docs/dalvik-constraints.html - platform/dalvik - Gitiles

 <!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-&lt;kind&gt;</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&lt;kind&gt;</code> and
           <code>iput&lt;kind&gt;</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&lt;kind&gt;</code> and
           <code>sput&lt;kind&gt;</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&lt;kind&gt;</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&lt;kind&gt;</code> instruction must be immediately
           preceded (in the <code>insns</code> array) by an
           <code>&lt;invoke-kind&gt;</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&lt;kind&gt;</code> instruction must be immediately
           preceded (in actual control flow) by a matching
           <code>return-&lt;kind&gt;</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>
	<!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>