done as the instruction immediately after an + <code>invoke-<i>kind</i></code> whose (single-word, non-object) result + is not to be ignored; anywhere else is invalid.</td> +</tr> +<tr> + <td>0b 11x</td> + <td>move-result-wide vAA</td> + <td><code>A:</code> destination register pair (8 bits)</td> + <td>Move the double-word result of the most recent + <code>invoke-<i>kind</i></code> into the indicated register pair. + This must be done as the instruction immediately after an + <code>invoke-<i>kind</i></code> whose (double-word) result + is not to be ignored; anywhere else is invalid.</td> +</tr> +<tr> + <td>0c 11x</td> + <td>move-result-object vAA</td> + <td><code>A:</code> destination register (8 bits)</td> + <td>Move the object result of the most recent <code>invoke-<i>kind</i></code> + into the indicated register. This must be done as the instruction + immediately after an <code>invoke-<i>kind</i></code> or + <code>filled-new-array</code> + whose (object) result is not to be ignored; anywhere else is invalid.</td> +</tr> +<tr> + <td>0d 11x</td> + <td>move-exception vAA</td> + <td><code>A:</code> destination register (8 bits)</td> + <td>Save a just-caught exception into the given register. This should + be the first instruction of any exception handler whose caught + exception is not to be ignored, and this instruction may <i>only</i> + ever occur as the first instruction of an exception handler; anywhere + else is invalid.</td> +</tr> +<tr> + <td>0e 10x</td> + <td>return-void</td> + <td> </td> + <td>Return from a <code>void</code> method.</td> +</tr> +<tr> + <td>0f 11x</td> + <td>return vAA</td> + <td><code>A:</code> return value register (8 bits)</td> + <td>Return from a single-width (32-bit) non-object value-returning + method. + </td> +</tr> +<tr> + <td>10 11x</td> + <td>return-wide vAA</td> + <td><code>A:</code> return value register-pair (8 bits)</td> + <td>Return from a double-width (64-bit) value-returning method.</td> +</tr> +<tr> + <td>11 11x</td> + <td>return-object vAA</td> + <td><code>A:</code> return value register (8 bits)</td> + <td>Return from an object-returning method.</td> +</tr> +<tr> + <td>12 11n</td> + <td>const/4 vA, #+B</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> signed int (4 bits)</td> + <td>Move the given literal value (sign-extended to 32 bits) into + the specified register.</td> +</tr> +<tr> + <td>13 21s</td> + <td>const/16 vAA, #+BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> + <td>Move the given literal value (sign-extended to 32 bits) into + the specified register.</td> +</tr> +<tr> + <td>14 31i</td> + <td>const vAA, #+BBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> arbitrary 32-bit constant</td> + <td>Move the given literal value into the specified register.</td> +</tr> +<tr> + <td>15 21h</td> + <td>const/high16 vAA, #+BBBB0000</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> + <td>Move the given literal value (right-zero-extended to 32 bits) into + the specified register.</td> +</tr> +<tr> + <td>16 21s</td> + <td>const-wide/16 vAA, #+BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> + <td>Move the given literal value (sign-extended to 64 bits) into + the specified register-pair.</td> +</tr> +<tr> + <td>17 31i</td> + <td>const-wide/32 vAA, #+BBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (32 bits)</td> + <td>Move the given literal value (sign-extended to 64 bits) into + the specified register-pair.</td> +</tr> +<tr> + <td>18 51l</td> + <td>const-wide vAA, #+BBBBBBBBBBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> arbitrary double-width (64-bit) constant</td> + <td>Move the given literal value into + the specified register-pair.</td> +</tr> +<tr> + <td>19 21h</td> + <td>const-wide/high16 vAA, #+BBBB000000000000</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> + <td>Move the given literal value (right-zero-extended to 64 bits) into + the specified register-pair.</td> +</tr> +<tr> + <td>1a 21c</td> + <td>const-string vAA, string@BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> string index</td> + <td>Move a reference to the string specified by the given index into the + specified register.</td> +</tr> +<tr> + <td>1b 31c</td> + <td>const-string/jumbo vAA, string@BBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> string index</td> + <td>Move a reference to the string specified by the given index into the + specified register.</td> +</tr> +<tr> + <td>1c 21c</td> + <td>const-class vAA, type@BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> type index</td> + <td>Move a reference to the class specified by the given index into the + specified register. In the case where the indicated type is primitive, + this will store a reference to the primitive type's degenerate + class.</td> +</tr> +<tr> + <td>1d 11x</td> + <td>monitor-enter vAA</td> + <td><code>A:</code> reference-bearing register (8 bits)</td> + <td>Acquire the monitor for the indicated object.</td> +</tr> +<tr> + <td>1e 11x</td> + <td>monitor-exit vAA</td> + <td><code>A:</code> reference-bearing register (8 bits)</td> + <td>Release the monitor for the indicated object. + <p><b>Note:</b> + If this instruction needs to throw an exception, it must do + so as if the pc has already advanced past the instruction. + It may be useful to think of this as the instruction successfully + executing (in a sense), and the exception getting thrown <i>after</i> + the instruction but <i>before</i> the next one gets a chance to + run. This definition makes it possible for a method to use + a monitor cleanup catch-all (e.g., <code>finally</code>) block as + the monitor cleanup for that block itself, as a way to handle the + arbitrary exceptions that might get thrown due to the historical + implementation of <code>Thread.stop()</code>, while still managing + to have proper monitor hygiene.</p> + </td> +</tr> +<tr> + <td>1f 21c</td> + <td>check-cast vAA, type@BBBB</td> + <td><code>A:</code> reference-bearing register (8 bits)<br/> + <code>B:</code> type index (16 bits)</td> + <td>Throw a <code>ClassCastException</code> if the reference in the + given register cannot be cast to the indicated type. + <p><b>Note:</b> Since <code>A</code> must always be a reference + (and not a primitive value), this will necessarily fail at runtime + (that is, it will throw an exception) if <code>B</code> refers to a + primitive type.</p> + </td> +</tr> +<tr> + <td>20 22c</td> + <td>instance-of vA, vB, type@CCCC</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> reference-bearing register (4 bits)<br/> + <code>C:</code> type index (16 bits)</td> + <td>Store in the given destination register <code>1</code> + if the indicated reference is an instance of the given type, + or <code>0</code> if not. + <p><b>Note:</b> Since <code>B</code> must always be a reference + (and not a primitive value), this will always result + in <code>0</code> being stored if <code>C</code> refers to a primitive + type.</td> +</tr> +<tr> + <td>21 12x</td> + <td>array-length vA, vB</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> array reference-bearing register (4 bits)</td> + <td>Store in the given destination register the length of the indicated + array, in entries</td> +</tr> +<tr> + <td>22 21c</td> + <td>new-instance vAA, type@BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> type index</td> + <td>Construct a new instance of the indicated type, storing a + reference to it in the destination. The type must refer to a + non-array class.</td> +</tr> +<tr> + <td>23 22c</td> + <td>new-array vA, vB, type@CCCC</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> size register<br/> + <code>C:</code> type index</td> + <td>Construct a new array of the indicated type and size. The type + must be an array type.</td> +</tr> +<tr> + <td>24 35c</td> + <td>filled-new-array {vD, vE, vF, vG, vA}, type@CCCC</td> + <td><code>B:</code> array size and argument word count (4 bits)<br/> + <code>C:</code> type index (16 bits)<br/> + <code>D..G, A:</code> argument registers (4 bits each)</td> + <td>Construct an array of the given type and size, filling it with the + supplied contents. The type must be an array type. The array's + contents must be single-word (that is, + no arrays of <code>long</code> or <code>double</code>). The constructed + instance is stored as a "result" in the same way that the method invocation + instructions store their results, so the constructed instance must + be moved to a register with a subsequent + <code>move-result-object</code> instruction (if it is to be used).</td> +</tr> +<tr> + <td>25 3rc</td> + <td>filled-new-array/range {vCCCC .. vNNNN}, type@BBBB</td> + <td><code>A:</code> array size and argument word count (8 bits)<br/> + <code>B:</code> type index (16 bits)<br/> + <code>C:</code> first argument register (16 bits)<br/> + <code>N = A + C - 1</code></td> + <td>Construct an array of the given type and size, filling it with + the supplied contents. Clarifications and restrictions are the same + as <code>filled-new-array</code>, described above.</td> +</tr> +<tr> + <td>26 31t</td> + <td>fill-array-data vAA, +BBBBBBBB <i>(with supplemental data as specified + below in "<code>fill-array-data</code> Format")</i></td> + <td><code>A:</code> array reference (8 bits)<br/> + <code>B:</code> signed "branch" offset to table data pseudo-instruction + (32 bits) + </td> + <td>Fill the given array with the indicated data. The reference must be + to an array of primitives, and the data table must match it in type and + must contain no more elements than will fit in the array. That is, + the array may be larger than the table, and if so, only the initial + elements of the array are set, leaving the remainder alone. + </td> +</tr> +<tr> + <td>27 11x</td> + <td>throw vAA</td> + <td><code>A:</code> exception-bearing register (8 bits)<br/></td> + <td>Throw the indicated exception.</td> +</tr> +<tr> + <td>28 10t</td> + <td>goto +AA</td> + <td><code>A:</code> signed branch offset (8 bits)</td> + <td>Unconditionally jump to the indicated instruction. + <p><b>Note:</b> + The branch offset may not be <code>0</code>. (A spin + loop may be legally constructed either with <code>goto/32</code> or + by including a <code>nop</code> as a target before the branch.)</p> + </td> +</tr> +<tr> + <td>29 20t</td> + <td>goto/16 +AAAA</td> + <td><code>A:</code> signed branch offset (16 bits)<br/></td> + <td>Unconditionally jump to the indicated instruction. + <p><b>Note:</b> + The branch offset may not be <code>0</code>. (A spin + loop may be legally constructed either with <code>goto/32</code> or + by including a <code>nop</code> as a target before the branch.)</p> + </td> +</tr> +<tr> + <td>2a 30t</td> + <td>goto/32 +AAAAAAAA</td> + <td><code>A:</code> signed branch offset (32 bits)<br/></td> + <td>Unconditionally jump to the indicated instruction.</td> +</tr> +<tr> + <td>2b 31t</td> + <td>packed-switch vAA, +BBBBBBBB <i>(with supplemental data as + specified below in "<code>packed-switch</code> Format")</i></td> + <td><code>A:</code> register to test<br/> + <code>B:</code> signed "branch" offset to table data pseudo-instruction + (32 bits) + </td> + <td>Jump to a new instruction based on the value in the + given register, using a table of offsets corresponding to each value + in a particular integral range, or fall through to the next + instruction if there is no match. + </td> +</tr> +<tr> + <td>2c 31t</td> + <td>sparse-switch vAA, +BBBBBBBB <i>(with supplemental data as + specified below in "<code>sparse-switch</code> Format")</i></td> + <td><code>A:</code> register to test<br/> + <code>B:</code> signed "branch" offset to table data pseudo-instruction + (32 bits) + </td> + <td>Jump to a new instruction based on the value in the given + register, using an ordered table of value-offset pairs, or fall + through to the next instruction if there is no match. + </td> +</tr> +<tr> + <td>2d..31 23x</td> + <td>cmp<i>kind</i> vAA, vBB, vCC<br/> + 2d: cmpl-float <i>(lt bias)</i><br/> + 2e: cmpg-float <i>(gt bias)</i><br/> + 2f: cmpl-double <i>(lt bias)</i><br/> + 30: cmpg-double <i>(gt bias)</i><br/> + 31: cmp-long + </td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> first source register or pair<br/> + <code>C:</code> second source register or pair</td> + <td>Perform the indicated floating point or <code>long</code> comparison, + storing <code>0</code> if the two arguments are equal, <code>1</code> + if the second argument is larger, or <code>-1</code> if the first + argument is larger. The "bias" listed for the floating point operations + indicates how <code>NaN</code> comparisons are treated: "Gt bias" + instructions return <code>1</code> for <code>NaN</code> comparisons, + and "lt bias" instructions return + <code>-1</code>. + <p>For example, to check to see if floating point + <code>a < b</code>, then it is advisable to use + <code>cmpg-float</code>; a result of <code>-1</code> indicates that + the test was true, and the other values indicate it was false either + due to a valid comparison or because one or the other values was + <code>NaN</code>.</p> + </td> +</tr> +<tr> + <td>32..37 22t</td> + <td>if-<i>test</i> vA, vB, +CCCC<br/> + 32: if-eq<br/> + 33: if-ne<br/> + 34: if-lt<br/> + 35: if-ge<br/> + 36: if-gt<br/> + 37: if-le<br/> + </td> + <td><code>A:</code> first register to test (4 bits)<br/> + <code>B:</code> second register to test (4 bits)<br/> + <code>C:</code> signed branch offset (16 bits)</td> + <td>Branch to the given destination if the given two registers' values + compare as specified. + <p><b>Note:</b> + The branch offset may not be <code>0</code>. (A spin + loop may be legally constructed either by branching around a + backward <code>goto</code> or by including a <code>nop</code> as + a target before the branch.)</p> + </td> +</tr> +<tr> + <td>38..3d 21t</td> + <td>if-<i>test</i>z vAA, +BBBB<br/> + 38: if-eqz<br/> + 39: if-nez<br/> + 3a: if-ltz<br/> + 3b: if-gez<br/> + 3c: if-gtz<br/> + 3d: if-lez<br/> + </td> + <td><code>A:</code> register to test (8 bits)<br/> + <code>B:</code> signed branch offset (16 bits)</td> + <td>Branch to the given destination if the given register's value compares + with 0 as specified. + <p><b>Note:</b> + The branch offset may not be <code>0</code>. (A spin + loop may be legally constructed either by branching around a + backward <code>goto</code> or by including a <code>nop</code> as + a target before the branch.)</p> + </td> +</tr> +<tr> + <td>3e..43 10x</td> + <td><i>(unused)</i></td> + <td> </td> + <td><i>(unused)</i></td> +</tr> +<tr> + <td>44..51 23x</td> + <td><i>arrayop</i> vAA, vBB, vCC<br/> + 44: aget<br/> + 45: aget-wide<br/> + 46: aget-object<br/> + 47: aget-boolean<br/> + 48: aget-byte<br/> + 49: aget-char<br/> + 4a: aget-short<br/> + 4b: aput<br/> + 4c: aput-wide<br/> + 4d: aput-object<br/> + 4e: aput-boolean<br/> + 4f: aput-byte<br/> + 50: aput-char<br/> + 51: aput-short + </td> + <td><code>A:</code> value register or pair; may be source or dest + (8 bits)<br/> + <code>B:</code> array register (8 bits)<br/> + <code>C:</code> index register (8 bits)</td> + <td>Perform the identified array operation at the identified index of + the given array, loading or storing into the value register.</td> +</tr> +<tr> + <td>52..5f 22c</td> + <td>i<i>instanceop</i> vA, vB, field@CCCC<br/> + 52: iget<br/> + 53: iget-wide<br/> + 54: iget-object<br/> + 55: iget-boolean<br/> + 56: iget-byte<br/> + 57: iget-char<br/> + 58: iget-short<br/> + 59: iput<br/> + 5a: iput-wide<br/> + 5b: iput-object<br/> + 5c: iput-boolean<br/> + 5d: iput-byte<br/> + 5e: iput-char<br/> + 5f: iput-short + </td> + <td><code>A:</code> value register or pair; may be source or dest + (4 bits)<br/> + <code>B:</code> object register (4 bits)<br/> + <code>C:</code> instance field reference index (16 bits)</td> + <td>Perform the identified object instance field operation with + the identified field, loading or storing into the value register. + <p><b>Note:</b> These opcodes are reasonable candidates for static linking, + altering the field argument to be a more direct offset.</p> + </td> +</tr> +<tr> + <td>60..6d 21c</td> + <td>s<i>staticop</i> vAA, field@BBBB<br/> + 60: sget<br/> + 61: sget-wide<br/> + 62: sget-object<br/> + 63: sget-boolean<br/> + 64: sget-byte<br/> + 65: sget-char<br/> + 66: sget-short<br/> + 67: sput<br/> + 68: sput-wide<br/> + 69: sput-object<br/> + 6a: sput-boolean<br/> + 6b: sput-byte<br/> + 6c: sput-char<br/> + 6d: sput-short + </td> + <td><code>A:</code> value register or pair; may be source or dest + (8 bits)<br/> + <code>B:</code> static field reference index (16 bits)</td> + <td>Perform the identified object static field operation with the identified + static field, loading or storing into the value register. + <p><b>Note:</b> These opcodes are reasonable candidates for static linking, + altering the field argument to be a more direct offset.</p> + </td> +</tr> +<tr> + <td>6e..72 35c</td> + <td>invoke-<i>kind</i> {vD, vE, vF, vG, vA}, meth@CCCC<br/> + 6e: invoke-virtual<br/> + 6f: invoke-super<br/> + 70: invoke-direct<br/> + 71: invoke-static<br/> + 72: invoke-interface + </td> + <td><code>B:</code> argument word count (4 bits)<br/> + <code>C:</code> method index (16 bits)<br/> + <code>D..G, A:</code> argument registers (4 bits each)</td> + <td>Call the indicated method. The result (if any) may be stored + with an appropriate <code>move-result*</code> variant as the immediately + subsequent instruction. + <p><code>invoke-virtual</code> is used to invoke a normal virtual + method (a method that is not <code>static</code> or <code>final</code>, + and is not a constructor).</p> + <p><code>invoke-super</code> is used to invoke the closest superclass's + virtual method (as opposed to the one with the same <code>method_id</code> + in the calling class).</p> + <p><code>invoke-direct</code> is used to invoke a non-<code>static</code> + direct method (that is, an instance method that is by its nature + non-overridable, namely either a <code>private</code> instance method + or a constructor).</p> + <p><code>invoke-static</code> is used to invoke a <code>static</code> + method (which is always considered a direct method).</p> + <p><code>invoke-interface</code> is used to invoke an + <code>interface</code> method, that is, on an object whose concrete + class isn't known, using a <code>method_id</code> that refers to + an <code>interface</code>.</p> + <p><b>Note:</b> These opcodes are reasonable candidates for static linking, + altering the method argument to be a more direct offset + (or pair thereof).</p> + </td> +</tr> +<tr> + <td>73 10x</td> + <td><i>(unused)</i></td> + <td> </td> + <td><i>(unused)</i></td> +</tr> +<tr> + <td>74..78 3rc</td> + <td>invoke-<i>kind</i>/range {vCCCC .. vNNNN}, meth@BBBB<br/> + 74: invoke-virtual/range<br/> + 75: invoke-super/range<br/> + 76: invoke-direct/range<br/> + 77: invoke-static/range<br/> + 78: invoke-interface/range + </td> + <td><code>A:</code> argument word count (8 bits)<br/> + <code>B:</code> method index (16 bits)<br/> + <code>C:</code> first argument register (16 bits)<br/> + <code>N = A + C - 1</code></td> + <td>Call the indicated method. See first <code>invoke-<i>kind</i></code> + description above for details, caveats, and suggestions. + </td> +</tr> +<tr> + <td>79..7a 10x</td> + <td><i>(unused)</i></td> + <td> </td> + <td><i>(unused)</i></td> +</tr> +<tr> + <td>7b..8f 12x</td> + <td><i>unop</i> vA, vB<br/> + 7b: neg-int<br/> + 7c: not-int<br/> + 7d: neg-long<br/> + 7e: not-long<br/> + 7f: neg-float<br/> + 80: neg-double<br/> + 81: int-to-long<br/> + 82: int-to-float<br/> + 83: int-to-double<br/> + 84: long-to-int<br/> + 85: long-to-float<br/> + 86: long-to-double<br/> + 87: float-to-int<br/> + 88: float-to-long<br/> + 89: float-to-double<br/> + 8a: double-to-int<br/> + 8b: double-to-long<br/> + 8c: double-to-float<br/> + 8d: int-to-byte<br/> + 8e: int-to-char<br/> + 8f: int-to-short + </td> + <td><code>A:</code> destination register or pair (4 bits)<br/> + <code>B:</code> source register or pair (4 bits)</td> + <td>Perform the identified unary operation on the source register, + storing the result in the destination register.</td> +</tr> + +<tr> + <td>90..af 23x</td> + <td><i>binop</i> vAA, vBB, vCC<br/> + 90: add-int<br/> + 91: sub-int<br/> + 92: mul-int<br/> + 93: div-int<br/> + 94: rem-int<br/> + 95: and-int<br/> + 96: or-int<br/> + 97: xor-int<br/> + 98: shl-int<br/> + 99: shr-int<br/> + 9a: ushr-int<br/> + 9b: add-long<br/> + 9c: sub-long<br/> + 9d: mul-long<br/> + 9e: div-long<br/> + 9f: rem-long<br/> + a0: and-long<br/> + a1: or-long<br/> + a2: xor-long<br/> + a3: shl-long<br/> + a4: shr-long<br/> + a5: ushr-long<br/> + a6: add-float<br/> + a7: sub-float<br/> + a8: mul-float<br/> + a9: div-float<br/> + aa: rem-float<br/> + ab: add-double<br/> + ac: sub-double<br/> + ad: mul-double<br/> + ae: div-double<br/> + af: rem-double + </td> + <td><code>A:</code> destination register or pair (8 bits)<br/> + <code>B:</code> first source register or pair (8 bits)<br/> + <code>C:</code> second source register or pair (8 bits)</td> + <td>Perform the identified binary operation on the two source registers, + storing the result in the first source register.</td> +</tr> +<tr> + <td>b0..cf 12x</td> + <td><i>binop</i>/2addr vA, vB<br/> + b0: add-int/2addr<br/> + b1: sub-int/2addr<br/> + b2: mul-int/2addr<br/> + b3: div-int/2addr<br/> + b4: rem-int/2addr<br/> + b5: and-int/2addr<br/> + b6: or-int/2addr<br/> + b7: xor-int/2addr<br/> + b8: shl-int/2addr<br/> + b9: shr-int/2addr<br/> + ba: ushr-int/2addr<br/> + bb: add-long/2addr<br/> + bc: sub-long/2addr<br/> + bd: mul-long/2addr<br/> + be: div-long/2addr<br/> + bf: rem-long/2addr<br/> + c0: and-long/2addr<br/> + c1: or-long/2addr<br/> + c2: xor-long/2addr<br/> + c3: shl-long/2addr<br/> + c4: shr-long/2addr<br/> + c5: ushr-long/2addr<br/> + c6: add-float/2addr<br/> + c7: sub-float/2addr<br/> + c8: mul-float/2addr<br/> + c9: div-float/2addr<br/> + ca: rem-float/2addr<br/> + cb: add-double/2addr<br/> + cc: sub-double/2addr<br/> + cd: mul-double/2addr<br/> + ce: div-double/2addr<br/> + cf: rem-double/2addr + </td> + <td><code>A:</code> destination and first source register or pair + (4 bits)<br/> + <code>B:</code> second source register or pair (4 bits)</td> + <td>Perform the identified binary operation on the two source registers, + storing the result in the first source register.</td> +</tr> +<tr> + <td>d0..d7 22s</td> + <td><i>binop</i>/lit16 vA, vB, #+CCCC<br/> + d0: add-int/lit16<br/> + d1: rsub-int (reverse subtract)<br/> + d2: mul-int/lit16<br/> + d3: div-int/lit16<br/> + d4: rem-int/lit16<br/> + d5: and-int/lit16<br/> + d6: or-int/lit16<br/> + d7: xor-int/lit16 + </td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> source register (4 bits)<br/> + <code>C:</code> signed int constant (16 bits)</td> + <td>Perform the indicated binary op on the indicated register (first + argument) and literal value (second argument), storing the result in + the destination register. + <p><b>Note:</b> + <code>rsub-int</code> does not have a suffix since this version is the + main opcode of its family. Also, see below for details on its semantics. + </p> + </td> +</tr> +<tr> + <td>d8..e2 22b</td> + <td><i>binop</i>/lit8 vAA, vBB, #+CC<br/> + d8: add-int/lit8<br/> + d9: rsub-int/lit8<br/> + da: mul-int/lit8<br/> + db: div-int/lit8<br/> + dc: rem-int/lit8<br/> + dd: and-int/lit8<br/> + de: or-int/lit8<br/> + df: xor-int/lit8<br/> + e0: shl-int/lit8<br/> + e1: shr-int/lit8<br/> + e2: ushr-int/lit8 + </td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> source register (8 bits)<br/> + <code>C:</code> signed int constant (8 bits)</td> + <td>Perform the indicated binary op on the indicated register (first + argument) and literal value (second argument), storing the result + in the destination register. + <p><b>Note:</b> See below for details on the semantics of + <code>rsub-int</code>.</p> + </td> +</tr> +<tr> + <td>e3..ff 10x</td> + <td><i>(unused)</i></td> + <td> </td> + <td><i>(unused)</i></td> +</tr> +</tbody> +</table> + +<h2><code>packed-switch</code> Format</h2> + +<table class="supplement"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>ident</td> + <td>ushort = 0x0100</td> + <td>identifying pseudo-opcode</td> +</tr> +<tr> + <td>size</td> + <td>ushort</td> + <td>number of entries in the table</td> +</tr> +<tr> + <td>first_key</td> + <td>int</td> + <td>first (and lowest) switch case value</td> +</tr> +<tr> + <td>targets</td> + <td>int[]</td> + <td>list of <code>size</code> relative branch targets. The targets are + relative to the address of the switch opcode, not of this table. + </td> +</tr> +</tbody> +</table> + +<p><b>Note:</b> The total number of code units for an instance of this +table is <code>(size * 2) + 4</code>.</p> + +<h2><code>sparse-switch</code> Format</h2> + +<table class="supplement"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>ident</td> + <td>ushort = 0x0200</td> + <td>identifying pseudo-opcode</td> +</tr> +<tr> + <td>size</td> + <td>ushort</td> + <td>number of entries in the table</td> +</tr> +<tr> + <td>keys</td> + <td>int[]</td> + <td>list of <code>size</code> key values, sorted low-to-high</td> +</tr> +<tr> + <td>targets</td> + <td>int[]</td> + <td>list of <code>size</code> relative branch targets, each corresponding + to the key value at the same index. The targets are + relative to the address of the switch opcode, not of this table. + </td> +</tr> +</tbody> +</table> + +<p><b>Note:</b> The total number of code units for an instance of this +table is <code>(size * 4) + 2</code>.</p> + +<h2><code>fill-array-data</code> Format</h2> + +<table class="supplement"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>ident</td> + <td>ushort = 0x0300</td> + <td>identifying pseudo-opcode</td> +</tr> +<tr> + <td>element_width</td> + <td>ushort</td> + <td>number of bytes in each element</td> +</tr> +<tr> + <td>size</td> + <td>uint</td> + <td>number of elements in the table</td> +</tr> +<tr> + <td>data</td> + <td>ubyte[]</td> + <td>data values</td> +</tr> +</tbody> +</table> + +<p><b>Note:</b> The total number of code units for an instance of this +table is <code>(size * element_width + 1) / 2 + 4</code>.</p> + + +<h2>Mathematical Operation Details</h2> + +<p><b>Note:</b> Floating point operations must follow IEEE 754 rules, using +round-to-nearest and gradual underflow, except where stated otherwise.</p> + +<table class="math"> +<thead> +<tr> + <th>Opcode</th> + <th>C Semantics</th> + <th>Notes</th> +</tr> +</thead> +<tbody> +<tr> + <td>neg-int</td> + <td>int32 a;<br/> + int32 result = -a; + </td> + <td>Unary twos-complement.</td> +</tr> +<tr> + <td>not-int</td> + <td>int32 a;<br/> + int32 result = ~a; + </td> + <td>Unary ones-complement.</td> +</tr> +<tr> + <td>neg-long</td> + <td>int64 a;<br/> + int64 result = -a; + </td> + <td>Unary twos-complement.</td> +</tr> +<tr> + <td>not-long</td> + <td>int64 a;<br/> + int64 result = ~a; + </td> + <td>Unary ones-complement.</td> +</tr> +<tr> + <td>neg-float</td> + <td>float a;<br/> + float result = -a; + </td> + <td>Floating point negation.</td> +</tr> +<tr> + <td>neg-double</td> + <td>double a;<br/> + double result = -a; + </td> + <td>Floating point negation.</td> +</tr> +<tr> + <td>int-to-long</td> + <td>int32 a;<br/> + int64 result = (int64) a; + </td> + <td>Sign extension of <code>int32</code> into <code>int64</code>.</td> +</tr> +<tr> + <td>int-to-float</td> + <td>int32 a;<br/> + float result = (float) a; + </td> + <td>Conversion of <code>int32</code> to <code>float</code>, using + round-to-nearest. This loses precision for some values. + </td> +</tr> +<tr> + <td>int-to-double</td> + <td>int32 a;<br/> + double result = (double) a; + </td> + <td>Conversion of <code>int32</code> to <code>double</code>.</td> +</tr> +<tr> + <td>long-to-int</td> + <td>int64 a;<br/> + int32 result = (int32) a; + </td> + <td>Truncation of <code>int64</code> into <code>int32</code>.</td> +</tr> +<tr> + <td>long-to-float</td> + <td>int64 a;<br/> + float result = (float) a; + </td> + <td>Conversion of <code>int64</code> to <code>float</code>, using + round-to-nearest. This loses precision for some values. + </td> +</tr> +<tr> + <td>long-to-double</td> + <td>int64 a;<br/> + double result = (double) a; + </td> + <td>Conversion of <code>int64</code> to <code>double</code>, using + round-to-nearest. This loses precision for some values. + </td> +</tr> +<tr> + <td>float-to-int</td> + <td>float a;<br/> + int32 result = (int32) a; + </td> + <td>Conversion of <code>float</code> to <code>int32</code>, using + round-toward-zero. <code>NaN</code> and <code>-0.0</code> (negative zero) + convert to the integer <code>0</code>. Infinities and values with + too large a magnitude to be represented get converted to either + <code>0x7fffffff</code> or <code>-0x80000000</code> depending on sign. + </td> +</tr> +<tr> + <td>float-to-long</td> + <td>float a;<br/> + int64 result = (int64) a; + </td> + <td>Conversion of <code>float</code> to <code>int64</code>, using + round-toward-zero. The same special case rules as for + <code>float-to-int</code> apply here, except that out-of-range values + get converted to either <code>0x7fffffffffffffff</code> or + <code>-0x8000000000000000</code> depending on sign. + </td> +</tr> +<tr> + <td>float-to-double</td> + <td>float a;<br/> + double result = (double) a; + </td> + <td>Conversion of <code>float</code> to <code>double</code>, preserving + the value exactly. + </td> +</tr> +<tr> + <td>double-to-int</td> + <td>double a;<br/> + int32 result = (int32) a; + </td> + <td>Conversion of <code>double</code> to <code>int32</code>, using + round-toward-zero. The same special case rules as for + <code>float-to-int</code> apply here. + </td> +</tr> +<tr> + <td>double-to-long</td> + <td>double a;<br/> + int64 result = (int64) a; + </td> + <td>Conversion of <code>double</code> to <code>int64</code>, using + round-toward-zero. The same special case rules as for + <code>float-to-long</code> apply here. + </td> +</tr> +<tr> + <td>double-to-float</td> + <td>double a;<br/> + float result = (float) a; + </td> + <td>Conversion of <code>double</code> to <code>float</code>, using + round-to-nearest. This loses precision for some values. + </td> +</tr> +<tr> + <td>int-to-byte</td> + <td>int32 a;<br/> + int32 result = (a << 24) >> 24; + </td> + <td>Truncation of <code>int32</code> to <code>int8</code>, sign + extending the result. + </td> +</tr> +<tr> + <td>int-to-char</td> + <td>int32 a;<br/> + int32 result = a & 0xffff; + </td> + <td>Truncation of <code>int32</code> to <code>uint16</code>, without + sign extension. + </td> +</tr> +<tr> + <td>int-to-short</td> + <td>int32 a;<br/> + int32 result = (a << 16) >> 16; + </td> + <td>Truncation of <code>int32</code> to <code>int16</code>, sign + extending the result. + </td> +</tr> +<tr> + <td>add-int</td> + <td>int32 a, b;<br/> + int32 result = a + b; + </td> + <td>Twos-complement addition.</td> +</tr> +<tr> + <td>sub-int</td> + <td>int32 a, b;<br/> + int32 result = a - b; + </td> + <td>Twos-complement subtraction.</td> +</tr> +<tr> + <td>rsub-int</td> + <td>int32 a, b;<br/> + int32 result = b - a; + </td> + <td>Twos-complement reverse subtraction.</td> +</tr> +<tr> + <td>mul-int</td> + <td>int32 a, b;<br/> + int32 result = a * b; + </td> + <td>Twos-complement multiplication.</td> +</tr> +<tr> + <td>div-int</td> + <td>int32 a, b;<br/> + int32 result = a / b; + </td> + <td>Twos-complement division, rounded towards zero (that is, truncated to + integer). This throws <code>ArithmeticException</code> if + <code>b == 0</code>. + </td> +</tr> +<tr> + <td>rem-int</td> + <td>int32 a, b;<br/> + int32 result = a % b; + </td> + <td>Twos-complement remainder after division. The sign of the result + is the same as that of <code>a</code>, and it is more precisely + defined as <code>result == a - (a / b) * b</code>. This throws + <code>ArithmeticException</code> if <code>b == 0</code>. + </td> +</tr> +<tr> + <td>and-int</td> + <td>int32 a, b;<br/> + int32 result = a & b; + </td> + <td>Bitwise AND.</td> +</tr> +<tr> + <td>or-int</td> + <td>int32 a, b;<br/> + int32 result = a | b; + </td> + <td>Bitwise OR.</td> +</tr> +<tr> + <td>xor-int</td> + <td>int32 a, b;<br/> + int32 result = a ^ b; + </td> + <td>Bitwise XOR.</td> +</tr> +<tr> + <td>shl-int</td> + <td>int32 a, b;<br/> + int32 result = a << (b & 0x1f); + </td> + <td>Bitwise shift left (with masked argument).</td> +</tr> +<tr> + <td>shr-int</td> + <td>int32 a, b;<br/> + int32 result = a >> (b & 0x1f); + </td> + <td>Bitwise signed shift right (with masked argument).</td> +</tr> +<tr> + <td>ushr-int</td> + <td>uint32 a, b;<br/> + int32 result = a >> (b & 0x1f); + </td> + <td>Bitwise unsigned shift right (with masked argument).</td> +</tr> +<tr> + <td>add-long</td> + <td>int64 a, b;<br/> + int64 result = a + b; + </td> + <td>Twos-complement addition.</td> +</tr> +<tr> + <td>sub-long</td> + <td>int64 a, b;<br/> + int64 result = a - b; + </td> + <td>Twos-complement subtraction.</td> +</tr> +<tr> + <td>mul-long</td> + <td>int64 a, b;<br/> + int64 result = a * b; + </td> + <td>Twos-complement multiplication.</td> +</tr> +<tr> + <td>div-long</td> + <td>int64 a, b;<br/> + int64 result = a / b; + </td> + <td>Twos-complement division, rounded towards zero (that is, truncated to + integer). This throws <code>ArithmeticException</code> if + <code>b == 0</code>. + </td> +</tr> +<tr> + <td>rem-long</td> + <td>int64 a, b;<br/> + int64 result = a % b; + </td> + <td>Twos-complement remainder after division. The sign of the result + is the same as that of <code>a</code>, and it is more precisely + defined as <code>result == a - (a / b) * b</code>. This throws + <code>ArithmeticException</code> if <code>b == 0</code>. + </td> +</tr> +<tr> + <td>and-long</td> + <td>int64 a, b;<br/> + int64 result = a & b; + </td> + <td>Bitwise AND.</td> +</tr> +<tr> + <td>or-long</td> + <td>int64 a, b;<br/> + int64 result = a | b; + </td> + <td>Bitwise OR.</td> +</tr> +<tr> + <td>xor-long</td> + <td>int64 a, b;<br/> + int64 result = a ^ b; + </td> + <td>Bitwise XOR.</td> +</tr> +<tr> + <td>shl-long</td> + <td>int64 a, b;<br/> + int64 result = a << (b & 0x3f); + </td> + <td>Bitwise shift left (with masked argument).</td> +</tr> +<tr> + <td>shr-long</td> + <td>int64 a, b;<br/> + int64 result = a >> (b & 0x3f); + </td> + <td>Bitwise signed shift right (with masked argument).</td> +</tr> +<tr> + <td>ushr-long</td> + <td>uint64 a, b;<br/> + int64 result = a >> (b & 0x3f); + </td> + <td>Bitwise unsigned shift right (with masked argument).</td> +</tr> +<tr> + <td>add-float</td> + <td>float a, b;<br/> + float result = a + b; + </td> + <td>Floating point addition.</td> +</tr> +<tr> + <td>sub-float</td> + <td>float a, b;<br/> + float result = a - b; + </td> + <td>Floating point subtraction.</td> +</tr> +<tr> + <td>mul-float</td> + <td>float a, b;<br/> + float result = a * b; + </td> + <td>Floating point multiplication.</td> +</tr> +<tr> + <td>div-float</td> + <td>float a, b;<br/> + float result = a / b; + </td> + <td>Floating point division.</td> +</tr> +<tr> + <td>rem-float</td> + <td>float a, b;<br/> + float result = a % b; + </td> + <td>Floating point remainder after division. This function is different + than IEEE 754 remainder and is defined as + <code>result == a - roundTowardZero(a / b) * b</code>. + </td> +</tr> +<tr> + <td>add-double</td> + <td>double a, b;<br/> + double result = a + b; + </td> + <td>Floating point addition.</td> +</tr> +<tr> + <td>sub-double</td> + <td>double a, b;<br/> + double result = a - b; + </td> + <td>Floating point subtraction.</td> +</tr> +<tr> + <td>mul-double</td> + <td>double a, b;<br/> + double result = a * b; + </td> + <td>Floating point multiplication.</td> +</tr> +<tr> + <td>div-double</td> + <td>double a, b;<br/> + double result = a / b; + </td> + <td>Floating point division.</td> +</tr> +<tr> + <td>rem-double</td> + <td>double a, b;<br/> + double result = a % b; + </td> + <td>Floating point remainder after division. This function is different + than IEEE 754 remainder and is defined as + <code>result == a - roundTowardZero(a / b) * b</code>. + </td> +</tr> +</tbody> +</table> + +</body> +</html> diff --git a/docs/dalvik-constraints.css b/docs/dalvik-constraints.css new file mode 100644 index 000000000..a315a736b --- /dev/null +++ b/docs/dalvik-constraints.css @@ -0,0 +1,59 @@ +h1 { + font-family: serif; + color: #222266; +} + +h2 { + font-family: serif; + border-top-style: solid; + border-top-width: 2px; + border-color: #ccccdd; + padding-top: 12px; + margin-top: 48px; + margin-bottom: 2px; + color: #222266; +} + +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + + +/* general for all tables */ + +table { + border-collapse: collapse; + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table th { + font-family: sans-serif; + background: #aabbff; + text-align: left; +} + +table td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + border-color: #aaaaff; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 4px; + padding-right: 6px; + background: #eeeeff; + margin-top: 4pt; + margin-bottom: 0pt; +} diff --git a/docs/dalvik-constraints.html b/docs/dalvik-constraints.html new file mode 100644 index 000000000..105225a7b --- /dev/null +++ b/docs/dalvik-constraints.html @@ -0,0 +1,897 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + <head> + <title>Dalvik bytecode constraints</title> + <link rel=stylesheet href="dalvik-constraints.css"> + </head> + + <body> + + <h1>Dalvik bytecode constraints</h1> + +<!-- + <h1>General integrity constraints</h1> + + <table> + <tr> + <th> + Identifier + </th> + + <th> + Description + </th> + </tr> + + <tr> + <td> + A1 + </td> + + <td> + The magic number of the DEX file must be "dex\n035\0". + </td> + </tr> + + <tr> + <td> + A1 + </td> + + <td> + The checksum must be an Adler-32 checksum of the whole file contents + except magic and checksum field. + </td> + </tr> + + +The signature must be a SHA-1 hash of the whole file contents except magic, +checksum, and signature. + +The file_size must match the actual file size in bytes. + +The header_size must have the value 0x70. + +The endian_tag must have either the value ENDIAN_CONSTANT or +REVERSE_ENDIAN_CONSTANT. + +For each of the link, string_ids, type_ids, proto_ids, field_ids, method_ids, class_defs +and data sections, the offset and size fields must be either both zero or both +non-zero. In the latter case, the offset must be four-byte-aligned. + +All offset fields in the header except map_off must be four-byte-aligned. + +The map_off field must be either zero or point into the data section. In the +latter case, the data section must exist. + +None of the link, string_ids, type_ids, proto_ids, field_ids, method_ids, class_defs +and data sections must overlap each other or the header. + +If a map exists, then each map entry must have a valid type. Each type may +appear at most once. + +If a map exists, then each map entry must have a nonzero offset and size. The +offset must point into the corresponding section of the file (i.e. a +string_id_item must point into the string_ids section) and the explicit or +implicit size of the item must match the actual contents and size of the +section. + +If a map exists, then the offset of map entry n+1 must be greater or equal to +the offset of map entry n plus then size of map entry n. This implies +non-overlapping entries and low-to-high ordering. + +The following types of entries must have an offset that is +four-byte-aligned: string_id_item, type_id_item, proto_id_item, field_id_item, +method_id_item, class_def_item, type_list, code_item, +annotations_directory_item. + +For each string_id_item, the string_data_off field must contain a valid +reference into the data section. For the referenced string_data_item, the data +field must contain a valid MUTF-8 string, and the utf16_size must match the +decoded length of the string. + +For each type_id_item, the desciptor_idx field must contain a valid reference +into the string_ids list. The referenced string must be a valid type descriptor. + +For each proto_id_item, the shorty_idx field must contain a valid reference +into the string_ids list. The referenced string must be a valid shorty descriptor. +Also, the return_type_idx field must be a valid index into the type_ids section, +and the parameters_off field must be either zero or a valid offset pointing +into the data section. If nonzero, the parameter list must not contain any void +entries. + +For each field_id_item, both the class_idx and type_idx fields must be a valid + indices into the +type_ids list. The entry referenced by class_idx must be a non-array reference type. +In addition, the name_idx field must be a valid reference into the string_ids +section, and the contents of the referenced entry must conform to the MemberName +specification. + +For each method_id_item, the class_idx field must be a valid index into the +type_ids section, and the +referenced entry must be a non-array reference type. The proto_id field must +be a valid reference into the proto_ids list. The name_idx field must be a +valid reference into the string_ids +section, and the contents of the referenced entry must conform to the MemberName +specification. + +For each class_def_item, ... + +For each field_id_item, the class_idx field must be a valid index into the +type_ids list. The referenced entry must be a non-array reference type. + +... + +--> + + <h2> + Static constraints + </h2> + + <p> + Static constraints are constraints on individual elements of the bytecode. + They usually can be checked without employing control or data-flow analysis + techniques. + </p> + + <table> + <tr> + <th> + Identifier + </th> + + <th> + Description + </th> + + <th> + Spec equivalent + </th> + </tr> + + <tr> + <td> + A1 + </td> + + <td> + The <code>insns</code> array must not be empty. + </td> + + <td> + 4.8.1.1 + </td> + </tr> + + <tr> + <td> + A2 + </td> + + <td> + The first opcode in the <code>insns</code> array must have index zero. + </td> + + <td> + 4.8.1.3 + </td> + </tr> + + <tr> + <td> + A3 + </td> + + <td> + The <code>insns</code> array must only contain valid Dalvik opcodes. + </td> + + <td> + 4.8.1.4 + </td> + </tr> + + <tr> + <td> + A4 + </td> + + <td> + The index of instruction <code>n+1</code> must equal the index of + instruction <code>n</code> plus the length of instruction + <code>n</code>, taking into account possible operands. + </td> + + <td> + 4.8.1.5 + </td> + </tr> + + <tr> + <td> + A5 + </td> + + <td> + The last instruction in the <code>insns</code> array must end at index + <code>insns_size-1</code>. + </td> + + <td> + 4.8.1.6 + </td> + </tr> + + <tr> + <td> + A6 + </td> + + <td> + All <code>goto</code> and <code>if-<kind></code> targets must + be opcodes within in the same method. + </td> + + <td> + 4.8.1.7 + </td> + </tr> + + <tr> + <td> + A7 + </td> + + <td> + All targets of a <code>packed-switch</code> instruction must be + opcodes within in the same method. The size and the list of targets + must be consistent. + </td> + + <td> + 4.8.1.8 + </td> + </tr> + + <tr> + <td> + A8 + </td> + + <td> + All targets of a <code>sparse-switch</code> instruction must be + opcodes within in the same method. The corresponding table must be + consistent and sorted low-to-high. + </td> + + <td> + 4.8.1.9 + </td> + </tr> + + <tr> + <td> + A9 + </td> + + <td> + The <code>B</code> operand of the <code>const-string</code> and + <code>const-string/jumbo</code> instructions must be a valid index + into the string constant pool. + </td> + + <td> + 4.8.1.10 + </td> + </tr> + + <tr> + <td> + A10 + </td> + + <td> + The <code>C</code> operand of the <code>iget<kind></code> and + <code>iput<kind></code> instructions must be a valid index into + the field constant pool. The referenced entry must represent an + instance field. + </td> + + <td> + 4.8.1.12 + </td> + </tr> + + <tr> + <td> + A11 + </td> + + <td> + The <code>C</code> operand of the <code>sget<kind></code> and + <code>sput<kind></code> instructions must be a valid index into + the field constant pool. The referenced entry must represent a static + field. + </td> + + <td> + 4.8.1.12 + </td> + </tr> + + <tr> + <td> + A12 + </td> + + <td> + The <code>C</code> operand of the <code>invoke-virtual</code>, + <code>invoke-super</code>, <code<invoke-direct</code> and + <code>invoke-static</code> instructions must be a valid index into the + method constant pool. In all cases, the referenced + <code>method_id</code> must belong to a class (not an interface). + </td> + + <td> + 4.8.1.13 + </td> + </tr> + + <tr> + <td> + A13 + </td> + + <td> + The <code>B</code> operand of the <code>invoke-virtual/range</code>, + <code>invoke-super/range</code>, <code>invoke-direct/range</code>, and + <code>invoke-static/range</code> instructions must be a valid index + into the method constant pool. In all cases, the referenced + <code>method_id</code> must belong to a class (not an interface). + </td> + + <td> + 4.8.1.13 + </td> + </tr> + + <tr> + <td> + A14 + </td> + + <td> + A method the name of which starts with a '<' must only be invoked + implicitly by the VM, not by code originating from a Dex file. The + only exception is the instance initializer, which may be invoked by + <code>invoke-direct</code>. + </td> + + <td> + 4.8.1.14 + </td> + </tr> + + <tr> + <td> + A15 + </td> + + <td> + The <code>C</code> operand of the <code>invoke-interface</code> + instruction must be a valid index into the method constant pool. The + referenced <code>method_id</code> must belong to an interface (not a + class). + </td> + + <td> + 4.8.1.15 + </td> + </tr> + + <tr> + <td> + A16 + </td> + + <td> + The <code>B</code> operand of the <code>invoke-interface/range</code> + instruction must be a valid index into the method constant pool. + The referenced <code>method_id</code> must belong to an interface (not + a class). + </td> + + <td> + 4.8.1.15 + </td> + </tr> + + <tr> + <td> + A17 + </td> + + <td> + The <code>B</code> operand of the <code>const-class</code>, + <code>check-cast</code>, <code>new-instance</code>, and + <code>filled-new-array/range</code> instructions must be a valid index + into the type constant pool. + </td> + + <td> + 4.8.1.16 + </td> + </tr> + + <tr> + <td> + A18 + </td> + + <td> + The <code>C</code> operand of the <code>instance-of</code>, + <code>new-array</code>, and <code>filled-new-array</code> + instructions must be a valid index into the type constant pool. + </td> + + <td> + 4.8.1.16 + </td> + </tr> + + <tr> + <td> + A19 + </td> + + <td> + The dimensions of an array created by a <code>new-array</code> + instruction must be less than <code>256</code>. + </td> + + <td> + 4.8.1.17 + </td> + </tr> + + <tr> + <td> + A20 + </td> + + <td> + The <code>new</code> instruction must not refer to array classes, + interfaces, or abstract classes. + </td> + + <td> + 4.8.1.18 + </td> + </tr> + + <tr> + <td> + A21 + </td> + + <td> + The type referred to by a <code>new-array</code> instruction must be + a valid, non-reference type. + </td> + + <td> + 4.8.1.20 + </td> + </tr> + + <tr> + <td> + A22 + </td> + + <td> + All registers referred to by an instruction in a single-width + (non-pair) fashion must be valid for the current method. That is, + their indices must be non-negative and smaller than + <code>registers_size</code>. + </td> + + <td> + 4.8.1.21 + </td> + </tr> + + <tr> + <td> + A23 + </td> + + <td> + All registers referred to by an instruction in a double-width (pair) + fashion must be valid for the current method. That is, their indices + must be non-negative and smaller than <code>registers_size-1</code>. + </td> + + <td> + 4.8.1.23 + </td> + </tr> + </table> + + <h2> + Structural constraints + </h2> + + <p> + Structural constraints are constraints on relationships between several + elements of the bytecode. They usually can't be checked without employing + control or data-flow analysis techniques. + </p> + + <table> + <tr> + <th> + Identifier + </th> + + <th> + Description + </th> + + <th> + Spec equivalent + </th> + </tr> + + <tr> + <td> + B1 + </td> + + <td> + The number and types of arguments (registers and immediate values) + must always match the instruction. + </td> + + <td> + 4.8.2.1 + </td> + </tr> + + <tr> + <td> + B2 + </td> + + <td> + Register pairs must never be broken up. + </td> + + <td> + 4.8.2.3 + </td> + </tr> + + <tr> + <td> + B3 + </td> + + <td> + A register (or pair) has to be assigned first before it can be + read. + </td> + + <td> + 4.8.2.4 + </td> + </tr> + + <tr> + <td> + B4 + </td> + + <td> + An <code>invoke-direct</code> instruction must only invoke an instance + initializer or a method in the current class or one of its + superclasses. + </td> + + <td> + 4.8.2.7 + </td> + </tr> + + <tr> + <td> + B5 + </td> + + <td> + An instance initializer must only be invoked on an uninitialized + instance. + </td> + + <td> + 4.8.2.8 + </td> + </tr> + + <tr> + <td> + B6 + </td> + + <td> + Instance methods may only be invoked on and instance fields may only + be accessed on already initialized instances. + </td> + + <td> + 4.8.2.9 + </td> + </tr> + + <tr> + <td> + B7 + </td> + + <td> + A register which holds the result of a <code>new-instance</code>code> + instruction must not be used if the same + <code>new-instance</code>code> instruction is again executed before + the instance is initialized. + </td> + + <td> + 4.8.2.10 + </td> + </tr> + + <tr> + <td> + B8 + </td> + + <td> + An instance initializer must call another instance initializer (same + class or superclass) before any instance members can be accessed. + Exceptions are non-inherited instance fields, which can be assigned + before calling another initializer, and the <code>Object</code> class + in general. + </td> + + <td> + 4.8.2.11 + </td> + </tr> + + <tr> + <td> + B9 + </td> + + <td> + All actual method arguments must be assignment-compatible with their + respective formal arguments. + </td> + + <td> + 4.8.2.12 + </td> + </tr> + + <tr> + <td> + B10 + </td> + + <td> + For each instance method invocation, the actual instance must be + assignment-compatible with the class or interface specified in the + instruction. + </td> + + <td> + 4.8.2.13 + </td> + </tr> + + <tr> + <td> + B11 + </td> + + <td> + A <code>return<kind></code> instruction must match its + method's return type. + </td> + + <td> + 4.8.2.14 + </td> + </tr> + + <tr> + <td> + B12 + </td> + + <td> + When accessing protected members of a superclass, the actual type of + the instance being accessed must be either the current class or one + of its subclasses. + </td> + + <td> + 4.8.2.15 + </td> + </tr> + + <tr> + <td> + B13 + </td> + + <td> + The type of a value stored into a static field must be + assignment-compatible with or convertible to the field's type. + </td> + + <td> + 4.8.2.16 + </td> + </tr> + + <tr> + <td> + B14 + </td> + + <td> + The type of a value stored into a field must be assignment-compatible + with or convertible to the field's type. + </td> + + <td> + 4.8.2.17 + </td> + </tr> + + <tr> + <td> + B15 + </td> + + <td> + The type of every value stored into an array must be + assignment-compatible with the array's component type. + </td> + + <td> + 4.8.2.18 + </td> + </tr> + + <tr> + <td> + B16 + </td> + + <td> + The <code>A</code> operand of a <code>throw</code> instruction must + be assignment-compatible with <code>java.lang.Throwable</code>. + </td> + + <td> + 4.8.2.19 + </td> + </tr> + + <tr> + <td> + B17 + </td> + + <td> + The last reachable instruction of a method must either be a backwards + <code>goto</code> or branch, a <code>return</code>, or a + <code>throw</code> instruction. It must not be possible to leave the + <code>insns</code> array at the bottom. + </td> + + <td> + 4.8.2.20 + </td> + </tr> + + <tr> + <td> + B18 + </td> + + <td> + The unassigned half of a former register pair may not be read (is + considered invalid) until it has been re-assigned by some other + instruction. + </td> + + <td> + 4.8.2.3, 4.8.2.4 + </td> + </tr> + + <tr> + <td> + B19 + </td> + + <td> + A <code>move-result<kind></code> instruction must be immediately + preceded (in the <code>insns</code> array) by an + <code><invoke-kind></code> instruction. The only exception is + the <code>move-result-object</code> instruction, which may also be + preceded by a <code>filled-new-array</code> instruction. + </td> + + <td> + - + </td> + </tr> + + <tr> + <td> + B20 + </td> + + <td> + A <code>move-result<kind></code> instruction must be immediately + preceded (in actual control flow) by a matching + <code>return-<kind></code> instruction (it must not be jumped + to). The only exception is the <code>move-result-object</code> + instruction, which may also be preceded by a + <code>filled-new-array</code> instruction. + </td> + + <td> + - + </td> + </tr> + + <tr> + <td> + B21 + </td> + + <td> + A <code>move-exception</code> instruction must only appear as the + first instruction in an exception handler. + </td> + + <td> + - + </td> + </tr> + + <tr> + <td> + B22 + </td> + + <td> + The <code>packed-switch-data</code>, <code>sparse-switch-data</code>, + and <code>fill-array-data</code> pseudo-instructions must not be + reachable by control flow. + </td> + + <td> + - + </td> + </tr> + </table> + + </body> +</html> diff --git a/docs/debugger.html b/docs/debugger.html new file mode 100644 index 000000000..6e23f0df2 --- /dev/null +++ b/docs/debugger.html @@ -0,0 +1,211 @@ +<html> +<head> +<title>Dalvik Debugger Support</title> +</head> + +<body> +<h1>Dalvik Debugger Support</h1> + +<p> +The Dalvik virtual machine supports source-level debugging with many popular +development environments. Any tool that allows remote debugging over JDWP +(the +<a href="http://java.sun.com/javase/6/docs/technotes/guides/jpda/jdwp-spec.html"> +Java Debug Wire Protocol</a>) is expected work. Supported debuggers +include jdb, Eclipse, IntelliJ, and JSwat. +</p><p> +The VM does not support tools based on JVMTI (Java Virtual +Machine Tool Interface). This is a relatively intrusive approach that +relies on bytecode insertion, something the Dalvik VM does not currently +support. +</p><p> +Dalvik's implementation of JDWP also includes hooks for supporting +DDM (Dalvik Debug Monitor) features, notably as implemented by DDMS +(Dalvik Debug Monitor Server) and the Eclipse ADT plugin. The protocol +and VM interaction is described in some detail +<a href="debugmon.html">here</a>. +</p><p> +All of the debugger support in the VM lives in the <code>dalvik/vm/jdwp</code> +directory, and is almost entirely isolated from the rest of the VM sources. +<code>dalvik/vm/Debugger.c</code> bridges the gap. The goal in doing so +was to make it easier to re-use the JDWP code in other projects. +</p><p> + + +<h2>Implementation</h2> + +<p> +Every VM that has debugging enabled starts a "JDWP" thread. The thread +typically sits idle until DDMS or a debugger connects. The thread is +only responsible for handling requests from the debugger; VM-initated +communication, such as notifying the debugger when the VM has stopped at +a breakpoint, are sent from the affected thread. +</p><p> +When the VM is embedded in the Android framework, +debugging is enabled in the VM unless the system property +<code>ro.secure</code> is set to </code>1</code>. On these +"secure" devices, debugging is only enabled in app processes whose +manifest contains <code>android:debuggable="true"</code> in the +<code><application></code> element. + +</p><p> +The VM recognizes the difference between a connection from DDMS and a +connection from a debugger (either directly or in concert with DDMS). +A connection from DDMS alone doesn't result in a change in VM behavior, +but when the VM sees debugger packets it allocates additional data +structures and may switch to a different implementation of the interpreter. +</p><p> +Because Dalvik maps bytecode into memory read-only, some common +techniques are difficult to implement without allocating additional memory. +For example, suppose the debugger sets a breakpoint in a method. The +quick way to handle this is to insert a breakpoint instruction directly +into the code. When the instruction is reached, the breakpoint handler +engages. Without this, it's necessary to perform an "is there a breakpoint +here" scan. Even with some optimizations, the debug-enabled interpreter +is much slower than the regular interpreter (perhaps 5x). +</p><p> +The JDWP protocol is stateless, so the VM handles individual debugger +requests as they arrive, and posts events to the debugger as they happen. +</p><p> + + +<h2>Debug Data</h2> +<p> Source code debug data, which includes mappings of source code to +bytecode and lists describing which registers are used to hold method +arguments and local variables, are optionally emitted by the Java compiler. +When <code>dx</code> converts Java bytecode to Dalvik bytecode, it must +also convert this debug data. +</p><p> +<code>dx</code> must also ensure that it doesn't perform operations +that confuse the debugger. For example, re-using registers that hold +method arguments and the "<code>this</code>" pointer is allowed in +Dalvik bytecode if the values are never used or no longer needed. +This can be very confusing for the debugger (and the programmer) +since the values have method scope and aren't expected to disappear. For +this reason, <code>dx</code> generates sub-optimal code in some situations +when debugging support is enabled. +</p><p> +Some of the debug data is used for other purposes; in particular, having +filename and line number data is necessary for generating useful exception +stack traces. This data can be omitted by <code>dx</code> to make the DEX +file smaller. +</p><p> + + +<h2>Usage</h2> + +<p> +The Dalvik VM supports many of the same command-line flags that other popular +desktop VMs do. To start a VM with debugging enabled, you add a command-line +flag with some basic options. The basic incantation looks something +like this: + +<pre>-Xrunjdwp:transport=dt_socket,address=8000,server=y,suspend=y</pre> +or +<pre>-agentlib:jdwp=transport=dt_socket,address=8000,server=y,suspend=y</pre> + +</p><p> +After the initial prefix, options are provided as name=value pairs. The +options currently supported by the Dalvik VM are: +<dl> + <dt>transport (no default)</dt> + <dd>Communication transport mechanism to use. Dalvik supports + TCP/IP sockets (<code>dt_socket</code>) and connection over USB + through ADB (<code>dt_android_adb</code>). + </dd> + + <dt>server (default='n')</dt> + <dd>Determines whether the VM acts as a client or a server. When + acting as a server, the VM waits for a debugger to connect to it. + When acting as a client, the VM attempts to connect to a waiting + debugger. + </dd> + + <dt>suspend (default='n')</dt> + <dd>If set to 'y', the VM will wait for a debugger connection + before executing application code. When the debugger connects (or + when the VM finishes connecting to the debugger), the VM tells the + debugger that it has suspended, and will not proceed until told + to resume. If set to 'n', the VM just plows ahead. + </dd> + + <dt>address (default="")</dt> + <dd>This must be <code>hostname:port</code> when <code>server=n</code>, + but can be just <code>port</code> when <code>server=y</code>. This + specifies the IP address and port number to connect or listen to. + <br> + Listening on port 0 has a special meaning: try to + listen on port 8000; if that fails, try 8001, 8002, and so on. (This + behavior is non-standard and may be removed from a future release.) + <br>This option has no meaning for <code>transport=dt_android_adb</code>. + </dd> + + <dt>help (no arguments)</dt> + <dd>If this is the only option, a brief usage message is displayed. + </dd> + + <dt>launch, onthrow, oncaught, timeout</dt> + <dd>These options are accepted but ignored. + </dd> +</dl> + +</p><p> +To debug a program on an Android device using DDMS over USB, you could +use a command like this: +<pre>% dalvikvm -agentlib:jdwp=transport=dt_android_adb,suspend=y,server=y -cp /data/foo.jar Foo</pre> + +This tells the Dalvik VM to run the program with debugging enabled, listening +for a connection from DDMS, and waiting for a debugger. The program will show +up with an app name of "?" in the process list, because it wasn't started +from the Android application framework. From here you would connect your +debugger to the appropriate DDMS listen port (e.g. +<code>jdb -attach localhost:8700</code> after selecting it in the app list). + +</p><p> +To debug a program on an Android device using TCP/IP bridged across ADB, +you would first need to set up forwarding: +<pre>% adb forward tcp:8000 tcp:8000 +% adb shell dalvikvm -agentlib:jdwp=transport=dt_socket,address=8000,suspend=y,server=y -cp /data/foo.jar Foo</pre> +and then <code>jdb -attach localhost:8000</code>. +</p><p> +(In the above examples, the VM will be suspended when you attach. In jdb, +type <code>cont</code> to continue.) +</p><p> +The DDMS integration makes the <code>dt_android_adb</code> transport much +more convenient when debugging on an Android device, but when working with +Dalvik on the desktop it makes sense to use the TCP/IP transport. +</p><p> + + +<h2>Known Issues and Limitations</h2> + +</p><p> +Most of the optional features JDWP allows are not implemented. These +include field access watchpoints and better tracking of monitors. +</p><p> +Not all JDWP requests are implemented. In particular, anything that +never gets emitted by the debuggers we've used is not supported and will +result in error messages being logged. Support will be added when a +use case is uncovered. +</p><p> + +</p><p> +The debugger and garbage collector are somewhat loosely +integrated at present. The VM currently guarantees that any object the +debugger is aware of will not be garbage collected until after the +debugger disconnects. This can result in a build-up over time while the +debugger is connected. +</p><p> +The situation is exacerbated by a flaw in the exception processing code, +which results in nearly all exceptions being added to the "do not discard" +list, even if the debugger never sees them. Having a debugger attached +to a program that throws lots of exceptions can result in out-of-memory +errors. This will be fixed in a future release. +</p><p> + + +<address>Copyright © 2009 The Android Open Source Project</address> +</p> + +</body> +</html> diff --git a/docs/debugmon.html b/docs/debugmon.html new file mode 100644 index 000000000..cf56ef514 --- /dev/null +++ b/docs/debugmon.html @@ -0,0 +1,736 @@ +<HTML> + + +<head> + <title>Dalvik VM Debug Monitor</title> + <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> + <link href="http://www.google.com/favicon.ico" type="image/x-icon" + rel="shortcut icon"> + <link href="../android.css" type="text/css" rel="stylesheet"> + <script language="JavaScript1.2" type="text/javascript"> +function highlight(name) { + if (document.getElementById) { + tags = [ 'span', 'div', 'tr', 'td' ]; + for (i in tags) { + elements = document.getElementsByTagName(tags[i]); + if (elements) { + for (j = 0; j < elements.length; j++) { + elementName = elements[j].getAttribute("id"); + if (elementName == name) { + elements[j].style.backgroundColor = "#C0F0C0"; + } else if (elementName && elementName.indexOf("rev") == 0) { + elements[j].style.backgroundColor = "#FFFFFF"; + } + } + } + } + } +} + </script> +</head> +<body onload="prettyPrint()"> + +<h1><a name="My_Project_"></a>Dalvik VM<br>Debug Monitor</h1> + +<!-- Status is one of: Draft, Current, Needs Update, Obsolete --> +<p style="text-align:center"><strong>Status:</strong><em>Draft</em> +<small>(as of March 6, 2007)</small></p> +<address> +[authors] +<address> + +<!-- last modified date can be different to the "Status date." It automatically +updates +whenever the file is modified. --> +<i>Modified:</i> + <!-- this script automatically sets the modified date,you don't need to modify +it --> + <script type=text/javascript> + <!-- + var lm = new Date(document.lastModified); + document.write(lm.toDateString()); + //--> + </script> +</address> + +<p><br> +<HR> + +<h2>Introduction</h2> + +<p>It's extremely useful to be able to monitor the live state of the +VM. For Android, we need to monitor multiple VMs running on a device +connected through USB or a wireless network connection. This document +describes a debug monitor server that interacts with multiple VMs, and +an API that VMs and applications can use to provide information +to the monitor. + +<p>Some things we can monitor with the Dalvik Debug Monitor ("DDM"): +<ul> + <li> Thread states. Track thread creation/exit, busy/idle status. + <li> Overall heap status, useful for a heap bitmap display or + fragmentation analysis. +</ul> + +<p>It is possible for something other than a VM to act as a DDM client, but +that is a secondary goal. Examples include "logcat" log extraction +and system monitors for virtual memory usage and load average. + +<p>It's also possible for the DDM server to be run on the device, with +the information presented through the device UI. However, the initial goal +is to provide a display tool that takes advantage of desktop tools and +screen real estate. + +<p>This work is necessary because we are unable to use standard JVMTI-based +tools with Dalvik. JVMTI relies on bytecode insertion, which is not +currently possible because Dalvik doesn't support Java bytecode. + +<p>The DDM server is written in the Java programming language +for portability. It uses a desktop +UI toolkit (SWT) for its interface. + + +<h2>Protocol</h2> + +<p>To take advantage of existing infrastructure we are piggy-backing the +DDM protocol on top of JDWP (the Java Debug Wire Protocol, normally spoken +between a VM and a debugger). To a +non-DDM client, the DDM server just looks like a debugger. + +<p>The JDWP protocol is very close to what we want to use. In particular: +<ul> + <li>It explicitly allows for vendor-defined packets, so there is no + need to "bend" the JDWP spec. + <li>Events may be posted from the VM at arbitrary points. Such + events do not elicit a response from the debugger, meaning the client + can post data and immediately resume work without worrying about the + eventual response. + <li>The basic protocol is stateless and asynchronous. Request packets + from the debugger side include a serial number, which the VM includes + in the response packet. This allows multiple simultaneous + conversations, which means the DDM traffic can be interleaved with + debugger traffic. +</ul> + +<p>There are a few issues with using JDWP for our purposes: +<ul> + <li>The VM only expects one connection from a debugger, so you couldn't + attach the monitor and a debugger at the same time. This will be + worked around by connecting the debugger to the monitor and passing the + traffic through. (We're already doing the pass-through with "jdwpspy"; + requires some management of our request IDs though.) This should + be more convenient than the current "guess the port + number" system when we're attached to a device. + <li>The VM behaves differently when a debugger is attached. It will + run more slowly, and any objects passed to the monitor or debugger are + immune to GC. We can work around this by not enabling the slow path + until non-DDM traffic is observed. We also want to have a "debugger + has connected/disconnected" message that allows the VM to release + debugger-related resources without dropping the net connection. + <li>Non-DDM VMs should not freak out when DDM connects. There are + no guarantees here for 3rd-party VMs (e.g. a certain mainstream VM, + which crashes instantly), but our older JamVM can be + configured to reject the "hello" packet. +</ul> + + +<h3>Connection Establishment</h3> + +<p>There are two basic approaches: have the server contact the VMs, and +have the VMs contact the server. The former is less "precise" than the +latter, because you have to scan for the clients, but it has some +advantages. + +<p>There are three interesting scenarios: +<ol> + <li>The DDM server is started, then the USB-attached device is booted + or the simulator is launched. + <li>The device or simulator is already running when the DDM server + is started. + <li>The DDM server is running when an already-started device is + attached to USB. +</ol> +<p>If we have the VMs connect to the DDM server on startup, we only handle +case #1. If the DDM server scans for VMs when it starts, we only handle +case #2. Neither handles case #3, which is probably the most important +of the bunch as the device matures. +<p>The plan is to have a drop-down menu with two entries, +"scan workstation" and "scan device". +The former causes the DDM server to search for VMs on "localhost", the +latter causes it to search for VMs on the other side of an ADB connection. +The DDM server will scan for VMs every few seconds, either checking a +range of known VM ports (e.g. 8000-8040) or interacting with some sort +of process database on the device. Changing modes causes all existing +connections to be dropped. +<p>When the DDM server first starts, it will try to execute "adb usb" +to ensure that the ADB server is running. (Note it will be necessary +to launch the DDM server from a shell with "adb" in the path.) If this +fails, talking to the device will still be possible so long as the ADB +daemon is already running. + +<h4>Connecting a Debugger</h4> + +<p>With the DDM server sitting on the JDWP port of all VMs, it will be +necessary to connect the debugger through the DDM server. Each VM being +debugged will have a separate port being listened to by the DDM server, +allowing you to connect a debugger to one or more VMs simultaneously. + +<p>In the common case, however, the developer will only want to debug +a single VM. One port (say 8700) will be listened to by the DDM server, +and anything connecting to it will be connected to the "current VM" +(selected in the UI). This should allow developers to focus on a +single application, which may otherwise shift around in the ordering, without +having to adjust their IDE settings to a different port every time they +restart the device. + + +<h3>Packet Format</h3> + +<p>Information is sent in chunks. Each chunk starts with: +<pre> +u4 type +u4 length +</pre> +and contains a variable amount of type-specific data. +Unrecognized types cause an empty response from the client and +are quietly ignored by the server. [Should probably return an error; +need an "error" chunk type and a handler on the server side.] + +<p>The same chunk type may have different meanings when sent in different +directions. For example, the same type may be used for both a query and +a response to the query. For sanity the type must always be used in +related transactions. + +<p>This is somewhat redundant with the JDWP framing, which includes a +4-byte length and a two-byte type code ("command set" and "command"; a +range of command set values is designated for "vendor-defined commands +and extensions"). Using the chunk format allows us to remain independent +of the underlying transport, avoids intrusive integration +with JDWP client code, and provides a way to send multiple chunks in a +single transmission unit. [I'm taking the multi-chunk packets into +account in the design, but do not plan to implement them unless the need +arises.] + +<p>Because we may be sending data over a slow USB link, the chunks may be +compressed. Compressed chunks are written as a chunk type that +indicates the compression, followed by the compressed length, followed +by the original chunk type and the uncompressed length. For zlib's deflate +algorithm, the chunk type is "ZLIB". + +<p>Following the JDWP model, packets sent from the server to the client +are always acknowledged, but packets sent from client to server never are. +The JDWP error code field is always set to "no error"; failure responses +from specific requests must be encoded into the DDM messages. + +<p>In what follows "u4" is an unsigned 32-bit value and "u1" is an +unsigned 8-bit value. Values are written in big-endian order to match +JDWP. + + +<h3>Initial Handshake</h3> + +<p>After the JDWP handshake, the server sends a HELO chunk to the client. +If the client's JDWP layer rejects it, the server assumes that the client +is not a DDM-aware VM, and does not send it any further DDM queries. +<p>On the client side, upon seeing a HELO it can know that a DDM server +is attached and prepare accordingly. The VM should not assume that a +debugger is attached until a non-DDM packet arrives. + +<h4>Chunk HELO (server --> client)</h4> +<p>Basic "hello" message. +<pre> +u4 DDM server protocol version +</pre> + + +<h4>Chunk HELO (client --> server, reply only)</h4> +Information about the client. Must be sent in response to the HELO message. +<pre> +u4 DDM client protocol version +u4 pid +u4 VM ident string len (in 16-bit units) +u4 application name len (in 16-bit units) +var VM ident string (UTF-16) +var application name (UTF-16) +</pre> + +<p>If the client does not wish to speak to the DDM server, it should respond +with a JDWP error packet. This is the same behavior you'd get from a VM +that doesn't support DDM. + + +<h3>Debugger Management</h3> +<p>VMs usually prepare for debugging when a JDWP connection is established, +and release debugger-related resources when the connection drops. We want +to open the JDWP connection early and hold it open after the debugger +disconnects. +<p>The VM can tell when a debugger attaches, because it will start seeing +non-DDM JDWP traffic, but it can't identify the disconnect. For this reason, +we need to send a packet to the client when the debugger disconnects. +<p>If the DDM server is talking to a non-DDM-aware client, it will be +necessary to drop and re-establish the connection when the debugger goes away. +(This also works with DDM-aware clients; this packet is an optimization.) + +<h4>Chunk DBGD (server --> client)</h4> +<p>Debugger has disconnected. The client responds with a DBGD to acknowledge +receipt. No data in request, no response required. + + +<h3>VM Info</h3> +<p>Update the server's info about the client. + +<h4>Chunk APNM (client --> server)</h4> + +<p>If a VM's application name changes -- possible in our environment because +of the "pre-initialized" app processes -- it must send up one of these. +<pre> +u4 application name len (in 16-bit chars) +var application name (UTF-16) +</pre> + +<h4>Chunk WAIT (client --> server)</h4> + +<p>This tells DDMS that one or more threads are waiting on an external +event. The simplest use is to tell DDMS that the VM is waiting for a +debugger to attach. +<pre> +u1 reason (0 = wait for debugger) +</pre> +If DDMS is attached, the client VM sends this up when waitForDebugger() +is called. If waitForDebugger() is called before DDMS attaches, the WAIT +chunk will be sent up at about the same time as the HELO response. + + +<h3>Thread Status</h3> + +<p>The client can send updates when their status changes, or periodically +send thread state info, e.g. 2x per +second to allow a "blinkenlights" display of thread activity. + +<h4>Chunk THEN (server --> client)</h4> + +<p>Enable thread creation/death notification. +<pre> +u1 boolean (true=enable, false=disable) +</pre> +<p>The response is empty. The client generates THCR packets for all +known threads. (Note the THCR packets may arrive before the THEN +response.) + +<h4>Chunk THCR (client --> server)</h4> +<p>Thread Creation notification. +<pre> +u4 VM-local thread ID (usually a small int) +u4 thread name len (in 16-bit chars) +var thread name (UTF-16) +</pre> + +<h4>Chunk THDE (client --> server)</h4> +<p>Thread Death notification. +<pre> +u4 VM-local thread ID +</pre> + +<h4>Chunk THST (server --> client)</h4> + +<p>Enable periodic thread activity updates. +Threads in THCR messages are assumed to be in the "initializing" state. A +THST message should follow closely on the heels of THCR. +<pre> +u4 interval, in msec +</pre> +<p>An interval of 0 disables the updates. This is done periodically, +rather than every time the thread state changes, to reduce the amount +of data that must be sent for an actively running VM. + +<h4>Chunk THST (client --> server)</h4> +<p>Thread Status, describing the state of one or more threads. This is +most useful when creation/death notifications are enabled first. The +overall layout is: +<pre> +u4 count +var thread data +</pre> +Then, for every thread: +<pre> +u4 VM-local thread ID +u1 thread state +u1 suspended +</pre> +<p>"thread state" must be one of: +<ul> <!-- don't use ol, we may need (-1) or sparse --> + <li> 1 - running (now executing or ready to do so) + <li> 2 - sleeping (in Thread.sleep()) + <li> 3 - monitor (blocked on a monitor lock) + <li> 4 - waiting (in Object.wait()) + <li> 5 - initializing + <li> 6 - starting + <li> 7 - native (executing native code) + <li> 8 - vmwait (waiting on a VM resource) +</ul> +<p>"suspended" will be 0 if the thread is running, 1 if not. +<p>[Any reason not to make "suspended" be the high bit of "thread state"? +Do we need to differentiate suspend-by-GC from suspend-by-debugger?] +<p>[We might be able to send the currently-executing method. This is a +little risky in a running VM, and increases the size of the messages +considerably, but might be handy.] + + +<h3>Heap Status</h3> + +<p>The client sends what amounts to a color-coded bitmap to the server, +indicating which stretches of memory are free and which are in use. For +compactness the bitmap is run-length encoded, and based on multi-byte +"allocation units" rather than byte counts. + +<p>In the future the server will be able to correlate the bitmap with more +detailed object data, so enough information is provided to associate the +bitmap data with virtual addresses. + +<p>Heaps may be broken into segments within the VM, and due to memory +constraints it may be desirable to send the bitmap in smaller pieces, +so the protocol allows the heap data to be sent in several chunks. +To avoid ambiguity, the client is required +to send explicit "start" and "end" messages during an update. + +<p>All messages include a "heap ID" that can be used to differentiate +between multiple independent virtual heaps or perhaps a native heap. The +client is allowed to send information about different heaps simultaneously, +so all heap-specific information is tagged with a "heap ID". + +<h4>Chunk HPIF (server --> client)</h4> +<p>Request heap info. +<pre> +u1 when to send +</pre> +<p>The "when" values are: +<pre> +0: never +1: immediately +2: at the next GC +3: at every GC +</pre> + +<h4>Chunk HPIF (client --> server, reply only)</h4> +<p>Heap Info. General information about the heap, suitable for a summary +display. +<pre> +u4 number of heaps +</pre> +For each heap: +<pre> +u4 heap ID +u8 timestamp in ms since Unix epoch +u1 capture reason (same as 'when' value from server) +u4 max heap size in bytes (-Xmx) +u4 current heap size in bytes +u4 current number of bytes allocated +u4 current number of objects allocated +</pre> +<p>[We can get some of this from HPSG, more from HPSO.] +<p>[Do we need a "heap overhead" stat here, indicating how much goes to +waste? e.g. (8 bytes per object * number of objects)] + +<h4>Chunk HPSG (server --> client)</h4> +<p>Request transmission of heap segment data. +<pre> +u1 when to send +u1 what to send +</pre> +<p>The "when" to send will be zero to disable transmission, 1 to send +during a GC. Other values are currently undefined. (Could use to pick +which part of the GC to send it, or cause periodic transmissions.) +<p>The "what" field is currently 0 for HPSG and 1 for HPSO. +<p>No reply is expected. + +<h4>Chunk NHSG (server --> client)</h4> +<p>Request transmission of native heap segment data. +<pre> +u1 when to send +u1 what to send +</pre> +<p>The "when" to send will be zero to disable transmission, 1 to send +during a GC. Other values are currently undefined. +<p>The "what" field is currently ignored. +<p>No reply is expected. + +<h4>Chunk HPST/NHST (client --> server)</h4> +<p>This is a Heap Start message. It tells the server to discard any +existing notion of what the client's heap looks like, and prepare for +new information. HPST indicates a virtual heap dump and must be followed +by zero or more HPSG/HPSO messages and an HPEN. NHST indicates a native +heap dump and must be followed by zero or more NHSG messages and an NHEN. + +<p>The only data item is: +<pre> +u4 heap ID +</pre> + +<h4>Chunk HPEN/NHEN (client --> server)</h4> +<p>Heap End, indicating that all information about the heap has been sent. +A HPST will be paired with an HPEN and an NHST will be paired with an NHEN. + +<p>The only data item is: +<pre> +u4 heap ID +</pre> + +<h4>Chunk HPSG (client --> server)</h4> +<p>Heap segment data. Each chunk describes all or part of a contiguous +stretch of heap memory. +<pre> +u4 heap ID +u1 size of allocation unit, in bytes (e.g. 8 bytes) +u4 virtual address of segment start +u4 offset of this piece (relative to the virtual address) +u4 length of piece, in allocation units +var usage data +</pre> +<p>The "usage data" indicates the status of each allocation unit. The data +is a stream of pairs of bytes, where the first byte indicates the state +of the allocation unit, and the second byte indicates the number of +consecutive allocation units with the same state. +<p>The bits in the "state" byte have the following meaning: +<pre> ++---------------------------------------+ +| 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 | ++---------------------------------------+ +| P | U0 | K2 | K1 | K0 | S2 | S1 | S0 | ++---------------------------------------+ +</pre> +<ul> + <li>'S': solidity + <ul> + <li>0=free + <li>1=has hard reference + <li>2=has soft reference + <li>3=has weak reference + <li>4=has phantom reference + <li>5=pending finalization + <li>6=marked, about to be swept + </ul> + <li>'K': kind + <ul> + <li>0=object + <li>1=class object + <li>2=array of byte/boolean + <li>3=array of char/short + <li>4=array of Object/int/float + <li>5=array of long/double + </ul> + <li>'P': partial flag (not used for HPSG) + <li>'U': unused, must be zero +</ul> + +<p>The use of the various 'S' types depends on when the information is +sent. The current plan is to send it either immediately after a GC, +or between the "mark" and "sweep" phases of the GC. For a fancy generational +collector, we may just want to send it up periodically. + +<p>The run-length byte indicates the number of allocation units minus one, so a +length of 255 means there are 256 consecutive units with this state. In +some cases, e.g. arrays of bytes, the actual size of the data is rounded +up the nearest allocation unit. +<p>For HPSG, the runs do not end at object boundaries. It is not possible +to tell from this bitmap whether a run contains one or several objects. +(But see HPSO, below.) +<p>[If we find that we have many long runs, we can overload the 'P' flag +or dedicate the 'U' flag to indicate that we have a 16-bit length instead +of 8-bit. We can also use a variable-width integer scheme for the length, +encoding 1-128 in one byte, 1-16384 in two bytes, etc.] +<p>[Alternate plan for 'K': array of byte, array of char, array of Object, +array of miscellaneous primitive type] +<p>To parse the data, the server runs through the usage data until either +(a) the end of the chunk is reached, or (b) all allocation units have been +accounted for. (If these two things don't happen at the same time, the +chunk is rejected.) +<p>Example: suppose a VM has a heap at 0x10000 that is 0x2000 bytes long +(with an 8-byte allocation unit size, that's 0x0400 units long). +The client could send one chunk (allocSize=8, virtAddr=0x10000, offset=0, +length=0x0400) or two (allocSize=8, virtAddr=0x10000, offset=0, length=0x300; +then allocSize=8, virtAddr=0x10000, offset=0x300, length=0x100). +<p>The client must encode the entire heap, including all free space at +the end, or the server will not have an accurate impression of the amount +of memory in the heap. This refers to the current heap size, not the +maximum heap size. + +<h4>Chunk HPSO (client --> server)</h4> +<p>This is essentially identical to HPSG, but the runs are terminated at +object boundaries. If an object is larger than 256 allocation units, the +"partial" flag is set in all runs except the last. +<p>The resulting unpacked bitmap is identical, but the object boundary +information can be used to gain insights into heap layout. +<p>[Do we want to have a separate message for this? Maybe just include +a "variant" flag in the HPST packet. Another possible form of output +would be one that indicates the age, in generations, of each block of +memory. That would provide a quick visual indication of "permanent vs. +transient residents", perhaps with a 16-level grey scale.] + +<h4>Chunk NHSG (client --> server)</h4> +<p>Native heap segment data. Each chunk describes all or part of a +contiguous stretch of native heap memory. The format is the same as +for HPSG, except that only solidity values 0 (= free) and 1 (= hard +reference) are used, and the kind value is always 0 for free chunks +and 7 for allocated chunks, indicating a non-VM object. +<pre> +u4 heap ID +u1 size of allocation unit, in bytes (e.g. 8 bytes) +u4 virtual address of segment start +u4 offset of this piece (relative to the virtual address) +u4 length of piece, in allocation units +var usage data +</pre> + +<h3>Generic Replies</h3> + +The client-side chunk handlers need a common way to report simple success +or failure. By convention, an empty reply packet indicates success. + +<h4>Chunk FAIL (client --> server, reply only)</h4> +<p>The chunk includes a machine-readable error code and a +human-readable error message. Server code can associate the failure +with the original request by comparing the JDWP packet ID. +<p>This allows a standard way of, for example, rejecting badly-formed +request packets. +<pre> +u4 error code +u4 error message len (in 16-bit chars) +var error message (UTF-16) +</pre> + +<h3>Miscellaneous</h3> + +<h4>Chunk EXIT (server --> client)</h4> +<p>Cause the client to exit with the specified status, using System.exit(). +Useful for certain kinds of testing. +<pre> +u4 exit status +</pre> + +<h4>Chunk DTRC (server --> client)</h4> +<p>[TBD] start/stop dmtrace; can send the results back over the wire. For +size reasons we probably need "sending", "data", "key", "finished" as +4 separate chunks/packets rather than one glob. + + +<h2>Client API</h2> + +<p>The API is written in the Java programming language +for convenience. The code is free to call native methods if appropriate. + +<h3>Chunk Handler API</h3> + +<p>The basic idea is that arbitrary code can register handlers for +specific chunk types. When a DDM chunk with that type arrives, the +appropriate handler is invoked. The handler's return value provides the +response to the server. + +<p>There are two packages. android.ddm lives in the "framework" library, +and has all of the chunk handlers and registration code. It can freely +use Android classes. org.apache.harmony.dalvik.ddmc lives in the "core" +library, and has +some base classes and features that interact with the VM. Nothing should +need to modify the org.apache.harmony.dalvik.ddmc classes. + +<p>The DDM classes pass chunks of data around with a simple class: + +<pre class=prettyprint> +class Chunk { + int type; + byte[] data; + int offset, length; +}; +</pre> + +<p>The chunk handlers accept and return them: +<pre class=prettyprint> +public Chunk handleChunk(Chunk request) +</pre> +<p>The code is free to parse the chunk and generate a response in any +way it chooses. Big-endian byte ordering is recommended but not mandatory. +<p>Chunk handlers will be notified when a DDM server connects or disconnects, +so that they can perform setup and cleanup operations: +<pre class=prettyprint> +public void connected() +public void disconnected() +</pre> + +<p>The method processes the request, formulates a response, and returns it. +If the method returns null, an empty JDWP success message will be returned. +<p>The request/response interaction is essentially asynchronous in the +protocol. The packets are linked together with the JDWP message ID. +<p>[We could use ByteBuffer here instead of byte[], but it doesn't gain +us much. Wrapping a ByteBuffer around an array is easy. We don't want +to pass the full packet in because we could have multiple chunks in one +request packet. The DDM code needs to collect and aggregate the responses +to all chunks into a single JDWP response packet. Parties wanting to +write multiple chunks in response to a single chunk should send a null +response back and use "sendChunk()" to send the data independently.] + +<h3>Unsolicited event API</h3> + +<p>If a piece of code wants to send a chunk of data to the server at some +arbitrary time, it may do so with a method provided by +org.apache.harmony.dalvik.DdmServer: + +<pre class=prettyprint> +public static void sendChunk(Chunk chunk) +</pre> + +<p>There is no response or status code. No exceptions are thrown. + + +<h2>Server API</h2> + +<p>This is similar to the client side in many ways, but makes extensive +use of ByteBuffer in a perhaps misguided attempt to use java.nio.channels +and avoid excessive thread creation and unnecessary data copying. + +<p>Upon receipt of a packet, the server will identify it as one of: +<ol> + <li>Message to be passed through to the debugger + <li>Response to an earlier request + <li>Unsolicited event packet +</ol> +<p>To handle (2), when messages are sent from the server to the client, +the message must be paired with a callback method. The response might be +delayed for a while -- or might never arrive -- so the server can't block +waiting for responses from the client. +<p>The chunk handlers look like this: +<pre class=prettyprint> +public void handleChunk(Client client, int type, + ByteBuffer data, boolean isReply, int msgId) +</pre> +<p>The arguments are: +<dl> + <dt>client + <dd>An object representing the client VM that send us the packet. + <dt>type + <dd>The 32-bit chunk type. + <dt>data + <dd>The data. The data's length can be determined by calling data.limit(). + <dt>isReply + <dd>Set to "true" if this was a reply to a message we sent earlier, + "false" if the client sent this unsolicited. + <dt>msgId + <dd>The JDWP message ID. Useful for connecting replies with requests. +</dl> +<p>If a handler doesn't like the contents of a packet, it should log an +error message and return. If the handler doesn't recognize the packet at +all, it can call the superclass' handleUnknownChunk() method. + +<p>As with the client, the server code can be notified when clients +connect or disconnect. This allows the handler to send initialization +code immediately after a connect, or clean up after a disconnect. +<p>Data associated with a client can be stored in a ClientData object, +which acts as a general per-client dumping around for VM and UI state. + + +<P><BR> + +<HR> + +<address>Copyright © 2007 The Android Open Source Project</address> + +</body> +</HTML> diff --git a/docs/dex-format.css b/docs/dex-format.css new file mode 100644 index 000000000..17e935f75 --- /dev/null +++ b/docs/dex-format.css @@ -0,0 +1,387 @@ +h1 { + font-family: serif; + border-top-style: solid; + border-top-width: 5px; + padding-top: 9pt; + margin-top: 40pt; + color: #222266; +} + +h1.title { + border: none; +} + +h2 { + font-family: serif; + border-top-style: solid; + border-top-width: 2px; + border-color: #ccccdd; + padding-top: 9pt; + margin-top: 40pt; + margin-bottom: 2pt; + color: #222266; +} + +h3 { + font-family: serif; + font-style: bold; + margin-top: 20pt; + margin-bottom: 2pt; + color: #222266; +} + +h4 { + font-family: serif; + font-style: italic; + margin-top: 2pt; + margin-bottom: 2pt; + color: #666688; +} + +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + +pre { + background: #eeeeff; + border-color: #aaaaff; + border-style: solid; + border-width: 1px; + margin-left: 40pt; + margin-right: 40pt; + padding: 6pt; +} + +table { + border-collapse: collapse; + margin-top: 10pt; + margin-left: 40pt; + margin-right: 40pt; +} + +table th { + font-family: sans-serif; + background: #aabbff; +} + +table td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + border-color: #aaaaff; + padding-top: 3pt; + padding-bottom: 3pt; + padding-left: 3pt; + padding-right: 4pt; + background: #eeeeff; +} + +table p { + margin-bottom: 0pt; +} + +/* for the bnf syntax sections */ + +table.bnf { + background: #eeeeff; + border-color: #aaaaff; + border-style: solid; + border-width: 1px; + margin-top: 3pt; + margin-bottom: 3pt; + padding-top: 2pt; + padding-bottom: 6pt; + padding-left: 6pt; + padding-right: 6pt; +} + +table.bnf td { + border: none; + padding-left: 6pt; + padding-right: 6pt; + padding-top: 1pt; + padding-bottom: 1pt; +} + +table.bnf td:first-child { + padding-right: 0pt; + width: 8pt; +} + +table.bnf td:first-child td { + padding-left: 0pt; +} + +table.bnf td.def { + padding-top: 6pt; +} + +table.bnf td.bar { + padding-left: 15pt; +} + +table.bnf code { + font-weight: bold; +} + + +/* for the type name guide */ + +table.guide { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.guide td:first-child { + font-family: monospace; + width: 15%; +} + +table.guide td:first-child + td { + font-family: sans-serif; + width: 85%; +} + + +/* for the LEB128 example tables */ + +table.leb128Bits { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.leb128Bits td { + border-left: solid #aaaaff 1px; + border-right: solid #aaaaff 1px; +} + +table.leb128Bits td.start1 { + border-left: none; +} + +table.leb128Bits td.start2 { + border-left: solid #000 2px; +} + +table.leb128Bits td.end2 { + border-right: none; +} + +table.leb128 { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.leb128 td:first-child { + font-family: monospace; + text-align: center; + width: 31%; +} + +table.leb128 td:first-child + td { + font-family: monospace; + text-align: center; + width: 23%; +} + +table.leb128 td:first-child + td + td { + font-family: monospace; + text-align: center; + width: 23%; +} + +table.leb128 td:first-child + td + td + td { + font-family: monospace; + text-align: center; + width: 23%; +} + + +/* for the general format tables */ + +table.format { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.format td:first-child { + font-family: monospace; + width: 20%; +} + +table.format td:first-child + td { + font-family: monospace; + width: 20%; +} + +table.format td:first-child + td + td { + width: 60%; +} + +table.format td i { + font-family: sans-serif; +} + + +/* for the type code table */ + +table.typeCodes { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.typeCodes td:first-child { + font-family: monospace; + width: 30%; +} + +table.typeCodes td:first-child + td { + font-family: monospace; + width: 30%; +} + +table.typeCodes td:first-child + td + td { + font-family: monospace; + width: 10%; +} + +table.typeCodes td:first-child + td + td + td { + font-family: monospace; + width: 30%; +} + +table.typeCodes td i { + font-family: sans-serif; +} + + +/* for the access flags table */ + +table.accessFlags { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.accessFlags td:first-child { + font-family: monospace; + width: 10%; +} + +table.accessFlags td:first-child + td { + font-family: monospace; + width: 6%; +} + +table.accessFlags td:first-child + td + td { + width: 28%; +} + +table.accessFlags td:first-child + td + td + td { + width: 28%; +} + +table.accessFlags td:first-child + td + td + td + td { + width: 28%; +} + +table.accessFlags i { + font-family: sans-serif; +} + + +/* for the descriptor table */ + +table.descriptor { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.descriptor td:first-child { + font-family: monospace; + width: 25%; +} + +table.descriptor td:first-child + td { + font-family: sans-serif; + width: 75%; +} + + +/* for the debug bytecode table */ + +table.debugByteCode { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.debugByteCode td:first-child { + font-family: monospace; + width: 20%; +} + +table.debugByteCode td:first-child + td { + font-family: monospace; + width: 5%; +} + +table.debugByteCode td:first-child + td + td{ + font-family: monospace; + width: 15%; +} + +table.debugByteCode td:first-child + td + td + td { + width: 25%; +} + +table.debugByteCode td:first-child + td + td + td + td { + width: 35%; +} + +table.debugByteCode i { + font-family: sans-serif; +} + + +/* for the encoded value table */ + +table.encodedValue { + margin-top: 20pt; + margin-bottom: 20pt; +} + +table.encodedValue td:first-child { + font-family: monospace; + width: 12%; +} + +table.encodedValue td:first-child + td { + font-family: monospace; + width: 10%; +} + +table.encodedValue td:first-child + td + td { + font-family: monospace; + width: 15%; +} + +table.encodedValue td:first-child + td + td + td { + font-family: monospace; + width: 15%; +} + +table.encodedValue td:first-child + td + td + td + td { + width: 48%; +} + +table.encodedValue td i { + font-family: sans-serif; +} diff --git a/docs/dex-format.html b/docs/dex-format.html new file mode 100644 index 000000000..88a7fb0c5 --- /dev/null +++ b/docs/dex-format.html @@ -0,0 +1,3043 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>.dex — Dalvik Executable Format</title> +<link rel=stylesheet href="dex-format.css"> +</head> + +<body> + +<h1 class="title"><code>.dex</code> — Dalvik Executable Format</h1> +<p>Copyright © 2007 The Android Open Source Project + +<p>This document describes the layout and contents of <code>.dex</code> +files, which are used to hold a set of class definitions and their associated +adjunct data.</p> + +<h1>Guide To Types</h1> + +<table class="guide"> +<thead> +<tr> + <th>Name</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>byte</td> + <td>8-bit signed int</td> +</tr> +<tr> + <td>ubyte</td> + <td>8-bit unsigned int</td> +</tr> +<tr> + <td>short</td> + <td>16-bit signed int, little-endian</td> +</tr> +<tr> + <td>ushort</td> + <td>16-bit unsigned int, little-endian</td> +</tr> +<tr> + <td>int</td> + <td>32-bit signed int, little-endian</td> +</tr> +<tr> + <td>uint</td> + <td>32-bit unsigned int, little-endian</td> +</tr> +<tr> + <td>long</td> + <td>64-bit signed int, little-endian</td> +</tr> +<tr> + <td>ulong</td> + <td>64-bit unsigned int, little-endian</td> +</tr> +<tr> + <td>sleb128</td> + <td>signed LEB128, variable-length (see below)</td> +</tr> +<tr> + <td>uleb128</td> + <td>unsigned LEB128, variable-length (see below)</td> +</tr> +<tr> + <td>uleb128p1</td> + <td>unsigned LEB128 plus <code>1</code>, variable-length (see below)</td> +</tr> +</tbody> +</table> + +<h3>LEB128</h3> + +<p>LEB128 ("<b>L</b>ittle-<b>E</b>ndian <b>B</b>ase <b>128</b>") is a +variable-length encoding for +arbitrary signed or unsigned integer quantities. The format was +borrowed from the <a href="http://dwarfstd.org/Dwarf3Std.php">DWARF3</a> +specification. In a <code>.dex</code> file, LEB128 is only ever used to +encode 32-bit quantities.</p> + +<p>Each LEB128 encoded value consists of one to five +bytes, which together represent a single 32-bit value. Each +byte has its most significant bit set except for the final byte in the +sequence, which has its most significant bit clear. The remaining +seven bits of each byte are payload, with the least significant seven +bits of the quantity in the first byte, the next seven in the second +byte and so on. In the case of a signed LEB128 (<code>sleb128</code>), +the most significant payload bit of the final byte in the sequence is +sign-extended to produce the final value. In the unsigned case +(<code>uleb128</code>), any bits not explicitly represented are +interpreted as <code>0</code>. + +<table class="leb128Bits"> +<thead> +<tr><th colspan="16">Bitwise diagram of a two-byte LEB128 value</th></tr> +<tr> + <th colspan="8">First byte</td> + <th colspan="8">Second byte</td> +</tr> +</thead> +<tbody> +<tr> + <td class="start1"><code>1</code></td> + <td>bit<sub>6</sub></td> + <td>bit<sub>5</sub></td> + <td>bit<sub>4</sub></td> + <td>bit<sub>3</sub></td> + <td>bit<sub>2</sub></td> + <td>bit<sub>1</sub></td> + <td>bit<sub>0</sub></td> + <td class="start2"><code>0</code></td> + <td>bit<sub>13</sub></td> + <td>bit<sub>12</sub></td> + <td>bit<sub>11</sub></td> + <td>bit<sub>10</sub></td> + <td>bit<sub>9</sub></td> + <td>bit<sub>8</sub></td> + <td class="end2">bit<sub>7</sub></td> +</tr> +</tbody> +</table> + +<p>The variant <code>uleb128p1</code> is used to represent a signed +value, where the representation is of the value <i>plus one</i> encoded +as a <code>uleb128</code>. This makes the encoding of <code>-1</code> +(alternatively thought of as the unsigned value <code>0xffffffff</code>) +— but no other negative number — a single byte, and is +useful in exactly those cases where the represented number must either +be non-negative or <code>-1</code> (or <code>0xffffffff</code>), +and where no other negative values are allowed (or where large unsigned +values are unlikely to be needed).</p> + +<p>Here are some examples of the formats:</p> + +<table class="leb128"> +<thead> +<tr> + <th>Encoded Sequence</th> + <th>As <code>sleb128</code></th> + <th>As <code>uleb128</code></th> + <th>As <code>uleb128p1</code></th> +</tr> +</thead> +<tbody> + <tr><td>00</td><td>0</td><td>0</td><td>-1</td></tr> + <tr><td>01</td><td>1</td><td>1</td><td>0</td></tr> + <tr><td>7f</td><td>-1</td><td>127</td><td>126</td></tr> + <tr><td>80 7f</td><td>-128</td><td>16256</td><td>16255</td></tr> +</tbody> +</table> + +<h1>Overall File Layout</h1> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>header</td> + <td>header_item</td> + <td>the header</td> +</tr> +<tr> + <td>string_ids</td> + <td>string_id_item[]</td> + <td>string identifiers list. These are identifiers for all the strings + used by this file, either for internal naming (e.g., type descriptors) + or as constant objects referred to by code. This list must be sorted + by string contents, using UTF-16 code point values (not in a + locale-sensitive manner). + </td> +</tr> +<tr> + <td>type_ids</td> + <td>type_id_item[]</td> + <td>type identifiers list. These are identifiers for all types (classes, + arrays, or primitive types) referred to by this file, whether defined + in the file or not. This list must be sorted by <code>string_id</code> + index. + </td> +</tr> +<tr> + <td>proto_ids</td> + <td>proto_id_item[]</td> + <td>method prototype identifiers list. These are identifiers for all + prototypes referred to by this file. This list must be sorted in + return-type (by <code>type_id</code> index) major order, and then + by arguments (also by <code>type_id</code> index). + </td> +</tr> +<tr> + <td>field_ids</td> + <td>field_id_item[]</td> + <td>field identifiers list. These are identifiers for all fields + referred to by this file, whether defined in the file or not. This + list must be sorted, where the defining type (by <code>type_id</code> + index) is the major order, field name (by <code>string_id</code> index) + is the intermediate order, and type (by <code>type_id</code> index) + is the minor order. + </td> +</tr> +<tr> + <td>method_ids</td> + <td>method_id_item[]</td> + <td>method identifiers list. These are identifiers for all methods + referred to by this file, whether defined in the file or not. This + list must be sorted, where the defining type (by <code>type_id</code> + index) is the major order, method name (by <code>string_id</code> + index) is the intermediate order, and method + prototype (by <code>proto_id</code> index) is the minor order. + </td> +</tr> +<tr> + <td>class_defs</td> + <td>class_def_item[]</td> + <td>class definitions list. The classes must be ordered such that a given + class's superclass and implemented interfaces appear in the + list earlier than the referring class. + </td> +</tr> +<tr> + <td>data</td> + <td>ubyte[]</td> + <td>data area, containing all the support data for the tables listed above. + Different items have different alignment requirements, and + padding bytes are inserted before each item if necessary to achieve + proper alignment. + </td> +</tr> +<tr> + <td>link_data</td> + <td>ubyte[]</td> + <td>data used in statically linked files. The format of the data in + this section is left unspecified by this document; + this section is empty in unlinked files, and runtime implementations + may use it as they see fit. + </td> +</tr> +</tbody> +</table> + +<h1>Bitfield, String, and Constant Definitions</h1> + +<h2><code>DEX_FILE_MAGIC</code></h2> +<h4>embedded in <code>header_item</code></h4> + +<p>The constant array/string <code>DEX_FILE_MAGIC</code> is the list of +bytes that must appear at the beginning of a <code>.dex</code> file +in order for it to be recognized as such. The value intentionally +contains a newline (<code>"\n"</code> or <code>0x0a</code>) and a +null byte (<code>"\0"</code> or <code>0x00</code>) in order to help +in the detection of certain forms of corruption. The value also +encodes a format version number as three decimal digits, which is +expected to increase monotonically over time as the format evolves.</p> + +<pre> +ubyte[8] DEX_FILE_MAGIC = { 0x64 0x65 0x78 0x0a 0x30 0x33 0x35 0x00 } + = "dex\n035\0" +</pre> + +<p><b>Note:</b> At least a couple earlier versions of the format have +been used in widely-available public software releases. For example, +version <code>009</code> was used for the M3 releases of the +Android platform (November-December 2007), +and version <code>013</code> was used for the M5 releases of the Android +platform (February-March 2008). In several respects, these earlier versions +of the format differ significantly from the version described in this +document.</p> + +<h2><code>ENDIAN_CONSTANT</code> and <code>REVERSE_ENDIAN_CONSTANT</code></h2> +<h4>embedded in <code>header_item</code></h4> + +<p>The constant <code>ENDIAN_CONSTANT</code> is used to indicate the +endianness of the file in which it is found. Although the standard +<code>.dex</code> format is little-endian, implementations may choose +to perform byte-swapping. Should an implementation come across a +header whose <code>endian_tag</code> is <code>REVERSE_ENDIAN_CONSTANT</code> +instead of <code>ENDIAN_CONSTANT</code>, it would know that the file +has been byte-swapped from the expected form.</p> + +<pre> +uint ENDIAN_CONSTANT = 0x12345678; +uint REVERSE_ENDIAN_CONSTANT = 0x78563412; +</pre> + +<h2><code>NO_INDEX</code></h2> +<h4>embedded in <code>class_def_item</code> and +<code>debug_info_item</code></h4> + +<p>The constant <code>NO_INDEX</code> is used to indicate that +an index value is absent.</p> + +<p><b>Note:</b> This value isn't defined to be +<code>0</code>, because that is in fact typically a valid index.</p> + +<p><b>Also Note:</b> The chosen value for <code>NO_INDEX</code> is +representable as a single byte in the <code>uleb128p1</code> encoding.</p> + +<pre> +uint NO_INDEX = 0xffffffff; // == -1 if treated as a signed int +</pre> + +<h2><code>access_flags</code> Definitions</h2> +<h4>embedded in <code>class_def_item</code>, +<code>field_item</code>, <code>method_item</code>, and +<code>InnerClass</code></h4> + +<p>Bitfields of these flags are used to indicate the accessibility and +overall properties of classes and class members.</p> + +<table class="accessFlags"> +<thead> +<tr> + <th>Name</th> + <th>Value</th> + <th>For Classes (and <code>InnerClass</code> annotations)</th> + <th>For Fields</th> + <th>For Methods</th> +</tr> +</thead> +<tbody> +<tr> + <td>ACC_PUBLIC</td> + <td>0x1</td> + <td><code>public</code>: visible everywhere</td> + <td><code>public</code>: visible everywhere</td> + <td><code>public</code>: visible everywhere</td> +</tr> +<tr> + <td>ACC_PRIVATE</td> + <td>0x2</td> + <td><super>*</super> + <code>private</code>: only visible to defining class + </td> + <td><code>private</code>: only visible to defining class</td> + <td><code>private</code>: only visible to defining class</td> +</tr> +<tr> + <td>ACC_PROTECTED</td> + <td>0x4</td> + <td><super>*</super> + <code>protected</code>: visible to package and subclasses + </td> + <td><code>protected</code>: visible to package and subclasses</td> + <td><code>protected</code>: visible to package and subclasses</td> +</tr> +<tr> + <td>ACC_STATIC</td> + <td>0x8</td> + <td><super>*</super> + <code>static</code>: is not constructed with an outer + <code>this</code> reference</td> + <td><code>static</code>: global to defining class</td> + <td><code>static</code>: does not take a <code>this</code> argument</td> +</tr> +<tr> + <td>ACC_FINAL</td> + <td>0x10</td> + <td><code>final</code>: not subclassable</td> + <td><code>final</code>: immutable after construction</td> + <td><code>final</code>: not overridable</td> +</tr> +<tr> + <td>ACC_SYNCHRONIZED</td> + <td>0x20</td> + <td> </td> + <td> </td> + <td><code>synchronized</code>: associated lock automatically acquired + around call to this method. <b>Note:</b> This is only valid to set when + <code>ACC_NATIVE</code> is also set.</td> +</tr> +<tr> + <td>ACC_VOLATILE</td> + <td>0x40</td> + <td> </td> + <td><code>volatile</code>: special access rules to help with thread + safety</td> + <td> </td> +</tr> +<tr> + <td>ACC_BRIDGE</td> + <td>0x40</td> + <td> </td> + <td> </td> + <td>bridge method, added automatically by compiler as a type-safe + bridge</td> +</tr> +<tr> + <td>ACC_TRANSIENT</td> + <td>0x80</td> + <td> </td> + <td><code>transient</code>: not to be saved by default serialization</td> + <td> </td> +</tr> +<tr> + <td>ACC_VARARGS</td> + <td>0x80</td> + <td> </td> + <td> </td> + <td>last argument should be treated as a "rest" argument by compiler</td> +</tr> +<tr> + <td>ACC_NATIVE</td> + <td>0x100</td> + <td> </td> + <td> </td> + <td><code>native</code>: implemented in native code</td> +</tr> +<tr> + <td>ACC_INTERFACE</td> + <td>0x200</td> + <td><code>interface</code>: multiply-implementable abstract class</td> + <td> </td> + <td> </td> +</tr> +<tr> + <td>ACC_ABSTRACT</td> + <td>0x400</td> + <td><code>abstract</code>: not directly instantiable</td> + <td> </td> + <td><code>abstract</code>: unimplemented by this class</td> +</tr> +<tr> + <td>ACC_STRICT</td> + <td>0x800</td> + <td> </td> + <td> </td> + <td><code>strictfp</code>: strict rules for floating-point arithmetic</td> +</tr> +<tr> + <td>ACC_SYNTHETIC</td> + <td>0x1000</td> + <td>not directly defined in source code</td> + <td>not directly defined in source code</td> + <td>not directly defined in source code</td> +</tr> +<tr> + <td>ACC_ANNOTATION</td> + <td>0x2000</td> + <td>declared as an annotation class</td> + <td> </td> + <td> </td> +</tr> +<tr> + <td>ACC_ENUM</td> + <td>0x4000</td> + <td>declared as an enumerated type</td> + <td>declared as an enumerated value</td> + <td> </td> +</tr> +<tr> + <td><i>(unused)</i></td> + <td>0x8000</td> + <td> </td> + <td> </td> + <td> </td> +</tr> +<tr> + <td>ACC_CONSTRUCTOR</td> + <td>0x10000</td> + <td> </td> + <td> </td> + <td>constructor method (class or instance initializer)</td> +</tr> +<tr> + <td>ACC_DECLARED_<br/>SYNCHRONIZED</td> + <td>0x20000</td> + <td> </td> + <td> </td> + <td>declared <code>synchronized</code>. <b>Note:</b> This has no effect on + execution (other than in reflection of this flag, per se). + </td> +</tr> +</tbody> +</table> + +<p><super>*</super> Only allowed on for <code>InnerClass</code> annotations, +and must not ever be on in a <code>class_def_item</code>.</p> + +<h2>MUTF-8 (Modified UTF-8) Encoding</h2> + +<p>As a concession to easier legacy support, the <code>.dex</code> format +encodes its string data in a de facto standard modified UTF-8 form, hereafter +referred to as MUTF-8. This form is identical to standard UTF-8, except:</p> + +<ul> + <li>Only the one-, two-, and three-byte encodings are used.</li> + <li>Code points in the range <code>U+10000</code> … + <code>U+10ffff</code> are encoded as a surrogate pair, each of + which is represented as a three-byte encoded value.</li> + <li>The code point <code>U+0000</code> is encoded in two-byte form.</li> + <li>A plain null byte (value <code>0</code>) indicates the end of + a string, as is the standard C language interpretation.</li> +</ul> + +<p>The first two items above can be summarized as: MUTF-8 +is an encoding format for UTF-16, instead of being a more direct +encoding format for Unicode characters.</p> + +<p>The final two items above make it simultaneously possible to include +the code point <code>U+0000</code> in a string <i>and</i> still manipulate +it as a C-style null-terminated string.</p> + +<p>However, the special encoding of <code>U+0000</code> means that, unlike +normal UTF-8, the result of calling the standard C function +<code>strcmp()</code> on a pair of MUTF-8 strings does not always +indicate the properly signed result of comparison of <i>unequal</i> strings. +When ordering (not just equality) is a concern, the most straightforward +way to compare MUTF-8 strings is to decode them character by character, +and compare the decoded values. (However, more clever implementations are +also possible.)</p> + +<p>Please refer to <a href="http://unicode.org">The Unicode +Standard</a> for further information about character encoding. +MUTF-8 is actually closer to the (relatively less well-known) encoding +<a href="http://www.unicode.org/reports/tr26/">CESU-8</a> than to UTF-8 +per se.</p> + +<h2><code>encoded_value</code> Encoding</h2> +<h4>embedded in <code>annotation_element</code> and +<code>encoded_array_item</code></h4> + +<p>An <code>encoded_value</code> is an encoded piece of (nearly) +arbitrary hierarchically structured data. The encoding is meant to +be both compact and straightforward to parse.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>(value_arg << 5) | value_type</td> + <td>ubyte</td> + <td>byte indicating the type of the immediately subsequent + <code>value</code> along + with an optional clarifying argument in the high-order three bits. + See below for the various <code>value</code> definitions. + In most cases, <code>value_arg</code> encodes the length of + the immediately-subsequent <code>value</code> in bytes, as + <code>(size - 1)</code>, e.g., <code>0</code> means that + the value requires one byte, and <code>7</code> means it requires + eight bytes; however, there are exceptions as noted below. + </td> +</tr> +<tr> + <td>value</td> + <td>ubyte[]</td> + <td>bytes representing the value, variable in length and interpreted + differently for different <code>value_type</code> bytes, though + always little-endian. See the various value definitions below for + details. + </td> +</tr> +</tbody> +</table> + +<h3>Value Formats</h3> + +<table class="encodedValue"> +<thead> +<tr> + <th>Type Name</th> + <th><code>value_type</code></th> + <th><code>value_arg</code> Format</th> + <th><code>value</code> Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>VALUE_BYTE</td> + <td>0x00</td> + <td><i>(none; must be <code>0</code>)</i></td> + <td>ubyte[1]</td> + <td>signed one-byte integer value</td> +</tr> +<tr> + <td>VALUE_SHORT</td> + <td>0x02</td> + <td>size - 1 (0…1)</td> + <td>ubyte[size]</td> + <td>signed two-byte integer value, sign-extended</td> +</tr> +<tr> + <td>VALUE_CHAR</td> + <td>0x03</td> + <td>size - 1 (0…1)</td> + <td>ubyte[size]</td> + <td>unsigned two-byte integer value, zero-extended</td> +</tr> +<tr> + <td>VALUE_INT</td> + <td>0x04</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>signed four-byte integer value, sign-extended</td> +</tr> +<tr> + <td>VALUE_LONG</td> + <td>0x06</td> + <td>size - 1 (0…7)</td> + <td>ubyte[size]</td> + <td>signed eight-byte integer value, sign-extended</td> +</tr> +<tr> + <td>VALUE_FLOAT</td> + <td>0x10</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>four-byte bit pattern, zero-extended <i>to the right</i>, and + interpreted as an IEEE754 32-bit floating point value + </td> +</tr> +<tr> + <td>VALUE_DOUBLE</td> + <td>0x11</td> + <td>size - 1 (0…7)</td> + <td>ubyte[size]</td> + <td>eight-byte bit pattern, zero-extended <i>to the right</i>, and + interpreted as an IEEE754 64-bit floating point value + </td> +</tr> +<tr> + <td>VALUE_STRING</td> + <td>0x17</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>unsigned (zero-extended) four-byte integer value, + interpreted as an index into + the <code>string_ids</code> section and representing a string value + </td> +</tr> +<tr> + <td>VALUE_TYPE</td> + <td>0x18</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>unsigned (zero-extended) four-byte integer value, + interpreted as an index into + the <code>type_ids</code> section and representing a reflective + type/class value + </td> +</tr> +<tr> + <td>VALUE_FIELD</td> + <td>0x19</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>unsigned (zero-extended) four-byte integer value, + interpreted as an index into + the <code>field_ids</code> section and representing a reflective + field value + </td> +</tr> +<tr> + <td>VALUE_METHOD</td> + <td>0x1a</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>unsigned (zero-extended) four-byte integer value, + interpreted as an index into + the <code>method_ids</code> section and representing a reflective + method value + </td> +</tr> +<tr> + <td>VALUE_ENUM</td> + <td>0x1b</td> + <td>size - 1 (0…3)</td> + <td>ubyte[size]</td> + <td>unsigned (zero-extended) four-byte integer value, + interpreted as an index into + the <code>field_ids</code> section and representing the value of + an enumerated type constant + </td> +</tr> +<tr> + <td>VALUE_ARRAY</td> + <td>0x1c</td> + <td><i>(none; must be <code>0</code>)</i></td> + <td>encoded_array</td> + <td>an array of values, in the format specified by + "<code>encoded_array</code> Format" below. The size + of the <code>value</code> is implicit in the encoding. + </td> +</tr> +<tr> + <td>VALUE_ANNOTATION</td> + <td>0x1d</td> + <td><i>(none; must be <code>0</code>)</i></td> + <td>encoded_annotation</td> + <td>a sub-annotation, in the format specified by + "<code>encoded_annotation</code> Format" below. The size + of the <code>value</code> is implicit in the encoding. + </td> +</tr> +<tr> + <td>VALUE_NULL</td> + <td>0x1e</td> + <td><i>(none; must be <code>0</code>)</i></td> + <td><i>(none)</i></td> + <td><code>null</code> reference value</td> +</tr> +<tr> + <td>VALUE_BOOLEAN</td> + <td>0x1f</td> + <td>boolean (0…1)</td> + <td><i>(none)</i></td> + <td>one-bit value; <code>0</code> for <code>false</code> and + <code>1</code> for <code>true</code>. The bit is represented in the + <code>value_arg</code>. + </td> +</tr> +</tbody> +</table> + +<h3><code>encoded_array</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>uleb128</td> + <td>number of elements in the array</td> +</tr> +<tr> + <td>values</td> + <td>encoded_value[size]</td> + <td>a series of <code>size</code> <code>encoded_value</code> byte + sequences in the format specified by this section, concatenated + sequentially. + </td> +</tr> +</tbody> +</table> + +<h3><code>encoded_annotation</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>type_idx</td> + <td>uleb128</td> + <td>type of the annotation. This must be a class (not array or primitive) + type. + </td> +</tr> +<tr> + <td>size</td> + <td>uleb128</td> + <td>number of name-value mappings in this annotation</td> +</tr> +<tr> + <td>elements</td> + <td>annotation_element[size]</td> + <td>elements of the annotataion, represented directly in-line (not as + offsets). Elements must be sorted in increasing order by + <code>string_id</code> index. + </td> +</tr> +</tbody> +</table> + +<h3><code>annotation_element</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>name_idx</td> + <td>uleb128</td> + <td>element name, represented as an index into the + <code>string_ids</code> section. The string must conform to the + syntax for <i>MemberName</i>, defined above. + </td> +</tr> +<tr> + <td>value</td> + <td>encoded_value</td> + <td>element value</td> +</tr> +</tbody> +</table> + +<h2>String Syntax</h2> + +<p>There are several kinds of item in a <code>.dex</code> file which +ultimately refer to a string. The following BNF-style definitions +indicate the acceptable syntax for these strings.</p> + +<h3><i>SimpleName</i></h3> + +<p>A <i>SimpleName</i> is the basis for the syntax of the names of other +things. The <code>.dex</code> format allows a fair amount of latitude +here (much more than most common source languages). In brief, a simple +name may consist of any low-ASCII alphabetic character or digit, a few +specific low-ASCII symbols, and most non-ASCII code points that are not +control, space, or special characters. Note that surrogate code points +(in the range <code>U+d800</code> … <code>U+dfff</code>) are not +considered valid name characters, per se, but Unicode supplemental +characters <i>are</i> valid (which are represented by the final +alternative of the rule for <i>SimpleNameChar</i>), and they should be +represented in a file as pairs of surrogate code points in the MUTF-8 +encoding.</p> + +<table class="bnf"> + <tr><td colspan="2" class="def"><i>SimpleName</i> →</td></tr> + <tr> + <td/> + <td><i>SimpleNameChar</i> (<i>SimpleNameChar</i>)*</td> + </tr> + + <tr><td colspan="2" class="def"><i>SimpleNameChar</i> →</td></tr> + <tr> + <td/> + <td><code>'A'</code> … <code>'Z'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'a'</code> … <code>'z'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'0'</code> … <code>'9'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'$'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'-'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'_'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>U+00a1</code> … <code>U+1fff</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>U+2010</code> … <code>U+2027</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>U+2030</code> … <code>U+d7ff</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>U+e000</code> … <code>U+ffef</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>U+10000</code> … <code>U+10ffff</code></td> + </tr> +</table> + +<h3><i>MemberName</i></h3> +<h4>used by <code>field_id_item</code> and <code>method_id_item</code></h4> + +<p>A <i>MemberName</i> is the name of a member of a class, members being +fields, methods, and inner classes.</p> + +<table class="bnf"> + <tr><td colspan="2" class="def"><i>MemberName</i> →</td></tr> + <tr> + <td/> + <td><i>SimpleName</i></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'<'</code> <i>SimpleName</i> <code>'>'</code></td> + </tr> +</table> + +<h3><i>FullClassName</i></h3> + +<p>A <i>FullClassName</i> is a fully-qualified class name, including an +optional package specifier followed by a required name.</p> + +<table class="bnf"> + <tr><td colspan="2" class="def"><i>FullClassName</i> →</td></tr> + <tr> + <td/> + <td><i>OptionalPackagePrefix</i> <i>SimpleName</i></td> + </tr> + + <tr><td colspan="2" class="def"><i>OptionalPackagePrefix</i> →</td></tr> + <tr> + <td/> + <td>(<i>SimpleName</i> <code>'/'</code>)*</td> + </tr> +</table> + +<h3><i>TypeDescriptor</i></h3> +<h4>used by <code>type_id_item</code></h4> + +<p>A <i>TypeDescriptor</i> is the representation of any type, including +primitives, classes, arrays, and <code>void</code>. See below for +the meaning of the various versions.</p> + +<table class="bnf"> + <tr><td colspan="2" class="def"><i>TypeDescriptor</i> →</td></tr> + <tr> + <td/> + <td><code>'V'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><i>FieldTypeDescriptor</i></td> + </tr> + + <tr><td colspan="2" class="def"><i>FieldTypeDescriptor</i> →</td></tr> + <tr> + <td/> + <td><i>NonArrayFieldTypeDescriptor</i></td> + </tr> + <tr> + <td class="bar">|</td> + <td>(<code>'['</code> * 1…255) + <i>NonArrayFieldTypeDescriptor</i></td> + </tr> + + <tr> + <td colspan="2" class="def"><i>NonArrayFieldTypeDescriptor</i>→</td> + </tr> + <tr> + <td/> + <td><code>'Z'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'B'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'S'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'C'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'I'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'J'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'F'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'D'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'L'</code> <i>FullClassName</i> <code>';'</code></td> + </tr> +</table> + +<h3><i>ShortyDescriptor</i></h3> +<h4>used by <code>proto_id_item</code></h4> + +<p>A <i>ShortyDescriptor</i> is the short form representation of a method +prototype, including return and parameter types, except that there is +no distinction between various reference (class or array) types. Instead, +all reference types are represented by a single <code>'L'</code> character.</p> + +<table class="bnf"> + <tr><td colspan="2" class="def"><i>ShortyDescriptor</i> →</td></tr> + <tr> + <td/> + <td><i>ShortyReturnType</i> (<i>ShortyFieldType</i>)*</td> + </tr> + + <tr><td colspan="2" class="def"><i>ShortyReturnType</i> →</td></tr> + <tr> + <td/> + <td><code>'V'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><i>ShortyFieldType</i></td> + </tr> + + <tr><td colspan="2" class="def"><i>ShortyFieldType</i> →</td></tr> + <tr> + <td/> + <td><code>'Z'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'B'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'S'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'C'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'I'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'J'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'F'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'D'</code></td> + </tr> + <tr> + <td class="bar">|</td> + <td><code>'L'</code></td> + </tr> +</table> + +<h2><i>TypeDescriptor</i> Semantics</h2> + +<p>This is the meaning of each of the variants of <i>TypeDescriptor</i>.</p> + +<table class="descriptor"> +<thead> +<tr> + <th>Syntax</th> + <th>Meaning</th> +</tr> +</thead> +<tbody> +<tr> + <td>V</td> + <td><code>void</code>; only valid for return types</td> +</tr> +<tr> + <td>Z</td> + <td><code>boolean</code></td> +</tr> +<tr> + <td>B</td> + <td><code>byte</code></td> +</tr> +<tr> + <td>S</td> + <td><code>short</code></td> +</tr> +<tr> + <td>C</td> + <td><code>char</code></td> +</tr> +<tr> + <td>I</td> + <td><code>int</code></td> +</tr> +<tr> + <td>J</td> + <td><code>long</code></td> +</tr> +<tr> + <td>F</td> + <td><code>float</code></td> +</tr> +<tr> + <td>D</td> + <td><code>double</code></td> +</tr> +<tr> + <td>L<i>fully/qualified/Name</i>;</td> + <td>the class <code><i>fully.qualified.Name</i></code></td> +</tr> +<tr> + <td>[<i>descriptor</i></td> + <td>array of <code><i>descriptor</i></code>, usable recursively for + arrays-of-arrays, though it is invalid to have more than 255 + dimensions. + </td> +</tr> +</tbody> +</table> + +<h1>Items and Related Structures</h1> + +<p>This section includes definitions for each of the top-level items that +may appear in a <code>.dex</code> file. + +<h2><code>header_item</code></h2> +<h4>appears in the <code>header</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>magic</td> + <td>ubyte[8] = DEX_FILE_MAGIC</td> + <td>magic value. See discussion above under "<code>DEX_FILE_MAGIC</code>" + for more details. + </td> +</tr> +<tr> + <td>checksum</td> + <td>uint</td> + <td>adler32 checksum of the rest of the file (everything but + <code>magic</code> and this field); used to detect file corruption + </td> +</tr> +<tr> + <td>signature</td> + <td>ubyte[20]</td> + <td>SHA-1 signature (hash) of the rest of the file (everything but + <code>magic</code>, <code>checksum</code>, and this field); used + to uniquely identify files + </td> +</tr> +<tr> + <td>file_size</td> + <td>uint</td> + <td>size of the entire file (including the header), in bytes +</tr> +<tr> + <td>header_size</td> + <td>uint = 0x70</td> + <td>size of the header (this entire section), in bytes. This allows for at + least a limited amount of backwards/forwards compatibility without + invalidating the format. + </td> +</tr> +<tr> + <td>endian_tag</td> + <td>uint = ENDIAN_CONSTANT</td> + <td>endianness tag. See discussion above under "<code>ENDIAN_CONSTANT</code> + and <code>REVERSE_ENDIAN_CONSTANT</code>" for more details. + </td> +</tr> +<tr> + <td>link_size</td> + <td>uint</td> + <td>size of the link section, or <code>0</code> if this file isn't + statically linked</td> +</tr> +<tr> + <td>link_off</td> + <td>uint</td> + <td>offset from the start of the file to the link section, or + <code>0</code> if <code>link_size == 0</code>. The offset, if non-zero, + should be to an offset into the <code>link_data</code> section. The + format of the data pointed at is left unspecified by this document; + this header field (and the previous) are left as hooks for use by + runtime implementations. + </td> +</tr> +<tr> + <td>map_off</td> + <td>uint</td> + <td>offset from the start of the file to the map item, or + <code>0</code> if this file has no map. The offset, if non-zero, + should be to an offset into the <code>data</code> section, + and the data should be in the format specified by "<code>map_list</code>" + below. + </td> +</tr> +<tr> + <td>string_ids_size</td> + <td>uint</td> + <td>count of strings in the string identifiers list</td> +</tr> +<tr> + <td>string_ids_off</td> + <td>uint</td> + <td>offset from the start of the file to the string identifiers list, or + <code>0</code> if <code>string_ids_size == 0</code> (admittedly a + strange edge case). The offset, if non-zero, + should be to the start of the <code>string_ids</code> section. + </td> +</tr> +<tr> + <td>type_ids_size</td> + <td>uint</td> + <td>count of elements in the type identifiers list</td> +</tr> +<tr> + <td>type_ids_off</td> + <td>uint</td> + <td>offset from the start of the file to the type identifiers list, or + <code>0</code> if <code>type_ids_size == 0</code> (admittedly a + strange edge case). The offset, if non-zero, + should be to the start of the <code>type_ids</code> + section. + </td> +</tr> +<tr> + <td>proto_ids_size</td> + <td>uint</td> + <td>count of elements in the prototype identifiers list</td> +</tr> +<tr> + <td>proto_ids_off</td> + <td>uint</td> + <td>offset from the start of the file to the prototype identifiers list, or + <code>0</code> if <code>proto_ids_size == 0</code> (admittedly a + strange edge case). The offset, if non-zero, + should be to the start of the <code>proto_ids</code> + section. + </td> +</tr> +<tr> + <td>field_ids_size</td> + <td>uint</td> + <td>count of elements in the field identifiers list</td> +</tr> +<tr> + <td>field_ids_off</td> + <td>uint</td> + <td>offset from the start of the file to the field identifiers list, or + <code>0</code> if <code>field_ids_size == 0</code>. The offset, if + non-zero, should be to the start of the <code>field_ids</code> + section.</td> +</td> +</tr> +<tr> + <td>method_ids_size</td> + <td>uint</td> + <td>count of elements in the method identifiers list</td> +</tr> +<tr> + <td>method_ids_off</td> + <td>uint</td> + <td>offset from the start of the file to the method identifiers list, or + <code>0</code> if <code>method_ids_size == 0</code>. The offset, if + non-zero, should be to the start of the <code>method_ids</code> + section.</td> +</tr> +<tr> + <td>class_defs_size</td> + <td>uint</td> + <td>count of elements in the class definitions list</td> +</tr> +<tr> + <td>class_defs_off</td> + <td>uint</td> + <td>offset from the start of the file to the class definitions list, or + <code>0</code> if <code>class_defs_size == 0</code> (admittedly a + strange edge case). The offset, if non-zero, + should be to the start of the <code>class_defs</code> section. + </td> +</tr> +<tr> + <td>data_size</td> + <td>uint</td> + <td>Size of <code>data</code> section in bytes. Must be an even + multiple of sizeof(uint).</td> +</tr> +<tr> + <td>data_off</td> + <td>uint</td> + <td>offset from the start of the file to the start of the + <code>data</code> section. + </td> +</tr> +</tbody> +</table> + +<h2><code>map_list</code></h2> +<h4>appears in the <code>data</code> section</h4> +<h4>referenced from <code>header_item</code></h4> +<h4>alignment: 4 bytes</h4> + +<p>This is a list of the entire contents of a file, in order. It +contains some redundancy with respect to the <code>header_item</code> +but is intended to be an easy form to use to iterate over an entire +file. A given type may appear at most once in a map, but there is no +restriction on what order types may appear in, other than the +restrictions implied by the rest of the format (e.g., a +<code>header</code> section must appear first, followed by a +<code>string_ids</code> section, etc.). Additionally, the map entries must +be ordered by initial offset and must not overlap.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>uint</td> + <td>size of the list, in entries</td> +</tr> +<tr> + <td>list</td> + <td>map_item[size]</td> + <td>elements of the list</td> +</tr> +</tbody> +</table> + +<h3><code>map_item</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>type</td> + <td>ushort</td> + <td>type of the items; see table below</td> +</tr> +<tr> + <td>unused</td> + <td>ushort</td> + <td><i>(unused)</i></td> +</tr> +<tr> + <td>size</td> + <td>uint</td> + <td>count of the number of items to be found at the indicated offset</td> +</tr> +<tr> + <td>offset</td> + <td>uint</td> + <td>offset from the start of the file to the items in question</td> +</tr> +</tbody> +</table> + + +<h3>Type Codes</h3> + +<table class="typeCodes"> +<thead> +<tr> + <th>Item Type</th> + <th>Constant</th> + <th>Value</th> + <th>Item Size In Bytes</th> +</tr> +</thead> +<tbody> +<tr> + <td>header_item</td> + <td>TYPE_HEADER_ITEM</td> + <td>0x0000</td> + <td>0x70</td> +</tr> +<tr> + <td>string_id_item</td> + <td>TYPE_STRING_ID_ITEM</td> + <td>0x0001</td> + <td>0x04</td> +</tr> +<tr> + <td>type_id_item</td> + <td>TYPE_TYPE_ID_ITEM</td> + <td>0x0002</td> + <td>0x04</td> +</tr> +<tr> + <td>proto_id_item</td> + <td>TYPE_PROTO_ID_ITEM</td> + <td>0x0003</td> + <td>0x0c</td> +</tr> +<tr> + <td>field_id_item</td> + <td>TYPE_FIELD_ID_ITEM</td> + <td>0x0004</td> + <td>0x08</td> +</tr> +<tr> + <td>method_id_item</td> + <td>TYPE_METHOD_ID_ITEM</td> + <td>0x0005</td> + <td>0x08</td> +</tr> +<tr> + <td>class_def_item</td> + <td>TYPE_CLASS_DEF_ITEM</td> + <td>0x0006</td> + <td>0x20</td> +</tr> +<tr> + <td>map_list</td> + <td>TYPE_MAP_LIST</td> + <td>0x1000</td> + <td>4 + (item.size * 12)</td> +</tr> +<tr> + <td>type_list</td> + <td>TYPE_TYPE_LIST</td> + <td>0x1001</td> + <td>4 + (item.size * 2)</td> +</tr> +<tr> + <td>annotation_set_ref_list</td> + <td>TYPE_ANNOTATION_SET_REF_LIST</td> + <td>0x1002</td> + <td>4 + (item.size * 4)</td> +</tr> +<tr> + <td>annotation_set_item</td> + <td>TYPE_ANNOTATION_SET_ITEM</td> + <td>0x1003</td> + <td>4 + (item.size * 4)</td> +</tr> +<tr> + <td>class_data_item</td> + <td>TYPE_CLASS_DATA_ITEM</td> + <td>0x2000</td> + <td><i>implicit; must parse</i></td> +</tr> +<tr> + <td>code_item</td> + <td>TYPE_CODE_ITEM</td> + <td>0x2001</td> + <td><i>implicit; must parse</i></td> +</tr> +<tr> + <td>string_data_item</td> + <td>TYPE_STRING_DATA_ITEM</td> + <td>0x2002</td> + <td><i>implicit; must parse</i></td> +</tr> +<tr> + <td>debug_info_item</td> + <td>TYPE_DEBUG_INFO_ITEM</td> + <td>0x2003</td> + <td><i>implicit; must parse</i></td> +</tr> +<tr> + <td>annotation_item</td> + <td>TYPE_ANNOTATION_ITEM</td> + <td>0x2004</td> + <td><i>implicit; must parse</i></td> +</tr> +<tr> + <td>encoded_array_item</td> + <td>TYPE_ENCODED_ARRAY_ITEM</td> + <td>0x2005</td> + <td><i>implicit; must parse</i></td> +</tr> +<tr> + <td>annotations_directory_item</td> + <td>TYPE_ANNOTATIONS_DIRECTORY_ITEM</td> + <td>0x2006</td> + <td><i>implicit; must parse</i></td> +</tr> +</tbody> +</table> + + +<h2><code>string_id_item</code></h2> +<h4>appears in the <code>string_ids</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>string_data_off</td> + <td>uint</td> + <td>offset from the start of the file to the string data for this + item. The offset should be to a location + in the <code>data</code> section, and the data should be in the + format specified by "<code>string_data_item</code>" below. + There is no alignment requirement for the offset. + </td> +</tr> +</tbody> +</table> + +<h2><code>string_data_item</code></h2> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: none (byte-aligned)</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>utf16_size</td> + <td>uleb128</td> + <td>size of this string, in UTF-16 code units (which is the "string + length" in many systems). That is, this is the decoded length of + the string. (The encoded length is implied by the position of + the <code>0</code> byte.)</td> +</tr> +<tr> + <td>data</td> + <td>ubyte[]</td> + <td>a series of MUTF-8 code units (a.k.a. octets, a.k.a. bytes) + followed by a byte of value <code>0</code>. See + "MUTF-8 (Modified UTF-8) Encoding" above for details and + discussion about the data format. + <p><b>Note:</b> It is acceptable to have a string which includes + (the encoded form of) UTF-16 surrogate code units (that is, + <code>U+d800</code> … <code>U+dfff</code>) + either in isolation or out-of-order with respect to the usual + encoding of Unicode into UTF-16. It is up to higher-level uses of + strings to reject such invalid encodings, if appropriate.</p> + </td> +</tr> +</tbody> +</table> + +<h2><code>type_id_item</code></h2> +<h4>appears in the <code>type_ids</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>descriptor_idx</td> + <td>uint</td> + <td>index into the <code>string_ids</code> list for the descriptor + string of this type. The string must conform to the syntax for + <i>TypeDescriptor</i>, defined above. + </td> +</tr> +</tbody> +</table> + +<h2><code>proto_id_item</code></h2> +<h4>appears in the <code>proto_ids</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>shorty_idx</td> + <td>uint</td> + <td>index into the <code>string_ids</code> list for the short-form + descriptor string of this prototype. The string must conform to the + syntax for <i>ShortyDescriptor</i>, defined above, and must correspond + to the return type and parameters of this item. + </td> +</tr> +<tr> + <td>return_type_idx</td> + <td>uint</td> + <td>index into the <code>type_ids</code> list for the return type + of this prototype + </td> +</tr> +<tr> + <td>parameters_off</td> + <td>uint</td> + <td>offset from the start of the file to the list of parameter types + for this prototype, or <code>0</code> if this prototype has no + parameters. This offset, if non-zero, should be in the + <code>data</code> section, and the data there should be in the + format specified by <code>"type_list"</code> below. Additionally, there + should be no reference to the type <code>void</code> in the list. + </td> +</tr> +</tbody> +</table> + +<h2><code>field_id_item</code></h2> +<h4>appears in the <code>field_ids</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>class_idx</td> + <td>ushort</td> + <td>index into the <code>type_ids</code> list for the definer of this + field. This must be a class type, and not an array or primitive type. + </td> +</tr> +<tr> + <td>type_idx</td> + <td>ushort</td> + <td>index into the <code>type_ids</code> list for the type of + this field + </td> +</tr> +<tr> + <td>name_idx</td> + <td>uint</td> + <td>index into the <code>string_ids</code> list for the name of this + field. The string must conform to the syntax for <i>MemberName</i>, + defined above. + </td> +</tr> +</tbody> +</table> + +<h2><code>method_id_item</code></h2> +<h4>appears in the <code>method_ids</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>class_idx</td> + <td>ushort</td> + <td>index into the <code>type_ids</code> list for the definer of this + method. This must be a class or array type, and not a primitive type. + </td> +</tr> +<tr> + <td>proto_idx</td> + <td>ushort</td> + <td>index into the <code>proto_ids</code> list for the prototype of + this method + </td> +</tr> +<tr> + <td>name_idx</td> + <td>uint</td> + <td>index into the <code>string_ids</code> list for the name of this + method. The string must conform to the syntax for <i>MemberName</i>, + defined above. + </td> +</tr> +</tbody> +</table> + +<h2><code>class_def_item</code></h2> +<h4>appears in the <code>class_defs</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>class_idx</td> + <td>uint</td> + <td>index into the <code>type_ids</code> list for this class. + This must be a class type, and not an array or primitive type. + </td> +</tr> +<tr> + <td>access_flags</td> + <td>uint</td> + <td>access flags for the class (<code>public</code>, <code>final</code>, + etc.). See "<code>access_flags</code> Definitions" for details. + </td> +</tr> +<tr> + <td>superclass_idx</td> + <td>uint</td> + <td>index into the <code>type_ids</code> list for the superclass, or + the constant value <code>NO_INDEX</code> if this class has no + superclass (i.e., it is a root class such as <code>Object</code>). + If present, this must be a class type, and not an array or primitive type. + </td> +</tr> +<tr> + <td>interfaces_off</td> + <td>uint</td> + <td>offset from the start of the file to the list of interfaces, or + <code>0</code> if there are none. This offset + should be in the <code>data</code> section, and the data + there should be in the format specified by + "<code>type_list</code>" below. Each of the elements of the list + must be a class type (not an array or primitive type), and there + must not be any duplicates. + </td> +</tr> +<tr> + <td>source_file_idx</td> + <td>uint</td> + <td>index into the <code>string_ids</code> list for the name of the + file containing the original source for (at least most of) this class, + or the special value <code>NO_INDEX</code> to represent a lack of + this information. The <code>debug_info_item</code> of any given method + may override this source file, but the expectation is that most classes + will only come from one source file. + </td> +</tr> +<tr> + <td>annotations_off</td> + <td>uint</td> + <td>offset from the start of the file to the annotations structure + for this class, or <code>0</code> if there are no annotations on + this class. This offset, if non-zero, should be in the + <code>data</code> section, and the data there should be in + the format specified by "<code>annotations_directory_item</code>" below, + with all items referring to this class as the definer. + </td> +</tr> +<tr> + <td>class_data_off</td> + <td>uint</td> + <td>offset from the start of the file to the associated + class data for this item, or <code>0</code> if there is no class + data for this class. (This may be the case, for example, if this class + is a marker interface.) The offset, if non-zero, should be in the + <code>data</code> section, and the data there should be in the + format specified by "<code>class_data_item</code>" below, with all + items referring to this class as the definer. + </td> +</tr> +<tr> + <td>static_values_off</td> + <td>uint</td> + <td>offset from the start of the file to the list of initial + values for <code>static</code> fields, or <code>0</code> if there + are none (and all <code>static</code> fields are to be initialized with + <code>0</code> or <code>null</code>). This offset should be in the + <code>data</code> section, and the data there should be in the + format specified by "<code>encoded_array_item</code>" below. The size + of the array must be no larger than the number of <code>static</code> + fields declared by this class, and the elements correspond to the + <code>static</code> fields in the same order as declared in the + corresponding <code>field_list</code>. The type of each array + element must match the declared type of its corresponding field. + If there are fewer elements in the array than there are + <code>static</code> fields, then the leftover fields are initialized + with a type-appropriate <code>0</code> or <code>null</code>. + </td> +</tr> +</tbody> +</table> + +<h2><code>class_data_item</code></h2> +<h4>referenced from <code>class_def_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: none (byte-aligned)</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>static_fields_size</td> + <td>uleb128</td> + <td>the number of static fields defined in this item</td> +</tr> +<tr> + <td>instance_fields_size</td> + <td>uleb128</td> + <td>the number of instance fields defined in this item</td> +</tr> +<tr> + <td>direct_methods_size</td> + <td>uleb128</td> + <td>the number of direct methods defined in this item</td> +</tr> +<tr> + <td>virtual_methods_size</td> + <td>uleb128</td> + <td>the number of virtual methods defined in this item</td> +</tr> +<tr> + <td>static_fields</td> + <td>encoded_field[static_fields_size]</td> + <td>the defined static fields, represented as a sequence of + encoded elements. The fields must be sorted by + <code>field_idx</code> in increasing order. + </td> +</tr> +<tr> + <td>instance_fields</td> + <td>encoded_field[instance_fields_size]</td> + <td>the defined instance fields, represented as a sequence of + encoded elements. The fields must be sorted by + <code>field_idx</code> in increasing order. + </td> +</tr> +<tr> + <td>direct_methods</td> + <td>encoded_method[direct_methods_size]</td> + <td>the defined direct (any of <code>static</code>, <code>private</code>, + or constructor) methods, represented as a sequence of + encoded elements. The methods must be sorted by + <code>method_idx</code> in increasing order. + </td> +</tr> +<tr> + <td>virtual_methods</td> + <td>encoded_method[virtual_methods_size]</td> + <td>the defined virtual (none of <code>static</code>, <code>private</code>, + or constructor) methods, represented as a sequence of + encoded elements. This list should <i>not</i> include inherited + methods unless overridden by the class that this item represents. The + methods must be sorted by <code>method_idx</code> in increasing order. + </td> +</tr> +</tbody> +</table> + +<p><b>Note:</b> All elements' <code>field_id</code>s and +<code>method_id</code>s must refer to the same defining class.</p> + +<h3><code>encoded_field</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>field_idx_diff</td> + <td>uleb128</td> + <td>index into the <code>field_ids</code> list for the identity of this + field (includes the name and descriptor), represented as a difference + from the index of previous element in the list. The index of the + first element in a list is represented directly. + </td> +</tr> +<tr> + <td>access_flags</td> + <td>uleb128</td> + <td>access flags for the field (<code>public</code>, <code>final</code>, + etc.). See "<code>access_flags</code> Definitions" for details. + </td> +</tr> +</tbody> +</table> + +<h3><code>encoded_method</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>method_idx_diff</td> + <td>uleb128</td> + <td>index into the <code>method_ids</code> list for the identity of this + method (includes the name and descriptor), represented as a difference + from the index of previous element in the list. The index of the + first element in a list is represented directly. + </td> +</tr> +<tr> + <td>access_flags</td> + <td>uleb128</td> + <td>access flags for the method (<code>public</code>, <code>final</code>, + etc.). See "<code>access_flags</code> Definitions" for details. + </td> +</tr> +<tr> + <td>code_off</td> + <td>uleb128</td> + <td>offset from the start of the file to the code structure for this + method, or <code>0</code> if this method is either <code>abstract</code> + or <code>native</code>. The offset should be to a location in the + <code>data</code> section. The format of the data is specified by + "<code>code_item</code>" below. + </td> +</tr> +</tbody> +</table> + +<h2><code>type_list</code></h2> +<h4>referenced from <code>class_def_item</code> and +<code>proto_id_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>uint</td> + <td>size of the list, in entries</td> +</tr> +<tr> + <td>list</td> + <td>type_item[size]</td> + <td>elements of the list</td> +</tr> +</tbody> +</table> + +<h3><code>type_item</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>type_idx</td> + <td>ushort</td> + <td>index into the <code>type_ids</code> list</td> +</tr> +</tbody> +</table> + +<h2><code>code_item</code></h2> +<h4>referenced from <code>method_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>registers_size</td> + <td>ushort</td> + <td>the number of registers used by this code</td> +</tr> +<tr> + <td>ins_size</td> + <td>ushort</td> + <td>the number of words of incoming arguments to the method that this + code is for</td> +</tr> +<tr> + <td>outs_size</td> + <td>ushort</td> + <td>the number of words of outgoing argument space required by this + code for method invocation + </td> +</tr> +<tr> + <td>tries_size</td> + <td>ushort</td> + <td>the number of <code>try_item</code>s for this instance. If non-zero, + then these appear as the <code>tries</code> array just after the + <code>insns</code> in this instance. + </td> +</tr> +<tr> + <td>debug_info_off</td> + <td>uint</td> + <td>offset from the start of the file to the debug info (line numbers + + local variable info) sequence for this code, or <code>0</code> if + there simply is no information. The offset, if non-zero, should be + to a location in the <code>data</code> section. The format of + the data is specified by "<code>debug_info_item</code>" below. + </td> +</tr> +<tr> + <td>insns_size</td> + <td>uint</td> + <td>size of the instructions list, in 16-bit code units</td> +</tr> +<tr> + <td>insns</td> + <td>ushort[insns_size]</td> + <td>actual array of bytecode. The format of code in an <code>insns</code> + array is specified by the companion document + <a href="dalvik-bytecode.html">"Bytecode for the Dalvik VM"</a>. Note + that though this is defined as an array of <code>ushort</code>, there + are some internal structures that prefer four-byte alignment. Also, + if this happens to be in an endian-swapped file, then the swapping is + <i>only</i> done on individual <code>ushort</code>s and not on the + larger internal structures. + </td> +</tr> +<tr> + <td>padding</td> + <td>ushort <i>(optional)</i> = 0</td> + <td>two bytes of padding to make <code>tries</code> four-byte aligned. + This element is only present if <code>tries_size</code> is non-zero + and <code>insns_size</code> is odd. + </td> +</tr> +<tr> + <td>tries</td> + <td>try_item[tries_size] <i>(optional)</i></td> + <td>array indicating where in the code exceptions may be caught and + how to handle them. Elements of the array must be non-overlapping in + range and in order from low to high address. This element is only + present if <code>tries_size</code> is non-zero. + </td> +</tr> +<tr> + <td>handlers</td> + <td>encoded_catch_handler_list <i>(optional)</i></td> + <td>bytes representing a list of lists of catch types and associated + handler addresses. Each <code>try_item</code> has a byte-wise offset + into this structure. This element is only present if + <code>tries_size</code> is non-zero. + </td> +</tr> +</tbody> +</table> + +<h3><code>try_item</code> Format </h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>start_addr</td> + <td>uint</td> + <td>start address of the block of code covered by this entry. The address + is a count of 16-bit code units to the start of the first covered + instruction. + </td> +</tr> +<tr> + <td>insn_count</td> + <td>ushort</td> + <td>number of 16-bit code units covered by this entry. The last code + unit covered (inclusive) is <code>start_addr + insn_count - 1</code>. + </td> +</tr> +<tr> + <td>handler_off</td> + <td>ushort</td> + <td>offset in bytes from the start of the associated encoded handler data + to the <code>catch_handler_item</code> for this entry + </td> +</tr> +</tbody> +</table> + +<h3><code>encoded_catch_handler_list</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>uleb128</td> + <td>size of this list, in entries</td> +</tr> +<tr> + <td>list</td> + <td>encoded_catch_handler[handlers_size]</td> + <td>actual list of handler lists, represented directly (not as offsets), + and concatenated sequentially</td> +</tr> +</tbody> +</table> + +<h3><code>encoded_catch_handler</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>sleb128</td> + <td>number of catch types in this list. If non-positive, then this is + the negative of the number of catch types, and the catches are followed + by a catch-all handler. For example: A <code>size</code> of <code>0</code> + means that there is a catch-all but no explicitly typed catches. + A <code>size</code> of <code>2</code> means that there are two explicitly + typed catches and no catch-all. And a <code>size</code> of <code>-1</code> + means that there is one typed catch along with a catch-all. + </td> +</tr> +<tr> + <td>handlers</td> + <td>encoded_type_addr_pair[abs(size)]</td> + <td>stream of <code>abs(size)</code> encoded items, one for each caught + type, in the order that the types should be tested. + </td> +</tr> +<tr> + <td>catch_all_addr</td> + <td>uleb128 <i>(optional)</i></td> + <td>bytecode address of the catch-all handler. This element is only + present if <code>size</code> is non-positive. + </td> +</tr> +</tbody> +</table> + +<h3><code>encoded_type_addr_pair</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>type_idx</td> + <td>uleb128</td> + <td>index into the <code>type_ids</code> list for the type of the + exception to catch + </td> +</tr> +<tr> + <td>addr</td> + <td>uleb128</td> + <td>bytecode address of the associated exception handler</td> +</tr> +</tbody> +</table> + +<h2><code>debug_info_item</code></h2> +<h4>referenced from <code>code_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: none (byte-aligned)</h4> + +<p>Each <code>debug_info_item</code> defines a DWARF3-inspired byte-coded +state machine that, when interpreted, emits the positions +table and (potentially) the local variable information for a +<code>code_item</code>. The sequence begins with a variable-length +header (the length of which depends on the number of method +parameters), is followed by the state machine bytecodes, and ends +with an <code>DBG_END_SEQUENCE</code> byte.</p> + +<p>The state machine consists of five registers. The +<code>address</code> register represents the instruction offset in the +associated <code>insns_item</code> in 16-bit code units. The +<code>address</code> register starts at <code>0</code> at the beginning of each +<code>debug_info</code> sequence and may only monotonically increase. +The <code>line</code> register represents what source line number +should be associated with the next positions table entry emitted by +the state machine. It is initialized in the sequence header, and may +change in positive or negative directions but must never be less than +<code>1</code>. The <code>source_file</code> register represents the +source file that the line number entries refer to. It is initialized to +the value of <code>source_file_idx</code> in <code>class_def_item</code>. +The other two variables, <code>prologue_end</code> and +<code>epilogue_begin</code>, are boolean flags (initialized to +<code>false</code>) that indicate whether the next position emitted +should be considered a method prologue or epilogue. The state machine +must also track the name and type of the last local variable live in +each register for the <code>DBG_RESTART_LOCAL</code> code.</p> + +<p>The header is as follows:</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>line_start</td> + <td>uleb128</td> + <td>the initial value for the state machine's <code>line</code> register. + Does not represent an actual positions entry. + </td> +</tr> +<tr> + <td>parameters_size</td> + <td>uleb128</td> + <td>the number of parameter names that are encoded. There should be + one per method parameter, excluding an instance method's <code>this</code>, + if any. + </td> +</tr> +<tr> + <td>parameter_names</td> + <td>uleb128p1[parameters_size]</td> + <td>string index of the method parameter name. An encoded value of + <code>NO_INDEX</code> indicates that no name + is available for the associated parameter. The type descriptor + and signature are implied from the method descriptor and signature. + </td> +</tr> +</tbody> +</table> + +<p>The byte code values are as follows:</p> + +<table class="debugByteCode"> +<thead> +<tr> + <th>Name</th> + <th>Value</th> + <th>Format</th> + <th>Arguments</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>DBG_END_SEQUENCE</td> + <td>0x00</td> + <td></td> + <td><i>(none)</i></td> + <td>terminates a debug info sequence for a <code>code_item</code></td> +</tr> +<tr> + <td>DBG_ADVANCE_PC</td> + <td>0x01</td> + <td>uleb128 addr_diff</td> + <td><code>addr_diff</code>: amount to add to address register</td> + <td>advances the address register without emitting a positions entry</td> +</tr> +<tr> + <td>DBG_ADVANCE_LINE</td> + <td>0x02</td> + <td>sleb128 line_diff</td> + <td><code>line_diff</code>: amount to change line register by</td> + <td>advances the line register without emitting a positions entry</td> +</tr> +<tr> + <td>DBG_START_LOCAL</td> + <td>0x03</td> + <td>uleb128 register_num<br/> + uleb128p1 name_idx<br/> + uleb128p1 type_idx + </td> + <td><code>register_num</code>: register that will contain local<br/> + <code>name_idx</code>: string index of the name<br/> + <code>type_idx</code>: type index of the type + </td> + <td>introduces a local variable at the current address. Either + <code>name_idx</code> or <code>type_idx</code> may be + <code>NO_INDEX</code> to indicate that that value is unknown. + </td> +</tr> +<tr> + <td>DBG_START_LOCAL_EXTENDED</td> + <td>0x04</td> + <td>uleb128 register_num<br/> + uleb128p1 name_idx<br/> + uleb128p1 type_idx<br/> + uleb128p1 sig_idx + </td> + <td><code>register_num</code>: register that will contain local<br/> + <code>name_idx</code>: string index of the name<br/> + <code>type_idx</code>: type index of the type<br/> + <code>sig_idx</code>: string index of the type signature + </td> + <td>introduces a local with a type signature at the current address. + Any of <code>name_idx</code>, <code>type_idx</code>, or + <code>sig_idx</code> may be <code>NO_INDEX</code> + to indicate that that value is unknown. (If <code>sig_idx</code> is + <code>-1</code>, though, the same data could be represented more + efficiently using the opcode <code>DBG_START_LOCAL</code>.) + <p><b>Note:</b> See the discussion under + "<code>dalvik.annotation.Signature</code>" below for caveats about + handling signatures.</p> + </td> +</tr> +<tr> + <td>DBG_END_LOCAL</td> + <td>0x05</td> + <td>uleb128 register_num</td> + <td><code>register_num</code>: register that contained local</td> + <td>marks a currently-live local variable as out of scope at the current + address + </td> +</tr> +<tr> + <td>DBG_RESTART_LOCAL</td> + <td>0x06</td> + <td>uleb128 register_num</td> + <td><code>register_num</code>: register to restart</td> + <td>re-introduces a local variable at the current address. The name + and type are the same as the last local that was live in the specified + register. + </td> +</tr> +<tr> + <td>DBG_SET_PROLOGUE_END</td> + <td>0x07</td> + <td></td> + <td><i>(none)</i></td> + <td>sets the <code>prologue_end</code> state machine register, + indicating that the next position entry that is added should be + considered the end of a method prologue (an appropriate place for + a method breakpoint). The <code>prologue_end</code> register is + cleared by any special (<code>>= 0x0a</code>) opcode. + </td> +</tr> +<tr> + <td>DBG_SET_EPILOGUE_BEGIN</td> + <td>0x08</td> + <td></td> + <td><i>(none)</i></td> + <td>sets the <code>epilogue_begin</code> state machine register, + indicating that the next position entry that is added should be + considered the beginning of a method epilogue (an appropriate place + to suspend execution before method exit). + The <code>epilogue_begin</code> register is cleared by any special + (<code>>= 0x0a</code>) opcode. + </td> +</tr> +<tr> + <td>DBG_SET_FILE</td> + <td>0x09</td> + <td>uleb128p1 name_idx</td> + <td><code>name_idx</code>: string index of source file name; + <code>NO_INDEX</code> if unknown + </td> + <td>indicates that all subsequent line number entries make reference to this + source file name, instead of the default name specified in + <code>code_item</code> + </td> +</tr> +<tr> + <td><i>Special Opcodes</i></td> + <!-- When updating the range below, make sure to search for other + instances of 0x0a in this section. --> + <td>0x0a…0xff</td> + <td></td> + <td><i>(none)</i></td> + <td>advances the <code>line</code> and <code>address</code> registers, + emits a position entry, and clears <code>prologue_end</code> and + <code>epilogue_begin</code>. See below for description. + </td> +</tr> +</tbody> +</table> + +<h3>Special Opcodes</h3> + +<p>Opcodes with values between <code>0x0a</code> and <code>0xff</code> +(inclusive) move both the <code>line</code> and <code>address</code> +registers by a small amount and then emit a new position table entry. +The formula for the increments are as follows:</p> + +<pre> +DBG_FIRST_SPECIAL = 0x0a // the smallest special opcode +DBG_LINE_BASE = -4 // the smallest line number increment +DBG_LINE_RANGE = 15 // the number of line increments represented + +adjusted_opcode = opcode - DBG_FIRST_SPECIAL + +line += DBG_LINE_BASE + (adjusted_opcode % DBG_LINE_RANGE) +address += (adjusted_opcode / DBG_LINE_RANGE) +</pre> + +<h2><code>annotations_directory_item</code></h2> +<h4>referenced from <code>class_def_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>class_annotations_off</td> + <td>uint</td> + <td>offset from the start of the file to the annotations made directly + on the class, or <code>0</code> if the class has no direct annotations. + The offset, if non-zero, should be to a location in the + <code>data</code> section. The format of the data is specified + by "<code>annotation_set_item</code>" below. + </td> +</tr> +<tr> + <td>fields_size</td> + <td>uint</td> + <td>count of fields annotated by this item</td> +</tr> +<tr> + <td>annotated_methods_off</td> + <td>uint</td> + <td>count of methods annotated by this item</td> +</tr> +<tr> + <td>annotated_parameters_off</td> + <td>uint</td> + <td>count of method parameter lists annotated by this item</td> +</tr> +<tr> + <td>field_annotations</td> + <td>field_annotation[fields_size] <i>(optional)</i></td> + <td>list of associated field annotations. The elements of the list must + be sorted in increasing order, by <code>field_idx</code>. + </td> +</tr> +<tr> + <td>method_annotations</td> + <td>method_annotation[methods_size] <i>(optional)</i></td> + <td>list of associated method annotations. The elements of the list must + be sorted in increasing order, by <code>method_idx</code>. + </td> +</tr> +<tr> + <td>parameter_annotations</td> + <td>parameter_annotation[parameters_size] <i>(optional)</i></td> + <td>list of associated method parameter annotations. The elements of the + list must be sorted in increasing order, by <code>method_idx</code>. + </td> +</tr> +</tbody> +</table> + +<p><b>Note:</b> All elements' <code>field_id</code>s and +<code>method_id</code>s must refer to the same defining class.</p> + +<h3><code>field_annotation</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>field_idx</td> + <td>uint</td> + <td>index into the <code>field_ids</code> list for the identity of the + field being annotated + </td> +</tr> +<tr> + <td>annotations_off</td> + <td>uint</td> + <td>offset from the start of the file to the list of annotations for + the field. The offset should be to a location in the <code>data</code> + section. The format of the data is specified by + "<code>annotation_set_item</code>" below. + </td> +</tr> +</tbody> +</table> + +<h3><code>method_annotation</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>method_idx</td> + <td>uint</td> + <td>index into the <code>method_ids</code> list for the identity of the + method being annotated + </td> +</tr> +<tr> + <td>annotations_off</td> + <td>uint</td> + <td>offset from the start of the file to the list of annotations for + the method. The offset should be to a location in the + <code>data</code> section. The format of the data is specified by + "<code>annotation_set_item</code>" below. + </td> +</tr> +</tbody> +</table> + +<h3><code>parameter_annotation</code> Format</h2> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>method_idx</td> + <td>uint</td> + <td>index into the <code>method_ids</code> list for the identity of the + method whose parameters are being annotated + </td> +</tr> +<tr> + <td>annotations_off</td> + <td>uint</td> + <td>offset from the start of the file to the list of annotations for + the method parameters. The offset should be to a location in the + <code>data</code> section. The format of the data is specified by + "<code>annotation_set_ref_list</code>" below. + </td> +</tr> +</tbody> +</table> + +<h2><code>annotation_set_ref_list</code></h2> +<h4>referenced from <code>parameter_annotations_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>uint</td> + <td>size of the list, in entries</td> +</tr> +<tr> + <td>list</td> + <td>annotation_set_ref_item[size]</td> + <td>elements of the list</td> +</tr> +</tbody> +</table> + +<h3><code>annotation_set_ref_item</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>annotations_off</td> + <td>uint</td> + <td>offset from the start of the file to the referenced annotation set + or <code>0</code> if there are no annotations for this element. + The offset, if non-zero, should be to a location in the <code>data</code> + section. The format of the data is specified by + "<code>annotation_set_item</code>" below. + </td> +</tr> +</tbody> +</table> + +<h2><code>annotation_set_item</code></h2> +<h4>referenced from <code>annotations_directory_item</code>, +<code>field_annotations_item</code>, +<code>method_annotations_item</code>, and +<code>annotation_set_ref_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: 4 bytes</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>size</td> + <td>uint</td> + <td>size of the set, in entries</td> +</tr> +<tr> + <td>entries</td> + <td>annotation_off_item[size]</td> + <td>elements of the set. The elements must be sorted in increasing order, + by <code>type_idx</code>. + </td> +</tr> +</tbody> +</table> + +<h3><code>annotation_off_item</code> Format</h3> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>annotation_off</td> + <td>uint</td> + <td>offset from the start of the file to an annotation. + The offset should be to a location in the <code>data</code> section, + and the format of the data at that location is specified by + "<code>annotation_item</code>" below. + </td> +</tr> +</tbody> +</table> + + +<h2><code>annotation_item</code></h2> +<h4>referenced from <code>annotation_set_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: none (byte-aligned)</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>visibility</td> + <td>ubyte</td> + <td>intended visibility of this annotation (see below)</td> +</tr> +<tr> + <td>annotation</td> + <td>encoded_annotation</td> + <td>encoded annotation contents, in the format described by + "<code>encoded_annotation</code> Format" under + "<code>encoded_value</code> Encoding" above. + </td> +</tr> +</tbody> +</table> + +<h3>Visibility values</h3> + +<p>These are the options for the <code>visibility</code> field in an +<code>annotation_item</code>:</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Value</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>VISIBILITY_BUILD</td> + <td>0x00</td> + <td>intended only to be visible at build time (e.g., during compilation + of other code) + </td> +</tr> +<tr> + <td>VISIBILITY_RUNTIME</td> + <td>0x01</td> + <td>intended to visible at runtime</td> +</tr> +<tr> + <td>VISIBILITY_SYSTEM</td> + <td>0x02</td> + <td>intended to visible at runtime, but only to the underlying system + (and not to regular user code) + </td> +</tr> +</tbody> +</table> + +<h2><code>encoded_array_item</code></h2> +<h4>referenced from <code>class_def_item</code></h4> +<h4>appears in the <code>data</code> section</h4> +<h4>alignment: none (byte-aligned)</h4> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>encoded_array</td> + <td>bytes representing the encoded array value, in the format specified + by "<code>encoded_array</code> Format" under "<code>encoded_value</code> + Encoding" above. + </td> +</tr> +</tbody> +</table> + +<h1>System Annotations</h1> + +<p>System annotations are used to represent various pieces of reflective +information about classes (and methods and fields). This information is +generally only accessed indirectly by client (non-system) code.</p> + +<p>System annotations are represented in <code>.dex</code> files as +annotations with visibility set to <code>VISIBILITY_SYSTEM</code>. + +<h2><code>dalvik.annotation.AnnotationDefault</code></h2> +<h4>appears on methods in annotation interfaces</h4> + +<p>An <code>AnnotationDefault</code> annotation is attached to each +annotation interface which wishes to indicate default bindings.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>Annotation</td> + <td>the default bindings for this annotation, represented as an annotation + of this type. The annotation need not include all names defined by the + annotation; missing names simply do not have defaults. + </td> +</tr> +</tbody> +</table> + +<h2><code>dalvik.annotation.EnclosingClass</code></h2> +<h4>appears on classes</h4> + +<p>An <code>EnclosingClass</code> annotation is attached to each class +which is either defined as a member of another class, per se, or is +anonymous but not defined within a method body (e.g., a synthetic +inner class). Every class that has this annotation must also have an +<code>InnerClass</code> annotation. Additionally, a class may not have +both an <code>EnclosingClass</code> and an +<code>EnclosingMethod</code> annotation.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>Class</td> + <td>the class which most closely lexically scopes this class</td> +</tr> +</tbody> +</table> + +<h2><code>dalvik.annotation.EnclosingMethod</code></h2> +<h4>appears on classes</h4> + +<p>An <code>EnclosingMethod</code> annotation is attached to each class +which is defined inside a method body. Every class that has this +annotation must also have an <code>InnerClass</code> annotation. +Additionally, a class may not have both an <code>EnclosingClass</code> +and an <code>EnclosingMethod</code> annotation.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>Method</td> + <td>the method which most closely lexically scopes this class</td> +</tr> +</tbody> +</table> + +<h2><code>dalvik.annotation.InnerClass</code></h2> +<h4>appears on classes</h4> + +<p>An <code>InnerClass</code> annotation is attached to each class +which is defined in the lexical scope of another class's definition. +Any class which has this annotation must also have <i>either</i> an +<code>EnclosingClass</code> annotation <i>or</i> an +<code>EnclosingMethod</code> annotation.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>name</td> + <td>String</td> + <td>the originally declared simple name of this class (not including any + package prefix). If this class is anonymous, then the name is + <code>null</code>. + </td> +</tr> +<tr> + <td>accessFlags</td> + <td>int</td> + <td>the originally declared access flags of the class (which may differ + from the effective flags because of a mismatch between the execution + models of the source language and target virtual machine) + </td> +</tr> +</tbody> +</table> + +<h2><code>dalvik.annotation.MemberClasses</code></h2> +<h4>appears on classes</h4> + +<p>A <code>MemberClasses</code> annotation is attached to each class +which declares member classes. (A member class is a direct inner class +that has a name.)</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>Class[]</td> + <td>array of the member classes</td> +</tr> +</tbody> +</table> + +<h2><code>dalvik.annotation.Signature</code></h2> +<h4>appears on classes, fields, and methods</h4> + +<p>A <code>Signature</code> annotation is attached to each class, +field, or method which is defined in terms of a more complicated type +than is representable by a <code>type_id_item</code>. The +<code>.dex</code> format does not define the format for signatures; it +is merely meant to be able to represent whatever signatures a source +language requires for successful implementation of that language's +semantics. As such, signatures are not generally parsed (or verified) +by virtual machine implementations. The signatures simply get handed +off to higher-level APIs and tools (such as debuggers). Any use of a +signature, therefore, should be written so as not to make any +assumptions about only receiving valid signatures, explicitly guarding +itself against the possibility of coming across a syntactically +invalid signature.</p> + +<p>Because signature strings tend to have a lot of duplicated content, +a <code>Signature</code> annotation is defined as an <i>array</i> of +strings, where duplicated elements naturally refer to the same +underlying data, and the signature is taken to be the concatenation of +all the strings in the array. There are no rules about how to pull +apart a signature into separate strings; that is entirely up to the +tools that generate <code>.dex</code> files.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>String[]</td> + <td>the signature of this class or member, as an array of strings that + is to be concatenated together</td> +</tr> +</tbody> +</table> + +<h2><code>dalvik.annotation.Throws</code></h2> +<h4>appears on methods</h4> + +<p>A <code>Throws</code> annotation is attached to each method which is +declared to throw one or more exception types.</p> + +<table class="format"> +<thead> +<tr> + <th>Name</th> + <th>Format</th> + <th>Description</th> +</tr> +</thead> +<tbody> +<tr> + <td>value</td> + <td>Class[]</td> + <td>the array of exception types thrown</td> +</tr> +</tbody> +</table> + +</body> +</html> diff --git a/docs/dexopt.html b/docs/dexopt.html new file mode 100644 index 000000000..7f0b4bcd2 --- /dev/null +++ b/docs/dexopt.html @@ -0,0 +1,326 @@ +<html> +<head> + <title>Dalvik Optimization and Verification</title> +</head> + +<body> +<h1>Dalvik Optimization and Verification With <i>dexopt</i></h1> + +<p> +The Dalvik virtual machine was designed specifically for the Android +mobile platform. The target systems have little RAM, store data on slow +internal flash memory, and generally have the performance characteristics +of decade-old desktop systems. They also run Linux, which provides +virtual memory, processes and threads, and UID-based security mechanisms. +<p> +The features and limitations caused us to focus on certain goals: + +<ul> + <li>Class data, notably bytecode, must be shared between multiple + processes to minimize total system memory usage. + <li>The overhead in launching a new app must be minimized to keep + the device responsive. + <li>Storing class data in individual files results in a lot of + redundancy, especially with respect to strings. To conserve disk + space we need to factor this out. + <li>Parsing class data fields adds unnecessary overhead during + class loading. Accessing data values (e.g. integers and strings) + directly as C types is better. + <li>Bytecode verification is necessary, but slow, so we want to verify + as much as possible outside app execution. + <li>Bytecode optimization (quickened instructions, method pruning) is + important for speed and battery life. + <li>For security reasons, processes may not edit shared code. +</ul> + +<p> +The typical VM implementation uncompresses individual classes from a +compressed archive and stores them on the heap. This implies a separate +copy of each class in every process, and slows application startup because +the code must be uncompressed (or at least read off disk in many small +pieces). On the other hand, having the bytecode on the local heap makes +it easy to rewrite instructions on first use, facilitating a number of +different optimizations. +<p> +The goals led us to make some fundamental decisions: + +<ul> + <li>Multiple classes are aggregated into a single "DEX" file. + <li>DEX files are mapped read-only and shared between processes. + <li>Byte ordering and word alignment are adjusted to suit the local + system. + <li>Bytecode verification is mandatory for all classes, but we want + to "pre-verify" whatever we can. + <li>Optimizations that require rewriting bytecode must be done ahead + of time. +</ul> + +<p> +The consequences of these decisions are explained in the following sections. + + +<h2>VM Operation</h2> + +<p> +Application code is delivered to the system in a <code>.jar</code> +or <code>.apk</code> file. These are really just <code>.zip</code> +archives with some meta-data files added. The Dalvik DEX data file +is always called <code>classes.dex</code>. +<p> +The bytecode cannot be memory-mapped and executed directly from the zip +file, because the data is compressed and the start of the file is not +guaranteed to be word-aligned. These problems could be addressed by +storing <code>classes.dex</code> without compression and padding out the zip +file, but that would increase the size of the package sent across the +data network. +<p> +We need to extract <code>classes.dex</code> from the zip archive before +we can use it. While we have the file available, we might as well perform +some of the other actions (realignment, optimization, verification) described +earlier. This raises a new question however: who is responsible for doing +this, and where do we keep the output? + +<h3>Preparation</h3> + +<p> +There are at least three different ways to create a "prepared" DEX file, +sometimes known as "ODEX" (for Optimized DEX): +<ol> + <li>The VM does it "just in time". The output goes into a special + <code>dalvik-cache</code> directory. This works on the desktop and + engineering-only device builds where the permissions on the + <code>dalvik-cache</code> directory are not restricted. On production + devices, this is not allowed. + <li>The system installer does it when an application is first added. + It has the privileges required to write to <code>dalvik-cache</code>. + <li>The build system does it ahead of time. The relevant <code>jar</code> + / <code>apk</code> files are present, but the <code>classes.dex</code> + is stripped out. The optimized DEX is stored next to the original + zip archive, not in <code>dalvik-cache</code>, and is part of the + system image. +</ol> +<p> +The <code>dalvik-cache</code> directory is more accurately +<code>$ANDROID_DATA/data/dalvik-cache</code>. The files inside it have +names derived from the full path of the source DEX. On the device the +directory is owned by <code>system</code> / <code>system</code> +and has 0771 permissions, and the optimized DEX files stored there are +owned by <code>system</code> and the +application's group, with 0644 permissions. DRM-locked applications will +use 640 permissions to prevent other user applications from examining them. +The bottom line is that you can read your own DEX file and those of most +other applications, but you cannot create, modify, or remove them. +<p> +Preparation of the DEX file for the "just in time" and "system installer" +approaches proceeds in three steps: +<p> +First, the dalvik-cache file is created. This must be done in a process +with appropriate privileges, so for the "system installer" case this is +done within <code>installd</code>, which runs as root. +<p> +Second, the <code>classes.dex</code> entry is extracted from the the zip +archive. A small amount of space is left at the start of the file for +the ODEX header. +<p> +Third, the file is memory-mapped for easy access and tweaked for use on +the current system. This includes byte-swapping and structure realigning, +but no meaningful changes to the DEX file. We also do some basic +structure checks, such as ensuring that file offsets and data indices +fall within valid ranges. +<p> +The build system uses a hairy process that involves starting the +emulator, forcing just-in-time optimization of all relevant DEX files, +and then extracting the results from <code>dalvik-cache</code>. The +reasons for doing this, rather than using a tool that runs on the desktop, +will become more apparent when the optimizations are explained. +<p> +Once the code is byte-swapped and aligned, we're ready to go. We append +some pre-computed data, fill in the ODEX header at the start of the file, +and start executing. (The header is filled in last, so that we don't +try to use a partial file.) If we're interested in verification and +optimization, however, we need to insert a step after the initial prep. + +<h3>dexopt</h3> + +<p> +We want to verify and optimize all of the classes in the DEX file. The +easiest and safest way to do this is to load all of the classes into +the VM and run through them. Anything that fails to load is simply not +verified or optimized. Unfortunately, this can cause allocation of some +resources that are difficult to release (e.g. loading of native shared +libraries), so we don't want to do it in the same virtual machine that +we're running applications in. +<p> +The solution is to invoke a program called <code>dexopt</code>, which +is really just a back door into the VM. It performs an abbreviated VM +initialization, loads zero or more DEX files from the bootstrap class +path, and then sets about verifying and optimizing whatever it can from +the target DEX. On completion, the process exits, freeing all resources. +<p> +It is possible for multiple VMs to want the same DEX file at the same +time. File locking is used to ensure that dexopt is only run once. + + +<h2>Verification</h2> + +<p> +The bytecode verification process involves scanning through the instructions +in every method in every class in a DEX file. The goal is to identify +illegal instruction sequences so that we don't have to check for them at +run time. Many of the computations involved are also necessary for "exact" +garbage collection. See +<a href="verifier.html">Dalvik Bytecode Verifier Notes</a> for more +information. +<p> +For performance reasons, the optimizer (described in the next section) +assumes that the verifier has run successfully, and makes some potentially +unsafe assumptions. By default, Dalvik insists upon verifying all classes, +and only optimizes classes that have been verified. If you want to +disable the verifier, you can use command-line flags to do so. See also +<a href="embedded-vm-control.html"> Controlling the Embedded VM</a> +for instructions on controlling these +features within the Android application framework. +<p> +Reporting of verification failures is a tricky issue. For example, +calling a package-scope method on a class in a different package is +illegal and will be caught by the verifier. We don't necessarily want +to report it during verification though -- we actually want to throw +an exception when the method call is attempted. Checking the access +flags on every method call is expensive though. The +<a href="verifier.html">Dalvik Bytecode Verifier Notes</a> document +addresses this issue. +<p> +Classes that have been verified successfully have a flag set in the ODEX. +They will not be re-verified when loaded. The Linux access permissions +are expected to prevent tampering; if you can get around those, installing +faulty bytecode is far from the easiest line of attack. The ODEX file has +a 32-bit checksum, but that's chiefly present as a quick check for +corrupted data. + + +<h2>Optimization</h2> + +<p> +Virtual machine interpreters typically perform certain optimizations the +first time a piece of code is used. Constant pool references are replaced +with pointers to internal data structures, operations that always succeed +or always work a certain way are replaced with simpler forms. Some of +these require information only available at runtime, others can be inferred +statically when certain assumptions are made. +<p> +The Dalvik optimizer does the following: +<ul> + <li>For virtual method calls, replace the method index with a + vtable index. + <li>For instance field get/put, replace the field index with + a byte offset. Also, merge the boolean / byte / char / short + variants into a single 32-bit form (less code in the interpreter + means more room in the CPU I-cache). + <li>Replace a handful of high-volume calls, like String.length(), + with "inline" replacements. This skips the usual method call + overhead, directly switching from the interpreter to a native + implementation. + <li>Prune empty methods. The simplest example is + <code>Object.<init></code>, which does nothing, but must be + called whenever any object is allocated. The instruction is + replaced with a new version that acts as a no-op unless a debugger + is attached. + <li>Append pre-computed data. For example, the VM wants to have a + hash table for lookups on class name. Instead of computing this + when the DEX file is loaded, we can compute it now, saving heap + space and computation time in every VM where the DEX is loaded. +</ul> + +<p> +All of the instruction modifications involve replacing the opcode with +one not defined by the Dalvik specification. This allows us to freely +mix optimized and unoptimized instructions. The set of optimized +instructions, and their exact representation, is tied closely to the VM +version. +<p> +Most of the optimizations are obvious "wins". The use of raw indices +and offsets not only allows us to execute more quickly, we can also +skip the initial symbolic resolution. Pre-computation eats up +disk space, and so must be done in moderation. +<p> +There are a couple of potential sources of trouble with these +optimizations. First, vtable indices and byte offsets are subject to +change if the VM is updated. Second, if a superclass is in a different +DEX, and that other DEX is updated, we need to ensure that our optimized +indices and offsets are updated as well. A similar but more subtle +problem emerges when user-defined class loaders are employed: the class +we actually call may not be the one we expected to call. +<p>These problems are addressed with dependency lists and some limitations +on what can be optimized. + + +<h2>Dependencies and Limitations</h2> + +<p> +The optimized DEX file includes a list of dependencies on other DEX files, +plus the CRC-32 and modification date from the originating +<code>classes.dex</code> zip file entry. The dependency list includes the +full path to the <code>dalvik-cache</code> file, and the file's SHA-1 +signature. The timestamps of files on the device are unreliable and +not used. The dependency area also includes the VM version number. +<p> +An optimized DEX is dependent upon all of the DEX files in the bootstrap +class path. DEX files that are part of the bootstrap class path depend +upon the DEX files that appeared earlier. To ensure that nothing outside +the dependent DEX files is available, <code>dexopt</code> only loads the +bootstrap classes. References to classes in other DEX files fail, which +causes class loading and/or verification to fail, and classes with +external dependencies are simply not optimized. +<p> +This means that splitting code out into many separate DEX files has a +disadvantage: virtual method calls and instance field lookups between +non-boot DEX files can't be optimized. Because verification is pass/fail +with class granularity, no method in a class that has any reliance on +classes in external DEX files can be optimized. This may be a bit +heavy-handed, but it's the only way to guarantee that nothing breaks +when individual pieces are updated. +<p> +Another negative consequence: any change to a bootstrap DEX will result +in rejection of all optimized DEX files. This makes it hard to keep +system updates small. +<p> +Despite our caution, there is still a possibility that a class in a DEX +file loaded by a user-defined class loader could ask for a bootstrap class +(say, String) and be given a different class with the same name. If a +class in the DEX file being processed has the same name as a class in the +bootstrap DEX files, the class will be flagged as ambiguous and references +to it will not be resolved during verification / optimization. The class +linking code in the VM does additional checks to plug another hole; +see the verbose description in the VM sources for details (vm/oo/Class.c). +<p> +If one of the dependencies is updated, we need to re-verify and +re-optimize the DEX file. If we can do a just-in-time <code>dexopt</code> +invocation, this is easy. If we have to rely on the installer daemon, or +the DEX was shipped only in ODEX, then the VM has to reject the DEX. +<p> +The output of <code>dexopt</code> is byte-swapped and struct-aligned +for the host, and contains indices and offsets that are highly VM-specific +(both version-wise and platform-wise). For this reason it's tricky to +write a version of <code>dexopt</code> that runs on the desktop but +generates output suitable for a particular device. The safest way to +invoke it is on the target device, or on an emulator for that device. + + +<h2>Generated DEX</h2> + +<p> +Some languages and frameworks rely on the ability to generate bytecode +and execute it. The rather heavy <code>dexopt</code> verification and +optimization model doesn't work well with that. +<p> +We intend to support this in a future release, but the exact method is +to be determined. We may allow individual classes to be added or whole +DEX files; may allow Java bytecode or Dalvik bytecode in instructions; +may perform the usual set of optimizations, or use a separate interpreter +that performs on-first-use optimizations directly on the bytecode (which +won't be mapped read-only, since it's locally defined). + +<address>Copyright © 2008 The Android Open Source Project</address> + +</body> +</html> diff --git a/docs/embedded-vm-control.html b/docs/embedded-vm-control.html new file mode 100644 index 000000000..f90f0e596 --- /dev/null +++ b/docs/embedded-vm-control.html @@ -0,0 +1,240 @@ +<html> +<head> + <title>Controlling the Embedded VM</title> + <link rel=stylesheet href="android.css"> +</head> + +<body> +<h1>Controlling the Embedded VM</h1> + +<ul> + <li><a href="#overview">Overview</a> + <li><a href="#checkjni">Extended JNI Checks</a> + <li><a href="#assertions">Assertions</a> + <li><a href="#verifier">Bytecode Verification and Optimization</a> + <li><a href="#execmode">Execution Mode</a> + <li><a href="#dp">Deadlock Prediction</a> + <li><a href="#stackdump">Stack Dumps</a> +</ul> + +<h2><a name="overview">Overview</a></h2> + +<p>The Dalvik VM supports a variety of command-line arguments +(use <code>adb shell dalvikvm -help</code> to get a summary), but +it's not possible to pass arbitrary arguments through the +Android application runtime. It is, however, possible to affect the +VM behavior through certain system properties. + +<p>For all of the features described below, you would set the system property +with <code>setprop</code>, +issuing a shell command on the device like this: +<pre>adb shell setprop <name> <value></pre> + +<p>The Android runtime must be restarted before the changes will take +effect (<code>adb shell stop; adb shell start</code>). This is because the +settings are processed in the "zygote" process, which starts early and stays +around "forever". + +<p>You could also add a line to <code>/data/local.prop</code> that looks like: +<pre><name> = <value></pre> + +<p>Such changes will survive reboots, but will be removed by anything +that wipes the data partition. (Hint: create a <code>local.prop</code> +on your workstation, then <code>adb push local.prop /data</code> .) + + +<h2><a name="checkjni">Extended JNI Checks</a></h2> + +<p>JNI, the Java Native Interface, provides a way for code written in the +Java programming language +interact with native (C/C++) code. The extended JNI checks will cause +the system to run more slowly, but they can spot a variety of nasty bugs +before they have a chance to cause problems. + +<p>There are two system properties that affect this feature, which is +enabled with the <code>-Xcheck:jni</code> command-line argument. The +first is <code>ro.kernel.android.checkjni</code>. This is set by the +Android build system for development builds. (It may also be set by +the Android emulator unless the <code>-nojni</code> flag is provided on the +emulator command line.) Because this is an "ro." property, the value cannot +be changed once the device has started. + +<p>To allow toggling of the CheckJNI flag, a second +property, <code>dalvik.vm.checkjni</code>, is also checked. The value +of this overrides the value from <code>ro.kernel.android.checkjni</code>. + +<p>If neither property is defined, or <code>dalvik.vm.checkjni</code> +is set to <code>false</code>, the <code>-Xcheck:jni</code> flag is +not passed in, and JNI checks will be disabled. + +<p>To enable JNI checking: +<pre>adb shell setprop dalvik.vm.checkjni true</pre> + +<p>You can also pass JNI-checking options into the VM through a system +property. The value set for <code>dalvik.vm.jniopts</code> will +be passed in as the <code>-Xjniopts</code> argument. + +<p>For more information about JNI checks, see +<a href="jni-tips.html">JNI Tips</a>. + + +<h2><a name="assertions">Assertions</a></h2> + +<p>Dalvik VM supports the Java programming language "assert" statement. +By default they are off, but the <code>dalvik.vm.enableassertions</code> +property provides a way to set the value for a <code>-ea</code> argument. + +<p>The argument behaves the same as it does in other desktop VMs. You +can provide a class name, a package name (followed by "..."), or the +special value "all". + +<p>For example, this: +<pre>adb shell setprop dalvik.vm.enableassertions all</pre> +enables assertions in all non-system classes. + +<p>The system property is much more limited than the full command line. +It is not possible to specify more than one <code>-ea</code> entry, and there +is no way to specify a <code>-da</code> entry. There is presently no +equivalent for <code>-esa</code>/<code>-dsa</code>. + + +<h2><a name="verifier">Bytecode Verification and Optimization</a></h2> + +<p>The system tries to pre-verify all classes in a DEX file to reduce +class load overhead, and performs a series of optimizations to improve +runtime performance. Both of these are done by the <code>dexopt</code> +command, either in the build system or by the installer. On a development +device, <code>dexopt</code> may be run the first time a DEX file is used +and whenever it or one of its dependencies is updated ("just-in-time" +optimization and verification). + +<p>There are two command-line flags that control the just-in-time +verification and optimization, +<code>-Xverify</code> and <code>-Xdexopt</code>. The Android framework +configures these based on the <code>dalvik.vm.dexopt-flags</code> +property. + +<p>If you set: +<pre>adb shell setprop dalvik.vm.dexopt-flags v=a,o=v</pre> +then the framework will pass <code>-Xverify:all -Xdexopt:verified</code> +to the VM. This enables verification, and only optimizes classes that +successfully verified. This is the safest setting, and is the default. +<p>You could also set <code>dalvik.vm.dexopt-flags</code> to <code>v=n</code> +to have the framework pass <code>-Xverify:none -Xdexopt:verified</code> +to disable verification. (We could pass in <code>-Xdexopt:all</code> to +allow optimization, but that wouldn't necessarily optimize more of the +code, since classes that fail verification may well be skipped by the +optimizer for the same reasons.) Classes will not be verified by +<code>dexopt</code>, and unverified code will be loaded and executed. + +<p>Enabling verification will make the <code>dexopt</code> command +take significantly longer, because the verification process is fairly slow. +Once the verified and optimized DEX files have been prepared, verification +incurs no additional overhead except when loading classes that failed +to pre-verify. + +<p>If your DEX files are processed with verification disabled, and you +later turn the verifier on, application loading will be noticeably +slower (perhaps 40% or more) as classes are verified on first use. + +<p>For best results you should force a re-dexopt of all DEX files when +this property changes. You can do this with: +<pre>adb shell "rm /data/dalvik-cache/*"</pre> +This removes the cached versions of the DEX files. Remember to +stop and restart the runtime (<code>adb shell stop; adb shell start</code>). + +<p>(Previous version of the runtime supported the boolean +<code>dalvik.vm.verify-bytecode</code> property, but that has been +superceded by <code>dalvik.vm.dexopt-flags</code>.)</p> + + +<h2><a name="execmode">Execution Mode</a></h2> + +<p>The current implementation of the Dalvik VM includes three distinct +interpreter cores. These are referred to as "fast", "portable", and +"debug". The "fast" interpreter is optimized for the current +platform, and might consist of hand-optimized assembly routines. In +constrast, the "portable" interpreter is written in C and expected to +run on a broad range of platforms. The "debug" interpreter is a variant +of "portable" that includes support for profiling and single-stepping. + +<p>The VM allows you to choose between "fast" and "portable" with an +extended form of the <code>-Xint</code> argument. The value of this +argument can be set through the <code>dalvik.vm.execution-mode</code> +system property. + +<p>To select the "portable" interpreter, you would use: +<pre>adb shell setprop dalvik.vm.execution-mode int:portable</pre> +If the property is not specified, the most appropriate interpreter +will be selected automatically. At some point this mechanism may allow +selection of other modes, such as JIT compilation. + +<p>Not all platforms have an optimized implementation. In such cases, +the "fast" interpreter is generated as a series of C stubs, and the +result will be slower than the +"portable" version. (When we have optimized versions for all popular +architectures the naming convention will be more accurate.) + +<p>If profiling is enabled or a debugger is attached, the VM +switches to the "debug" interpreter. When profiling ends or the debugger +disconnects, the original interpreter is resumed. (The "debug" interpreter +is substantially slower, something to keep in mind when evaluating +profiling data.) + + +<h2><a name="dp">Deadlock Prediction</a></h2> + +<p>If the VM is built with <code>WITH_DEADLOCK_PREDICTION</code>, the deadlock +predictor can be enabled with the <code>-Xdeadlockpredict</code> argument. +(The output from <code>dalvikvm -help</code> will tell you if the VM was +built appropriately -- look for <code>deadlock_prediction</code> on the +<code>Configured with:</code> line.) +This feature tells the VM to keep track of the order in which object +monitor locks are acquired. If the program attempts to acquire a set +of locks in a different order from what was seen earlier, the VM logs +a warning and optionally throws an exception. + +<p>The command-line argument is set based on the +<code>dalvik.vm.deadlock-predict</code> property. Valid values are +<code>off</code> to disable it (default), <code>warn</code> to log the +problem but continue executing, <code>err</code> to cause a +<code>dalvik.system.PotentialDeadlockError</code> to be thrown from the +<code>monitor-enter</code> instruction, and <code>abort</code> to have +the entire VM abort. + +<p>You will usually want to use: +<pre>adb shell setprop dalvik.vm.deadlock-predict err</pre> +unless you are keeping an eye on the logs as they scroll by. + +<p>Please note that this feature is deadlock prediction, not deadlock +detection -- in the current implementation, the computations are performed +after the lock is acquired (this simplifies the code, reducing the +overhead added to every mutex operation). You can spot a deadlock in a +hung process by sending a <code>kill -3</code> and examining the stack +trace written to the log. + +<p>This only takes monitors into account. Native mutexes and other resources +can also be the cause of deadlocks, but will not be detected by this. + + +<h2><a name="stackdump">Stack Dumps</a></h2> + +<p>Like other desktop VMs, when the Dalvik VM receives a SIGQUIT +(Ctrl-\ or <code>kill -3</code>), it dumps stack traces for all threads. +By default this goes to the Android log, but it can also be written to a file. + +<p>The <code>dalvik.vm.stack-trace-file</code> property allows you to +specify the name of the file where the thread stack traces will be written. +The file will be created (world writable) if it doesn't exist, and the +new information will be appended to the end of the file. The filename +is passed into the VM via the <code>-Xstacktracefile</code> argument. + +<p>For example: +<pre>adb shell setprop dalvik.vm.stack-trace-file /tmp/stack-traces.txt</pre> + +<p>If the property is not defined, the VM will write the stack traces to +the Android log when the signal arrives. + +<address>Copyright © 2008 The Android Open Source Project</address> + +</body></html> diff --git a/docs/instruction-formats.css b/docs/instruction-formats.css new file mode 100644 index 000000000..ee23c5c64 --- /dev/null +++ b/docs/instruction-formats.css @@ -0,0 +1,129 @@ +h1 { + font-family: serif; + color: #222266; +} + +h2 { + font-family: serif; + border-top-style: solid; + border-top-width: 2px; + border-color: #ccccdd; + padding-top: 12px; + margin-top: 48px; + margin-bottom: 2px; + color: #222266; +} + +h3 { + font-family: serif; + color: #222266; +} + +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + +table th { + font-family: sans-serif; + background: #aaaaff; +} + +table { + border-collapse: collapse; +} + +table td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + border-color: #aaaaff; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 2px; + padding-right: 2px; + background: #eeeeff; +} + + +/* the mnemonic guide */ + +table.letters { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.letters td:first-child { + font-family: monospace; + width: 10%; + text-align: center; +} + +table.letters td:first-child + td { + width: 10%; + text-align: center; +} + +table.letters td:first-child + td + td { + width: 80%; +} + + +/* the formats, per se */ + +table.format { + background: #aaaaaa; + border-collapse: collapse; + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.format td { + font-family: monospace; +} + +table.format td + td i { + font-family: sans-serif; +} + +table.format td sub { + font-family: sans-serif; +} + +table.format td sub { + font-family: sans-serif; + font-style: italic; + font-size: 70% +} + +table.format th:first-child { + width: 28%; +} + +table.format th:first-child + th { + width: 5%; +} + +table.format th:first-child + th + th { + width: 45%; +} + +table.format th:first-child + th + th + th { + width: 22%; +} + +table.format p { + margin-bottom: 0pt; +}
\ No newline at end of file diff --git a/docs/instruction-formats.html b/docs/instruction-formats.html new file mode 100644 index 000000000..d7bf69057 --- /dev/null +++ b/docs/instruction-formats.html @@ -0,0 +1,430 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>Dalvik VM Instruction Formats</title> +<link rel=stylesheet href="instruction-formats.css"> +</head> + +<body> + +<h1>Dalvik VM Instruction Formats</h1> +<p>Copyright © 2007 The Android Open Source Project + +<h2>Introduction and Overview</h2> + +<p>This document lists the instruction formats used by Dalvik bytecode +and is meant to be used in conjunction with the +<a href="dalvik-bytecode.html">bytecode reference document</a>.</p> + +<h3>Bitwise descriptions</h3> + +<p>The first column in the format table lists the bitwise layout of +the format. It consists of one or more space-separated "words" each of +which describes a 16-bit code unit. Each character in a word +represents four bits, read from high bits to low, with vertical bars +("<code>|</code>") interspersed to aid in reading. Uppercase letters +in sequence from "<code>A</code>" are used to indicate fields within +the format (which then get defined further by the syntax column). The term +"<code>op</code>" is used to indicate the position of the eight-bit +opcode within the format. A slashed zero ("<code>Ø</code>") is +used to indicate that all bits should be zero in the indicated +position.</p> + +<p>For example, the format "<code>B|A|<i>op</i> CCCC</code>" indicates +that the format consists of two 16-bit code units. The first word +consists of the opcode in the low eight bits and a pair of four-bit +values in the high eight bits; and the second word consists of a single +16-bit value.</p> + +<h3>Format IDs</h3> + +<p>The second column in the format table indicates the short identifier +for the format, which is used in other documents and in code to identify +the format.</p> + +<p>Format IDs consist of three characters, two digits followed by a +letter. The first digit indicates the number of 16-bit code units in the +format. The second digit indicates the maximum number of registers that the +format contains (maximum, since some formats can accomodate a variable +number of registers), with the special designation "<code>r</code>" indicating +that a range of registers is encoded. The final letter semi-mnemonically +indicates the type of any extra data encoded by the format. For example, +format "<code>21t</code>" is of length two, contains one register reference, +and additionally contains a branch target.</p> + +<p>Suggested static linking formats have an additional "<code>s</code>" suffix, +making them four characters total.</p> + +<p>The full list of typecode letters are as follows. Note that some +forms have different sizes, depending on the format:</p> + +<table class="letters"> +<thead> +<tr> + <th>Mnemonic</th> + <th>Bit Sizes</th> + <th>Meaning</th> +</tr> +</thead> +<tbody> +<tr> + <td>b</td> + <td>8</td> + <td>immediate signed <b>b</b>yte</td> +</tr> +<tr> + <td>c</td> + <td>16, 32</td> + <td><b>c</b>onstant pool index</td> +</tr> +<tr> + <td>f</td> + <td>16</td> + <td>inter<b>f</b>ace constants (only used in statically linked formats) + </td> +</tr> +<tr> + <td>h</td> + <td>16</td> + <td>immediate signed <b>h</b>at (high-order bits of a 32- or 64-bit + value; low-order bits are all <code>0</code>) + </td> +</tr> +<tr> + <td>i</td> + <td>32</td> + <td>immediate signed <b>i</b>nt, or 32-bit float</td> +</tr> +<tr> + <td>l</td> + <td>64</td> + <td>immediate signed <b>l</b>ong, or 64-bit double</td> +</tr> +<tr> + <td>m</td> + <td>16</td> + <td><b>m</b>ethod constants (only used in statically linked formats)</td> +</tr> +<tr> + <td>n</td> + <td>4</td> + <td>immediate signed <b>n</b>ibble</td> +</tr> +<tr> + <td>s</td> + <td>16</td> + <td>immediate signed <b>s</b>hort</td> +</tr> +<tr> + <td>t</td> + <td>8, 16, 32</td> + <td>branch <b>t</b>arget</td> +</tr> +<tr> + <td>x</td> + <td>0</td> + <td>no additional data</td> +</tr> +</tbody> +</table> + +<h3>Syntax</h3> + +<p>The third column of the format table indicates the human-oriented +syntax for instructions which use the indicated format. Each instruction +starts with the named opcode and is optionally followed by one or +more arguments, themselves separated with commas.</p> + +<p>Wherever an argument refers to a field from the first column, the +letter for that field is indicated in the syntax, repeated once for +each four bits of the field. For example, an eight-bit field labeled +"<code>BB</code>" in the first column would also be labeled +"<code>BB</code>" in the syntax column.</p> + +<p>Arguments which name a register have the form "<code>v<i>X</i></code>". +The prefix "<code>v</code>" was chosen instead of the more common +"<code>r</code>" exactly to avoid conflicting with (non-virtual) architectures +on which a Dalvik virtual machine might be implemented which themselves +use the prefix "<code>r</code>" for their registers. (That is, this +decision makes it possible to talk about both virtual and real registers +together without the need for circumlocution.)</p> + +<p>Arguments which indicate a literal value have the form +"<code>#+<i>X</i></code>". Some formats indicate literals that only +have non-zero bits in their high-order bits; for these, the zeroes +are represented explicitly in the syntax, even though they do not +appear in the bitwise representation.</p> + +<p>Arguments which indicate a relative instruction address offset have the +form "<code>+<i>X</i></code>".</p> + +<p>Arguments which indicate a literal constant pool index have the form +"<code><i>kind</i>@<i>X</i></code>", where "<code><i>kind</i></code>" +indicates which constant pool is being referred to. Each opcode that +uses such a format explicitly allows only one kind of constant; see +the opcode reference to figure out the correspondence. The four +kinds of constant pool are "<code>string</code>" (string pool index), +"<code>type</code>" (type pool index), "<code>field</code>" (field +pool index), and "<code>meth</code>" (method pool index).</p> + +<p>Similar to the representation of constant pool indices, there are +also suggested (optional) forms that indicate prelinked offsets or +indices. These prelinked values include "<code>vtaboff</code>" +(vtable offset), "<code>fieldoff</code>" (field offset), and +"<code>iface</code>" (interface pool index).</p> + +<p>In the cases where a format value isn't explictly part of the syntax +but instead picks a variant, each variant is listed with the prefix +"<code>[<i>X</i>=<i>N</i>]</code>" (e.g., "<code>[B=2]</code>") to indicate +the correspondence.</p> + +<h2>The Formats</h2> + +<table class="format"> +<thead> +<tr> + <th>Format</th> + <th>ID</th> + <th>Syntax</th> + <th>Notable Opcodes Covered</th> +</tr> +</thead> +<tbody> +<tr> + <td>ØØ|<i>op</i></td> + <td>10x</td> + <td><i><code>op</code></i></td> + <td> </td> +</tr> +<tr> + <td rowspan="2">B|A|<i>op</i></td> + <td>12x</td> + <td><i><code>op</code></i> vA, vB</td> + <td> </td> +</tr> +<tr> + <td>11n</td> + <td><i><code>op</code></i> vA, #+B</td> + <td> </td> +</tr> +<tr> + <td rowspan="2">AA|<i>op</i></td> + <td>11x</td> + <td><i><code>op</code></i> vAA</td> + <td> </td> +</tr> +<tr> + <td>10t</td> + <td><i><code>op</code></i> +AA</td> + <td>goto</td> +</tr> +<tr> + <td>ØØ|<i>op</i> AAAA</td></td> + <td>20t</td> + <td><i><code>op</code></i> +AAAA</td> + <td>goto/16</td> +</tr> +<tr> + <td rowspan="5">AA|<i>op</i> BBBB</td> + <td>22x</td> + <td><i><code>op</code></i> vAA, vBBBB</td> + <td> </td> +</tr> +<tr> + <td>21t</td> + <td><i><code>op</code></i> vAA, +BBBB</td> + <td> </td> +</tr> +<tr> + <td>21s</td> + <td><i><code>op</code></i> vAA, #+BBBB</td> + <td> </td> +</tr> +<tr> + <td>21h</td> + <td><i><code>op</code></i> vAA, #+BBBB0000<br/> + <i><code>op</code></i> vAA, #+BBBB000000000000 + </td> + <td> </td> +</tr> +<tr> + <td>21c</td> + <td><i><code>op</code></i> vAA, type@BBBB<br/> + <i><code>op</code></i> vAA, field@BBBB<br/> + <i><code>op</code></i> vAA, string@BBBB + </td> + <td>check-cast<br/> + const-class<br/> + const-string + </td> +</tr> +<tr> + <td rowspan="2">AA|<i>op</i> CC|BB</td> + <td>23x</td> + <td><i><code>op</code></i> vAA, vBB, vCC</td> + <td> </td> +</tr> +<tr> + <td>22b</td> + <td><i><code>op</code></i> vAA, vBB, #+CC</td> + <td> </td> +</tr> +<tr> + <td rowspan="4">B|A|<i>op</i> CCCC</td> + <td>22t</td> + <td><i><code>op</code></i> vA, vB, +CCCC</td> + <td> </td> +</tr> +<tr> + <td>22s</td> + <td><i><code>op</code></i> vA, vB, #+CCCC</td> + <td> </td> +</tr> +<tr> + <td>22c</td> + <td><i><code>op</code></i> vA, vB, type@CCCC<br/> + <i><code>op</code></i> vA, vB, field@CCCC + </td> + <td>instance-of</td> +</tr> +<tr> + <td>22cs</td> + <td><i><code>op</code></i> vA, vB, fieldoff@CCCC</td> + <td><i>(suggested format for statically linked field access instructions of + format 22c)</i> + </td> +</tr> +<tr> + <td>ØØ|<i>op</i> AAAA<sub>lo</sub> AAAA<sub>hi</sub></td></td> + <td>30t</td> + <td><i><code>op</code></i> +AAAAAAAA</td> + <td>goto/32</td> +</tr> +<tr> + <td>ØØ|<i>op</i> AAAA BBBB</td> + <td>32x</td> + <td><i><code>op</code></i> vAAAA, vBBBB</td> + <td> </td> +</tr> +<tr> + <td rowspan="3">AA|<i>op</i> BBBB<sub>lo</sub> BBBB<sub>hi</sub></td> + <td>31i</td> + <td><i><code>op</code></i> vAA, #+BBBBBBBB</td> + <td> </td> +</tr> +<tr> + <td>31t</td> + <td><i><code>op</code></i> vAA, +BBBBBBBB</td> + <td> </td> +</tr> +<tr> + <td>31c</td> + <td><i><code>op</code></i> vAA, string@BBBBBBBB</td> + <td>const-string/jumbo</td> +</tr> +<tr> + <td>B|A|<i>op</i> CCCC G|F|E|D</td> + <td>35c</td> + <td><i>[<code>B=5</code>] <code>op</code></i> {vD, vE, vF, vG, vA}, + meth@CCCC<br/> + <i>[<code>B=5</code>] <code>op</code></i> {vD, vE, vF, vG, vA}, + type@CCCC<br/> + <i>[<code>B=4</code>] <code>op</code></i> {vD, vE, vF, vG}, + <i><code>kind</code></i>@CCCC<br/> + <i>[<code>B=3</code>] <code>op</code></i> {vD, vE, vF}, + <i><code>kind</code></i>@CCCC<br/> + <i>[<code>B=2</code>] <code>op</code></i> {vD, vE}, + <i><code>kind</code></i>@CCCC<br/> + <i>[<code>B=1</code>] <code>op</code></i> {vD}, + <i><code>kind</code></i>@CCCC<br/> + <i>[<code>B=0</code>] <code>op</code></i> {}, + <i><code>kind</code></i>@CCCC + </td> + <td> </td> +</tr> +<tr> + <td>B|A|<i>op</i> CCCC G|F|E|D</td> + <td>35ms</td> + + <td><i>[<code>B=5</code>] <code>op</code></i> {vD, vE, vF, vG, vA}, + vtaboff@CCCC<br/> + <i>[<code>B=4</code>] <code>op</code></i> {vD, vE, vF, vG}, + vtaboff@CCCC<br/> + <i>[<code>B=3</code>] <code>op</code></i> {vD, vE, vF}, + vtaboff@CCCC<br/> + <i>[<code>B=2</code>] <code>op</code></i> {vD, vE}, + vtaboff@CCCC<br/> + <i>[<code>B=1</code>] <code>op</code></i> {vD}, + vtaboff@CCCC<br/> + </td> + <td><i>(suggested format for statically linked <code>invoke-virtual</code> + and <code>invoke-super</code> instructions of format 35c)</i> + </td> +</tr> +<tr> + <td>B|A|<i>op</i> DDCC H|G|F|E</td> + <td>35fs</td> + <td><i>[<code>B=5</code>] <code>op</code></i> {vE, vF, vG, vH, vA}, + vtaboff@CC, iface@DD<br/> + <i>[<code>B=4</code>] <code>op</code></i> {vE, vF, vG, vH}, + vtaboff@CC, iface@DD<br/> + <i>[<code>B=3</code>] <code>op</code></i> {vE, vF, vG}, + vtaboff@CC, iface@DD<br/> + <i>[<code>B=2</code>] <code>op</code></i> {vE, vF}, + vtaboff@CC, iface@DD<br/> + <i>[<code>B=1</code>] <code>op</code></i> {vE}, + vtaboff@CC, iface@DD<br/> + </td> + <td><i>(suggested format for statically linked <code>invoke-interface</code> + instructions of format 35c)</i> + </td> +</tr> +<tr> + <td>AA|<i>op</i> BBBB CCCC</td> + <td>3rc</td> + <td><i><code>op</code></i> {vCCCC .. vNNNN}, meth@BBBB<br/> + <i><code>op</code></i> {vCCCC .. vNNNN}, type@BBBB<br/> + <p><i>(where <code>NNNN = CCCC+AA-1</code>, that is <code>A</code> + determines the count <code>0..255</code>, and <code>C</code> + determines the first register)</i></p> + </td> + <td> </td> +</tr> +<tr> + <td>AA|<i>op</i> BBBB CCCC</td> + <td>3rms</td> + <td><i><code>op</code></i> {vCCCC .. vNNNN}, vtaboff@BBBB<br/> + <p><i>(where <code>NNNN = CCCC+AA-1</code>, that is <code>A</code> + determines the count <code>0..255</code>, and <code>C</code> + determines the first register)</i></p> + </td> + <td><i>(suggested format for statically linked <code>invoke-virtual</code> + and <code>invoke-super</code> instructions of format <code>3rc</code>)</i> + </td> +</tr> +<tr> + <td>AA|<i>op</i> CCBB DDDD</td> + <td>3rfs</td> + <td><i><code>op</code></i> {vDDDD .. vNNNN}, vtaboff@BB, + iface@CC<br/> + <p><i>(where <code>NNNN = DDDD+AA-1</code>, that is <code>A</code> + determines the count <code>0..255</code>, and <code>D</code> + determines the first register)</i></p> + </td> + <td><i>(suggested format for statically linked <code>invoke-interface</code> + instructions of format <code>3rc</code>)</i> + </td> +</tr> +<tr> + <td>AA|<i>op</i> BBBB<sub>lo</sub> BBBB BBBB BBBB<sub>hi</sub></td> + <td>51l</td> + <td><i><code>op</code></i> vAA, #+BBBBBBBBBBBBBBBB</td> + <td>const-wide</td> +</tr> +</tbody> +</table> + +</body> +</html> diff --git a/docs/java-bytecode.css b/docs/java-bytecode.css new file mode 100644 index 000000000..6075c0d69 --- /dev/null +++ b/docs/java-bytecode.css @@ -0,0 +1,54 @@ +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + +h1 { + text-align: center; +} + +table { + vertical-align: top; + border-collapse: collapse; + font-family: sans-serif; +} + +td { + vertical-align: top; + background: #f8f8f8; + border-width: 0; +} + +td.outer { + width: 25%; + padding: 0; +} + +td.outer table { + width: 100%; +} + +td.outer td { + border-width: 0; + background: #f8f8f8; + padding: 1pt; + padding-left: 10pt; + padding-right: 2pt; +} + +tr.d td { + background: #dddddd; +} + +td.outer td + td + td { + font-family: monospace; + font-weight: bold; + padding-right: 5pt; +}
\ No newline at end of file diff --git a/docs/java-bytecode.html b/docs/java-bytecode.html new file mode 100644 index 000000000..691ae5490 --- /dev/null +++ b/docs/java-bytecode.html @@ -0,0 +1,228 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>Java Bytecode At A Glance</title> +<link rel="stylesheet" href="java-bytecode.css"> +</head> + +<body> + +<h1>Java Bytecode At A Glance</h1> + +<table align="center"> +<tr><td class="outer"><table> +<tr><td>0x00</td><td>0</td><td>nop</td></tr> +<tr><td>0x01</td><td>1</td><td>aconst_null</td></tr> +<tr class="d"><td>0x02</td><td>2</td><td>iconst_m1</td></tr> +<tr class="d"><td>0x03</td><td>3</td><td>iconst_0</td></tr> +<tr><td>0x04</td><td>4</td><td>iconst_1</td></tr> +<tr><td>0x05</td><td>5</td><td>iconst_2</td></tr> +<tr class="d"><td>0x06</td><td>6</td><td>iconst_3</td></tr> +<tr class="d"><td>0x07</td><td>7</td><td>iconst_4</td></tr> +<tr><td>0x08</td><td>8</td><td>iconst_5</td></tr> +<tr><td>0x09</td><td>9</td><td>lconst_0</td></tr> +<tr class="d"><td>0x0a</td><td>10</td><td>lconst_1</td></tr> +<tr class="d"><td>0x0b</td><td>11</td><td>fconst_0</td></tr> +<tr><td>0x0c</td><td>12</td><td>fconst_1</td></tr> +<tr><td>0x0d</td><td>13</td><td>fconst_2</td></tr> +<tr class="d"><td>0x0e</td><td>14</td><td>dconst_0</td></tr> +<tr class="d"><td>0x0f</td><td>15</td><td>dconst_1</td></tr> +<tr><td>0x10</td><td>16</td><td>bipush</td></tr> +<tr><td>0x11</td><td>17</td><td>sipush</td></tr> +<tr class="d"><td>0x12</td><td>18</td><td>ldc</td></tr> +<tr class="d"><td>0x13</td><td>19</td><td>ldc_w</td></tr> +<tr><td>0x14</td><td>20</td><td>ldc2_w</td></tr> +<tr><td>0x15</td><td>21</td><td>iload</td></tr> +<tr class="d"><td>0x16</td><td>22</td><td>lload</td></tr> +<tr class="d"><td>0x17</td><td>23</td><td>fload</td></tr> +<tr><td>0x18</td><td>24</td><td>dload</td></tr> +<tr><td>0x19</td><td>25</td><td>aload</td></tr> +<tr class="d"><td>0x1a</td><td>26</td><td>iload_0</td></tr> +<tr class="d"><td>0x1b</td><td>27</td><td>iload_1</td></tr> +<tr><td>0x1c</td><td>28</td><td>iload_2</td></tr> +<tr><td>0x1d</td><td>29</td><td>iload_3</td></tr> +<tr class="d"><td>0x1e</td><td>30</td><td>lload_0</td></tr> +<tr class="d"><td>0x1f</td><td>31</td><td>lload_1</td></tr> +<tr><td>0x20</td><td>32</td><td>lload_2</td></tr> +<tr><td>0x21</td><td>33</td><td>lload_3</td></tr> +<tr class="d"><td>0x22</td><td>34</td><td>fload_0</td></tr> +<tr class="d"><td>0x23</td><td>35</td><td>fload_1</td></tr> +<tr><td>0x24</td><td>36</td><td>fload_2</td></tr> +<tr><td>0x25</td><td>37</td><td>fload_3</td></tr> +<tr class="d"><td>0x26</td><td>38</td><td>dload_0</td></tr> +<tr class="d"><td>0x27</td><td>39</td><td>dload_1</td></tr> +<tr><td>0x28</td><td>40</td><td>dload_2</td></tr> +<tr><td>0x29</td><td>41</td><td>dload_3</td></tr> +<tr class="d"><td>0x2a</td><td>42</td><td>aload_0</td></tr> +<tr class="d"><td>0x2b</td><td>43</td><td>aload_1</td></tr> +<tr><td>0x2c</td><td>44</td><td>aload_2</td></tr> +<tr><td>0x2d</td><td>45</td><td>aload_3</td></tr> +<tr class="d"><td>0x2e</td><td>46</td><td>iaload</td></tr> +<tr class="d"><td>0x2f</td><td>47</td><td>laload</td></tr> +<tr><td>0x30</td><td>48</td><td>faload</td></tr> +<tr><td>0x31</td><td>49</td><td>daload</td></tr> +<tr class="d"><td>0x32</td><td>50</td><td>aaload</td></tr> +</table></td> +<td class="outer"><table> +<tr><td>0x33</td><td>51</td><td>baload</td></tr> +<tr><td>0x34</td><td>52</td><td>caload</td></tr> +<tr class="d"><td>0x35</td><td>53</td><td>saload</td></tr> +<tr class="d"><td>0x36</td><td>54</td><td>istore</td></tr> +<tr><td>0x37</td><td>55</td><td>lstore</td></tr> +<tr><td>0x38</td><td>56</td><td>fstore</td></tr> +<tr class="d"><td>0x39</td><td>57</td><td>dstore</td></tr> +<tr class="d"><td>0x3a</td><td>58</td><td>astore</td></tr> +<tr><td>0x3b</td><td>59</td><td>istore_0</td></tr> +<tr><td>0x3c</td><td>60</td><td>istore_1</td></tr> +<tr class="d"><td>0x3d</td><td>61</td><td>istore_2</td></tr> +<tr class="d"><td>0x3e</td><td>62</td><td>istore_3</td></tr> +<tr><td>0x3f</td><td>63</td><td>lstore_0</td></tr> +<tr><td>0x40</td><td>64</td><td>lstore_1</td></tr> +<tr class="d"><td>0x41</td><td>65</td><td>lstore_2</td></tr> +<tr class="d"><td>0x42</td><td>66</td><td>lstore_3</td></tr> +<tr><td>0x43</td><td>67</td><td>fstore_0</td></tr> +<tr><td>0x44</td><td>68</td><td>fstore_1</td></tr> +<tr class="d"><td>0x45</td><td>69</td><td>fstore_2</td></tr> +<tr class="d"><td>0x46</td><td>70</td><td>fstore_3</td></tr> +<tr><td>0x47</td><td>71</td><td>dstore_0</td></tr> +<tr><td>0x48</td><td>72</td><td>dstore_1</td></tr> +<tr class="d"><td>0x49</td><td>73</td><td>dstore_2</td></tr> +<tr class="d"><td>0x4a</td><td>74</td><td>dstore_3</td></tr> +<tr><td>0x4b</td><td>75</td><td>astore_0</td></tr> +<tr><td>0x4c</td><td>76</td><td>astore_1</td></tr> +<tr class="d"><td>0x4d</td><td>77</td><td>astore_2</td></tr> +<tr class="d"><td>0x4e</td><td>78</td><td>astore_3</td></tr> +<tr><td>0x4f</td><td>79</td><td>iastore</td></tr> +<tr><td>0x50</td><td>80</td><td>lastore</td></tr> +<tr class="d"><td>0x51</td><td>81</td><td>fastore</td></tr> +<tr class="d"><td>0x52</td><td>82</td><td>dastore</td></tr> +<tr><td>0x53</td><td>83</td><td>aastore</td></tr> +<tr><td>0x54</td><td>84</td><td>bastore</td></tr> +<tr class="d"><td>0x55</td><td>85</td><td>castore</td></tr> +<tr class="d"><td>0x56</td><td>86</td><td>sastore</td></tr> +<tr><td>0x57</td><td>87</td><td>pop</td></tr> +<tr><td>0x58</td><td>88</td><td>pop2</td></tr> +<tr class="d"><td>0x59</td><td>89</td><td>dup</td></tr> +<tr class="d"><td>0x5a</td><td>90</td><td>dup_x1</td></tr> +<tr><td>0x5b</td><td>91</td><td>dup_x2</td></tr> +<tr><td>0x5c</td><td>92</td><td>dup2</td></tr> +<tr class="d"><td>0x5d</td><td>93</td><td>dup2_x1</td></tr> +<tr class="d"><td>0x5e</td><td>94</td><td>dup2_x2</td></tr> +<tr><td>0x5f</td><td>95</td><td>swap</td></tr> +<tr><td>0x60</td><td>96</td><td>iadd</td></tr> +<tr class="d"><td>0x61</td><td>97</td><td>ladd</td></tr> +<tr class="d"><td>0x62</td><td>98</td><td>fadd</td></tr> +<tr><td>0x63</td><td>99</td><td>dadd</td></tr> +<tr><td>0x64</td><td>100</td><td>isub</td></tr> +<tr class="d"><td>0x65</td><td>101</td><td>lsub</td></tr> +</table></td> +<td class="outer"><table> +<tr><td>0x66</td><td>102</td><td>fsub</td></tr> +<tr><td>0x67</td><td>103</td><td>dsub</td></tr> +<tr class="d"><td>0x68</td><td>104</td><td>imul</td></tr> +<tr class="d"><td>0x69</td><td>105</td><td>lmul</td></tr> +<tr><td>0x6a</td><td>106</td><td>fmul</td></tr> +<tr><td>0x6b</td><td>107</td><td>dmul</td></tr> +<tr class="d"><td>0x6c</td><td>108</td><td>idiv</td></tr> +<tr class="d"><td>0x6d</td><td>109</td><td>ldiv</td></tr> +<tr><td>0x6e</td><td>110</td><td>fdiv</td></tr> +<tr><td>0x6f</td><td>111</td><td>ddiv</td></tr> +<tr class="d"><td>0x70</td><td>112</td><td>irem</td></tr> +<tr class="d"><td>0x71</td><td>113</td><td>lrem</td></tr> +<tr><td>0x72</td><td>114</td><td>frem</td></tr> +<tr><td>0x73</td><td>115</td><td>drem</td></tr> +<tr class="d"><td>0x74</td><td>116</td><td>ineg</td></tr> +<tr class="d"><td>0x75</td><td>117</td><td>lneg</td></tr> +<tr><td>0x76</td><td>118</td><td>fneg</td></tr> +<tr><td>0x77</td><td>119</td><td>dneg</td></tr> +<tr class="d"><td>0x78</td><td>120</td><td>ishl</td></tr> +<tr class="d"><td>0x79</td><td>121</td><td>lshl</td></tr> +<tr><td>0x7a</td><td>122</td><td>ishr</td></tr> +<tr><td>0x7b</td><td>123</td><td>lshr</td></tr> +<tr class="d"><td>0x7c</td><td>124</td><td>iushr</td></tr> +<tr class="d"><td>0x7d</td><td>125</td><td>lushr</td></tr> +<tr><td>0x7e</td><td>126</td><td>iand</td></tr> +<tr><td>0x7f</td><td>127</td><td>land</td></tr> +<tr class="d"><td>0x80</td><td>128</td><td>ior</td></tr> +<tr class="d"><td>0x81</td><td>129</td><td>lor</td></tr> +<tr><td>0x82</td><td>130</td><td>ixor</td></tr> +<tr><td>0x83</td><td>131</td><td>lxor</td></tr> +<tr class="d"><td>0x84</td><td>132</td><td>iinc</td></tr> +<tr class="d"><td>0x85</td><td>133</td><td>i2l</td></tr> +<tr><td>0x86</td><td>134</td><td>i2f</td></tr> +<tr><td>0x87</td><td>135</td><td>i2d</td></tr> +<tr class="d"><td>0x88</td><td>136</td><td>l2i</td></tr> +<tr class="d"><td>0x89</td><td>137</td><td>l2f</td></tr> +<tr><td>0x8a</td><td>138</td><td>l2d</td></tr> +<tr><td>0x8b</td><td>139</td><td>f2i</td></tr> +<tr class="d"><td>0x8c</td><td>140</td><td>f2l</td></tr> +<tr class="d"><td>0x8d</td><td>141</td><td>f2d</td></tr> +<tr><td>0x8e</td><td>142</td><td>d2i</td></tr> +<tr><td>0x8f</td><td>143</td><td>d2l</td></tr> +<tr class="d"><td>0x90</td><td>144</td><td>d2f</td></tr> +<tr class="d"><td>0x91</td><td>145</td><td>i2b</td></tr> +<tr><td>0x92</td><td>146</td><td>i2c</td></tr> +<tr><td>0x93</td><td>147</td><td>i2s</td></tr> +<tr class="d"><td>0x94</td><td>148</td><td>lcmp</td></tr> +<tr class="d"><td>0x95</td><td>149</td><td>fcmpl</td></tr> +<tr><td>0x96</td><td>150</td><td>fcmpg</td></tr> +<tr><td>0x97</td><td>151</td><td>dcmpl</td></tr> +<tr class="d"><td>0x98</td><td>152</td><td>dcmpg</td></tr> +</table></td> +<td class="outer"><table> +<tr><td>0x99</td><td>153</td><td>ifeq</td></tr> +<tr><td>0x9a</td><td>154</td><td>ifne</td></tr> +<tr class="d"><td>0x9b</td><td>155</td><td>iflt</td></tr> +<tr class="d"><td>0x9c</td><td>156</td><td>ifge</td></tr> +<tr><td>0x9d</td><td>157</td><td>ifgt</td></tr> +<tr><td>0x9e</td><td>158</td><td>ifle</td></tr> +<tr class="d"><td>0x9f</td><td>159</td><td>if_icmpeq</td></tr> +<tr class="d"><td>0xa0</td><td>160</td><td>if_icmpne</td></tr> +<tr><td>0xa1</td><td>161</td><td>if_icmplt</td></tr> +<tr><td>0xa2</td><td>162</td><td>if_icmpge</td></tr> +<tr class="d"><td>0xa3</td><td>163</td><td>if_icmpgt</td></tr> +<tr class="d"><td>0xa4</td><td>164</td><td>if_icmple</td></tr> +<tr><td>0xa5</td><td>165</td><td>if_acmpeq</td></tr> +<tr><td>0xa6</td><td>166</td><td>if_acmpne</td></tr> +<tr class="d"><td>0xa7</td><td>167</td><td>goto</td></tr> +<tr class="d"><td>0xa8</td><td>168</td><td>jsr</td></tr> +<tr><td>0xa9</td><td>169</td><td>ret</td></tr> +<tr><td>0xaa</td><td>170</td><td>tableswitch</td></tr> +<tr class="d"><td>0xab</td><td>171</td><td>lookupswitch</td></tr> +<tr class="d"><td>0xac</td><td>172</td><td>ireturn</td></tr> +<tr><td>0xad</td><td>173</td><td>lreturn</td></tr> +<tr><td>0xae</td><td>174</td><td>freturn</td></tr> +<tr class="d"><td>0xaf</td><td>175</td><td>dreturn</td></tr> +<tr class="d"><td>0xb0</td><td>176</td><td>areturn</td></tr> +<tr><td>0xb1</td><td>177</td><td>return</td></tr> +<tr><td>0xb2</td><td>178</td><td>getstatic</td></tr> +<tr class="d"><td>0xb3</td><td>179</td><td>putstatic</td></tr> +<tr class="d"><td>0xb4</td><td>180</td><td>getfield</td></tr> +<tr><td>0xb5</td><td>181</td><td>putfield</td></tr> +<tr><td>0xb6</td><td>182</td><td>invokevirtual</td></tr> +<tr class="d"><td>0xb7</td><td>183</td><td>invokespecial</td></tr> +<tr class="d"><td>0xb8</td><td>184</td><td>invokestatic</td></tr> +<tr><td>0xb9</td><td>185</td><td>invokeinterface</td></tr> +<tr><td>0xba</td><td>186</td><td><i>(unused)</i></td></tr> +<tr class="d"><td>0xbb</td><td>187</td><td>new</td></tr> +<tr class="d"><td>0xbc</td><td>188</td><td>newarray</td></tr> +<tr><td>0xbd</td><td>189</td><td>anewarray</td></tr> +<tr><td>0xbe</td><td>190</td><td>arraylength</td></tr> +<tr class="d"><td>0xbf</td><td>191</td><td>athrow</td></tr> +<tr class="d"><td>0xc0</td><td>192</td><td>checkcast</td></tr> +<tr><td>0xc1</td><td>193</td><td>instanceof</td></tr> +<tr><td>0xc2</td><td>194</td><td>monitorenter</td></tr> +<tr class="d"><td>0xc3</td><td>195</td><td>monitorexit</td></tr> +<tr class="d"><td>0xc4</td><td>196</td><td>wide</td></tr> +<tr><td>0xc5</td><td>197</td><td>multianewarray</td></tr> +<tr><td>0xc6</td><td>198</td><td>ifnull</td></tr> +<tr class="d"><td>0xc7</td><td>199</td><td>ifnonnull</td></tr> +<tr class="d"><td>0xc8</td><td>200</td><td>goto_w</td></tr> +<tr><td>0xc9</td><td>201</td><td>jsr_w</td></tr> +</table></td></tr> +</table> + +</body> +</html> diff --git a/docs/java-constraints.css b/docs/java-constraints.css new file mode 100644 index 000000000..a315a736b --- /dev/null +++ b/docs/java-constraints.css @@ -0,0 +1,59 @@ +h1 { + font-family: serif; + color: #222266; +} + +h2 { + font-family: serif; + border-top-style: solid; + border-top-width: 2px; + border-color: #ccccdd; + padding-top: 12px; + margin-top: 48px; + margin-bottom: 2px; + color: #222266; +} + +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + + +/* general for all tables */ + +table { + border-collapse: collapse; + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table th { + font-family: sans-serif; + background: #aabbff; + text-align: left; +} + +table td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + border-color: #aaaaff; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 4px; + padding-right: 6px; + background: #eeeeff; + margin-top: 4pt; + margin-bottom: 0pt; +} diff --git a/docs/java-constraints.html b/docs/java-constraints.html new file mode 100644 index 000000000..2410a1e7f --- /dev/null +++ b/docs/java-constraints.html @@ -0,0 +1,1080 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + <head> + <title>Java bytecode constraints</title> + <link rel=stylesheet href="java-constraints.css"> + </head> + + <body> + <h1> + Bytecode constraints + </h1> + + <p> + From the point of view of a piece of code written in the Java + programming language or targeted in the same way to <code>.class</code> + files, the Dalvik VM aims to behave in a way + that is fully consistent with the language's definition. + That is, the code running in Dalvik will behave the same as it + would have running in any other virtual machine. This includes + verification failures. + The Dx/Dalvik system will check roughly the same + constraints that any other VM would, except as noted in the file + <a href="verifier.html">verifier.html</a>. The following table briefly + lists all Dx/Dalvik verification constraints together their analogs + from the book <i>The Java<super>TM</super> Language Specification</i>, + second edition. In the numbering scheme, the first three + elements refer to the specification chapter, the fourth one to the + bullet inside that chapter. The failure mode specifies whether the + constraint will fail during the Dx conversion or during verification in + the VM itself. + </p> + + <h2> + Static constraints + </h2> + + <p> + Static constraints are constraints on individual elements of the bytecode. + They usually can be checked without employing control or data-flow analysis + techniques. + </p> + + <table> + <tr> + <th> + Identifier + </th> + + <th> + Description + </th> + + <th> + Spec equivalent + </th> + + <th> + Failure mode + </th> + </tr> + + <tr> + <td> + A1 + </td> + + <td> + The <code>code</code> array must not be empty. + </td> + + <td> + 4.8.1.1 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A2 + </td> + + <td> + The <code>code</code> array must not be larger than 65535 bytes. + </td> + + <td> + 4.8.1.2 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A3 + </td> + + <td> + The first opcode in <code>code</code> array must have index + <code>0</code>. + </td> + + <td> + 4.8.1.3 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A4 + </td> + + <td> + The <code>code</code> array must only contain valid opcodes. + </td> + + <td> + 4.8.1.4 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A5 + </td> + + <td> + The index of instruction <code>n+1</code> must equal the index of + instruction <code>n</code> plus the length of instruction + <code>n</code>, taking into account a possible <code>wide</code> + instruction. Opcodes modified by a <code>wide</code> instruction must + not be directly reachable. + </td> + + <td> + 4.8.1.5 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A6 + </td> + + <td> + The last instruction in <code>code</code> array must end at index + <code>code_length-1</code>. + </td> + + <td> + 4.8.1.6 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A7 + </td> + + <td> + All jump and branch targets must be opcodes within the same method. + Opcodes modified by a <code>wide</code> instruction must not be + directly reachable via a jump or branch instruction. + </td> + + <td> + 4.8.1.7 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A8 + </td> + + <td> + All targets of a <code>tableswitch</code> instruction must be opcodes + within the same method. Upper and lower bounds must be consistent. + Opcodes modified by a <code>wide</code> instruction must not be + directly reachable via a <code>tableswitch</code> instruction. + </td> + + <td> + 4.8.1.8 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A9 + </td> + + <td> + All targets of a <code>lookupswitch</code> instruction must be opcodes + within the same method. Its table must be consistent and sorted + low-to-high. Opcodes modified by a <code>wide</code> instruction must + not be directly reachable via a <code>lookupswitch</code> instruction. + </td> + + <td> + 4.8.1.9 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A10 + </td> + + <td> + The operands of <code>ldc</code> and <code>ldc_w</code> instructions + must be valid indices into the constant pool. The respective entries + must be of type <code>CONSTANT_Integer</code>, + <code>CONSTANT_Float</code>, or <code>CONSTANT_String</code>. + </td> + + <td> + 4.8.1.10 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A11 + </td> + + <td> + The operands of <code>ldc2_w</code> instructions must be valid indices + into the constant pool. The respective entries must be of type + <code>CONSTANT_Long</code> or <code>CONSTANT_Double</code>. The + subsequent constant pool entry must be valid and remain unused. + </td> + + <td> + 4.8.1.11 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A12 + </td> + + <td> + The Operands of <code>get<kind></code> and + <code>put<kind></code> instructions must be valid indices into + constant pool. The respective entries must be of type + <code>CONSTANT_Fieldref</code>. + </td> + + <td> + 4.8.1.12 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A13 + </td> + + <td> + The first two operands of <code>invokevirtual</code>, + <code>invokespecial</code>, and <code>invokestatic</code> must form a + valid 16-bit index into the constant pool. The respective entries must + be of type <code>CONSTANT_Methodref</code>. + </td> + + <td> + 4.8.1.13 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A14 + </td> + + <td> + Methods whose names start with '<' must only be invoked implicitly by + the VM, not by class file code. The only exception is the instance + initializer, which may be invoked by <code>invokespecial</code>. + </td> + + <td> + 4.8.1.14 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A15 + </td> + + <td> + The first two operands of <code>invokeinterface</code> must form a + valid 16-bit index into the constant pool. The entry must be of type + <code>CONSTANT_Interface_Methodref</code>. The third operand must + specify number of local variables and the fourth operand must always + be zero. + </td> + + <td> + 4.8.1.15 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A16 + </td> + + <td> + The operands of <code>instanceof</code>, <code>checkcast</code>, + <code>new</code>, and <code>anewarray</code> instructions must + be a valid index into the constant pool. The first two operands of + <code>multianewarray</code> instruction must form a valid 16-bit index + into the constant pool. All respective entries must be of type + <code>CONSTANT_Class</code>. + </td> + + <td> + 4.8.1.16 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A17 + </td> + + <td> + The dimensions of an array created by <code>anewarray</code> + instructions must be less than <code>256</code>. + </td> + + <td> + 4.8.1.17 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A18 + </td> + + <td> + The <code>new</code> instruction must not reference array classes, + interfaces, or abstract classes. + </td> + + <td> + 4.8.1.18 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A19 + </td> + + <td> + The type referenced by a <code>multinewarray</code> instruction must + have at least as many dimensions as specified in the instruction. The + dimensions operand must not be <code>0</code> + </td> + + <td> + 4.8.1.19 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A20 + </td> + + <td> + The type referenced by a <code>newarray</code> instruction must be a + valid, non-reference type. + </td> + + <td> + 4.8.1.20 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A21 + </td> + + <td> + The index operand of instructions explicitly referencing single-width + local variables must be non-negative and smaller than + <code>max_locals</code>. + </td> + + <td> + 4.8.1.21 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A22 + </td> + + <td> + The index operand of instructions implicitly referencing single-width + local variables must be non-negative and smaller than + <code>max_locals</code>. + </td> + + <td> + 4.8.1.22 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A23 + </td> + + <td> + The index operand of instructions explicitly referencing double-width + local variables must be non-negative and smaller than + <code>max_locals-1</code>. + </td> + + <td> + 4.8.1.23 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A24 + </td> + + <td> + The index operand of instructions implicitly referencing double-width + local variables must be non-negative and smaller than + <code>max_locals-1</code>. + </td> + + <td> + 4.8.1.24 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A25 + </td> + + <td> + The index operand of <code>wide</code> instructions explicitly + referencing single-width local variables must be non-negative and + smaller than <code>max_locals</code>. + </td> + + <td> + 4.8.1.25 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + A26 + </td> + + <td> + The index operand of <code>wide</code> instructions explicitly + referencing double-width local variables must be non-negative and + smaller than <code>max_locals-1</code>. + </td> + + <td> + 4.8.1.25 + </td> + + <td> + DX + </td> + </tr> + </table> + + <h2> + Structural constraints + </h2> + + <p> + Structural constraints are constraints on relationships between several + elements of the bytecode. They usually can't be checked without employing + control or data-flow analysis techniques. + </p> + + <table> + <tr> + <th> + Identifier + </th> + + <th> + Description + </th> + + <th> + Spec equivalent + </th> + + <th> + Failure mode + </th> + </tr> + + <tr> + <td> + B1 + </td> + + <td> + The number and types of arguments (operands and local variables) must + always match the instruction. + </td> + + <td> + 4.8.2.1 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B2 + </td> + + <td> + The operand stack must have the same depth for all executions paths + leading to an instruction. + </td> + + <td> + 4.8.2.2 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B3 + </td> + + <td> + Local variable pairs must never be broken up. + </td> + + <td> + 4.8.2.3 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B4 + </td> + + <td> + A local variable (or pair) has to be assigned first before it can be + read. + </td> + + <td> + 4.8.2.4 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B5 + </td> + + <td> + The operand stack must never grow beyond <code>max_stack</code>. + </td> + + <td> + 4.8.2.5 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B6 + </td> + + <td> + The operand stack must never underflow. + </td> + + <td> + 4.8.2.6 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B7 + </td> + + <td> + An <code>invokespecial</code> instruction must only invoke an instance + initializer or a method in the current class or one of its + superclasses. + </td> + + <td> + 4.8.2.7 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B8 + </td> + + <td> + An instance initializer must only be invoked on an uninitialized + instance residing on the operand stack. + </td> + + <td> + 4.8.2.8 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B9 + </td> + + <td> + Instance methods may only be invoked on and instance fields may only + be accessed on already initialized instances. + </td> + + <td> + 4.8.2.9 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B10 + </td> + + <td> + The must be no backwards branches with uninitialized instances on the + operand stack or in local variables. There must be no code protected + by an exception handler that contains local variables with + uninitialized instances. + </td> + + <td> + 4.8.2.10 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B11 + </td> + + <td> + An instance initializer must call another instance initializer (same + class or superclass) before any instance members can be accessed. + Exceptions are non-inherited instance fields, which can be assigned + before calling another initializer, and the <code>Object</code> class + in general. + </td> + + <td> + 4.8.2.11 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B12 + </td> + + <td> + All actual method arguments must be assignment-compatible with formal + arguments. + </td> + + <td> + 4.8.2.12 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B13 + </td> + + <td> + For each instance method invocation, the actual instance must be + assignment-compatible with the class or interface specified in the + instruction. + </td> + + <td> + 4.8.2.13 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B14 + </td> + + <td> + A returns instruction must match its method's return type. + </td> + + <td> + 4.8.2.14 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B15 + </td> + + <td> + When accessing protected members of a superclass, the actual type of + the instance being accessed must be either the current class or one + of its subclasses. + </td> + + <td> + 4.8.2.15 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B16 + </td> + + <td> + The type of a value stored into a static field must be + assignment-compatible with or convertible to the field's type. + </td> + + <td> + 4.8.2.16 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B17 + </td> + + <td> + The type of a value stored into a field must be assignment-compatible + with or convertible to the field's type. + </td> + + <td> + 4.8.2.17 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B18 + </td> + + <td> + The type of every value stored into an array must be + assignment-compatible with the array's component type. + </td> + + <td> + 4.8.2.18 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B19 + </td> + + <td> + The operand of an <code>athrow</code> instruction must be + assignment-compatible with <code>java.lang.Throwable</code>. + </td> + + <td> + 4.8.2.19 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B20 + </td> + + <td> + The last reachable instruction of a method must either be a backwards + jump or branch, a return, or an <code>athrow</code> instruction. It + must not be possible to leave the <code>code</code> array at the + bottom. + </td> + + <td> + 4.8.2.20 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B21 + </td> + + <td> + Local variable values must not be used as return addresses. + </td> + + <td> + 4.8.2.21 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B22 + </td> + + <td> + There must be a single, uniquely determined return instruction per + subroutine call. + </td> + + <td> + 4.8.2.22 + </td> + + <td> + VM + </td> + </tr> + + <tr> + <td> + B23 + </td> + + <td> + Subroutine calls must not be directly or indirectly self-recursive. + </td> + + <td> + 4.8.2.23 + </td> + + <td> + DX + </td> + </tr> + + <tr> + <td> + B24 + </td> + + <td> + <code>ReturnAddress</code> instances must not be reused. If a + subroutine returns to a <code>ReturnAddress</code> further up the + stack than where its original call instruction is located, then all + <code>ReturnAddress</code> instances further down the stack must + never be used. + </td> + + <td> + 4.8.2.24 + </td> + + <td> + DX + </td> + </tr> + + </table> + </body> +</html> diff --git a/docs/jni-tips.html b/docs/jni-tips.html new file mode 100644 index 000000000..e2c3b8509 --- /dev/null +++ b/docs/jni-tips.html @@ -0,0 +1,512 @@ +<html> + <head> + <title>Android JNI Tips</title> + <link rel=stylesheet href="android.css"> + </head> + + <body> + <h1><a name="JNI_Tips"></a>Android JNI Tips</h1> +<p> +</p><p> +</p><ul> +<li> <a href="#What_s_JNI_">What's JNI?</a> +</li> +<li> <a href="#JavaVM_and_JNIEnv">JavaVM and JNIEnv</a> + +</li> +<li> <a href="#jclassID_jmethodID_and_jfieldID">jclassID, jmethodID, and jfieldID</a> +</li> +<li> <a href="#local_vs_global_references">Local vs. Global References</a> +</li> +<li> <a href="#UTF_8_and_UTF_16_strings">UTF-8 and UTF-16 Strings</a> +</li> +<li> <a href="#Arrays">Primitive Arrays</a> +</li> +<li> <a href="#RegionCalls">Region Calls</a> +</li> +<li> <a href="#Exceptions">Exceptions</a> +</li> + +<li> <a href="#Extended_checking">Extended Checking</a> +</li> +<li> <a href="#Native_Libraries">Native Libraries</a> +</li> +<li> <a href="#64bit">64-bit Considerations</a> +</li> + +<li> <a href="#Unsupported">Unsupported Features</a> +</ul> +<p> +<noautolink> +</noautolink></p><p> +</p><h2><a name="What_s_JNI_"> </a> What's JNI? </h2> +<p> + +JNI is the Java Native Interface. It defines a way for code written in the +Java programming language to interact with native +code, e.g. functions written in C/C++. It's VM-neutral, has support for loading code from +dynamic shared libraries, and while cumbersome at times is reasonably efficient. +</p><p> +You really should read through the +<a href="http://java.sun.com/javase/6/docs/technotes/guides/jni/spec/jniTOC.html">JNI spec for J2SE 1.6</a> +to get a sense for how JNI works and what features are available. Some +aspects of the interface aren't immediately obvious on +first reading, so you may find the next few sections handy. +The more detailed <i>JNI Programmer's Guide and Specification</i> can be found +<a href="http://java.sun.com/docs/books/jni/html/jniTOC.html">here</a>. +</p><p> +</p><p> +</p><h2><a name="JavaVM_and_JNIEnv"> </a> JavaVM and JNIEnv </h2> +<p> +JNI defines two key data structures, "JavaVM" and "JNIEnv". Both of these are essentially +pointers to pointers to function tables. (In the C++ version, it's a class whose sole member +is a pointer to a function table.) The JavaVM provides the "invocation interface" functions, +which allow you to create and destroy the VM. In theory you can have multiple VMs per process, +but Android's VMs only allow one. +</p><p> +The JNIEnv provides most of the JNI functions. Your native functions all receive a JNIEnv as +the first argument. +</p><p> + +On some VMs, the JNIEnv is used for thread-local storage. For this reason, <strong>you cannot share a JNIEnv between threads</strong>. +If a piece of code has no other way to get its JNIEnv, you should share +the JavaVM, and use JavaVM->GetEnv to discover the thread's JNIEnv. +</p><p> +The C and C++ declarations of JNIEnv and JavaVM are different. "jni.h" provides different typedefs +depending on whether it's included into ".c" or ".cpp". For this reason it's a bad idea to +include JNIEnv arguments in header files included by both languages. (Put another way: if your +header file requires "#ifdef __cplusplus", you may have to do some extra work if anything in +that header refers to JNIEnv.) +</p><p> +</p><p> +</p><h2><a name="jclassID_jmethodID_and_jfieldID"> jclassID, jmethodID, and jfieldID </a></h2> +<p> +If you want to access an object's field from native code, you would do the following: +</p><p> +</p><ul> +<li> Get the class object reference for the class with <code>FindClass</code> +</li> +<li> Get the field ID for the field with <code>GetFieldID</code> +</li> +<li> Get the contents of the field with something appropriate, e.g. +<code>GetIntField</code> +</li> +</ul> +<p> +Similarly, to call a method, you'd first get a class object reference and then a method ID. The IDs are often just +pointers to internal VM data structures. Looking them up may require several string +comparisons, but once you have them the actual call to get the field or invoke the method +is very quick. +</p><p> +If performance is important, it's useful to look the values up once and cache the results +in your native code. Because we are limiting ourselves to one VM per process, it's reasonable +to store this data in a static local structure. +</p><p> +The class references, field IDs, and method IDs are guaranteed valid until the class is unloaded. Classes +are only unloaded if all classes associated with a ClassLoader can be garbage collected, +which is rare but will not be impossible in our system. The jclassID +is a class reference and <strong>must be protected</strong> with a call +to <code>NewGlobalRef</code> (see the next section). +</p><p> +If you would like to cache the IDs when a class is loaded, and automatically re-cache them +if the class is ever unloaded and reloaded, the correct way to initialize +the IDs is to add a piece of code that looks like this to the appropriate class: +</p><p> + +</p><pre> /* + * We use a class initializer to allow the native code to cache some + * field offsets. + */ + + /* + * A native function that looks up and caches interesting + * class/field/method IDs for this class. Returns false on failure. + */ + native private static boolean nativeClassInit(); + + /* + * Invoke the native initializer when the class is loaded. + */ + static { + if (!nativeClassInit()) + throw new RuntimeException("native init failed"); + } +</pre> +<p> +Create a nativeClassInit method in your C/C++ code that performs the ID lookups. The code +will be executed once, when the class is initialized. If the class is ever unloaded and +then reloaded, it will be executed again. (See the implementation of java.io.FileDescriptor +for an example in our source tree.) +</p><p> +</p><p> +</p><p> +</p><h2><a name="local_vs_global_references"> Local vs. Global References </a></h2> +<p> +Every object that JNI returns is a "local reference". This means that it's valid for the +duration of the current native method in the current thread. +<strong>Even if the object itself continues to live on after the native method returns, the reference is not valid.</strong> +This applies to all sub-classes of jobject, including jclass and jarray. +(Dalvik VM will warn you about this when -Xcheck:jni is enabled.) +</p><p> + +If you want to hold on to a reference for a longer period, you must use a "global" reference. +The <code>NewGlobalRef</code> function takes the local reference as +an argument and returns a global one: + +<p><pre>jobject* localRef = [...]; +jobject* globalRef; +globalRef = env->NewGlobalRef(localRef); +</pre> + +The global reference is guaranteed to be valid until you call +<code>DeleteGlobalRef</code>. +</p><p> +All JNI methods accept both local and global references as arguments. +</p><p> +Programmers are required to "not excessively allocate" local references. In practical terms this means +that if you're creating large numbers of local references, perhaps while running through an array of +Objects, you should free them manually with +<code>DeleteLocalRef</code> instead of letting JNI do it for you. The +VM is only required to reserve slots for +16 local references, so if you need more than that you should either delete as you go or use +<code>EnsureLocalCapacity</code> to reserve more. +</p><p> +Note: method and field IDs are just 32-bit identifiers, not object +references, and should not be passed to <code>NewGlobalRef</code>. The raw data +pointers returned by functions like <code>GetStringUTFChars</code> +and <code>GetByteArrayElements</code> are also not objects. +</p><p> +One unusual case deserves separate mention. If you attach a native +thread to the VM with AttachCurrentThread, the code you are running will +never "return" to the VM until the thread detaches from the VM. Any local +references you create will have to be deleted manually unless the thread +is about to exit or detach. +</p><p> +</p><p> +</p><p> +</p><h2><a name="UTF_8_and_UTF_16_strings"> </a> UTF-8 and UTF-16 Strings </h2> +<p> +The Java programming language uses UTF-16. For convenience, JNI provides methods that work with "modified UTF-8" encoding +as well. (Some VMs use the modified UTF-8 internally to store strings; ours do not.) The +modified encoding only supports the 8- and 16-bit forms, and stores ASCII NUL values in a 16-bit encoding. +The nice thing about it is that you can count on having C-style zero-terminated strings, +suitable for use with standard libc string functions. The down side is that you cannot pass +arbitrary UTF-8 data into the VM and expect it to work correctly. +</p><p> +It's usually best to operate with UTF-16 strings. With our current VMs, the +<code>GetStringChars</code> method +does not require a copy, whereas <code>GetStringUTFChars</code> requires a malloc and a UTF conversion. Note that +<strong>UTF-16 strings are not zero-terminated</strong>, and \u0000 is allowed, +so you need to hang on to the string length as well as +the string pointer. + +</p><p> +<strong>Don't forget to Release the strings you Get</strong>. The string functions return <code>jchar*</code> or <code>jbyte*</code>, which +are pointers to primitive types rather than local references. They are +guaranteed valid until Release is called, which means they are not +released when the native method returns. +</p><p> +</p><p> + + +</p><h2><a name="Arrays"> </a> Primitive Arrays </h2> +<p> +JNI provides functions for accessing the contents of array objects. +While arrays of objects must be accessed one entry at a time, arrays of +primitives can be read and written directly as if they were declared in C. +</p><p> +To make the interface as efficient as possible without constraining +the VM implementation, +the <code>Get<PrimitiveType>ArrayElements</code> family of calls +allows the VM to either return a pointer to the actual elements, or +allocate some memory and make a copy. Either way, the raw pointer returned +is guaranteed to be valid until the corresponding <code>Release</code> call +is issued (which implies that, if the data wasn't copied, the array object +will be pinned down and can't be relocated as part of compacting the heap). +<strong>You must Release every array you Get.</strong> Also, if the Get +call fails, you must ensure that your code doesn't try to Release a NULL +pointer later. +</p><p> +You can determine whether or not the data was copied by passing in a +non-NULL pointer for the <code>isCopy</code> argument. This is rarely +useful. +</p><p> +The <code>Release</code> call takes a <code>mode</code> argument that can +have one of three values. The actions performed by the VM depend upon +whether it returned a pointer to the actual data or a copy of it: +<ul> + <li><code>0</code> + <ul> + <li>Actual: the array object is un-pinned. + <li>Copy: data is copied back. The buffer with the copy is freed. + </ul> + <li><code>JNI_COMMIT</code> + <ul> + <li>Actual: does nothing. + <li>Copy: data is copied back. The buffer with the copy + <strong>is not freed</strong>. + </ul> + <li><code>JNI_ABORT</code> + <ul> + <li>Actual: the array object is un-pinned. Earlier + writes are <strong>not</strong> aborted. + <li>Copy: the buffer with the copy is freed; any changes to it are lost. + </ul> +</ul> +</p><p> +One reason for checking the <code>isCopy</code> flag is to know if +you need to call <code>Release</code> with <code>JNI_COMMIT</code> +after making changes to an array -- if you're alternating between making +changes and executing code that uses the contents of the array, you may be +able to +skip the no-op commit. Another possible reason for checking the flag is for +efficient handling of <code>JNI_ABORT</code>. For example, you might want +to get an array, modify it in place, pass pieces to other functions, and +then discard the changes. If you know that JNI is making a new copy for +you, there's no need to create another "editable" copy. If JNI is passing +you the original, then you do need to make your own copy. +</p><p> +Some have asserted that you can skip the <code>Release</code> call if +<code>*isCopy</code> is false. This is not the case. If no copy buffer was +allocated, then the original memory must be pinned down and can't be moved by +the garbage collector. +</p><p> +Also note that the <code>JNI_COMMIT</code> flag does NOT release the array, +and you will need to call <code>Release</code> again with a different flag +eventually. +</p><p> +</p><p> + + +</p><h2><a name="RegionCalls"> Region Calls </a></h2> + +<p> +There is an alternative to calls like <code>Get<Type>ArrayElements</code> +and <code>GetStringChars</code> that may be very helpful when all you want +to do is copy data in or out. Consider the following: +<pre> + jbyte* data = env->GetByteArrayElements(array, NULL); + if (data != NULL) { + memcpy(buffer, data, len); + env->ReleaseByteArrayElements(array, data, JNI_ABORT); + } +</pre> +<p> +This grabs the array, copies the first <code>len</code> byte +elements out of it, and then releases the array. Depending upon the VM +policies the <code>Get</code> call will either pin or copy the array contents. +We copy the data (for perhaps a second time), then call Release; in this case +we use <code>JNI_ABORT</code> so there's no chance of a third copy. +</p><p> +We can accomplish the same thing with this: +<pre> + env->GetByteArrayRegion(array, 0, len, buffer); +</pre> +</p><p> +This accomplishes the same thing, with several advantages: +<ul> + <li>Requires one JNI call instead of 3, reducing overhead. + <li>Doesn't require pinning or extra data copies. + <li>Reduces the risk of programmer error -- no need to match up + <code>Get</code> and <code>Release</code> calls. +</ul> +</p><p> +Similarly, you can use the <code>Set<Type>ArrayRegion</code> call +to copy data into an array, and <code>GetStringRegion</code> or +<code>GetStringUTFRegion</code> to copy characters out of a +<code>String</code>. + + +</p><h2><a name="Exceptions"> Exceptions </a></h2> +<p> +<strong>You may not call most JNI functions while an exception is pending.</strong> +Your code is expected to notice the exception (via the function's return value, +<code>ExceptionCheck()</code>, or <code>ExceptionOccurred()</code>) and return, +or clear the exception and handle it. +</p><p> +The only JNI functions that you are allowed to call while an exception is +pending are: +<font size="-1"><ul> + <li>DeleteGlobalRef + <li>DeleteLocalRef + <li>DeleteWeakGlobalRef + <li>ExceptionCheck + <li>ExceptionClear + <li>ExceptionDescribe + <li>ExceptionOccurred + <li>MonitorExit + <li>PopLocalFrame + <li>PushLocalFrame + <li>Release<PrimitiveType>ArrayElements + <li>ReleasePrimitiveArrayCritical + <li>ReleaseStringChars + <li>ReleaseStringCritical + <li>ReleaseStringUTFChars +</ul></font> +</p><p> +Note that exceptions thrown by interpreted code do not "leap over" native code, +and C++ exceptions thrown by native code are not handled by Dalvik. +The JNI <code>Throw</code> and <code>ThrowNew</code> instructions just +set an exception pointer in the current thread. Upon returning to the VM from +native code, the exception will be noted and handled appropriately. +</p><p> +Native code can "catch" an exception by calling <code>ExceptionCheck</code> or +<code>ExceptionOccurred</code>, and clear it with +<code>ExceptionClear</code>. As usual, +discarding exceptions without handling them can lead to problems. +</p><p> +There are no built-in functions for manipulating the Throwable object +itself, so if you want to (say) get the exception string you will need to +find the Throwable class, look up the method ID for +<code>getMessage "()Ljava/lang/String;"</code>, invoke it, and if the result +is non-NULL use <code>GetStringUTFChars</code> to get something you can +hand to printf or a LOG macro. + +</p><p> +</p><p> +</p><h2><a name="Extended_checking"> Extended Checking </a></h2> +<p> +JNI does very little error checking. Calling <code>SetFieldInt</code> +on an Object field will succeed, even if the field is marked +<code>private</code> and <code>final</code>. The +goal is to minimize the overhead on the assumption that, if you've written it in native code, +you probably did it for performance reasons. +</p><p> +Some VMs support extended checking with the "<code>-Xcheck:jni</code>" flag. If the flag is set, the VM +puts a different table of functions into the JavaVM and JNIEnv pointers. These functions do +an extended series of checks before calling the standard implementation. + +</p><p> +Some things that may be verified: +</p><p> +</p> +<ul> +<li> Check for null pointers where not allowed. +<li> +<li> Verify argument type correctness (jclass is a class object, +jfieldID points to field data, jstring is a java.lang.String). +</li> +<li> Field type correctness, e.g. don't store a HashMap in a String field. +</li> +<li> Check to see if an exception is pending on calls where pending exceptions are not legal. +</li> +<li> Check for calls to inappropriate functions between Critical get/release calls. +</li> +<li> Check that JNIEnv structs aren't being shared between threads. + +</li> +<li> Make sure local references aren't used outside their allowed lifespan. +</li> +<li> UTF-8 strings contain valid "modified UTF-8" data. +</li> +</ul> +<p>Accessibility of methods and fields (i.e. public vs. private) is not +checked. +<p> +The Dalvik VM supports the <code>-Xcheck:jni</code> flag. For a +description of how to enable it for Android apps, see +<a href="embedded-vm-control.html">Controlling the Embedded VM</a>. +It's currently enabled by default in the Android emulator and on +"engineering" device builds. + +</p><p> +JNI checks can be modified with the <code>-Xjniopts</code> command-line +flag. Currently supported values include: +</p> +<blockquote><dl> +<dt>forcecopy +<dd>When set, any function that can return a copy of the original data +(array of primitive values, UTF-16 chars) will always do so. The buffers +are over-allocated and surrounded with a guard pattern to help identify +code writing outside the buffer, and the contents are erased before the +storage is freed to trip up code that uses the data after calling Release. +<dt>warnonly +<dd>By default, JNI "warnings" cause the VM to abort. With this flag +it continues on. +</dl></blockquote> + + +</p><p> +</p><h2><a name="Native_Libraries"> Native Libraries </a></h2> +<p> +You can load native code from shared libraries with the standard +<code>System.loadLibrary()</code> call. The +preferred way to get at your native code is: +</p><p> +</p><ul> +<li> Call <code>System.loadLibrary()</code> from a static class initializer. (See the earlier example, where one is used to call nativeClassInit().) The argument is the "undecorated" library name, e.g. to load "libfubar.so" you would pass in "fubar". + +</li> +<li> Provide a native function: <code><strong>jint JNI_OnLoad(JavaVM* vm, void* reserved)</strong></code> +</li> +<li>In <code>JNI_OnLoad</code>, register all of your native methods. You +should declare +the methods "static" so the names don't take up space in the symbol table +on the device. +</li> +</ul> +<p> +The <code>JNI_OnLoad</code> function should look something like this if +written in C: +</p><blockquote><pre>jint JNI_OnLoad(JavaVM* vm, void* reserved) +{ + JNIEnv* env; + if ((*vm)->GetEnv(vm, (void**) &env, JNI_VERSION_1_4) != JNI_OK) + return -1; + + /* get class with (*env)->FindClass */ + /* register methods with (*env)->RegisterNatives */ + + return JNI_VERSION_1_4; +} +</pre></blockquote> +</p><p> +You can also call <code>System.load()</code> with the full path name of the +shared library. For Android apps, you can get the full path to the +application's private data storage area from the context object. +</p><p> +Dalvik does support "discovery" of native methods that are named in a +specific way (see <a href="http://java.sun.com/javase/6/docs/technotes/guides/jni/spec/design.html#wp615"> + the JNI spec</a> for details), but this is a less desirable +approach. It requires more space in the shared object symbol table, +loading is slower because it requires string searches through all of the +loaded shared libraries, and if a method signature is wrong you won't know +about it until the first time the method is actually used. +</p><p> + + +</p><h2><a name="64bit"> 64-bit Considerations </a></h2> + +<p> +Android is currently expected to run on 32-bit platforms. In theory it +could be built for a 64-bit system, but that is not a goal at this time. +For the most part this isn't something that you will need to worry about +when interacting with native code, +but it becomes significant if you plan to store pointers to native +structures in integer fields in an object. To support architectures +that use 64-bit pointers, <strong>you need to stash your native pointers in a +<code>long</code> field rather than an <code>int</code></strong>. + + +</p><h2><a name="Unsupported"> Unsupported Features </a></h2> +<p>All JNI 1.6 features are supported, with the following exceptions: +<ul> + <li><code>DefineClass</code> is not implemented. Dalvik does not use + Java bytecodes or class files, so passing in binary class data + doesn't work. Translation facilities may be added in a future + version of the VM.</li> + <li><code>NewWeakGlobalRef</code> and <code>DeleteWeakGlobalRef</code> + are not implemented. The + VM supports weak references, but not JNI "weak global" references. + These will be supported in a future release.</li> + <li><code>GetObjectRefType</code> (new in 1.6) is implemented but not fully + functional -- it can't always tell the difference between "local" and + "global" references.</li> +</ul> + +</p> + +<address>Copyright © 2008 The Android Open Source Project</address> + + </body> +</html> diff --git a/docs/libraries.html b/docs/libraries.html new file mode 100644 index 000000000..e1c3035f1 --- /dev/null +++ b/docs/libraries.html @@ -0,0 +1,165 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" + "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> + +<title>Dalvik Libraries</title> + +<link rel=stylesheet href="dex-format.css"> +<link href="prettify.css" type="text/css" rel="stylesheet" /> +<script type="text/javascript" src="prettify.js"></script> + +<style> +ul.code li { + font-family: monospace; +} +</style> + +</head> + +<body onload="prettyPrint()"> + +<h1 class="title">Dalvik Libraries</h1> + +<p>The Dalvik Libraries, also known as the <i>Android core libraries</i>, +implement general purpose APIs used by code written in the Java programming +language. While the libraries themselves don't depend on Android, they do form +the foundation of the Android framework. Android applications use the Dalvik +libraries both directly and indirectly for data structures, networking, +concurrency, I/O, and more.</p> + +<p>The Dalvik libraries break down into two categories:</p> + +<ul> + <li><a href="#vm-specific">Dalvik VM-specific libraries</a></li> + <li><a href="#interop">Java programming language interoperability + libraries</a></li> +</ul> + +<p>Any system claiming to be Android-compatible must implement these libraries. +Unless otherwise noted, both the signatures and the behavior of such a system +need to conform to the Android 1.0 reference implementation. Both types of +conformance will be checked by the upcoming Android Compatibility Test Suite +(CTS).</p> + +<a name="vm-specific"/><h2>Dalvik VM-specific libraries</h2> + +<p>The VM-specific libraries enable requesting or modifying VM-specific +information. Code that uses these classes is only portable across Dalvik-based +systems. The VM-specific Dalvik packages include:</p> + +<ul class="code"> + <li>dalvik.annotation</li> + <li>dalvik.bytecode</li> + <li>dalvik.system</li> +</ul> + +<a name="interop"/><h2>Java programming language interoperability libraries</h2> + +<p>This category of library provides a familiar environment for programmers +writing code in the Java programming language. Much of the implementation of +this code comes from <a href="http://harmony.apache.org/">Apache Harmony</a>. +Sometimes, we have to change the Harmony code to make it more suitable for the +memory and CPU-constrained environments targeted by Dalvik. We delineate +Dalvik-specific changes like so: + +<pre class="prettyprint"> + private static final long serialVersionUID = 8683452581122892189L; + +// BEGIN android-added + /** zero-element array */ + private static final Object[] emptyArray = new Object[0]; +// END android-added + + private transient int firstIndex; +</pre> + +<p>If you change existing Harmony code instead of just inserting new code, use +<code>android-changed</code> instead of <code>android-added</code>. These +markers help us keep track of our own changes when we pull down updates from +Harmony.</p> + +<p>Packages in this category include:</p> + +<ul class="code"> + <li>java.io</li> + <li>java.lang</li> + <li>java.lang.annotation</li> + <li>java.lang.ref</li> + <li>java.lang.reflect</li> + <li>java.math</li> + <li>java.net</li> + <li>java.nio</li> + <li>java.nio.channels</li> + <li>java.nio.channels.spi</li> + <li>java.nio.charset</li> + <li>java.nio.charset.spi</li> + <li>java.security</li> + <li>java.security.acl</li> + <li>java.security.cert</li> + <li>java.security.interfaces</li> + <li>java.security.spec</li> + <li>java.sql</li> + <li>java.text</li> + <li>java.util</li> + <li>java.util.concurrent</li> + <li>java.util.concurrent.atomic</li> + <li>java.util.concurrent.locks</li> + <li>java.util.jar</li> + <li>java.util.logging</li> + <li>java.util.prefs</li> + <li>java.util.regex</li> + <li>java.util.zip</li> + <li>javax.crypto</li> + <li>javax.crypto.interfaces</li> + <li>javax.crypto.spec</li> + <li>javax.net</li> + <li>javax.net.ssl</li> + <li>javax.security.auth</li> + <li>javax.security.auth.callback</li> + <li>javax.security.auth.login</li> + <li>javax.security.auth.x500</li> + <li>javax.security.cert</li> + <li>javax.sql</li> + <li>javax.xml</li> + <li>javax.xml.parsers</li> + <li>org.w3c.dom</li> + <li>org.xml.sax</li> + <li>org.xml.sax.ext</li> + <li>org.xml.sax.helpers</li> +</ul> + +<p>We only provide the core functionality of <code>XMLParser</code> and +<code>DocumentBuilder</code> in the XML packages. Some methods dealing with XML +schema were left out because we don't provide the corresponding packages.</p> + +<p>In addition to the aforementioned packages, we plan to support the following +packages some time in the future. We currently have an unfinished +implementation of 2D drawing and image processing.</p> + +<ul class="code"> + <li>java.awt</li> + <li>java.awt.color</li> + <li>java.awt.event</li> + <li>java.awt.font</li> + <li>java.awt.geom</li> + <li>java.awt.im</li> + <li>java.awt.im.spi</li> + <li>java.awt.image</li> + <li>java.awt.image.renderable</li> + <li>javax.imageio</li> + <li>javax.imageio.event</li> + <li>javax.imageio.metadata</li> + <li>javax.imageio.plugins.bmp</li> + <li>javax.imageio.plugins.jpeg</li> + <li>javax.imageio.spi</li> + <li>javax.imageio.stream</li> +</ul> + +<p style="margin-top: 50px">Copyright © 2008 The Android Open Source +Project</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-00-nop.html b/docs/opcodes/opcode-00-nop.html new file mode 100644 index 000000000..726f560ae --- /dev/null +++ b/docs/opcodes/opcode-00-nop.html @@ -0,0 +1,59 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>nop</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>nop</h1> + +<h2>Purpose</h2> + +<p> +Waste cycles. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>00 10x</td> + <td>nop</td> + <td> </td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<p> +None. +</p> + +<h2>Behavior</h2> + +<p> +No externally observable effects, that is, all registers and object state(s) +stay the same. The program counter silently advances to the next instruction. +</p> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-01-move.html b/docs/opcodes/opcode-01-move.html new file mode 100644 index 000000000..13c11502e --- /dev/null +++ b/docs/opcodes/opcode-01-move.html @@ -0,0 +1,92 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move</h1> + +<h2>Purpose</h2> + +<p> +Move the contents of one non-object register to another. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>01 12x</td> + <td>move vA, vB</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> source register (4 bits)</td> +</tr> +<tr> + <td>02 22x</td> + <td>move/from16 vAA, vBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> source register (16 bits)</td> +</tr> +<tr> + <td>03 32x</td> + <td>move/16 vAAAA, vBBBB</td> + <td><code>A:</code> destination register (16 bits)<br/> + <code>B:</code> source register (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be valid register indices in the current stack frame. + </li> + <li> + Register vB must be defined. + </li> + <li> + Register vB must not contain a reference value. + </li> + <li> + Register vB must not be part of a register pair. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of register vB is moved to register vA, that is, vA' = vB. + </li> + <li> + If register v(A-1) is the first half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+1) is the second half of a register pair, register v(A+1)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-04-move-wide.html b/docs/opcodes/opcode-04-move-wide.html new file mode 100644 index 000000000..42202be39 --- /dev/null +++ b/docs/opcodes/opcode-04-move-wide.html @@ -0,0 +1,106 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move-wide</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move</h1> + +<h2>Purpose</h2> + +<p> +Move the contents of one register-pair to another. +</p> +<p> +Note: It is legal to move from vN to either vN-1 or vN+1, so implementations +must arrange for both halves of a register pair to be read before anything is +written. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>04 12x</td> + <td>move-wide vA, vB</td> + <td><code>A:</code> destination register pair (4 bits)<br/> + <code>B:</code> source register pair (4 bits)</td> +</tr> +<tr> + <td>05 22x</td> + <td>move-wide/from16 vAA, vBBBB</td> + <td><code>A:</code> destination register pair (8 bits)<br/> + <code>B:</code> source register pair (16 bits)</td> +</tr> +<tr> + <td>06 32x</td> + <td>move-wide/16 vAAAA, vBBBB</td> + <td><code>A:</code> destination register pair (16 bits)<br/> + <code>B:</code> source register pair (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A+1 and B+1 must be valid register indices in the current stackframe + (which includes A and B being valid). + </li> + <li> + Register vB must be the lower half of a register pair (which excludes the + case of it containing a reference). + </li> + <li> + Both register vB and v(B+1) must be defined. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of register vB is moved to register vA, that is, vA' = vB. + </li> + <li> + The value of register v(B+1) is moved to register v(A+1), that is, v(A+1)' + = v(B+1). + </li> + <li> + If register v(A-1) is the lower half of a register pair, then v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+2) is the upper half of a register pair, then v(A+2)' + becomes undefined. + </li> + <li> + If A = B-1, then v(B+1)' becomes undefined. + </li> + <li> + If A = B+1, then v(B)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-07-move-object.html b/docs/opcodes/opcode-07-move-object.html new file mode 100644 index 000000000..e62750dc7 --- /dev/null +++ b/docs/opcodes/opcode-07-move-object.html @@ -0,0 +1,90 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move-object</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move-object</h1> + +<h2>Purpose</h2> + +<p> +Move the contents of one object-bearing register to another. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>07 12x</td> + <td>move-object vA, vB</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> source register (4 bits)</td> +</tr> +<tr> + <td>08 22x</td> + <td>move-object/from16 vAA, vBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> source register (16 bits)</td> +</tr> +<tr> + <td>09 32x</td> + <td>move-object/16 vAAAA, vBBBB</td> + <td><code>A:</code> destination register (16 bits)<br/> + <code>B:</code> source register (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be legal register indices in the current stack frame. + </li> + <li> + Register vB must be defined. + </li> + <li> + Register vB must contain a reference value (which excludes the case of it + being part of a register pair). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of register vB is moved to register vA, that is, vA' = vB. + </li> + <li> + If register v(A-1) is the lower half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+1) is the upper half of a register pair, register v(A+1)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-0a-move-result.html b/docs/opcodes/opcode-0a-move-result.html new file mode 100644 index 000000000..f043d1373 --- /dev/null +++ b/docs/opcodes/opcode-0a-move-result.html @@ -0,0 +1,96 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move-result</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move-result</h1> + +<h2>Purpose</h2> + +<p> +Move the single-word non-object result of the most recent invoke-kind into the +indicated register. This must be done as the instruction immediately after an +invoke-kind whose (single-word, non-object) result is not to be ignored; +anywhere else is invalid. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>0a 11x</td> + <td>move-result vAA</td> + <td><code>A:</code> destination register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + The instruction must be immediately preceded (in the code array) by an + invoke-kind instruction. + </li> + <li> + The instruction must be immediately reached (in the actual control flow) + through returning from this invoke-kind instruction (it must not be jumped + to). + </li> + <li> + The result delivered by the invoke-kind instruction must not be a reference + value or require a register pair. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The result delivered by the invoke-kind instruction is moved to register + vA, that is, vA' = result. + </li> + <li> + If register v(A-1) is the lower half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+1) is the upper half of a register pair, register v(A+1)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +<h2>Notes</h2> + +<p> +This instruction can also be thought of as reading the contents of a special +"result" register that is made valid and defined by executing a non-void return +instruction or a filled-new-array instruction. The execution of any other +instruction (including this one) renders this special register invalid. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-0b-move-result-wide.html b/docs/opcodes/opcode-0b-move-result-wide.html new file mode 100644 index 000000000..c3e57793e --- /dev/null +++ b/docs/opcodes/opcode-0b-move-result-wide.html @@ -0,0 +1,101 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move-result-wide</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move-result-wide</h1> + +<h2>Purpose</h2> + +<p> +Move the double-word result of the most recent invoke-kind into the indicated +register pair. This must be done as the instruction immediately after an +invoke-kind whose (double-word) result is not to be ignored; anywhere else is +invalid. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>0b 11x</td> + <td>move-result-wide vAA</td> + <td><code>A:</code> destination register pair (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A+1 must be a valid register index in the current stack frame (which + includes A itself being valid). + </li> + <li> + The instruction must be immediately preceded (in the code array) by an + invoke-kind instruction. + </li> + <li> + The instruction must be immediately reached (in the actual control flow) + through returning from this invoke-kind instruction (it must not be jumped + to). + </li> + <li> + The result delivered by the invoke-kind instruction must be either a long + or a double value. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The upper 32 bits of the result delivered by the invoke-kind instruction are + moved to register vA, that is, vA' = result >> 0x20. + </li> + <li> + The lower 32 bits of the result delivered by the invoke-kind instruction are + moved to register v(A+1), that is, v(A+1)' = result & 0xffffffff. + </li> + <li> + If register v(A-1) is the lower half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+2) is the upper half of a register pair, register v(A+2)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +<h2>Notes</h2> + +<p> +This instruction can also be thought of as reading the contents of a special +"result" register that is made valid and defined by executing a non-void return +instruction or a filled-new-array instruction. The execution of any other +instruction (including this one) renders this special register invalid. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-0c-move-result-object.html b/docs/opcodes/opcode-0c-move-result-object.html new file mode 100644 index 000000000..f065fed2e --- /dev/null +++ b/docs/opcodes/opcode-0c-move-result-object.html @@ -0,0 +1,97 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move-result-object</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move-result-object</h1> + +<h2>Purpose</h2> + +<p> +Move the object result of the most recent invoke-kind into the indicated +register. This must be done as the instruction immediately after an invoke-kind +or filled-new-array whose (object) result is not to be ignored; anywhere else +is invalid. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>0c 11x</td> + <td>move-result-object vAA</td> + <td><code>A:</code> destination register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + The instruction must be immediately preceded (in the code array) by an + invoke-kind, filled-new-array, or filled-new-array/range instruction. + </li> + <li> + The instruction must be immediately reached (in the actual control flow) + through returning from this invoke-kind instruction or by passing a + filled-new-array or filled-new-array/range instruction (it must not be + jumped to). + </li> + <li> + The result delivered by the invoke-kind instruction must be a reference + value (which excludes the case of a long and double values). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The result delivered by the invoke-kind instruction is moved to register + vA, that, is vA' = result. + </li> + <li> + If register v(A-1) is the lower half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+1) is the upper half of a register pair, register v(A+1)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +<h2>Notes</h2> + +<p> +This instruction can also be thought of as reading the contents of a special +"result" register that is made valid and defined by executing a non-void return +instruction or a filled-new-array instruction. The execution of any other +instruction (including this one) renders this special register invalid. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-0d-move-exception.html b/docs/opcodes/opcode-0d-move-exception.html new file mode 100644 index 000000000..3c3bc83d1 --- /dev/null +++ b/docs/opcodes/opcode-0d-move-exception.html @@ -0,0 +1,79 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>move-exception</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>move-exception</h1> + +<h2>Purpose</h2> + +<p> +Save a just-caught exception into the given register. This should be the first +instruction of any exception handler whose caught exception is not to be +ignored, and this instruction may only ever occur as the first instruction of an +exception handler; anywhere else is invalid. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>0d 11x</td> + <td>move-exception vAA</td> + <td><code>A:</code> destination register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + The instruction must be the first instruction (in the code array) of an + instruction handler, that is, its offset in the code array must match one of + the handlers defined for the method in the Dex file. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The active exception of the current thread is moved to register vA, that is, + vA' = exception. + </li> + <li> + If register v(A-1) is the lower half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+1) is the upper half of a register pair, register v(A+1)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-0e-return-void.html b/docs/opcodes/opcode-0e-return-void.html new file mode 100644 index 000000000..7a9431e3e --- /dev/null +++ b/docs/opcodes/opcode-0e-return-void.html @@ -0,0 +1,84 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>return-void</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>return-void</h1> + +<h2>Purpose</h2> + +<p> +Return from a void method. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>0e 10x</td> + <td>return-void</td> + <td> </td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + The return type of the current method must be void. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + If the method is synchronized, the object's monitor is released in a way + similar to the monitor-exit instruction. + </li> + <li> + The stack frame of the current method invocation is removed from the stack. + This includes all its registers becoming invalid. + </li> + <li> + If the stack is now empty, the current thread terminates. + </li> + <li> + Otherwise, the following happens: + <ul> + <li> + The stack frame that caused this method invocation becomes valid. This + includes all its registers and their old values. + </li> + <li> + Execution continues at the bytecode instruction immediately following + the invoke-kind or invoke-kind/range instruction that caused this method + invocation. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-0f-return.html b/docs/opcodes/opcode-0f-return.html new file mode 100644 index 000000000..33e0b7c2e --- /dev/null +++ b/docs/opcodes/opcode-0f-return.html @@ -0,0 +1,100 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>return</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>return</h1> + +<h2>Purpose</h2> + +<p> +Return from a single-width (32-bit) non-object value-returning method. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>0f 11x</td> + <td>return vAA</td> + <td><code>A:</code> return value register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + The return type of the current method must not be double, long, or a + reference. + </li> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + Register vA must not be part of a register pair. + </li> + <li> + The type of vA must match the return type of the method. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + If the method is synchronized, the object's monitor is released in a way + similar to the monitor-exit instruction. + </li> + <li> + The stack frame of the current method invocation is removed from the stack. + This includes all its registers becoming invalid. + </li> + <li> + If the stack is now empty, the current thread terminates. + </li> + <li> + Otherwise, the following happens: + <ul> + <li> + The stack frame that caused this method invocation becomes valid. This + includes all its registers and their old values. + </li> + <li> + Execution continues at the bytecode instruction immediately following + the invoke-kind or invoke-kind/range instruction that caused this + method invocation. + </li> + <li> + The return value can be consumed by (exactly) the first instruction + following the invoke-kind or invoke-kind/range instruction that caused + this method invocation, and this instructions needs to be a move-result + instruction. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-10-return-wide.html b/docs/opcodes/opcode-10-return-wide.html new file mode 100644 index 000000000..f1a6b83c7 --- /dev/null +++ b/docs/opcodes/opcode-10-return-wide.html @@ -0,0 +1,99 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>return-wide</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>return-wide</h1> + +<h2>Purpose</h2> + +<p> +Return from a double-width (64-bit) value-returning method. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>10 11x</td> + <td>return-wide vAA</td> + <td><code>A:</code> return value register-pair (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + The return type of the current method must be double or long. + </li> + <li> + A+1 must be a valid register index in the current stack frame (which + includes A being valid). + </li> + <li> + Register vA must be the lower half of a register pair. + </li> + <li> + The type of vA must match the return type of the method. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + If the method is synchronized, the object's monitor is released in a way + similar to the monitor-exit instruction. + </li> + <li> + The stack frame of the current method invocation is removed from the stack. + This includes all its registers becoming invalid. + </li> + <li> + If the stack is now empty, the current thread terminates. + </li> + <li> + Otherwise, the following happens: + <ul> + <li> + The stack frame that caused this method invocation becomes valid. This + includes all its registers and their old values. + </li> + <li> + Execution continues at the bytecode instruction immediately following + the invoke instruction that caused this method invocation. + </li> + <li> + The return value can be consumed by (exactly) the first instruction + following the invoke-kind or invoke-kind/range instruction that caused + this method invocation, and this instructions needs to be a + move-result-wide instruction. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-11-return-object.html b/docs/opcodes/opcode-11-return-object.html new file mode 100644 index 000000000..369233ca8 --- /dev/null +++ b/docs/opcodes/opcode-11-return-object.html @@ -0,0 +1,96 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>return-object</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>return-object</h1> + +<h2>Purpose</h2> + +<p> +Return from an object-returning method. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>11 11x</td> + <td>return-object vAA</td> + <td><code>A:</code> return value register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + The return type of the current method must be a reference. + </li> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + Register vA must be known to be reference-bearing. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + If the method is synchronized, the object's monitor is released in a way + similar to the monitor-exit instruction. + </li> + <li> + The stack frame of the current method invocation is removed from the stack. + This includes all its registers becoming invalid. + </li> + <li> + If the stack is now empty, the current thread terminates. + </li> + <li> + Otherwise, the following happens: + </li> + <ul> + <li> + The stack frame that caused this method invocation becomes valid. This + includes all its registers and their old values. + </li> + <li> + Execution continues at the bytecode instruction immediately following + the invoke instruction that caused this method invocation. + </li> + <li> + The return value can be consumed by (exactly) the first instruction + following the invoke-kind or invoke-kind/range instruction that caused + this method invocation, and this instructions needs to be a + move-result-object instruction. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-12-const.html b/docs/opcodes/opcode-12-const.html new file mode 100644 index 000000000..b2d894fe3 --- /dev/null +++ b/docs/opcodes/opcode-12-const.html @@ -0,0 +1,102 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>const</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>const</h1> + +<h2>Purpose</h2> + +<p> +Move the given literal value (sign-extended to 32 bits, if necessary) into the +specified register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>12 11n</td> + <td>const/4 vA, #+B</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> signed int (4 bits)</td> +</tr> +<tr> + <td>13 21s</td> + <td>const/16 vAA, #+BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> +</tr> +<tr> + <td>14 31i</td> + <td>const vAA, #+BBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> arbitrary 32-bit constant</td> +</tr> +<tr> + <td>15 21h</td> + <td>const/high16 vAA, #+BBBB0000</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stackframe. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + First, an adjusted value B' is determined as follows: + <ul> + <li> + If we are executing the /high16 variant, then B is left-shifted by 16 + bits, that is, B'=B << 0x10 + <li> + Otherwise, if B is a 4 bit or 16 bit constant, it is sign-extended to 32 + bits, that is, B'=sign-extended(B). + </li> + <li> + Otherwise, B'=B. + </li> + </ul> + <li> + Then, the adjusted value B' is moved into the register A, that is, vA'=B' + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper half of a register pair, v(A+1)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-16-const-wide.html b/docs/opcodes/opcode-16-const-wide.html new file mode 100644 index 000000000..ac92b04cd --- /dev/null +++ b/docs/opcodes/opcode-16-const-wide.html @@ -0,0 +1,110 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>const-wide</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>const-wide</h1> + +<h2>Purpose</h2> + +<p> +Move the given literal value (sign-extended to 64 bits) into the specified +register-pair. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>16 21s</td> + <td>const-wide/16 vAA, #+BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> +</tr> +<tr> + <td>17 31i</td> + <td>const-wide/32 vAA, #+BBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (32 bits)</td> +</tr> +<tr> + <td>18 51l</td> + <td>const-wide vAA, #+BBBBBBBBBBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> arbitrary double-width (64-bit) constant</td> +</tr> +<tr> + <td>19 21h</td> + <td>const-wide/high16 vAA, #+BBBB000000000000</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> signed int (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + First, an adjusted value B' is determined as follows: + <ul> + <li> + If we are executing the /high16 variant, then B is left-shifted by 40 + bits, that is, B'=B << 0x28 + <li> + Otherwise, if B is a 16 bit or 32 bit constant, it is sign-extended to + 64 bits, that is, B'=sign-extended(B). + </li> + <li> + Otherwise, B'=B. + </li> + </ul> + <li> + The immediate value B is moved into the register pair (vA, v(A+1)), that is, + <ul> + <li> + vA' = B << 0x20 + </li> + <li> + v(A+1)' = B & 0xffffffff + </li> + </ul> + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+2) is the upper half of a register pair, v(A+2)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-1a-const-string.html b/docs/opcodes/opcode-1a-const-string.html new file mode 100644 index 000000000..bd89b70c6 --- /dev/null +++ b/docs/opcodes/opcode-1a-const-string.html @@ -0,0 +1,85 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>const-string</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>const-string</h1> + +<h2>Purpose</h2> + +<p> +Move a reference to the string specified by the given index into the specified +register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>1a 21c</td> + <td>const-string vAA, string@BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> string index</td> +</tr> +<tr> + <td>1b 31c</td> + <td>const-string/jumbo vAA, string@BBBBBBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> string index</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + B must be a valid index into the string constant pool. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + A new java.lang.String object S is allocated on the heap and filled with the + contents of string pool entry B. + </li> + <li> + A reference to an internalized version of the new object is moved into + register vA, that is, the instruction behaves as if vA' = S.intern() was + called. + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper half of a register pair, v(A+1)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-1b-const-class.html b/docs/opcodes/opcode-1b-const-class.html new file mode 100644 index 000000000..7a33e771b --- /dev/null +++ b/docs/opcodes/opcode-1b-const-class.html @@ -0,0 +1,92 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>const-class</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>const-class</h1> + +<h2>Purpose</h2> + +<p> +Move a reference to the class specified by the given index into the specified +register. In the case where the indicated type is primitive, this will store a +reference to the primitive type's degenerate class. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>1c 21c</td> + <td>const-class vAA, type@BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> type index</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + B must be a valid index into the type constant pool. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the class C the name of which is + contained in type pool entry B. + </li> + <li> + If B refers to a primitive type, the corresponding degenerate class is used + instead. + </li> + <li> + If C has not been loaded and resolved before, it is being loaded and + resolved. All exceptions that are possible during class loading can occur at + this point. + </li> + <li> + A reference to C is moved into register vA, that is, vA' = C. + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper half of a register pair, v(A+1)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ClassNotFoundException is thrown if the class does not exist at all. + </li> + <li> + VerifyError is thrown if the class does exist, but could not be verified. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-1d-monitor-enter.html b/docs/opcodes/opcode-1d-monitor-enter.html new file mode 100644 index 000000000..cecc939a1 --- /dev/null +++ b/docs/opcodes/opcode-1d-monitor-enter.html @@ -0,0 +1,93 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>monitor-enter</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>monitor-enter</h1> + +<h2>Purpose</h2> + +<p> +Acquire the monitor for the indicated object. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>1d 11x</td> + <td>monitor-enter vAA</td> + <td><code>A:</code> reference-bearing register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index for the current stack frame. + </li> + <li> + Register vA must contain a reference to an object. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made for the current thread to acquire the monitor of the + indicated object. Various results are possible: + <ul> + <li> + If the monitor is not owned by any thread at this point, then the + current thread becomes owner of the monitor. The entry count of the + indicated object is set to 1. + </li> + <li> + Otherwise, if the monitor is owned by the same thread that attempts the + acquiration, then the entry count of the indicated object is increased + by 1. + </li> + <li> + Otherwise the monitor is owned by a different thread. The current thread + sleeps until the monitor of the object is released. Once that happens, a + new attempt to acquire the monitor is made, as described here. There is + no guarantee that the second attempt (or any subsequent attempt) will be + successful. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vA is null. + </li> + <li> + IllegalMonitorStateException if the entry count exceeds an + (implementation-dependent) upper bound for recursive monitor entries. Note + that it is unlikely this bound is ever hit, since for most implementations + the call stack will overflow before. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-1e-monitor-exit.html b/docs/opcodes/opcode-1e-monitor-exit.html new file mode 100644 index 000000000..bee711d1d --- /dev/null +++ b/docs/opcodes/opcode-1e-monitor-exit.html @@ -0,0 +1,101 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>monitor-exit</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>monitor-exit</h1> + +<h2>Purpose</h2> + +<p> +Release the monitor for the indicated object. +</p> +<p> +Note: If this instruction needs to throw an exception, it must do so as if the +pc has already advanced past the instruction. It may be useful to think of this +as the instruction successfully executing (in a sense), and the exception +getting thrown after the instruction but before the next one gets a chance to +run. This definition makes it possible for a method to use a monitor cleanup +catch-all (e.g., finally) block as the monitor cleanup for that block itself, +as a way to handle the arbitrary exceptions that might get thrown due to the +historical implementation of Thread.stop(), while still managing to have proper +monitor hygiene. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>1e 11x</td> + <td>monitor-exit vAA</td> + <td><code>A:</code> reference-bearing register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index for the current stack frame. + </li> + <li> + Register vA must contain a reference to an object. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made for the current thread to release the monitor of the + indicated object. + </li> + <li> + If the current thread is the owner, the following happens: + <ul> + <li> + The monitor's entry count is decreased by one. + </li> + <li> + If the entry count has reached zero, the monitor is released. Other + threads waiting for the same monitor have a chance to acquire it. + </li> + </ul> + </li> + <li> + Any exception that gets thrown by this instruction bears the PC of the + instruction following the monitor-exit. That is, from the point of view of + an exception handler it cannot be distinguished from the same type of + exception being thrown immediately after the monitor-exit instruction. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException is thrown if vA is null. + </li> + <li> + IllegalMonitorStateException is thrown if the current thread is not the + owner of that monitor. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-1f-check-cast.html b/docs/opcodes/opcode-1f-check-cast.html new file mode 100644 index 000000000..6325ca47b --- /dev/null +++ b/docs/opcodes/opcode-1f-check-cast.html @@ -0,0 +1,95 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>check-cast</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>check-cast</h1> + +<h2>Purpose</h2> + +<p> +Throw if the reference in the given register cannot be cast to the indicated +type. The type must be a reference type (not a primitive type). +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>1f 21c</td> + <td>check-cast vAA, type@BBBB</td> + <td><code>A:</code> reference-bearing register (8 bits)<br/> + <code>B:</code> type index (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + Register vA must contain a reference value. + </li> + <li> + B must be a valid index into the type pool. + </li> + <li> + Type pool entry B must contain a valid type descriptor for a reference type. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the class C the name of which is + contained in type pool entry B. + </li> + <li> + If C has not been loaded and resolved before, it is being loaded and + resolved. All exceptions that are possible during class loading can occur at + this point. + </li> + <li> + The run-time type of the object reference vA is compared against C. + <ul> + <li> + If vA is null, the instruction succeeds (without further effects). + </li> + <li> + If vA is assignment compatible with C according to the usual rules of + the Java programming language, the instruction succeeds (without further + effects). + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ClassCastException is thrown if vA is either not null or not assignment + compatible with C. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-20-instance-of.html b/docs/opcodes/opcode-20-instance-of.html new file mode 100644 index 000000000..5dbfef7c6 --- /dev/null +++ b/docs/opcodes/opcode-20-instance-of.html @@ -0,0 +1,100 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>instance-of</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>instance-of</h1> + +<h2>Purpose</h2> + +<p> +Store in the given destination register 1 if the indicated reference is an +instance of the given type, or 0 if not. The type must be a reference type (not +a primitive type). +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>20 22c</td> + <td>instance-of vA, vB, type@CCCC</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> reference-bearing register (4 bits)<br/> + <code>C:</code> type index (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be a valid register indices for the current stack frame. + </li> + <li> + Register vB must contain a reference value. + </li> + <li> + C must be a valid index into the type constant pool. + </li> + <li> + Type constant pool entry C must contain a valid type descriptor for a + reference type. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the class K the name of which is + contained in type pool entry C. + </li> + <li> + If K has not been loaded and resolved before, it is being loaded and + resolved. All exceptions that are possible during class loading can occur at + this point. + </li> + <li> + The run-time type of the object reference vB is compared against K. The + register vA reflects the result: + <ul> + <li> + vA' = 1 if (and only if) vB is not null and vB is assignment compatible + with K according to the usual rules of the Java programming language. + </li> + <li> + vA' = 0 otherwise + </li> + </ul> + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper half of a register pair, v(A+1)' becomes undefined. + </li> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-21-array-length.html b/docs/opcodes/opcode-21-array-length.html new file mode 100644 index 000000000..576774215 --- /dev/null +++ b/docs/opcodes/opcode-21-array-length.html @@ -0,0 +1,78 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>array-length</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>array-length</h1> + +<h2>Purpose</h2> + +<p> +Store in the given destination register the length of the indicated array, +in entries. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>21 12x</td> + <td>array-length vA, vB</td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> array reference-bearing register (4 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be valid register indices for the current stack frame. + </li> + <li> + Register vB must contain a reference to an array. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The length of the array referenced by vB is stored in vA, that is + vA' = length(vB). + </li> + <li> + If register v(A-1) is the lower half of a register pair, register v(A-1)' + becomes undefined. + </li> + <li> + If register v(A+1) is the upper half of a register pair, register v(A+1)' + becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException is thrown if the value of register vB is null. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-22-new-instance.html b/docs/opcodes/opcode-22-new-instance.html new file mode 100644 index 000000000..ec0a34719 --- /dev/null +++ b/docs/opcodes/opcode-22-new-instance.html @@ -0,0 +1,95 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>new-instance</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>new-instance</h1> + +<h2>Purpose</h2> + +<p> +Construct a new instance of the indicated type, storing a reference to it in the +destination. The type must refer to a non-array class. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>22 21c</td> + <td>new-instance vAA, type@BBBB</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> type index</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index for the current stack frame. + </li> + <li> + B must be a valid index into the type pool. + </li> + <li> + Type constant pool entry B must contain a valid type descriptor for a + non-array class. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the class C the name of which is + contained in type pool entry B. + </li> + <li> + If C has not been loaded and resolved before, it is being loaded and + resolved. All exceptions that are possible during class loading can occur at + this point. + </li> + <li> + An attempt is made to create a new instance I of C. All exceptions that are + possible during instantiation can occur at this point. + </li> + <li> + A reference to the new instance is stored in register vA, that is vA' = I. + </li> + <li> + If v(A-1) is the lower part of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper part of a register pair, v(A+1)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + All exceptions that are possible during class loading can occur. + </li> + <li> + All exceptions that are possible during instantiation can occur. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-23-new-array.html b/docs/opcodes/opcode-23-new-array.html new file mode 100644 index 000000000..be46a49b3 --- /dev/null +++ b/docs/opcodes/opcode-23-new-array.html @@ -0,0 +1,108 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>new-array</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>new-array</h1> + +<h2>Purpose</h2> + +<p> +Construct a new array of the indicated type and size. The type must be an array +type. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>23 22c</td> + <td>new-array vA, vB, type@CCCC</td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> size register<br/> + <code>C:</code> type index</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A and B must be valid register indices for the current stack frame. + </li> + <li> + Register vB must not contain a reference value. + </li> + <li> + Register vB must not be part of a register pair. + </li> + <li> + C must be a valid index into the type pool. + </li> + <li> + Type constant pool entry C must contain a valid array type descriptor. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the class K the name of which is + contained in type pool entry C. + </li> + <li> + If K has not been loaded and resolved before, it is being loaded and + resolved. All exceptions that are possible during class loading can occur at + this point. + </li> + <li> + An attempt is made to create a new instance I of K and length B. All + exceptions that are possible during instantiation can occur at this point. + </li> + <li> + All elements of the new array are initialized to null (for object arrays) or + 0 (for numeric arrays) or false (for boolean arrays). + </li> + <li> + A reference to the new array is moved to register vA, that is, vA' = I. + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper half of a register pair, v(A+1)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NegativeArraySizeException if vB < 0 + </li> + <li> + All exceptions that are possible during class loading can occur. + </li> + <li> + All exceptions that are possible during instantiation can occur. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-24-filled-new-array.html b/docs/opcodes/opcode-24-filled-new-array.html new file mode 100644 index 000000000..cba0c5d61 --- /dev/null +++ b/docs/opcodes/opcode-24-filled-new-array.html @@ -0,0 +1,144 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>filled-new-array</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>filled-new-array</h1> + +<h2>Purpose</h2> + +<p> +Construct an array of the given type and size, filling it with the supplied +contents. The type must be an array type. The array's contents must be +single-word (that is, no arrays of long or double). The constructed instance is +stored as a "result" in the same way that the method invocation instructions +store their results, so the constructed instance must be moved to a register +with a subsequent move-result-object instruction (if it is to be used). +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>24 35c</td> + <td>filled-new-array {vD, vE, vF, vG, vA}, type@CCCC</td> + <td><code>B:</code> array size and argument word count (4 bits)<br/> + <code>C:</code> type index (16 bits)<br/> + <code>D..G, A:</code> argument registers (4 bits each)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + The value B must not be greater than 5. + </li> + <li> + If B > 0, then D must be a valid register index in the current stack frame. + </li> + <li> + If B > 1, then E must be a valid register index in the current stack frame. + </li> + <li> + If B > 2, then F must be a valid register index in the current stack frame. + </li> + <li> + If B > 3, then G must be a valid register index in the current stack frame. + </li> + <li> + If B > 4, then A must be a valid register index in the current stack frame. + </li> + <li> + C must be a valid index into the type pool. + </li> + <li> + The type denoted by C must be a valid array type descriptor. + </li> + <li> + The element size of the type denoted by C must be no larger than 32 bits. + </li> + <li> + If the element type is a primitive type, then all actual arguments + (vD .. vA, depending on B) must be primitive, too. + </li> + <li> + If the element type is a reference type, then all actual arguments + (vD .. vA, depending on B) must be references, too. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the type T the name of which is + contained in type pool entry C. + </li> + <li> + If T is a reference type and it has not been loaded and resolved before, it + is being loaded and resolved. All exceptions that are possible during class + loading can occur at this point. + </li> + <li> + An attempt is made to create a new array R of type T and length B. All + exceptions that are possible during instantiation can occur at this point. + </li> + <li> + The elements of R are filled according to the following rules: + <ul> + <li> + If B > 0 then R[0] = vD + </li> + <li> + If B > 1 then R[1] = vE + </li> + <li> + If B > 2 then R[2] = vF + </li> + <li> + If B > 3 then R[3] = vG + </li> + <li> + If B > 4 then R[4] = vA + </li> + </ul> + </li> + <li> + No reference to R is stored in any register. Instead, R can be accessed by a + move-result-object instruction immediately following this filled-new-array + instruction. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NegativeArraySizeException if vB < 0 + </li> + <li> + All exceptions that are possible during class loading can occur. + </li> + <li> + All exceptions that are possible during instantiation can occur. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-25-filled-new-array-range.html b/docs/opcodes/opcode-25-filled-new-array-range.html new file mode 100644 index 000000000..944a8b164 --- /dev/null +++ b/docs/opcodes/opcode-25-filled-new-array-range.html @@ -0,0 +1,125 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>filled-new-array/range</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>filled-new-array/range</h1> + +<h2>Purpose</h2> + +<p> +Construct an array of the given type and size, filling it with the supplied +contents. Clarifications and restrictions are the same as filled-new-array, +described above. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>25 3rc</td> + <td>filled-new-array/range {vCCCC .. vNNNN}, type@BBBB</td> + <td><code>A:</code> array size and argument word count (8 bits)<br/> + <code>B:</code> type index (16 bits)<br/> + <code>C:</code> first argument register (16 bits)<br/> + <code>N = A + C - 1</code></td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + vN must be a valid register index in the current stack frame (this means + everything below vN is valid, too). + </li> + <li> + For all values I in the interval [C .. N] the following must hold: + <ul> + <li> + vI must not be part of a register pair + </li> + <li> + If the array type is a simple type, vI must be a simple type, too. + </li> + <li> + If the array type is a reference type, vI must be a reference type, too. + </li> + </ul> + </li> + <li> + B must be a valid index into the type pool. + </li> + <li> + The type denoted by B must be an array type. + </li> + <li> + The element size of the type denoted by B must be no larger than 32 bits. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + An attempt is made to get a reference to the type T the name of which is + contained in type pool entry B. + </li> + <li> + If T has not been loaded and resolved before, it is being loaded and + resolved. All exceptions that are possible during class loading can occur at + this point. + </li> + <li> + An attempt is made to create a new instance J of type T and length vA. All + exceptions that are possible during instantiation can occur at this point. + </li> + <li> + The elements of R are filled according to the following rules: + <ul> + <li> + J[0] = vC + </li> + <li> + J[1] = v(C+1) + </li> + <li> + ... + </li> + <li> + J[vA] = vN + </li> + </ul> + </li> + <li> + No reference to J is stored in any register. Instead, J can be accessed by a + move-result-object instruction immediately following this filled-new-array + instruction. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NegativeArraySizeException if vA < 0 + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-26-fill-array-data.html b/docs/opcodes/opcode-26-fill-array-data.html new file mode 100644 index 000000000..c5040a5ba --- /dev/null +++ b/docs/opcodes/opcode-26-fill-array-data.html @@ -0,0 +1,96 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>fill-array-data</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>fill-array-data</h1> + +<h2>Purpose</h2> + +<p> +Fill the given array with the indicated data. The reference must be to an array +of primitives, and the data table must match it in type and size. +</p> +<p> +Note: The address of the table is guaranteed to be even (that is, 4-byte +aligned). If the code size of the method is otherwise odd, then an extra code +unit is inserted between the main code and the table whose value is the same as +a nop. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> + <td>26 31t</td> + <td>fill-array-data vAA, +BBBBBBBB <i>(with supplemental data as specified + below in "<code>fill-array-data</code> Format")</i></td> + <td><code>A:</code> array reference (8 bits)<br/> + <code>B:</code> signed "branch" offset to table data (32 bits)</td> + </td> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + vA must be a reference-bearing register (according to data flow) and contain + an array-reference. + </li> + <li> + B must be branch offset in the same method. + </li> + <li> + The target address (PC+B) must be 4-byte aligned. + </li> + <li> + The target address must hold the pseudo-opcode 0x300. + </li> + <li> + The table entry size must match the size of the data type of the array. + </li> + <li> + The table size must be equal or smaller than the array length. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The contents of the array referenced by vA are filled with the table data, + starting from array index 0 and in the given order. + </li> + <li> + If there are less elements in the table than the array provides space for, + the remaining array elements stay untouched. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vA is null. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-27-throw.html b/docs/opcodes/opcode-27-throw.html new file mode 100644 index 000000000..f4416ce3c --- /dev/null +++ b/docs/opcodes/opcode-27-throw.html @@ -0,0 +1,80 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>throw</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>throw</h1> + +<h2>Purpose</h2> + +<p> +Throw the indicated exception. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>27 11x</td> + <td>throw vAA</td> + <td><code>A:</code> exception-bearing register (8 bits)<br/></td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + Register vA must be a reference-bearing register + </li> + <li> + Register vA must be assignment-compatible with java.lang.Throwable according + to the usual rules of the Java programming language. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + Throws the given exception vA, resulting in a search for a matching handler + according to the usual rules of the Java programming language. + </li> + <li> + If no matching handler is found for the current thread, the thread + terminates, possibly notifying its uncaught exception handler or thread + group before. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vA is null. + </li> + <li> + Otherwise, the indicated exception. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-28-goto.html b/docs/opcodes/opcode-28-goto.html new file mode 100644 index 000000000..5ca92b6f5 --- /dev/null +++ b/docs/opcodes/opcode-28-goto.html @@ -0,0 +1,73 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>goto</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>goto</h1> + +<h2>Purpose</h2> + +<p> +Unconditionally jump to the indicated instruction. +</p> +<p> +Note: The branch offset may not be 0. (A spin loop may be legally constructed +either with goto/32 or by including a nop as a target before the branch.) +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>28 10t</td> + <td>goto +AA</td> + <td><code>A:</code> signed branch offset (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must point to a valid bytecode instruction inside the current method. + </li> + <li> + A must not be 0. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The (otherwise invisible) program counter PC is set to the address of the + instruction plus the given offset, that is, PC' = PC(goto) + A. + </li> + <li> + Executions resumes at PC'. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-29-goto-16.html b/docs/opcodes/opcode-29-goto-16.html new file mode 100644 index 000000000..abe4aece7 --- /dev/null +++ b/docs/opcodes/opcode-29-goto-16.html @@ -0,0 +1,73 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>goto/16</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>goto/16</h1> + +<h2>Purpose</h2> + +<p> +Unconditionally jump to the indicated instruction. +</p> +<p> +Note: The branch offset may not be 0. (A spin loop may be legally constructed +either with goto/32 or by including a nop as a target before the branch.) +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>29 20t</td> + <td>goto/16 +AAAA</td> + <td><code>A:</code> signed branch offset (16 bits)<br/></td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must point to a valid bytecode instruction inside the current method. + </li> + <li> + A must not be 0. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The (otherwise invisible) program counter PC is set to the address of the + instruction plus the given offset, that is, PC' = PC(goto) + A. + </li> + <li> + Executions resumes at PC'. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-2a-goto-32.html b/docs/opcodes/opcode-2a-goto-32.html new file mode 100644 index 000000000..7d64ac979 --- /dev/null +++ b/docs/opcodes/opcode-2a-goto-32.html @@ -0,0 +1,66 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>goto/32</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>goto/32</h1> + +<h2>Purpose</h2> + +<p> +Unconditionally jump to the indicated instruction. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>2a 30t</td> + <td>goto/32 +AAAAAAAA</td> + <td><code>A:</code> signed branch offset (32 bits)<br/></td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must point to a valid bytecode instruction inside the current method. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The (otherwise invisible) program counter PC is set to the address of the + instruction plus the given offset, that is, PC' = PC(goto) + A. + </li> + <li> + Executions resumes at PC'. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-2b-packed-switch.html b/docs/opcodes/opcode-2b-packed-switch.html new file mode 100644 index 000000000..73125bdcc --- /dev/null +++ b/docs/opcodes/opcode-2b-packed-switch.html @@ -0,0 +1,103 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>packed-switch</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>packed-switch</h1> + +<h2>Purpose</h2> + +<p> +Jump to a new instruction based on the value in the given register, using a +table of offsets corresponding to each value in a particular integral range, or +fall through to the next instruction if there is no match. +</p> +<p> +Note: The address of the table is guaranteed to be even (that is, 4-byte +aligned). If the code size of the method is otherwise odd, then an extra code +unit is inserted between the main code and the table whose value is the same as +a nop. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>2b 31t</td> + <td>packed-switch vAA, +BBBBBBBB <i>(with supplemental data as + specified below in "<code>packed-switch</code> Format")</i></td> + <td><code>A:</code> register to test<br/> + <code>B:</code> signed "branch" offset to table data (32 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + Let PC be the address of the packed-switch instruction in the code array of + the current method. Then T = PC + B with the following properties: + <ul> + <li> + T must be 4-byte-aligned. + </li + <li> + T must be in the same method. + </li> + <li> + T must point to a packed-switch data table. + </li> + </ul> + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of vA is used as an index into the given table data. + </li> + <li> + If vA is in the range of the table, that is, if vA >= table.first_key and + vA < first_key + size, then the jump target is determined as follows: + <ul> + <li> + PC' = PC + table.targets[vA - table.firstKey]. + </li> + <li> + Execution resumes at this address. + </li> + </ul> + </li> + <li> + Otherwise execution continues at the instruction following the packed-switch + statement. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-2c-sparse-switch.html b/docs/opcodes/opcode-2c-sparse-switch.html new file mode 100644 index 000000000..cec581e59 --- /dev/null +++ b/docs/opcodes/opcode-2c-sparse-switch.html @@ -0,0 +1,110 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>sparse-switch</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>sparse-switch</h1> + +<h2>Purpose</h2> + +<p> +Jump to a new instruction based on the value in the given register, using an +ordered table of value-offset pairs, or fall through to the next instruction if +there is no match. +</p> +<p> +Note: The address of the table is guaranteed to be even (that is, 4-byte +aligned). If the code size of the method is otherwise odd, then an extra code +unit is inserted between the main code and the table whose value is the same as +a nop. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>2c 31t</td> + <td>sparse-switch vAA, +BBBBBBBB <i>(with supplemental data as + specified below in "<code>sparse-switch</code> Format")</i></td> + <td><code>A:</code> register to test<br/> + <code>B:</code> signed "branch" offset to table data (32 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stack frame. + </li> + <li> + Let PC be the address of the packed-switch instruction in the code array of + the current method. Then T = PC + B with the following properties: + <ul> + <li> + T must be 4-byte-aligned. + </li> + <li> + T must be in the same method. + </li> + <li> + T must point to a sparse-switch data table. + </li> + </ul> + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of vA is used as a lookup key inside the sparse table data. + </li> + <li> + If there exists an I with 0 <= I < table.size such that table.keys[I] = vA, + then the jump target is determined as follows: + <ul> + <li> + PC' = PC + table.targets[I]. + </li> + <li> + Execution will resume at this address. + </li> + </ul> + </li> + <li> + Otherwise execution continues at the instruction following the sparse-switch + statement. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +<h2>Notes</h2> + +<p> +The low-to-high ordering of the keys allows the VM to employ binary search for +the lookup, resulting in O(log table.size) comparisons. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-2d-cmp-kind.html b/docs/opcodes/opcode-2d-cmp-kind.html new file mode 100644 index 000000000..431ccd40d --- /dev/null +++ b/docs/opcodes/opcode-2d-cmp-kind.html @@ -0,0 +1,121 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>cmp<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>cmp<kind></h1> + +<h2>Purpose</h2> + +<p> +Perform the indicated floating point or long comparison, storing 0 if the two +arguments are equal, 1 if the second argument is larger, or -1 if the first +argument is larger. The "bias" listed for the floating point operations +indicates how NaN comparisons are treated: "Gt bias" instructions return 1 for +NaN comparisons, and "lt bias" instructions return -1. +</p> +<p> +For example, to check to see if floating point a < b, then it is advisable to +use cmpg-float; a result of -1 indicates that the test was true, and the other +values indicate it was false either due to a valid comparison or because one +or the other values was NaN. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>2d..31 23x</td> + <td>cmp<i>kind</i> vAA, vBB, vCC<br/> + 2d: cmpl-float <i>(lt bias)</i><br/> + 2e: cmpg-float <i>(gt bias)</i><br/> + 2f: cmpl-double <i>(lt bias)</i><br/> + 30: cmpg-double <i>(gt bias)</i><br/> + 31: cmp-long + </td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> first source register or pair<br/> + <code>C:</code> second source register or pair</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A, B and C must be valid register indices in the current stack frame. + </li> + <li> + For the two -float variants, both vB and vC must be of type float. + </li> + <li> + For the two -double variants, both vB and vC must be the lower part of a + register pair holding a double value. + </li> + <li> + For the -long variant, both both vB and vC must be the lower part of a + register pair holding a long value. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The values of registers vB and vC are compared. The result, which is stored + in vA, is one of the following three: + <ul> + <li> + If vB < vC, then vA'=-1. + </li> + <li> + If vB == vC, then vA'=0. + </li> + <li> + If vC > vC, then vA'=1. + </li> + </ul> + </li> + <li> + For the -float and -double variants, an addition "bias" specifies what + happens if one or both of the arguments are NaN: + <ul> + <li> + A "lt bias" results in vA'=-1. + </li> + <li> + A "gt bias" results in vA'=1. + </li> + </ul> + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + If v(A+1) is the upper half of a register pair, v(A+1)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-32-if-test.html b/docs/opcodes/opcode-32-if-test.html new file mode 100644 index 000000000..af0adb265 --- /dev/null +++ b/docs/opcodes/opcode-32-if-test.html @@ -0,0 +1,101 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>if-test</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>if-test</h1> + +<h2>Purpose</h2> + +<p> +Branch to the given destination if the given two registers' values compare as +specified. +</p> +<p> +Note: The branch offset may not be 0. (A spin loop may be legally constructed +either by branching around a backward goto or by including a nop as a target +before the branch.) +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>32..37 22t</td> + <td>if-<i>test</i> vA, vB, +CCCC<br/> + 32: if-eq<br/> + 33: if-ne<br/> + 34: if-lt<br/> + 35: if-ge<br/> + 36: if-gt<br/> + 37: if-le<br/> + </td> + <td><code>A:</code> first register to test (4 bits)<br/> + <code>B:</code> second register to test (4 bits)<br/> + <code>C:</code> signed branch offset (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A and B must be valid register indices for the current stack frame. + </li> + <li> + Registers vA and vB must not contain a reference value. + </li> + <li> + Registers vA and vB must not be part of a register pair. + </li> + <li> + Registers vA and vB must not contain a floating point value (???). + </li> + C must of a signed offset that, when added to the PC of the instruction, + points to a valid bytecode instruction inside the same method. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The values of registers vA and vB are compared according to the <test> + condition. Two results are possible: + <ul> + <li> + The condition holds. The value of C is used as a signed offset to the + address of the if-<test> instruction. Execution continues at the + resulting address. + </li> + <li> + The condition does not hold. Execution continues at the instruction + following the if-<test> instruction. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-38-if-testz.html b/docs/opcodes/opcode-38-if-testz.html new file mode 100644 index 000000000..354e89472 --- /dev/null +++ b/docs/opcodes/opcode-38-if-testz.html @@ -0,0 +1,100 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>if-<test>z</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>if-<test>z</h1> + +<h2>Purpose</h2> + +<p> +Branch to the given destination if the given register's value compares with 0 +as specified. +</p> +<p> + Note: The branch offset may not be 0. (A spin loop may be legally constructed + either by branching around a backward goto or by including a nop as a target + before the branch.) +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>38..3d 21t</td> + <td>if-<i>test</i>z vAA, +BBBB<br/> + 38: if-eqz<br/> + 39: if-nez<br/> + 3a: if-ltz<br/> + 3b: if-gez<br/> + 3c: if-gtz<br/> + 3d: if-lez<br/> + </td> + <td><code>A:</code> register to test (8 bits)<br/> + <code>B:</code> signed branch offset (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index for the current stackframe. + </li> + <li> + Register vA must not contain a reference value. + </li> + <li> + Register vA must not be part of a register pair. + </li> + <li> + Register vA must not contain a floating point value (???). + </li> + <li> + B must not be 0. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of register vA is compared to zero according to the <test> + condition. Two results are possible: + <ul> + <li> + The condition holds. The value of B is used as a signed offset to the + address of the if-<test>z instruction. Execution continues at the + resulting address. + </li> + <li> + The condition does not hold. Execution continues at the instruction + following the if-<test>z instruction. + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<p> +None. +</p> + +</body> +</html> diff --git a/docs/opcodes/opcode-44-aget.html b/docs/opcodes/opcode-44-aget.html new file mode 100644 index 000000000..b9d6f23c2 --- /dev/null +++ b/docs/opcodes/opcode-44-aget.html @@ -0,0 +1,113 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>aget<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>aget<kind></h1> + +<h2>Purpose</h2> + +<p> +Perform the identified array operation at the identified index of the given +array, storing into the value register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>44..51 23x</td> + <td><i>arrayop</i> vAA, vBB, vCC<br/> + 44: aget<br/> + 45: aget-wide<br/> + 46: aget-object<br/> + 47: aget-boolean<br/> + 48: aget-byte<br/> + 49: aget-char<br/> + 4a: aget-short<br/> + </td> + <td><code>A:</code> dest value register or pair; (8 bits)<br/> + <code>B:</code> array register (8 bits)<br/> + <code>C:</code> index register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A, B and C must be valid register indices in the current stackframe. + </li> + <li> + For the aget-wide variant, also A+1 must be a valid register index in the + current stackframe. + </li> + <li> + Register vB must contain an array reference. The component type of the + array must match the variant of the instruction. + </li> + <li> + Register vC must contain an integer value. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + For all but the -wide variant, the array element at the given index is moved + into register vA, that is, vA'=array[index]. + </li> + <li> + For the -wide variant, the array element at the given index is moved into + registers vA and v(A+1) as follows: + <ul> + <li> + vA'=array[index] >> 0x20 + </li> + <li> + v(A+1)'=array[index] & 0xffffffff; + </li> + </ul> + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + For all but the -wide variant, if v(A+1) is the upper half of a register + pair, v(A+1)' becomes undefined. + </li> + <li> + For the -wide variant, if v(A+2) is the upper half of a register pair, + v(A+2)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vB=null. + </li> + <li> + ArrayIndexOutOfBoundsException if vC < 0 or vC >= array.length. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-4b-aput.html b/docs/opcodes/opcode-4b-aput.html new file mode 100644 index 000000000..089c1cad9 --- /dev/null +++ b/docs/opcodes/opcode-4b-aput.html @@ -0,0 +1,98 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>aput<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>aput<kind></h1> + +<h2>Purpose</h2> + +<p> +Waste cycles. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>4b..51 23x</td> + <td><i>arrayop</i> vAA, vBB, vCC<br/> + 4b: aput<br/> + 4c: aput-wide<br/> + 4d: aput-object<br/> + 4e: aput-boolean<br/> + 4f: aput-byte<br/> + 50: aput-char<br/> + 51: aput-short + </td> + <td><code>A:</code> source value register or pair; (8 bits)<br/> + <code>B:</code> array register (8 bits)<br/> + <code>C:</code> index register (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A, B and C must be valid register indices in the current stack frame. + </li> + <li> + For the aget-wide variant, also A+1 must be a valid register index in the + current stack frame. + </li> + <li> + Register vB must contain an array reference. The component type of the array + must match the variant of the instruction. + </li> + <li> + Register vC must contain an integer value. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + For all but the -wide variant, the value of register vA is move into the + array element at the given index, that is, array[index]'=vA. + </li> + <li> + For the -wide variant, the registers vA and v(A+1) are moved into the array + element at the given index as follows: + <ul> + <li> + array[index]' = vA << 0x20 | v(A+1) + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vB=null. + </li> + <li> + ArrayIndexOutOfBoundsException if vC < 0 or vC >= array.length. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-52-iget.html b/docs/opcodes/opcode-52-iget.html new file mode 100644 index 000000000..837b511ce --- /dev/null +++ b/docs/opcodes/opcode-52-iget.html @@ -0,0 +1,109 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>iget<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>iget<kind></h1> + +<h2>Purpose</h2> + +<p> +Perform the identified object instance field operation with the identified +field, loading or storing into the value register. +</p> +<p> +Note: These opcodes are reasonable candidates for static linking, altering the +field argument to be a more direct offset. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>52..58 22c</td> + <td>i<i>instanceop</i> vA, vB, field@CCCC<br/> + 52: iget<br/> + 53: iget-wide<br/> + 54: iget-object<br/> + 55: iget-boolean<br/> + 56: iget-byte<br/> + 57: iget-char<br/> + 58: iget-short<br/> + </td> + <td><code>A:</code> dest value register or pair; (4 bits)<br/> + <code>B:</code> object register (4 bits)<br/> + <code>C:</code> instance field reference index (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A and B must be valid register indices in the current stackframe. + </li> + <li> + For the -wide variant, also A+1 must be a valid register index in the + current stackframe. + </li> + <li> + Register vB must contain an object reference. + </li> + <li> + C must be a valid index into the field reference pool. + </li> + <li> + The field must be an instance field. The type of the field denoted by C must + match the variant of the instruction. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of the given instance field is read from the given object and + moved into the given register vA, that is, vA'=<object>.<field>. + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li>For all but the -wide variant, if v(A+1) is the upper half of a register + pair, v(A+1)' becomes undefined. + </li> + <li> + For the -wide variant, if v(A+2) is the upper half of a register pair, + v(A+2)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if object is null. + </li> + <li> + IllegalAccessException if <object>.<field> is not visible from + the current context according to the usual visibility and access rules of + the Java programming language. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-59-iput.html b/docs/opcodes/opcode-59-iput.html new file mode 100644 index 000000000..22a3479b5 --- /dev/null +++ b/docs/opcodes/opcode-59-iput.html @@ -0,0 +1,112 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>iget<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>iget<kind></h1> + +<h2>Purpose</h2> + +<p> +Perform the identified object instance field operation with the identified +field, loading or storing into the value register. +</p> +<p> +Note: These opcodes are reasonable candidates for static linking, altering the +field argument to be a more direct offset. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>59..5f 22c</td> + <td>i<i>instanceop</i> vA, vB, field@CCCC<br/> + 59: iput<br/> + 5a: iput-wide<br/> + 5b: iput-object<br/> + 5c: iput-boolean<br/> + 5d: iput-byte<br/> + 5e: iput-char<br/> + 5f: iput-short + </td> + <td><code>A:</code> source value register or pair; (4 bits)<br/> + <code>B:</code> object register (4 bits)<br/> + <code>C:</code> instance field reference index (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A and B must be valid register indices in the current stack frame. + </li> + <li> + For the -wide variant, also A+1 must be a valid register index in the + current stack frame. + </li> + <li> + Register vB must contain an object reference. + </li> + <li> + C must be a valid index into the field reference pool. + </li> + <li> + The field must be an instance field. The type of the field denoted by C must + match the variant of the instruction. + </li> + <li> + For the -object variant, the instance referenced by register vA must be + assignment-compatible to the type of the field. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + For all but the -wide variant, the value of register vA is move into the + field, that is, <object>.<field>'=vA. + </li> + <li> + For the -wide variant, the registers vA and v(A+1) are moved into the + field as follows: + <ul> + <li> + <object>.<field>' = vA << 0x20 | v(A+1) + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vB=null. + </li> + <li> + IllegalAccessException if <object>.<field> is not visible from + the current context according to the usual visibility and access rules of + the Java programming language, or final. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-60-sget.html b/docs/opcodes/opcode-60-sget.html new file mode 100644 index 000000000..820886ef1 --- /dev/null +++ b/docs/opcodes/opcode-60-sget.html @@ -0,0 +1,106 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>sget<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>sget<kind></h1> + +<h2>Purpose</h2> + +<p> +Perform the identified object static field operation with the identified static +field, loading or storing into the value register. +</p> +<p> +Note: These opcodes are reasonable candidates for static linking, altering the +field argument to be a more direct offset. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>60..6d 21c</td> + <td>s<i>staticop</i> vAA, field@BBBB<br/> + 60: sget<br/> + 61: sget-wide<br/> + 62: sget-object<br/> + 63: sget-boolean<br/> + 64: sget-byte<br/> + 65: sget-char<br/> + 66: sget-short<br/> + </td> + <td><code>A:</code> dest value register or pair; (8 bits)<br/> + <code>B:</code> static field reference index (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stackframe. + </li> + <li> + For the -wide variant, also A+1 must be a valid register index in the + current stackframe. + </li> + <li> + B must be a valid index into the field reference pool. + </li> + <li> + The field denoted by B must be static. The type of the field denoted by B + must match the variant of the instruction. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The value of the given instance field is read from the given object and + moved into the given register vA, that is, vA'=<class>.<field>. + </li> + <li> + If v(A-1) is the lower half of a register pair, v(A-1)' becomes undefined. + </li> + <li> + For all but the -wide variant, if v(A+1) is the upper half of a register + pair, v(A+1)' becomes undefined. + </li> + <li> + For the -wide variant, if v(A+2) is the upper half of a register pair, + v(A+2)' becomes undefined. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if object is null. + </li> + <li> + IllegalAccessException if <object>.<field> is not visible from + the current context according to the usual visibility and access rules of + the Java programming language. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-67-sput.html b/docs/opcodes/opcode-67-sput.html new file mode 100644 index 000000000..b4d88bba3 --- /dev/null +++ b/docs/opcodes/opcode-67-sput.html @@ -0,0 +1,108 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>sput<kind></title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>sput<kind></h1> + +<h2>Purpose</h2> + +<p> +Perform the identified object static field operation with the identified static +field, loading or storing into the value register. +</p> +<p> +Note: These opcodes are reasonable candidates for static linking, altering the +field argument to be a more direct offset. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>67..6d 21c</td> + <td>s<i>staticop</i> vAA, field@BBBB<br/> + 67: sput<br/> + 68: sput-wide<br/> + 69: sput-object<br/> + 6a: sput-boolean<br/> + 6b: sput-byte<br/> + 6c: sput-char<br/> + 6d: sput-short + </td> + <td><code>A:</code> source value register or pair; (8 bits)<br/> + <code>B:</code> static field reference index (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + A must be a valid register index in the current stackframe. + </li> + <li> + For the -wide variant, also A+1 must be a valid register index in the + current stackframe. + </li> + <li> + B must be a valid index into the field reference pool. + </li> + <li> + The field must be static. The type of the field denoted by C must match the + variant of the instruction. + </li> + <li> + For the -object variant, the instance referenced by register vA must be + assignment-compatible to the type of the field. + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + For all but the -wide variant, the value of register vA is move into the + field, that is, <class>.<field>'=vA. + </li> + <li> + For the -wide variant, the registers vA and v(A+1) are moved into the field + as follows: + <ul> + <li> + <class>.<field>' = vA << 0x20 | v(A+1) + </li> + </ul> + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + NullPointerException if vB=null. + </li> + <li> + IllegalAccessException if <object>.<field> is not visible from + the current context according to the usual visibility and access rules of + the Java programming language, or final. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-7b-unop.html b/docs/opcodes/opcode-7b-unop.html new file mode 100644 index 000000000..1068a2555 --- /dev/null +++ b/docs/opcodes/opcode-7b-unop.html @@ -0,0 +1,108 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>unop</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>unop</h1> + +<h2>Purpose</h2> + +<p> +Perform the identified unary operation on the source register, storing the +result in the destination register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>7b..8f 12x</td> + <td><i>unop</i> vA, vB<br/> + 7b: neg-int<br/> + 7c: not-int<br/> + 7d: neg-long<br/> + 7e: not-long<br/> + 7f: neg-float<br/> + 80: neg-double<br/> + 81: int-to-long<br/> + 82: int-to-float<br/> + 83: int-to-double<br/> + 84: long-to-int<br/> + 85: long-to-float<br/> + 86: long-to-double<br/> + 87: float-to-int<br/> + 88: float-to-long<br/> + 89: float-to-double<br/> + 8a: double-to-int<br/> + 8b: double-to-long<br/> + 8c: double-to-float<br/> + 8d: int-to-byte<br/> + 8e: int-to-char<br/> + 8f: int-to-short + </td> + <td><code>A:</code> destination register or pair (4 bits)<br/> + <code>B:</code> source register or pair (4 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be valid register indices for the current stackframe. + </li> + <li> + If the input type of <unop> is double or long, also B+1 must be a + valid register index in the current stackframe. + </li> + <li> + If the output type of <unop> is double or long, also A+1 must be a + valid register index in the current stackframe. + </li> + <li> + The type of register vB must match the source type of the instruction (this + probably needs more detail). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The given operation <unop> is performed according to the semantics + specified in table XXX. + </li> + <li> + The result is stored in register vA, that is, vA'=<unop> vB. + </li> + <li> + It gets a bit messy if we want to describe all the combinations of input and + output with and without pairs here. Probably it's better to split it up. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ArithmeticException if an arithmetic error occurs during the instruction. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-90-binop.html b/docs/opcodes/opcode-90-binop.html new file mode 100644 index 000000000..cdc08a8f4 --- /dev/null +++ b/docs/opcodes/opcode-90-binop.html @@ -0,0 +1,120 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>binop</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>binop</h1> + +<h2>Purpose</h2> + +<p> +Perform the identified binary operation on the two source registers, storing +the result in the first source register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>90..af 23x</td> + <td><i>binop</i> vAA, vBB, vCC<br/> + 90: add-int<br/> + 91: sub-int<br/> + 92: mul-int<br/> + 93: div-int<br/> + 94: rem-int<br/> + 95: and-int<br/> + 96: or-int<br/> + 97: xor-int<br/> + 98: shl-int<br/> + 99: shr-int<br/> + 9a: ushr-int<br/> + 9b: add-long<br/> + 9c: sub-long<br/> + 9d: mul-long<br/> + 9e: div-long<br/> + 9f: rem-long<br/> + a0: and-long<br/> + a1: or-long<br/> + a2: xor-long<br/> + a3: shl-long<br/> + a4: shr-long<br/> + a5: ushr-long<br/> + a6: add-float<br/> + a7: sub-float<br/> + a8: mul-float<br/> + a9: div-float<br/> + aa: rem-float<br/> + ab: add-double<br/> + ac: sub-double<br/> + ad: mul-double<br/> + ae: div-double<br/> + af: rem-double + </td> + <td><code>A:</code> destination register or pair (8 bits)<br/> + <code>B:</code> first source register or pair (8 bits)<br/> + <code>C:</code> second source register or pair (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + All A, B and C must be valid register indices in the current stackframe. + </li> + <li> + For the -long and -double variants, also A+1, B+1 and C+1 must be valid + register indices. + </li> + <li> + Registers vB and vC must be defined. They must both contain values that + match the variant of the instruction (it's probably better to split this up + into multiple pages again). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The given operation <binop> is performed according to the semantics + specified in table XXX. + </li> + <li> + The result is stored in register vA, that is, vA'=<biop> vB. + </li> + <li> + For the -double and -long variants, (vA+1) is also affected. + </li> + <li> + As usual, neighboring registers might get undefined, if vA (and vA+1) were + part of a register pair originally. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ArithmeticException if an error occurs during the instruction. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-b0-binop-2addr.html b/docs/opcodes/opcode-b0-binop-2addr.html new file mode 100644 index 000000000..b3374f4a0 --- /dev/null +++ b/docs/opcodes/opcode-b0-binop-2addr.html @@ -0,0 +1,120 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>binop/2addr</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>binop/2addr</h1> + +<h2>Purpose</h2> + +<p> +Perform the identified binary operation on the two source registers, storing the +result in the first source register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>b0..cf 12x</td> + <td><i>binop</i>/2addr vA, vB<br/> + b0: add-int/2addr<br/> + b1: sub-int/2addr<br/> + b2: mul-int/2addr<br/> + b3: div-int/2addr<br/> + b4: rem-int/2addr<br/> + b5: and-int/2addr<br/> + b6: or-int/2addr<br/> + b7: xor-int/2addr<br/> + b8: shl-int/2addr<br/> + b9: shr-int/2addr<br/> + ba: ushr-int/2addr<br/> + bb: add-long/2addr<br/> + bc: sub-long/2addr<br/> + bd: mul-long/2addr<br/> + be: div-long/2addr<br/> + bf: rem-long/2addr<br/> + c0: and-long/2addr<br/> + c1: or-long/2addr<br/> + c2: xor-long/2addr<br/> + c3: shl-long/2addr<br/> + c4: shr-long/2addr<br/> + c5: ushr-long/2addr<br/> + c6: add-float/2addr<br/> + c7: sub-float/2addr<br/> + c8: mul-float/2addr<br/> + c9: div-float/2addr<br/> + ca: rem-float/2addr<br/> + cb: add-double/2addr<br/> + cc: sub-double/2addr<br/> + cd: mul-double/2addr<br/> + ce: div-double/2addr<br/> + cf: rem-double/2addr + </td> + <td><code>A:</code> destination and first source register or pair + (4 bits)<br/> + <code>B:</code> second source register or pair (4 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be valid register indices in the current stackframe. + </li> + <li> + For the -long and -double variants, also A+1 and B+1 must be valid register + indices. + </li> + <li> + Registers vA and vB must be defined. They must both contain values that + match the variant of the instruction (it's probably better to split this up + into multiple pages again). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The given operation <binop> is performed according to the semantics + specified in table XXX. + </li> + <li> + The result is stored in register vA, that is, vA'=vA <binop> vB. + </li> + <li> + For the -double and -long variants, (vA+1) is also affected. + </li> + <li> + As usual, neighboring registers might get undefined, if vA (and vA+1) were + part of a register pair originally. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ArithmeticException if an error occurs during the instruction. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-d0-binop-lit16.html b/docs/opcodes/opcode-d0-binop-lit16.html new file mode 100644 index 000000000..f9d3327b9 --- /dev/null +++ b/docs/opcodes/opcode-d0-binop-lit16.html @@ -0,0 +1,94 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>binop/lit16</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>binop/lit16</h1> + +<h2>Purpose</h2> + +<p> +Perform the indicated binary op on the indicated register (first argument) and +literal value (second argument), storing the result in the destination register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>d0..d7 22s</td> + <td><i>binop</i>/lit16 vA, vB, #+CCCC<br/> + d0: add-int/lit16<br/> + d1: rsub-int (reverse subtract)<br/> + d2: mul-int/lit16<br/> + d3: div-int/lit16<br/> + d4: rem-int/lit16<br/> + d5: and-int/lit16<br/> + d6: or-int/lit16<br/> + d7: xor-int/lit16 + </td> + <td><code>A:</code> destination register (4 bits)<br/> + <code>B:</code> source register (4 bits)<br/> + <code>C:</code> signed int constant (16 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be valid register indices in the current stackframe. + </li> + <li> + Registers vA and vB must be defined. They must both contain integer values. + </li> + <li> + C is an immediate, signed integer constant taken from the instruction stream + (actually this means there are no special requirements for C at all). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The given operation <binop> is performed according to the semantics + specified in table XXX. + </li> + <li> + Argument C is sign-extended to 32 bits before. + </li> + <li> + The result is stored in register vA, that is, vA'=vB <binop> vC. + </li> + <li> + As usual, neighboring registers might get undefined, if vA was part of a + register pair originally. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ArithmeticException if an error occurs during the instruction. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode-d8-binop-lit8.html b/docs/opcodes/opcode-d8-binop-lit8.html new file mode 100644 index 000000000..26005e9fc --- /dev/null +++ b/docs/opcodes/opcode-d8-binop-lit8.html @@ -0,0 +1,97 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> + +<html> + +<head> +<title>binop/lit8</title> +<link rel=stylesheet href="opcode.css"> +</head> + +<body> + +<h1>binop/lit8</h1> + +<h2>Purpose</h2> + +<p> +Perform the indicated binary op on the indicated register (first argument) and +literal value (second argument), storing the result in the destination register. +</p> + +<h2>Details</h2> + +<table class="instruc"> +<thead> +<tr> + <th>Op & Format</th> + <th>Mnemonic / Syntax</th> + <th>Arguments</th> +</tr> +</thead> +<tbody> +<tr> + <td>d8..e2 22b</td> + <td><i>binop</i>/lit8 vAA, vBB, #+CC<br/> + d8: add-int/lit8<br/> + d9: rsub-int/lit8<br/> + da: mul-int/lit8<br/> + db: div-int/lit8<br/> + dc: rem-int/lit8<br/> + dd: and-int/lit8<br/> + de: or-int/lit8<br/> + df: xor-int/lit8<br/> + e0: shl-int/lit8<br/> + e1: shr-int/lit8<br/> + e2: ushr-int/lit8 + </td> + <td><code>A:</code> destination register (8 bits)<br/> + <code>B:</code> source register (8 bits)<br/> + <code>C:</code> signed int constant (8 bits)</td> +</tr> +</tbody> +</table> + +<h2>Constraints</h2> + +<ul> + <li> + Both A and B must be valid register indices in the current stackframe. + </li> + <li> + Registers vA and vB must be defined. They must both contain integer values. + </li> + <li> + C is an immediate, signed integer constant taken from the instruction stream + (actually this means there are no special requirements for C at all). + </li> +</ul> + +<h2>Behavior</h2> + +<ul> + <li> + The given operation <binop> is performed according to the semantics + specified in table XXX. + </li> + <li> + Argument C is sign-extended to 32 bits before. + </li> + <li> + The result is stored in register vA, that is, vA'=vB <binop> vC. + </li> + <li> + As usual, neighboring registers might get undefined, if vA was part of a + register pair originally. + </li> +</ul> + +<h2>Exceptions</h2> + +<ul> + <li> + ArithmeticException if an error occurs during the instruction. + </li> +</ul> + +</body> +</html> diff --git a/docs/opcodes/opcode.css b/docs/opcodes/opcode.css new file mode 100644 index 000000000..c3c130424 --- /dev/null +++ b/docs/opcodes/opcode.css @@ -0,0 +1,166 @@ +h1 { + font-family: serif; + color: #222266; +} + +h2 { + font-family: serif; + border-top-style: solid; + border-top-width: 2px; + border-color: #ccccdd; + padding-top: 12px; + margin-top: 48px; + margin-bottom: 2px; + color: #222266; +} + +@media print { + table { + font-size: 8pt; + } +} + +@media screen { + table { + font-size: 10pt; + } +} + + +/* general for all tables */ + +table { + border-collapse: collapse; + margin-top: 12px; +} + +table th { + font-family: sans-serif; + background: #aabbff; + text-align: left; +} + +table td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + border-color: #aaaaff; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 4px; + padding-right: 6px; + background: #eeeeff; +} + +table td p { + margin-top: 4pt; + margin-bottom: 0pt; +} + + + +/* opcodes table */ + +table.instruc { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.instruc td { + font-family: sans-serif; + border-top-style: solid; + border-bottom-style: solid; + border-width: 1px; + padding-top: 4px; + padding-bottom: 4px; + padding-left: 2px; + padding-right: 2px; +} + +table.instruc td:first-child { + font-family: monospace; + font-size: 90%; + vertical-align: top; + width: 12%; +} + +table.instruc td:first-child + td { + font-family: monospace; + font-size: 90%; + vertical-align: top; + width: 23%; +} + +table.instruc td:first-child + td i { + font-family: sans-serif; + font-size: 90%; +} + +table.instruc td:first-child + td + td { + vertical-align: top; + width: 28%; +} + +table.instruc td:first-child + td + td + td { + vertical-align: top; + width: 37%; +} + + +/* supplemental opcode format table */ + +table.supplement { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.supplement td:first-child { + font-family: monospace; + vertical-align: top; + width: 20%; +} + +table.supplement td:first-child + td { + font-family: monospace; + vertical-align: top; + width: 20%; +} + +table.supplement td:first-child + td + td { + font-family: sans-serif; + vertical-align: top; + width: 60%; +} + + +/* math details table */ + +table.math { + margin-top: 24px; + margin-bottom: 24px; + margin-left: 48px; + margin-right: 48px; +} + +table.math td:first-child { + font-family: monospace; + vertical-align: top; + width: 10%; +} + +table.math td:first-child + td { + font-family: monospace; + vertical-align: top; + width: 30%; +} + +table.math td:first-child + td + td { + font-family: sans-serif; + vertical-align: top; + width: 60%; +} diff --git a/docs/prettify.css b/docs/prettify.css new file mode 100644 index 000000000..351152b82 --- /dev/null +++ b/docs/prettify.css @@ -0,0 +1,27 @@ +/* Pretty printing styles. Used with prettify.js. */ + +.str { color: #080; } +.kwd { color: #008; } +.com { color: #800; } +.typ { color: #606; } +.lit { color: #066; } +.pun { color: #660; } +.pln { color: #000; } +.tag { color: #008; } +.atn { color: #606; } +.atv { color: #080; } +.dec { color: #606; } +pre.prettyprint { padding: 2px; border: 1px solid #888; } + +@media print { + .str { color: #060; } + .kwd { color: #006; font-weight: bold; } + .com { color: #600; font-style: italic; } + .typ { color: #404; font-weight: bold; } + .lit { color: #044; } + .pun { color: #440; } + .pln { color: #000; } + .tag { color: #006; font-weight: bold; } + .atn { color: #404; } + .atv { color: #060; } +} diff --git a/docs/prettify.js b/docs/prettify.js new file mode 100644 index 000000000..c9541185f --- /dev/null +++ b/docs/prettify.js @@ -0,0 +1,1280 @@ +// Copyright (C) 2006 Google Inc. +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + + +/** + * @fileoverview + * some functions for browser-side pretty printing of code contained in html. + * + * The lexer should work on a number of languages including C and friends, + * Java, Python, Bash, SQL, HTML, XML, CSS, Javascript, and Makefiles. + * It works passably on Ruby, PHP and Awk and a decent subset of Perl, but, + * because of commenting conventions, doesn't work on Smalltalk, Lisp-like, or + * CAML-like languages. + * + * If there's a language not mentioned here, then I don't know it, and don't + * know whether it works. If it has a C-like, Bash-like, or XML-like syntax + * then it should work passably. + * + * Usage: + * 1) include this source file in an html page via + * <script type="text/javascript" src="/path/to/prettify.js"></script> + * 2) define style rules. See the example page for examples. + * 3) mark the <pre> and <code> tags in your source with class=prettyprint. + * You can also use the (html deprecated) <xmp> tag, but the pretty printer + * needs to do more substantial DOM manipulations to support that, so some + * css styles may not be preserved. + * That's it. I wanted to keep the API as simple as possible, so there's no + * need to specify which language the code is in. + * + * Change log: + * cbeust, 2006/08/22 + * Java annotations (start with "@") are now captured as literals ("lit") + */ + +// JSLint declarations +/*global console, document, navigator, setTimeout, window */ + +/** + * Split {@code prettyPrint} into multiple timeouts so as not to interfere with + * UI events. + * If set to {@code false}, {@code prettyPrint()} is synchronous. + */ +var PR_SHOULD_USE_CONTINUATION = true; + +/** the number of characters between tab columns */ +var PR_TAB_WIDTH = 8; + +/** Walks the DOM returning a properly escaped version of innerHTML. + * @param {Node} node + * @param {Array.<string>} out output buffer that receives chunks of HTML. + */ +var PR_normalizedHtml; + +/** Contains functions for creating and registering new language handlers. + * @type {Object} + */ +var PR; + +/** Pretty print a chunk of code. + * + * @param {string} sourceCodeHtml code as html + * @return {string} code as html, but prettier + */ +var prettyPrintOne; +/** find all the < pre > and < code > tags in the DOM with class=prettyprint + * and prettify them. + * @param {Function} opt_whenDone if specified, called when the last entry + * has been finished. + */ +var prettyPrint; + +/** browser detection. @extern */ +function _pr_isIE6() { + var isIE6 = navigator && navigator.userAgent && + /\bMSIE 6\./.test(navigator.userAgent); + _pr_isIE6 = function () { return isIE6; }; + return isIE6; +} + + +(function () { + /** Splits input on space and returns an Object mapping each non-empty part to + * true. + */ + function wordSet(words) { + words = words.split(/ /g); + var set = {}; + for (var i = words.length; --i >= 0;) { + var w = words[i]; + if (w) { set[w] = null; } + } + return set; + } + + // Keyword lists for various languages. + var FLOW_CONTROL_KEYWORDS = + "break continue do else for if return while "; + var C_KEYWORDS = FLOW_CONTROL_KEYWORDS + "auto case char const default " + + "double enum extern float goto int long register short signed sizeof " + + "static struct switch typedef union unsigned void volatile "; + var COMMON_KEYWORDS = C_KEYWORDS + "catch class delete false import " + + "new operator private protected public this throw true try "; + var CPP_KEYWORDS = COMMON_KEYWORDS + "alignof align_union asm axiom bool " + + "concept concept_map const_cast constexpr decltype " + + "dynamic_cast explicit export friend inline late_check " + + "mutable namespace nullptr reinterpret_cast static_assert static_cast " + + "template typeid typename typeof using virtual wchar_t where "; + var JAVA_KEYWORDS = COMMON_KEYWORDS + + "boolean byte extends final finally implements import instanceof null " + + "native package strictfp super synchronized throws transient "; + var CSHARP_KEYWORDS = JAVA_KEYWORDS + + "as base by checked decimal delegate descending event " + + "fixed foreach from group implicit in interface internal into is lock " + + "object out override orderby params readonly ref sbyte sealed " + + "stackalloc string select uint ulong unchecked unsafe ushort var "; + var JSCRIPT_KEYWORDS = COMMON_KEYWORDS + + "debugger eval export function get null set undefined var with " + + "Infinity NaN "; + var PERL_KEYWORDS = "caller delete die do dump elsif eval exit foreach for " + + "goto if import last local my next no our print package redo require " + + "sub undef unless until use wantarray while BEGIN END "; + var PYTHON_KEYWORDS = FLOW_CONTROL_KEYWORDS + "and as assert class def del " + + "elif except exec finally from global import in is lambda " + + "nonlocal not or pass print raise try with yield " + + "False True None "; + var RUBY_KEYWORDS = FLOW_CONTROL_KEYWORDS + "alias and begin case class def" + + " defined elsif end ensure false in module next nil not or redo rescue " + + "retry self super then true undef unless until when yield BEGIN END "; + var SH_KEYWORDS = FLOW_CONTROL_KEYWORDS + "case done elif esac eval fi " + + "function in local set then until "; + var ALL_KEYWORDS = ( + CPP_KEYWORDS + CSHARP_KEYWORDS + JSCRIPT_KEYWORDS + PERL_KEYWORDS + + PYTHON_KEYWORDS + RUBY_KEYWORDS + SH_KEYWORDS); + + // token style names. correspond to css classes + /** token style for a string literal */ + var PR_STRING = 'str'; + /** token style for a keyword */ + var PR_KEYWORD = 'kwd'; + /** token style for a comment */ + var PR_COMMENT = 'com'; + /** token style for a type */ + var PR_TYPE = 'typ'; + /** token style for a literal value. e.g. 1, null, true. */ + var PR_LITERAL = 'lit'; + /** token style for a punctuation string. */ + var PR_PUNCTUATION = 'pun'; + /** token style for a punctuation string. */ + var PR_PLAIN = 'pln'; + + /** token style for an sgml tag. */ + var PR_TAG = 'tag'; + /** token style for a markup declaration such as a DOCTYPE. */ + var PR_DECLARATION = 'dec'; + /** token style for embedded source. */ + var PR_SOURCE = 'src'; + /** token style for an sgml attribute name. */ + var PR_ATTRIB_NAME = 'atn'; + /** token style for an sgml attribute value. */ + var PR_ATTRIB_VALUE = 'atv'; + + /** + * A class that indicates a section of markup that is not code, e.g. to allow + * embedding of line numbers within code listings. + */ + var PR_NOCODE = 'nocode'; + + function isWordChar(ch) { + return (ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z'); + } + + /** Splice one array into another. + * Like the python <code> + * container[containerPosition:containerPosition + countReplaced] = inserted + * </code> + * @param {Array} inserted + * @param {Array} container modified in place + * @param {Number} containerPosition + * @param {Number} countReplaced + */ + function spliceArrayInto( + inserted, container, containerPosition, countReplaced) { + inserted.unshift(containerPosition, countReplaced || 0); + try { + container.splice.apply(container, inserted); + } finally { + inserted.splice(0, 2); + } + } + + /** A set of tokens that can precede a regular expression literal in + * javascript. + * http://www.mozilla.org/js/language/js20/rationale/syntax.html has the full + * list, but I've removed ones that might be problematic when seen in + * languages that don't support regular expression literals. + * + * <p>Specifically, I've removed any keywords that can't precede a regexp + * literal in a syntactically legal javascript program, and I've removed the + * "in" keyword since it's not a keyword in many languages, and might be used + * as a count of inches. + * @private + */ + var REGEXP_PRECEDER_PATTERN = function () { + var preceders = [ + "!", "!=", "!==", "#", "%", "%=", "&", "&&", "&&=", + "&=", "(", "*", "*=", /* "+", */ "+=", ",", /* "-", */ "-=", + "->", /*".", "..", "...", handled below */ "/", "/=", ":", "::", ";", + "<", "<<", "<<=", "<=", "=", "==", "===", ">", + ">=", ">>", ">>=", ">>>", ">>>=", "?", "@", "[", + "^", "^=", "^^", "^^=", "{", "|", "|=", "||", + "||=", "~" /* handles =~ and !~ */, + "break", "case", "continue", "delete", + "do", "else", "finally", "instanceof", + "return", "throw", "try", "typeof" + ]; + var pattern = '(?:' + + '(?:(?:^|[^0-9.])\\.{1,3})|' + // a dot that's not part of a number + '(?:(?:^|[^\\+])\\+)|' + // allow + but not ++ + '(?:(?:^|[^\\-])-)'; // allow - but not -- + for (var i = 0; i < preceders.length; ++i) { + var preceder = preceders[i]; + if (isWordChar(preceder.charAt(0))) { + pattern += '|\\b' + preceder; + } else { + pattern += '|' + preceder.replace(/([^=<>:&])/g, '\\$1'); + } + } + pattern += '|^)\\s*$'; // matches at end, and matches empty string + return new RegExp(pattern); + // CAVEAT: this does not properly handle the case where a regular + // expression immediately follows another since a regular expression may + // have flags for case-sensitivity and the like. Having regexp tokens + // adjacent is not + // valid in any language I'm aware of, so I'm punting. + // TODO: maybe style special characters inside a regexp as punctuation. + }(); + + // Define regexps here so that the interpreter doesn't have to create an + // object each time the function containing them is called. + // The language spec requires a new object created even if you don't access + // the $1 members. + var pr_amp = /&/g; + var pr_lt = /</g; + var pr_gt = />/g; + var pr_quot = /\"/g; + /** like textToHtml but escapes double quotes to be attribute safe. */ + function attribToHtml(str) { + return str.replace(pr_amp, '&') + .replace(pr_lt, '<') + .replace(pr_gt, '>') + .replace(pr_quot, '"'); + } + + /** escapest html special characters to html. */ + function textToHtml(str) { + return str.replace(pr_amp, '&') + .replace(pr_lt, '<') + .replace(pr_gt, '>'); + } + + + var pr_ltEnt = /</g; + var pr_gtEnt = />/g; + var pr_aposEnt = /'/g; + var pr_quotEnt = /"/g; + var pr_ampEnt = /&/g; + var pr_nbspEnt = / /g; + /** unescapes html to plain text. */ + function htmlToText(html) { + var pos = html.indexOf('&'); + if (pos < 0) { return html; } + // Handle numeric entities specially. We can't use functional substitution + // since that doesn't work in older versions of Safari. + // These should be rare since most browsers convert them to normal chars. + for (--pos; (pos = html.indexOf('&#', pos + 1)) >= 0;) { + var end = html.indexOf(';', pos); + if (end >= 0) { + var num = html.substring(pos + 3, end); + var radix = 10; + if (num && num.charAt(0) === 'x') { + num = num.substring(1); + radix = 16; + } + var codePoint = parseInt(num, radix); + if (!isNaN(codePoint)) { + html = (html.substring(0, pos) + String.fromCharCode(codePoint) + + html.substring(end + 1)); + } + } + } + + return html.replace(pr_ltEnt, '<') + .replace(pr_gtEnt, '>') + .replace(pr_aposEnt, "'") + .replace(pr_quotEnt, '"') + .replace(pr_ampEnt, '&') + .replace(pr_nbspEnt, ' '); + } + + /** is the given node's innerHTML normally unescaped? */ + function isRawContent(node) { + return 'XMP' === node.tagName; + } + + function normalizedHtml(node, out) { + switch (node.nodeType) { + case 1: // an element + var name = node.tagName.toLowerCase(); + out.push('<', name); + for (var i = 0; i < node.attributes.length; ++i) { + var attr = node.attributes[i]; + if (!attr.specified) { continue; } + out.push(' '); + normalizedHtml(attr, out); + } + out.push('>'); + for (var child = node.firstChild; child; child = child.nextSibling) { + normalizedHtml(child, out); + } + if (node.firstChild || !/^(?:br|link|img)$/.test(name)) { + out.push('<\/', name, '>'); + } + break; + case 2: // an attribute + out.push(node.name.toLowerCase(), '="', attribToHtml(node.value), '"'); + break; + case 3: case 4: // text + out.push(textToHtml(node.nodeValue)); + break; + } + } + + var PR_innerHtmlWorks = null; + function getInnerHtml(node) { + // inner html is hopelessly broken in Safari 2.0.4 when the content is + // an html description of well formed XML and the containing tag is a PRE + // tag, so we detect that case and emulate innerHTML. + if (null === PR_innerHtmlWorks) { + var testNode = document.createElement('PRE'); + testNode.appendChild( + document.createTextNode('<!DOCTYPE foo PUBLIC "foo bar">\n<foo />')); + PR_innerHtmlWorks = !/</.test(testNode.innerHTML); + } + + if (PR_innerHtmlWorks) { + var content = node.innerHTML; + // XMP tags contain unescaped entities so require special handling. + if (isRawContent(node)) { + content = textToHtml(content); + } + return content; + } + + var out = []; + for (var child = node.firstChild; child; child = child.nextSibling) { + normalizedHtml(child, out); + } + return out.join(''); + } + + /** returns a function that expand tabs to spaces. This function can be fed + * successive chunks of text, and will maintain its own internal state to + * keep track of how tabs are expanded. + * @return {function (string) : string} a function that takes + * plain text and return the text with tabs expanded. + * @private + */ + function makeTabExpander(tabWidth) { + var SPACES = ' '; + var charInLine = 0; + + return function (plainText) { + // walk over each character looking for tabs and newlines. + // On tabs, expand them. On newlines, reset charInLine. + // Otherwise increment charInLine + var out = null; + var pos = 0; + for (var i = 0, n = plainText.length; i < n; ++i) { + var ch = plainText.charAt(i); + + switch (ch) { + case '\t': + if (!out) { out = []; } + out.push(plainText.substring(pos, i)); + // calculate how much space we need in front of this part + // nSpaces is the amount of padding -- the number of spaces needed + // to move us to the next column, where columns occur at factors of + // tabWidth. + var nSpaces = tabWidth - (charInLine % tabWidth); + charInLine += nSpaces; + for (; nSpaces >= 0; nSpaces -= SPACES.length) { + out.push(SPACES.substring(0, nSpaces)); + } + pos = i + 1; + break; + case '\n': + charInLine = 0; + break; + default: + ++charInLine; + } + } + if (!out) { return plainText; } + out.push(plainText.substring(pos)); + return out.join(''); + }; + } + + // The below pattern matches one of the following + // (1) /[^<]+/ : A run of characters other than '<' + // (2) /<!--.*?-->/: an HTML comment + // (3) /<!\[CDATA\[.*?\]\]>/: a cdata section + // (3) /<\/?[a-zA-Z][^>]*>/ : A probably tag that should not be highlighted + // (4) /</ : A '<' that does not begin a larger chunk. Treated as 1 + var pr_chunkPattern = + /(?:[^<]+|<!--[\s\S]*?-->|<!\[CDATA\[([\s\S]*?)\]\]>|<\/?[a-zA-Z][^>]*>|<)/g; + var pr_commentPrefix = /^<!--/; + var pr_cdataPrefix = /^<\[CDATA\[/; + var pr_brPrefix = /^<br\b/i; + var pr_tagNameRe = /^<(\/?)([a-zA-Z]+)/; + + /** split markup into chunks of html tags (style null) and + * plain text (style {@link #PR_PLAIN}), converting tags which are + * significant for tokenization (<br>) into their textual equivalent. + * + * @param {string} s html where whitespace is considered significant. + * @return {Object} source code and extracted tags. + * @private + */ + function extractTags(s) { + // since the pattern has the 'g' modifier and defines no capturing groups, + // this will return a list of all chunks which we then classify and wrap as + // PR_Tokens + var matches = s.match(pr_chunkPattern); + var sourceBuf = []; + var sourceBufLen = 0; + var extractedTags = []; + if (matches) { + for (var i = 0, n = matches.length; i < n; ++i) { + var match = matches[i]; + if (match.length > 1 && match.charAt(0) === '<') { + if (pr_commentPrefix.test(match)) { continue; } + if (pr_cdataPrefix.test(match)) { + // strip CDATA prefix and suffix. Don't unescape since it's CDATA + sourceBuf.push(match.substring(9, match.length - 3)); + sourceBufLen += match.length - 12; + } else if (pr_brPrefix.test(match)) { + // <br> tags are lexically significant so convert them to text. + // This is undone later. + sourceBuf.push('\n'); + ++sourceBufLen; + } else { + if (match.indexOf(PR_NOCODE) >= 0 && isNoCodeTag(match)) { + // A <span class="nocode"> will start a section that should be + // ignored. Continue walking the list until we see a matching end + // tag. + var name = match.match(pr_tagNameRe)[2]; + var depth = 1; + end_tag_loop: + for (var j = i + 1; j < n; ++j) { + var name2 = matches[j].match(pr_tagNameRe); + if (name2 && name2[2] === name) { + if (name2[1] === '/') { + if (--depth === 0) { break end_tag_loop; } + } else { + ++depth; + } + } + } + if (j < n) { + extractedTags.push( + sourceBufLen, matches.slice(i, j + 1).join('')); + i = j; + } else { // Ignore unclosed sections. + extractedTags.push(sourceBufLen, match); + } + } else { + extractedTags.push(sourceBufLen, match); + } + } + } else { + var literalText = htmlToText(match); + sourceBuf.push(literalText); + sourceBufLen += literalText.length; + } + } + } + return { source: sourceBuf.join(''), tags: extractedTags }; + } + + /** True if the given tag contains a class attribute with the nocode class. */ + function isNoCodeTag(tag) { + return !!tag + // First canonicalize the representation of attributes + .replace(/\s(\w+)\s*=\s*(?:\"([^\"]*)\"|'([^\']*)'|(\S+))/g, + ' $1="$2$3$4"') + // Then look for the attribute we want. + .match(/[cC][lL][aA][sS][sS]=\"[^\"]*\bnocode\b/); + } + + /** Given triples of [style, pattern, context] returns a lexing function, + * The lexing function interprets the patterns to find token boundaries and + * returns a decoration list of the form + * [index_0, style_0, index_1, style_1, ..., index_n, style_n] + * where index_n is an index into the sourceCode, and style_n is a style + * constant like PR_PLAIN. index_n-1 <= index_n, and style_n-1 applies to + * all characters in sourceCode[index_n-1:index_n]. + * + * The stylePatterns is a list whose elements have the form + * [style : string, pattern : RegExp, context : RegExp, shortcut : string]. + & + * Style is a style constant like PR_PLAIN. + * + * Pattern must only match prefixes, and if it matches a prefix and context + * is null or matches the last non-comment token parsed, then that match is + * considered a token with the same style. + * + * Context is applied to the last non-whitespace, non-comment token + * recognized. + * + * Shortcut is an optional string of characters, any of which, if the first + * character, gurantee that this pattern and only this pattern matches. + * + * @param {Array} shortcutStylePatterns patterns that always start with + * a known character. Must have a shortcut string. + * @param {Array} fallthroughStylePatterns patterns that will be tried in + * order if the shortcut ones fail. May have shortcuts. + * + * @return {function (string, number?) : Array.<number|string>} a + * function that takes source code and returns a list of decorations. + */ + function createSimpleLexer(shortcutStylePatterns, + fallthroughStylePatterns) { + var shortcuts = {}; + (function () { + var allPatterns = shortcutStylePatterns.concat(fallthroughStylePatterns); + for (var i = allPatterns.length; --i >= 0;) { + var patternParts = allPatterns[i]; + var shortcutChars = patternParts[3]; + if (shortcutChars) { + for (var c = shortcutChars.length; --c >= 0;) { + shortcuts[shortcutChars.charAt(c)] = patternParts; + } + } + } + })(); + + var nPatterns = fallthroughStylePatterns.length; + var notWs = /\S/; + + return function (sourceCode, opt_basePos) { + opt_basePos = opt_basePos || 0; + var decorations = [opt_basePos, PR_PLAIN]; + var lastToken = ''; + var pos = 0; // index into sourceCode + var tail = sourceCode; + + while (tail.length) { + var style; + var token = null; + var match; + + var patternParts = shortcuts[tail.charAt(0)]; + if (patternParts) { + match = tail.match(patternParts[1]); + token = match[0]; + style = patternParts[0]; + } else { + for (var i = 0; i < nPatterns; ++i) { + patternParts = fallthroughStylePatterns[i]; + var contextPattern = patternParts[2]; + if (contextPattern && !contextPattern.test(lastToken)) { + // rule can't be used + continue; + } + match = tail.match(patternParts[1]); + if (match) { + token = match[0]; + style = patternParts[0]; + break; + } + } + + if (!token) { // make sure that we make progress + style = PR_PLAIN; + token = tail.substring(0, 1); + } + } + + decorations.push(opt_basePos + pos, style); + pos += token.length; + tail = tail.substring(token.length); + if (style !== PR_COMMENT && notWs.test(token)) { lastToken = token; } + } + return decorations; + }; + } + + var PR_MARKUP_LEXER = createSimpleLexer([], [ + [PR_PLAIN, /^[^<]+/, null], + [PR_DECLARATION, /^<!\w[^>]*(?:>|$)/, null], + [PR_COMMENT, /^<!--[\s\S]*?(?:-->|$)/, null], + [PR_SOURCE, /^<\?[\s\S]*?(?:\?>|$)/, null], + [PR_SOURCE, /^<%[\s\S]*?(?:%>|$)/, null], + [PR_SOURCE, + // Tags whose content is not escaped, and which contain source code. + /^<(script|style|xmp)\b[^>]*>[\s\S]*?<\/\1\b[^>]*>/i, null], + [PR_TAG, /^<\/?\w[^<>]*>/, null] + ]); + // Splits any of the source|style|xmp entries above into a start tag, + // source content, and end tag. + var PR_SOURCE_CHUNK_PARTS = /^(<[^>]*>)([\s\S]*)(<\/[^>]*>)$/; + /** split markup on tags, comments, application directives, and other top + * level constructs. Tags are returned as a single token - attributes are + * not yet broken out. + * @private + */ + function tokenizeMarkup(source) { + var decorations = PR_MARKUP_LEXER(source); + for (var i = 0; i < decorations.length; i += 2) { + if (decorations[i + 1] === PR_SOURCE) { + var start, end; + start = decorations[i]; + end = i + 2 < decorations.length ? decorations[i + 2] : source.length; + // Split out start and end script tags as actual tags, and leave the + // body with style SCRIPT. + var sourceChunk = source.substring(start, end); + var match = sourceChunk.match(PR_SOURCE_CHUNK_PARTS); + if (match) { + decorations.splice( + i, 2, + start, PR_TAG, // the open chunk + start + match[1].length, PR_SOURCE, + start + match[1].length + (match[2] || '').length, PR_TAG); + } + } + } + return decorations; + } + + var PR_TAG_LEXER = createSimpleLexer([ + [PR_ATTRIB_VALUE, /^\'[^\']*(?:\'|$)/, null, "'"], + [PR_ATTRIB_VALUE, /^\"[^\"]*(?:\"|$)/, null, '"'], + [PR_PUNCTUATION, /^[<>\/=]+/, null, '<>/='] + ], [ + [PR_TAG, /^[\w:\-]+/, /^</], + [PR_ATTRIB_VALUE, /^[\w\-]+/, /^=/], + [PR_ATTRIB_NAME, /^[\w:\-]+/, null], + [PR_PLAIN, /^\s+/, null, ' \t\r\n'] + ]); + /** split tags attributes and their values out from the tag name, and + * recursively lex source chunks. + * @private + */ + function splitTagAttributes(source, decorations) { + for (var i = 0; i < decorations.length; i += 2) { + var style = decorations[i + 1]; + if (style === PR_TAG) { + var start, end; + start = decorations[i]; + end = i + 2 < decorations.length ? decorations[i + 2] : source.length; + var chunk = source.substring(start, end); + var subDecorations = PR_TAG_LEXER(chunk, start); + spliceArrayInto(subDecorations, decorations, i, 2); + i += subDecorations.length - 2; + } + } + return decorations; + } + + /** returns a function that produces a list of decorations from source text. + * + * This code treats ", ', and ` as string delimiters, and \ as a string + * escape. It does not recognize perl's qq() style strings. + * It has no special handling for double delimiter escapes as in basic, or + * the tripled delimiters used in python, but should work on those regardless + * although in those cases a single string literal may be broken up into + * multiple adjacent string literals. + * + * It recognizes C, C++, and shell style comments. + * + * @param {Object} options a set of optional parameters. + * @return {function (string) : Array.<string|number>} a + * decorator that takes sourceCode as plain text and that returns a + * decoration list + */ + function sourceDecorator(options) { + var shortcutStylePatterns = [], fallthroughStylePatterns = []; + if (options.tripleQuotedStrings) { + // '''multi-line-string''', 'single-line-string', and double-quoted + shortcutStylePatterns.push( + [PR_STRING, /^(?:\'\'\'(?:[^\'\\]|\\[\s\S]|\'{1,2}(?=[^\']))*(?:\'\'\'|$)|\"\"\"(?:[^\"\\]|\\[\s\S]|\"{1,2}(?=[^\"]))*(?:\"\"\"|$)|\'(?:[^\\\']|\\[\s\S])*(?:\'|$)|\"(?:[^\\\"]|\\[\s\S])*(?:\"|$))/, + null, '\'"']); + } else if (options.multiLineStrings) { + // 'multi-line-string', "multi-line-string" + shortcutStylePatterns.push( + [PR_STRING, /^(?:\'(?:[^\\\']|\\[\s\S])*(?:\'|$)|\"(?:[^\\\"]|\\[\s\S])*(?:\"|$)|\`(?:[^\\\`]|\\[\s\S])*(?:\`|$))/, + null, '\'"`']); + } else { + // 'single-line-string', "single-line-string" + shortcutStylePatterns.push( + [PR_STRING, + /^(?:\'(?:[^\\\'\r\n]|\\.)*(?:\'|$)|\"(?:[^\\\"\r\n]|\\.)*(?:\"|$))/, + null, '"\'']); + } + fallthroughStylePatterns.push( + [PR_PLAIN, /^(?:[^\'\"\`\/\#]+)/, null, ' \r\n']); + if (options.hashComments) { + shortcutStylePatterns.push([PR_COMMENT, /^#[^\r\n]*/, null, '#']); + } + if (options.cStyleComments) { + fallthroughStylePatterns.push([PR_COMMENT, /^\/\/[^\r\n]*/, null]); + fallthroughStylePatterns.push( + [PR_COMMENT, /^\/\*[\s\S]*?(?:\*\/|$)/, null]); + } + if (options.regexLiterals) { + var REGEX_LITERAL = ( + // A regular expression literal starts with a slash that is + // not followed by * or / so that it is not confused with + // comments. + '^/(?=[^/*])' + // and then contains any number of raw characters, + + '(?:[^/\\x5B\\x5C]' + // escape sequences (\x5C), + + '|\\x5C[\\s\\S]' + // or non-nesting character sets (\x5B\x5D); + + '|\\x5B(?:[^\\x5C\\x5D]|\\x5C[\\s\\S])*(?:\\x5D|$))+' + // finally closed by a /. + + '(?:/|$)'); + fallthroughStylePatterns.push( + [PR_STRING, new RegExp(REGEX_LITERAL), REGEXP_PRECEDER_PATTERN]); + } + + var keywords = wordSet(options.keywords); + + options = null; + + /** splits the given string into comment, string, and "other" tokens. + * @param {string} sourceCode as plain text + * @return {Array.<number|string>} a decoration list. + * @private + */ + var splitStringAndCommentTokens = createSimpleLexer( + shortcutStylePatterns, fallthroughStylePatterns); + + var styleLiteralIdentifierPuncRecognizer = createSimpleLexer([], [ + [PR_PLAIN, /^\s+/, null, ' \r\n'], + // TODO(mikesamuel): recognize non-latin letters and numerals in idents + [PR_PLAIN, /^[a-z_$@][a-z_$@0-9]*/i, null], + // A hex number + [PR_LITERAL, /^0x[a-f0-9]+[a-z]/i, null], + // An octal or decimal number, possibly in scientific notation + [PR_LITERAL, + /^(?:\d(?:_\d+)*\d*(?:\.\d*)?|\.\d+)(?:e[+\-]?\d+)?[a-z]*/i, + null, '123456789'], + [PR_PUNCTUATION, /^[^\s\w\.$@]+/, null] + // Fallback will handle decimal points not adjacent to a digit + ]); + + /** splits plain text tokens into more specific tokens, and then tries to + * recognize keywords, and types. + * @private + */ + function splitNonStringNonCommentTokens(source, decorations) { + for (var i = 0; i < decorations.length; i += 2) { + var style = decorations[i + 1]; + if (style === PR_PLAIN) { + var start, end, chunk, subDecs; + start = decorations[i]; + end = i + 2 < decorations.length ? decorations[i + 2] : source.length; + chunk = source.substring(start, end); + subDecs = styleLiteralIdentifierPuncRecognizer(chunk, start); + for (var j = 0, m = subDecs.length; j < m; j += 2) { + var subStyle = subDecs[j + 1]; + if (subStyle === PR_PLAIN) { + var subStart = subDecs[j]; + var subEnd = j + 2 < m ? subDecs[j + 2] : chunk.length; + var token = source.substring(subStart, subEnd); + if (token === '.') { + subDecs[j + 1] = PR_PUNCTUATION; + } else if (token in keywords) { + subDecs[j + 1] = PR_KEYWORD; + } else if (/^@?[A-Z][A-Z$]*[a-z][A-Za-z$]*$/.test(token)) { + // classify types and annotations using Java's style conventions + subDecs[j + 1] = token.charAt(0) === '@' ? PR_LITERAL : PR_TYPE; + } + } + } + spliceArrayInto(subDecs, decorations, i, 2); + i += subDecs.length - 2; + } + } + return decorations; + } + + return function (sourceCode) { + // Split into strings, comments, and other. + // We do this because strings and comments are easily recognizable and can + // contain stuff that looks like other tokens, so we want to mark those + // early so we don't recurse into them. + var decorations = splitStringAndCommentTokens(sourceCode); + + // Split non comment|string tokens on whitespace and word boundaries + decorations = splitNonStringNonCommentTokens(sourceCode, decorations); + + return decorations; + }; + } + + var decorateSource = sourceDecorator({ + keywords: ALL_KEYWORDS, + hashComments: true, + cStyleComments: true, + multiLineStrings: true, + regexLiterals: true + }); + + /** identify regions of markup that are really source code, and recursivley + * lex them. + * @private + */ + function splitSourceNodes(source, decorations) { + for (var i = 0; i < decorations.length; i += 2) { + var style = decorations[i + 1]; + if (style === PR_SOURCE) { + // Recurse using the non-markup lexer + var start, end; + start = decorations[i]; + end = i + 2 < decorations.length ? decorations[i + 2] : source.length; + var subDecorations = decorateSource(source.substring(start, end)); + for (var j = 0, m = subDecorations.length; j < m; j += 2) { + subDecorations[j] += start; + } + spliceArrayInto(subDecorations, decorations, i, 2); + i += subDecorations.length - 2; + } + } + return decorations; + } + + /** identify attribute values that really contain source code and recursively + * lex them. + * @private + */ + function splitSourceAttributes(source, decorations) { + var nextValueIsSource = false; + for (var i = 0; i < decorations.length; i += 2) { + var style = decorations[i + 1]; + var start, end; + if (style === PR_ATTRIB_NAME) { + start = decorations[i]; + end = i + 2 < decorations.length ? decorations[i + 2] : source.length; + nextValueIsSource = /^on|^style$/i.test(source.substring(start, end)); + } else if (style === PR_ATTRIB_VALUE) { + if (nextValueIsSource) { + start = decorations[i]; + end = i + 2 < decorations.length ? decorations[i + 2] : source.length; + var attribValue = source.substring(start, end); + var attribLen = attribValue.length; + var quoted = + (attribLen >= 2 && /^[\"\']/.test(attribValue) && + attribValue.charAt(0) === attribValue.charAt(attribLen - 1)); + + var attribSource; + var attribSourceStart; + var attribSourceEnd; + if (quoted) { + attribSourceStart = start + 1; + attribSourceEnd = end - 1; + attribSource = attribValue; + } else { + attribSourceStart = start + 1; + attribSourceEnd = end - 1; + attribSource = attribValue.substring(1, attribValue.length - 1); + } + + var attribSourceDecorations = decorateSource(attribSource); + for (var j = 0, m = attribSourceDecorations.length; j < m; j += 2) { + attribSourceDecorations[j] += attribSourceStart; + } + + if (quoted) { + attribSourceDecorations.push(attribSourceEnd, PR_ATTRIB_VALUE); + spliceArrayInto(attribSourceDecorations, decorations, i + 2, 0); + } else { + spliceArrayInto(attribSourceDecorations, decorations, i, 2); + } + } + nextValueIsSource = false; + } + } + return decorations; + } + + /** returns a decoration list given a string of markup. + * + * This code recognizes a number of constructs. + * <!-- ... --> comment + * <!\w ... > declaration + * <\w ... > tag + * </\w ... > tag + * <?...?> embedded source + * <%...%> embedded source + * &[#\w]...; entity + * + * It does not recognizes %foo; doctype entities from . + * + * It will recurse into any <style>, <script>, and on* attributes using + * PR_lexSource. + */ + function decorateMarkup(sourceCode) { + // This function works as follows: + // 1) Start by splitting the markup into text and tag chunks + // Input: string s + // Output: List<PR_Token> where style in (PR_PLAIN, null) + // 2) Then split the text chunks further into comments, declarations, + // tags, etc. + // After each split, consider whether the token is the start of an + // embedded source section, i.e. is an open <script> tag. If it is, find + // the corresponding close token, and don't bother to lex in between. + // Input: List<string> + // Output: List<PR_Token> with style in + // (PR_TAG, PR_PLAIN, PR_SOURCE, null) + // 3) Finally go over each tag token and split out attribute names and + // values. + // Input: List<PR_Token> + // Output: List<PR_Token> where style in + // (PR_TAG, PR_PLAIN, PR_SOURCE, NAME, VALUE, null) + var decorations = tokenizeMarkup(sourceCode); + decorations = splitTagAttributes(sourceCode, decorations); + decorations = splitSourceNodes(sourceCode, decorations); + decorations = splitSourceAttributes(sourceCode, decorations); + return decorations; + } + + /** + * @param {string} sourceText plain text + * @param {Array.<number|string>} extractedTags chunks of raw html preceded + * by their position in sourceText in order. + * @param {Array.<number|string>} decorations style classes preceded by their + * position in sourceText in order. + * @return {string} html + * @private + */ + function recombineTagsAndDecorations(sourceText, extractedTags, decorations) { + var html = []; + // index past the last char in sourceText written to html + var outputIdx = 0; + + var openDecoration = null; + var currentDecoration = null; + var tagPos = 0; // index into extractedTags + var decPos = 0; // index into decorations + var tabExpander = makeTabExpander(PR_TAB_WIDTH); + + var adjacentSpaceRe = /([\r\n ]) /g; + var startOrSpaceRe = /(^| ) /gm; + var newlineRe = /\r\n?|\n/g; + var trailingSpaceRe = /[ \r\n]$/; + var lastWasSpace = true; // the last text chunk emitted ended with a space. + + // A helper function that is responsible for opening sections of decoration + // and outputing properly escaped chunks of source + function emitTextUpTo(sourceIdx) { + if (sourceIdx > outputIdx) { + if (openDecoration && openDecoration !== currentDecoration) { + // Close the current decoration + html.push('</span>'); + openDecoration = null; + } + if (!openDecoration && currentDecoration) { + openDecoration = currentDecoration; + html.push('<span class="', openDecoration, '">'); + } + // This interacts badly with some wikis which introduces paragraph tags + // into pre blocks for some strange reason. + // It's necessary for IE though which seems to lose the preformattedness + // of <pre> tags when their innerHTML is assigned. + // http://stud3.tuwien.ac.at/~e0226430/innerHtmlQuirk.html + // and it serves to undo the conversion of <br>s to newlines done in + // chunkify. + var htmlChunk = textToHtml( + tabExpander(sourceText.substring(outputIdx, sourceIdx))) + .replace(lastWasSpace + ? startOrSpaceRe + : adjacentSpaceRe, '$1 '); + // Keep track of whether we need to escape space at the beginning of the + // next chunk. + lastWasSpace = trailingSpaceRe.test(htmlChunk); + html.push(htmlChunk.replace(newlineRe, '<br />')); + outputIdx = sourceIdx; + } + } + + while (true) { + // Determine if we're going to consume a tag this time around. Otherwise + // we consume a decoration or exit. + var outputTag; + if (tagPos < extractedTags.length) { + if (decPos < decorations.length) { + // Pick one giving preference to extractedTags since we shouldn't open + // a new style that we're going to have to immediately close in order + // to output a tag. + outputTag = extractedTags[tagPos] <= decorations[decPos]; + } else { + outputTag = true; + } + } else { + outputTag = false; + } + // Consume either a decoration or a tag or exit. + if (outputTag) { + emitTextUpTo(extractedTags[tagPos]); + if (openDecoration) { + // Close the current decoration + html.push('</span>'); + openDecoration = null; + } + html.push(extractedTags[tagPos + 1]); + tagPos += 2; + } else if (decPos < decorations.length) { + emitTextUpTo(decorations[decPos]); + currentDecoration = decorations[decPos + 1]; + decPos += 2; + } else { + break; + } + } + emitTextUpTo(sourceText.length); + if (openDecoration) { + html.push('</span>'); + } + + return html.join(''); + } + + /** Maps language-specific file extensions to handlers. */ + var langHandlerRegistry = {}; + /** Register a language handler for the given file extensions. + * @param {function (string) : Array.<number|string>} handler + * a function from source code to a list of decorations. + * @param {Array.<string>} fileExtensions + */ + function registerLangHandler(handler, fileExtensions) { + for (var i = fileExtensions.length; --i >= 0;) { + var ext = fileExtensions[i]; + if (!langHandlerRegistry.hasOwnProperty(ext)) { + langHandlerRegistry[ext] = handler; + } else if ('console' in window) { + console.log('cannot override language handler %s', ext); + } + } + } + registerLangHandler(decorateSource, ['default-code']); + registerLangHandler(decorateMarkup, + ['default-markup', 'html', 'htm', 'xhtml', 'xml', 'xsl']); + registerLangHandler(sourceDecorator({ + keywords: CPP_KEYWORDS, + hashComments: true, + cStyleComments: true + }), ['c', 'cc', 'cpp', 'cxx', 'cyc']); + registerLangHandler(sourceDecorator({ + keywords: CSHARP_KEYWORDS, + hashComments: true, + cStyleComments: true + }), ['cs']); + registerLangHandler(sourceDecorator({ + keywords: JAVA_KEYWORDS, + cStyleComments: true + }), ['java']); + registerLangHandler(sourceDecorator({ + keywords: SH_KEYWORDS, + hashComments: true, + multiLineStrings: true + }), ['bsh', 'csh', 'sh']); + registerLangHandler(sourceDecorator({ + keywords: PYTHON_KEYWORDS, + hashComments: true, + multiLineStrings: true, + tripleQuotedStrings: true + }), ['cv', 'py']); + registerLangHandler(sourceDecorator({ + keywords: PERL_KEYWORDS, + hashComments: true, + multiLineStrings: true, + regexLiterals: true + }), ['perl', 'pl', 'pm']); + registerLangHandler(sourceDecorator({ + keywords: RUBY_KEYWORDS, + hashComments: true, + multiLineStrings: true, + regexLiterals: true + }), ['rb']); + registerLangHandler(sourceDecorator({ + keywords: JSCRIPT_KEYWORDS, + cStyleComments: true, + regexLiterals: true + }), ['js']); + + function prettyPrintOne(sourceCodeHtml, opt_langExtension) { + try { + // Extract tags, and convert the source code to plain text. + var sourceAndExtractedTags = extractTags(sourceCodeHtml); + /** Plain text. @type {string} */ + var source = sourceAndExtractedTags.source; + + /** Even entries are positions in source in ascending order. Odd entries + * are tags that were extracted at that position. + * @type {Array.<number|string>} + */ + var extractedTags = sourceAndExtractedTags.tags; + + // Pick a lexer and apply it. + if (!langHandlerRegistry.hasOwnProperty(opt_langExtension)) { + // Treat it as markup if the first non whitespace character is a < and + // the last non-whitespace character is a >. + opt_langExtension = + /^\s*</.test(source) ? 'default-markup' : 'default-code'; + } + + /** Even entries are positions in source in ascending order. Odd enties + * are style markers (e.g., PR_COMMENT) that run from that position until + * the end. + * @type {Array.<number|string>} + */ + var decorations = langHandlerRegistry[opt_langExtension].call({}, source); + + // Integrate the decorations and tags back into the source code to produce + // a decorated html string. + return recombineTagsAndDecorations(source, extractedTags, decorations); + } catch (e) { + if ('console' in window) { + console.log(e); + console.trace(); + } + return sourceCodeHtml; + } + } + + function prettyPrint(opt_whenDone) { + var isIE6 = _pr_isIE6(); + + // fetch a list of nodes to rewrite + var codeSegments = [ + document.getElementsByTagName('pre'), + document.getElementsByTagName('code'), + document.getElementsByTagName('xmp') ]; + var elements = []; + for (var i = 0; i < codeSegments.length; ++i) { + for (var j = 0; j < codeSegments[i].length; ++j) { + elements.push(codeSegments[i][j]); + } + } + codeSegments = null; + + // the loop is broken into a series of continuations to make sure that we + // don't make the browser unresponsive when rewriting a large page. + var k = 0; + + function doWork() { + var endTime = (PR_SHOULD_USE_CONTINUATION ? + new Date().getTime() + 250 /* ms */ : + Infinity); + for (; k < elements.length && new Date().getTime() < endTime; k++) { + var cs = elements[k]; + if (cs.className && cs.className.indexOf('prettyprint') >= 0) { + // If the classes includes a language extensions, use it. + // Language extensions can be specified like + // <pre class="prettyprint lang-cpp"> + // the language extension "cpp" is used to find a language handler as + // passed to PR_registerLangHandler. + var langExtension = cs.className.match(/\blang-(\w+)\b/); + if (langExtension) { langExtension = langExtension[1]; } + + // make sure this is not nested in an already prettified element + var nested = false; + for (var p = cs.parentNode; p; p = p.parentNode) { + if ((p.tagName === 'pre' || p.tagName === 'code' || + p.tagName === 'xmp') && + p.className && p.className.indexOf('prettyprint') >= 0) { + nested = true; + break; + } + } + if (!nested) { + // fetch the content as a snippet of properly escaped HTML. + // Firefox adds newlines at the end. + var content = getInnerHtml(cs); + content = content.replace(/(?:\r\n?|\n)$/, ''); + + // do the pretty printing + var newContent = prettyPrintOne(content, langExtension); + + // push the prettified html back into the tag. + if (!isRawContent(cs)) { + // just replace the old html with the new + cs.innerHTML = newContent; + } else { + // we need to change the tag to a <pre> since <xmp>s do not allow + // embedded tags such as the span tags used to attach styles to + // sections of source code. + var pre = document.createElement('PRE'); + for (var i = 0; i < cs.attributes.length; ++i) { + var a = cs.attributes[i]; + if (a.specified) { + var aname = a.name.toLowerCase(); + if (aname === 'class') { + pre.className = a.value; // For IE 6 + } else { + pre.setAttribute(a.name, a.value); + } + } + } + pre.innerHTML = newContent; + + // remove the old + cs.parentNode.replaceChild(pre, cs); + cs = pre; + } + + // Replace <br>s with line-feeds so that copying and pasting works + // on IE 6. + // Doing this on other browsers breaks lots of stuff since \r\n is + // treated as two newlines on Firefox, and doing this also slows + // down rendering. + if (isIE6 && cs.tagName === 'PRE') { + var lineBreaks = cs.getElementsByTagName('br'); + for (var j = lineBreaks.length; --j >= 0;) { + var lineBreak = lineBreaks[j]; + lineBreak.parentNode.replaceChild( + document.createTextNode('\r\n'), lineBreak); + } + } + } + } + } + if (k < elements.length) { + // finish up in a continuation + setTimeout(doWork, 250); + } else if (opt_whenDone) { + opt_whenDone(); + } + } + + doWork(); + } + + window['PR_normalizedHtml'] = normalizedHtml; + window['prettyPrintOne'] = prettyPrintOne; + window['prettyPrint'] = prettyPrint; + window['PR'] = { + 'createSimpleLexer': createSimpleLexer, + 'registerLangHandler': registerLangHandler, + 'sourceDecorator': sourceDecorator, + 'PR_ATTRIB_NAME': PR_ATTRIB_NAME, + 'PR_ATTRIB_VALUE': PR_ATTRIB_VALUE, + 'PR_COMMENT': PR_COMMENT, + 'PR_DECLARATION': PR_DECLARATION, + 'PR_KEYWORD': PR_KEYWORD, + 'PR_LITERAL': PR_LITERAL, + 'PR_NOCODE': PR_NOCODE, + 'PR_PLAIN': PR_PLAIN, + 'PR_PUNCTUATION': PR_PUNCTUATION, + 'PR_SOURCE': PR_SOURCE, + 'PR_STRING': PR_STRING, + 'PR_TAG': PR_TAG, + 'PR_TYPE': PR_TYPE + }; +})(); diff --git a/docs/verifier.html b/docs/verifier.html new file mode 100644 index 000000000..656b83273 --- /dev/null +++ b/docs/verifier.html @@ -0,0 +1,159 @@ +<html> +<head> +<title>Dalvik Bytecode Verifier Notes</title> +</head> + +<body> +<h1>Dalvik Bytecode Verifier Notes</h1> + +<p> +The bytecode verifier in the Dalvik VM attempts to provide the same sorts +of checks and guarantees that other popular virtual machines do. We +perform generally the same set of checks as are described in _The Java +Virtual Machine Specification, Second Edition_, including the updates +planned for the Third Edition. + +<p> +Verification can be enabled for all classes, disabled for all, or enabled +only for "remote" (non-bootstrap) classes. It should be performed for any +class that will be processed with the DEX optimizer, and in fact the +default VM behavior is to only optimize verified classes. + + +<h2>Why Verify?</h2> + +<p> +The verification process adds additional time to the build and to +the installation of new applications. It's fairly quick for app-sized +DEX files, but rather slow for the big "core" and "framework" files. +Why do it all, when our system relies on UNIX processes for security? +<p> +<ol> + <li>Optimizations. The interpreter can ignore a lot of potential + error cases because the verifier guarantees that they are impossible. + Also, we can optimize the DEX file more aggressively if we start + with a stronger set of assumptions about the bytecode. + <li>"Exact" GC. The work peformed during verification has significant + overlap with the work required to compute register use maps for exact + GC. Improper register use, caught by the verifier, could lead to + subtle problems with an "exact" GC. + <li>Intra-application security. If an app wants to download bits + of interpreted code over the network and execute them, it can safely + do so using well-established security mechanisms. + <li>3rd party app failure analysis. We have no way to control the + tools and post-processing utilities that external developers employ, + so when we get bug reports with a weird exception or native crash + it's very helpful to start with the assumption that the bytecode + is valid. +</ol> + + +<h2>Verifier Differences</h2> + +<p> +There are a few checks that the Dalvik bytecode verifier does not perform, +because they're not relevant. For example: +<ul> + <li>Type restrictions on constant pool references are not enforced, + because Dalvik does not have a pool of typed constants. (Dalvik + uses a simple index into type-specific pools.) + <li>Verification of the operand stack size is not performed, because + Dalvik does not have an operand stack. + <li>Limitations on <code>jsr</code> and <code>ret</code> do not apply, + because Dalvik doesn't support subroutines. +</ul> + +In some cases they are implemented differently, e.g.: +<ul> + <li>In a conventional VM, backward branches and exceptions are + forbidden when a local variable holds an uninitialized reference. The + restriction was changed to mark registers as invalid when they hold + references to the uninitialized result of a previous invocation of the + same <code>new-instance</code> instruction. + This solves the same problem -- trickery potentially allowing + uninitialized objects to slip past the verifier -- without unduly + limiting branches. +</ul> + +There are also some new ones, such as: +<ul> + <li>The <code>move-exception</code> instruction can only appear as + the first instruction in an exception handler. + <li>The <code>move-result*</code> instructions can only appear + immediately after an appropriate <code>invoke-*</code> + or <code>filled-new-array</code> instruction. +</ul> + +<p> +The Dalvik verifier is more restrictive than other VMs in one area: +type safety on sub-32-bit integer widths. These additional restrictions +should make it impossible to, say, pass a value outside the range +[-128, 127] to a function that takes a <code>byte</code> as an argument. + + +<h2>Verification Failures</h2> + +<p> +When the verifier rejects a class, it always throws a VerifyError. +This is different in some cases from other implementations. For example, +if a class attempts to perform an illegal access on a field, the expected +behavior is to receive an IllegalAccessError at runtime the first time +the field is actually accessed. The Dalvik verifier will reject the +entire class immediately. + +<p> +It's difficult to throw the error on first use in Dalvik. Possible ways +to implement this behavior include: + +<ol> +<li>We could replace the invalid field access instruction with a special +instruction that generates an illegal access error, and allow class +verification to complete successfully. This type of verification must +often be deferred to first class load, rather than be performed ahead of time +during DEX optimization, which means the bytecode instructions will be +mapped read-only during verification. So this won't work. +</li> + +<li>We can perform the access checks when the field/method/class is +resolved. In a typical VM implementation we would do the check when the +entry is resolved in the context of the current classfile, but our DEX +files combine multiple classfiles together, merging the field/method/class +resolution results into a single large table. Once one class successfully +resolves the field, every other class in the same DEX file would be able +to access the field. This is bad. +</li> + +<li>Perform the access checks on every field/method/class access. +This adds significant overhead. This is mitigated somewhat by the DEX +optimizer, which will convert many field/method/class accesses into a +simpler form after performing the access check. However, not all accesses +can be optimized (e.g. accesses to classes unknown at dexopt time), +and we don't currently have an optimized form of certain instructions +(notably static field operations). +</li> +</ol> + +<p> +Other implementations are possible, but they all involve allocating +some amount of additional memory or spending additional cycles +on non-DEX-optimized instructions. We don't want to throw an +IllegalAccessError at verification time, since that would indicate that +access to the class being verified was illegal. +<p> +One approach that might be worth pursuing: for situations like illegal +accesses, the verifier makes an in-RAM private copy of the method, and +alters the instructions there. The class object is altered to point at +the new copy of the instructions. This requires minimal memory overhead +and provides a better experience for developers. + +<p> +The VerifyError is accompanied by detailed, if somewhat cryptic, +information in the log file. From this it's possible to determine the +exact instruction that failed, and the reason for the failure. We can +also constructor the VerifyError with an IllegalAccessError passed in as +the cause. + +<address>Copyright © 2008 The Android Open Source Project</address> + +</body> +</html> |