View Javadoc
1   /*
2    * reserved comment block
3    * DO NOT REMOVE OR ALTER!
4    */
5   package com.sun.org.apache.bcel.internal.util;
6   
7   /* ====================================================================
8    * The Apache Software License, Version 1.1
9    *
10   * Copyright (c) 2001 The Apache Software Foundation.  All rights
11   * reserved.
12   *
13   * Redistribution and use in source and binary forms, with or without
14   * modification, are permitted provided that the following conditions
15   * are met:
16   *
17   * 1. Redistributions of source code must retain the above copyright
18   *    notice, this list of conditions and the following disclaimer.
19   *
20   * 2. Redistributions in binary form must reproduce the above copyright
21   *    notice, this list of conditions and the following disclaimer in
22   *    the documentation and/or other materials provided with the
23   *    distribution.
24   *
25   * 3. The end-user documentation included with the redistribution,
26   *    if any, must include the following acknowledgment:
27   *       "This product includes software developed by the
28   *        Apache Software Foundation (http://www.apache.org/)."
29   *    Alternately, this acknowledgment may appear in the software itself,
30   *    if and wherever such third-party acknowledgments normally appear.
31   *
32   * 4. The names "Apache" and "Apache Software Foundation" and
33   *    "Apache BCEL" must not be used to endorse or promote products
34   *    derived from this software without prior written permission. For
35   *    written permission, please contact apache@apache.org.
36   *
37   * 5. Products derived from this software may not be called "Apache",
38   *    "Apache BCEL", nor may "Apache" appear in their name, without
39   *    prior written permission of the Apache Software Foundation.
40   *
41   * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
42   * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
43   * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
44   * DISCLAIMED.  IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
45   * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
46   * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
47   * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
48   * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
49   * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
50   * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
51   * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
52   * SUCH DAMAGE.
53   * ====================================================================
54   *
55   * This software consists of voluntary contributions made by many
56   * individuals on behalf of the Apache Software Foundation.  For more
57   * information on the Apache Software Foundation, please see
58   * <http://www.apache.org/>.
59   */
60  
61  import java.util.*;
62  import com.sun.org.apache.bcel.internal.Constants;
63  import com.sun.org.apache.bcel.internal.generic.*;
64  import com.sun.org.apache.regexp.internal.*;
65  
66  /**
67   * InstructionFinder is a tool to search for given instructions patterns,
68   * i.e., match sequences of instructions in an instruction list via
69   * regular expressions. This can be used, e.g., in order to implement
70   * a peep hole optimizer that looks for code patterns and replaces
71   * them with faster equivalents.
72   *
73   * <p>This class internally uses the <a href="http://jakarta.apache.org/regexp/">
74   * Regexp</a> package to search for regular expressions.
75   *
76   * A typical application would look like this:
77  <pre>
78      InstructionFinder f   = new InstructionFinder(il);
79      String            pat = "IfInstruction ICONST_0 GOTO ICONST_1 NOP (IFEQ|IFNE)";
80  
81      for(Iterator i = f.search(pat, constraint); i.hasNext(); ) {
82        InstructionHandle[] match = (InstructionHandle[])i.next();
83        ...
84        il.delete(match[1], match[5]);
85        ...
86      }
87  </pre>
88   * @author  <A HREF="http://www.berlin.de/~markus.dahm/">M. Dahm</A>
89   * @see Instruction
90   * @see InstructionList
91   */
92  public class InstructionFinder {
93    private static final int OFFSET     = 32767; // char + OFFSET is outside of LATIN-1
94    private static final int NO_OPCODES = 256;   // Potential number, some are not used
95  
96    private static final HashMap map = new HashMap(); // Map<String,Pattern>
97  
98    private InstructionList     il;
99    private String              il_string;    // instruction list as string
100   private InstructionHandle[] handles;      // map instruction list to array
101 
102   /**
103    * @param il instruction list to search for given patterns
104    */
105   public InstructionFinder(InstructionList il) {
106     this.il = il;
107     reread();
108   }
109 
110   /**
111    * Reread the instruction list, e.g., after you've altered the list upon a match.
112    */
113   public final void reread() {
114     int    size  = il.getLength();
115     char[] buf   = new char[size]; // Create a string with length equal to il length
116     handles      = il.getInstructionHandles();
117 
118     // Map opcodes to characters
119     for(int i=0; i < size; i++)
120       buf[i] = makeChar(handles[i].getInstruction().getOpcode());
121 
122     il_string = new String(buf);
123   }
124 
125   /**
126    * Map symbolic instruction names like "getfield" to a single character.
127    *
128    * @param pattern instruction pattern in lower case
129    * @return encoded string for a pattern such as "BranchInstruction".
130    */
131   private static final String mapName(String pattern) {
132     String result = (String)map.get(pattern);
133 
134     if(result != null)
135       return result;
136 
137     for(short i=0; i < NO_OPCODES; i++)
138       if(pattern.equals(Constants.OPCODE_NAMES[i]))
139         return "" + makeChar(i);
140 
141     throw new RuntimeException("Instruction unknown: " + pattern);
142   }
143 
144   /**
145    * Replace symbolic names of instructions with the appropiate character and remove
146    * all white space from string. Meta characters such as +, * are ignored.
147    *
148    * @param pattern The pattern to compile
149    * @return translated regular expression string
150    */
151   private static final String compilePattern(String pattern) {
152     String       lower      = pattern.toLowerCase();
153     StringBuffer buf        = new StringBuffer();
154     int          size       = pattern.length();
155 
156     for(int i=0; i < size; i++) {
157       char ch = lower.charAt(i);
158 
159       if(Character.isLetterOrDigit(ch)) {
160         StringBuffer name = new StringBuffer();
161 
162         while((Character.isLetterOrDigit(ch) || ch == '_') && i < size) {
163           name.append(ch);
164 
165           if(++i < size)
166             ch = lower.charAt(i);
167           else
168             break;
169         }
170 
171         i--;
172 
173         buf.append(mapName(name.toString()));
174       } else if(!Character.isWhitespace(ch))
175         buf.append(ch);
176     }
177 
178     return buf.toString();
179   }
180 
181   /**
182    * @return the matched piece of code as an array of instruction (handles)
183    */
184   private InstructionHandle[] getMatch(int matched_from, int match_length) {
185     InstructionHandle[] match = new InstructionHandle[match_length];
186     System.arraycopy(handles, matched_from, match, 0, match_length);
187 
188     return match;
189   }
190 
191   /**
192    * Search for the given pattern in the instruction list. You can search for any valid
193    * opcode via its symbolic name, e.g. "istore". You can also use a super class or
194    * an interface name to match a whole set of instructions, e.g. "BranchInstruction" or
195    * "LoadInstruction". "istore" is also an alias for all "istore_x" instructions. Additional
196    * aliases are "if" for "ifxx", "if_icmp" for "if_icmpxx", "if_acmp" for "if_acmpxx".
197    *
198    * Consecutive instruction names must be separated by white space which will be removed
199    * during the compilation of the pattern.
200    *
201    * For the rest the usual pattern matching rules for regular expressions apply.<P>
202    * Example pattern:
203    * <pre>
204      search("BranchInstruction NOP ((IfInstruction|GOTO)+ ISTORE Instruction)*");
205    * </pre>
206    *
207    * <p>If you alter the instruction list upon a match such that other
208    * matching areas are affected, you should call reread() to update
209    * the finder and call search() again, because the matches are cached.
210    *
211    * @param pattern the instruction pattern to search for, where case is ignored
212    * @param from where to start the search in the instruction list
213    * @param constraint optional CodeConstraint to check the found code pattern for
214    * user-defined constraints
215    * @return iterator of matches where e.nextElement() returns an array of instruction handles
216    * describing the matched area
217    */
218   public final Iterator search(String pattern, InstructionHandle from,
219                                CodeConstraint constraint)
220   {
221     String search = compilePattern(pattern);
222     int  start    = -1;
223 
224     for(int i=0; i < handles.length; i++) {
225       if(handles[i] == from) {
226         start = i; // Where to start search from (index)
227         break;
228       }
229     }
230 
231     if(start == -1)
232       throw new ClassGenException("Instruction handle " + from +
233                                   " not found in instruction list.");
234     try {
235       RE regex = new RE(search);
236       ArrayList matches = new ArrayList();
237 
238       while(start < il_string.length() && regex.match(il_string, start)) {
239         int startExpr = regex.getParenStart(0);
240         int endExpr   = regex.getParenEnd(0);
241         int lenExpr   = regex.getParenLength(0);
242 
243         InstructionHandle[] match = getMatch(startExpr, lenExpr);
244 
245         if((constraint == null) || constraint.checkCode(match))
246           matches.add(match);
247         start = endExpr;
248       }
249 
250       return matches.iterator();
251     } catch(RESyntaxException e) {
252       System.err.println(e);
253     }
254 
255     return null;
256   }
257 
258   /**
259    * Start search beginning from the start of the given instruction list.
260    *
261    * @param pattern the instruction pattern to search for, where case is ignored
262    * @return iterator of matches where e.nextElement()
263    * returns an array of instruction handles describing the matched
264    * area
265    */
266   public final Iterator search(String pattern) {
267     return search(pattern, il.getStart(), null);
268   }
269 
270   /**
271    * Start search beginning from `from'.
272    *
273    * @param pattern the instruction pattern to search for, where case is ignored
274    * @param from where to start the search in the instruction list
275    * @return  iterator of matches where e.nextElement() returns an array of instruction handles
276    * describing the matched area
277    */
278   public final Iterator search(String pattern, InstructionHandle from) {
279     return search(pattern, from, null);
280   }
281 
282   /**
283    * Start search beginning from the start of the given instruction list.
284    * Check found matches with the constraint object.
285    *
286    * @param pattern the instruction pattern to search for, case is ignored
287    * @param constraint constraints to be checked on matching code
288    * @return instruction handle or `null' if the match failed
289    */
290   public final Iterator search(String pattern, CodeConstraint constraint) {
291     return search(pattern, il.getStart(), constraint);
292   }
293 
294   /**
295    * Convert opcode number to char.
296    */
297   private static final char makeChar(short opcode) {
298     return (char)(opcode + OFFSET);
299   }
300 
301   /**
302    * @return the inquired instruction list
303    */
304   public final InstructionList getInstructionList() { return il; }
305 
306   /**
307    * Code patterns found may be checked using an additional
308    * user-defined constraint object whether they really match the needed criterion.
309    * I.e., check constraints that can not expressed with regular expressions.
310    *
311    */
312   public interface CodeConstraint {
313     /**
314      * @param match array of instructions matching the requested pattern
315      * @return true if the matched area is really useful
316      */
317     public boolean checkCode(InstructionHandle[] match);
318   }
319 
320   // Initialize pattern map
321 
322   static {
323     map.put("arithmeticinstruction", "(irem|lrem|iand|ior|ineg|isub|lneg|fneg|fmul|ldiv|fadd|lxor|frem|idiv|land|ixor|ishr|fsub|lshl|fdiv|iadd|lor|dmul|lsub|ishl|imul|lmul|lushr|dneg|iushr|lshr|ddiv|drem|dadd|ladd|dsub)");
324     map.put("invokeinstruction", "(invokevirtual|invokeinterface|invokestatic|invokespecial)");
325     map.put("arrayinstruction", "(baload|aastore|saload|caload|fastore|lastore|iaload|castore|iastore|aaload|bastore|sastore|faload|laload|daload|dastore)");
326     map.put("gotoinstruction", "(goto|goto_w)");
327     map.put("conversioninstruction", "(d2l|l2d|i2s|d2i|l2i|i2b|l2f|d2f|f2i|i2d|i2l|f2d|i2c|f2l|i2f)");
328     map.put("localvariableinstruction", "(fstore|iinc|lload|dstore|dload|iload|aload|astore|istore|fload|lstore)");
329     map.put("loadinstruction", "(fload|dload|lload|iload|aload)");
330     map.put("fieldinstruction", "(getfield|putstatic|getstatic|putfield)");
331     map.put("cpinstruction", "(ldc2_w|invokeinterface|multianewarray|putstatic|instanceof|getstatic|checkcast|getfield|invokespecial|ldc_w|invokestatic|invokevirtual|putfield|ldc|new|anewarray)");
332     map.put("stackinstruction", "(dup2|swap|dup2_x2|pop|pop2|dup|dup2_x1|dup_x2|dup_x1)");
333     map.put("branchinstruction", "(ifle|if_acmpne|if_icmpeq|if_acmpeq|ifnonnull|goto_w|iflt|ifnull|if_icmpne|tableswitch|if_icmple|ifeq|if_icmplt|jsr_w|if_icmpgt|ifgt|jsr|goto|ifne|ifge|lookupswitch|if_icmpge)");
334     map.put("returninstruction", "(lreturn|ireturn|freturn|dreturn|areturn|return)");
335     map.put("storeinstruction", "(istore|fstore|dstore|astore|lstore)");
336     map.put("select", "(tableswitch|lookupswitch)");
337     map.put("ifinstruction", "(ifeq|ifgt|if_icmpne|if_icmpeq|ifge|ifnull|ifne|if_icmple|if_icmpge|if_acmpeq|if_icmplt|if_acmpne|ifnonnull|iflt|if_icmpgt|ifle)");
338     map.put("jsrinstruction", "(jsr|jsr_w)");
339     map.put("variablelengthinstruction", "(tableswitch|jsr|goto|lookupswitch)");
340     map.put("unconditionalbranch", "(goto|jsr|jsr_w|athrow|goto_w)");
341     map.put("constantpushinstruction", "(dconst|bipush|sipush|fconst|iconst|lconst)");
342     map.put("typedinstruction", "(imul|lsub|aload|fload|lor|new|aaload|fcmpg|iand|iaload|lrem|idiv|d2l|isub|dcmpg|dastore|ret|f2d|f2i|drem|iinc|i2c|checkcast|frem|lreturn|astore|lushr|daload|dneg|fastore|istore|lshl|ldiv|lstore|areturn|ishr|ldc_w|invokeinterface|aastore|lxor|ishl|l2d|i2f|return|faload|sipush|iushr|caload|instanceof|invokespecial|putfield|fmul|ireturn|laload|d2f|lneg|ixor|i2l|fdiv|lastore|multianewarray|i2b|getstatic|i2d|putstatic|fcmpl|saload|ladd|irem|dload|jsr_w|dconst|dcmpl|fsub|freturn|ldc|aconst_null|castore|lmul|ldc2_w|dadd|iconst|f2l|ddiv|dstore|land|jsr|anewarray|dmul|bipush|dsub|sastore|d2i|i2s|lshr|iadd|l2i|lload|bastore|fstore|fneg|iload|fadd|baload|fconst|ior|ineg|dreturn|l2f|lconst|getfield|invokevirtual|invokestatic|iastore)");
343     map.put("popinstruction", "(fstore|dstore|pop|pop2|astore|putstatic|istore|lstore)");
344     map.put("allocationinstruction", "(multianewarray|new|anewarray|newarray)");
345     map.put("indexedinstruction", "(lload|lstore|fload|ldc2_w|invokeinterface|multianewarray|astore|dload|putstatic|instanceof|getstatic|checkcast|getfield|invokespecial|dstore|istore|iinc|ldc_w|ret|fstore|invokestatic|iload|putfield|invokevirtual|ldc|new|aload|anewarray)");
346     map.put("pushinstruction", "(dup|lload|dup2|bipush|fload|ldc2_w|sipush|lconst|fconst|dload|getstatic|ldc_w|aconst_null|dconst|iload|ldc|iconst|aload)");
347     map.put("stackproducer", "(imul|lsub|aload|fload|lor|new|aaload|fcmpg|iand|iaload|lrem|idiv|d2l|isub|dcmpg|dup|f2d|f2i|drem|i2c|checkcast|frem|lushr|daload|dneg|lshl|ldiv|ishr|ldc_w|invokeinterface|lxor|ishl|l2d|i2f|faload|sipush|iushr|caload|instanceof|invokespecial|fmul|laload|d2f|lneg|ixor|i2l|fdiv|getstatic|i2b|swap|i2d|dup2|fcmpl|saload|ladd|irem|dload|jsr_w|dconst|dcmpl|fsub|ldc|arraylength|aconst_null|tableswitch|lmul|ldc2_w|iconst|dadd|f2l|ddiv|land|jsr|anewarray|dmul|bipush|dsub|d2i|newarray|i2s|lshr|iadd|lload|l2i|fneg|iload|fadd|baload|fconst|lookupswitch|ior|ineg|lconst|l2f|getfield|invokevirtual|invokestatic)");
348     map.put("stackconsumer", "(imul|lsub|lor|iflt|fcmpg|if_icmpgt|iand|ifeq|if_icmplt|lrem|ifnonnull|idiv|d2l|isub|dcmpg|dastore|if_icmpeq|f2d|f2i|drem|i2c|checkcast|frem|lreturn|astore|lushr|pop2|monitorexit|dneg|fastore|istore|lshl|ldiv|lstore|areturn|if_icmpge|ishr|monitorenter|invokeinterface|aastore|lxor|ishl|l2d|i2f|return|iushr|instanceof|invokespecial|fmul|ireturn|d2f|lneg|ixor|pop|i2l|ifnull|fdiv|lastore|i2b|if_acmpeq|ifge|swap|i2d|putstatic|fcmpl|ladd|irem|dcmpl|fsub|freturn|ifgt|castore|lmul|dadd|f2l|ddiv|dstore|land|if_icmpne|if_acmpne|dmul|dsub|sastore|ifle|d2i|i2s|lshr|iadd|l2i|bastore|fstore|fneg|fadd|ior|ineg|ifne|dreturn|l2f|if_icmple|getfield|invokevirtual|invokestatic|iastore)");
349     map.put("exceptionthrower", "(irem|lrem|laload|putstatic|baload|dastore|areturn|getstatic|ldiv|anewarray|iastore|castore|idiv|saload|lastore|fastore|putfield|lreturn|caload|getfield|return|aastore|freturn|newarray|instanceof|multianewarray|athrow|faload|iaload|aaload|dreturn|monitorenter|checkcast|bastore|arraylength|new|invokevirtual|sastore|ldc_w|ireturn|invokespecial|monitorexit|invokeinterface|ldc|invokestatic|daload)");
350     map.put("loadclass", "(multianewarray|invokeinterface|instanceof|invokespecial|putfield|checkcast|putstatic|invokevirtual|new|getstatic|invokestatic|getfield|anewarray)");
351     map.put("instructiontargeter", "(ifle|if_acmpne|if_icmpeq|if_acmpeq|ifnonnull|goto_w|iflt|ifnull|if_icmpne|tableswitch|if_icmple|ifeq|if_icmplt|jsr_w|if_icmpgt|ifgt|jsr|goto|ifne|ifge|lookupswitch|if_icmpge)");
352 
353     // Some aliases
354     map.put("if_icmp", "(if_icmpne|if_icmpeq|if_icmple|if_icmpge|if_icmplt|if_icmpgt)");
355     map.put("if_acmp", "(if_acmpeq|if_acmpne)");
356     map.put("if", "(ifeq|ifne|iflt|ifge|ifgt|ifle)");
357 
358     // Precompile some aliases first
359     map.put("iconst", precompile(Constants.ICONST_0, Constants.ICONST_5, Constants.ICONST_M1));
360     map.put("lconst", new String(new char[] { '(', makeChar(Constants.LCONST_0), '|',
361                                               makeChar(Constants.LCONST_1), ')' }));
362     map.put("dconst", new String(new char[] { '(', makeChar(Constants.DCONST_0), '|',
363                                               makeChar(Constants.DCONST_1), ')' }));
364     map.put("fconst", new String(new char[] { '(', makeChar(Constants.FCONST_0), '|',
365                                               makeChar(Constants.FCONST_1), ')' }));
366 
367     map.put("iload", precompile(Constants.ILOAD_0, Constants.ILOAD_3, Constants.ILOAD));
368     map.put("dload", precompile(Constants.DLOAD_0, Constants.DLOAD_3, Constants.DLOAD));
369     map.put("fload", precompile(Constants.FLOAD_0, Constants.FLOAD_3, Constants.FLOAD));
370     map.put("aload", precompile(Constants.ALOAD_0, Constants.ALOAD_3, Constants.ALOAD));
371 
372     map.put("istore", precompile(Constants.ISTORE_0, Constants.ISTORE_3, Constants.ISTORE));
373     map.put("dstore", precompile(Constants.DSTORE_0, Constants.DSTORE_3, Constants.DSTORE));
374     map.put("fstore", precompile(Constants.FSTORE_0, Constants.FSTORE_3, Constants.FSTORE));
375     map.put("astore", precompile(Constants.ASTORE_0, Constants.ASTORE_3, Constants.ASTORE));
376 
377     // Compile strings
378 
379     for(Iterator i = map.keySet().iterator(); i.hasNext(); ) {
380       String key   = (String)i.next();
381       String value = (String)map.get(key);
382 
383       char ch = value.charAt(1); // Omit already precompiled patterns
384       if(ch < OFFSET) {
385         map.put(key, compilePattern(value)); // precompile all patterns
386       }
387     }
388 
389     // Add instruction alias to match anything
390 
391     StringBuffer buf = new StringBuffer("(");
392 
393     for(short i=0; i < NO_OPCODES; i++) {
394       if(Constants.NO_OF_OPERANDS[i] != Constants.UNDEFINED) { // Not an invalid opcode
395         buf.append(makeChar(i));
396 
397         if(i < NO_OPCODES - 1)
398           buf.append('|');
399       }
400     }
401     buf.append(')');
402 
403     map.put("instruction", buf.toString());
404   }
405 
406   private static String precompile(short from, short to, short extra) {
407     StringBuffer buf = new StringBuffer("(");
408 
409     for(short i=from; i <= to; i++) {
410       buf.append(makeChar(i));
411       buf.append('|');
412     }
413 
414     buf.append(makeChar(extra));
415     buf.append(")");
416     return buf.toString();
417   }
418 
419   /*
420    * Internal debugging routines.
421    */
422   private static final String pattern2string(String pattern) {
423     return pattern2string(pattern, true);
424   }
425 
426   private static final String pattern2string(String pattern, boolean make_string) {
427     StringBuffer buf = new StringBuffer();
428 
429     for(int i=0; i < pattern.length(); i++) {
430       char ch = pattern.charAt(i);
431 
432       if(ch >= OFFSET) {
433         if(make_string)
434           buf.append(Constants.OPCODE_NAMES[ch - OFFSET]);
435         else
436           buf.append((int)(ch - OFFSET));
437       } else
438         buf.append(ch);
439     }
440 
441     return buf.toString();
442   }
443 }