1 /* 2 * reserved comment block 3 * DO NOT REMOVE OR ALTER! 4 */ 5 package com.sun.org.apache.bcel.internal.util; 6 7 /* 8 * Licensed to the Apache Software Foundation (ASF) under one or more 9 * contributor license agreements. See the NOTICE file distributed with 10 * this work for additional information regarding copyright ownership. 11 * The ASF licenses this file to You under the Apache License, Version 2.0 12 * (the "License"); you may not use this file except in compliance with 13 * the License. You may obtain a copy of the License at 14 * 15 * http://www.apache.org/licenses/LICENSE-2.0 16 * 17 * Unless required by applicable law or agreed to in writing, software 18 * distributed under the License is distributed on an "AS IS" BASIS, 19 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 20 * See the License for the specific language governing permissions and 21 * limitations under the License. 22 * 23 */ 24 25 import com.sun.org.apache.bcel.internal.Constants; 26 import com.sun.org.apache.bcel.internal.generic.*; 27 import java.util.*; 28 import java.util.regex.Matcher; 29 import java.util.regex.Pattern; 30 31 /** 32 * InstructionFinder is a tool to search for given instructions patterns, 33 * i.e., match sequences of instructions in an instruction list via 34 * regular expressions. This can be used, e.g., in order to implement 35 * a peep hole optimizer that looks for code patterns and replaces 36 * them with faster equivalents. 37 * 38 * <p>This class internally uses the <a href="http://jakarta.apache.org/regexp/"> 39 * Regexp</a> package to search for regular expressions. 40 * 41 * A typical application would look like this: 42 <pre> 43 InstructionFinder f = new InstructionFinder(il); 44 String pat = "IfInstruction ICONST_0 GOTO ICONST_1 NOP (IFEQ|IFNE)"; 45 46 for(Iterator i = f.search(pat, constraint); i.hasNext(); ) { 47 InstructionHandle[] match = (InstructionHandle[])i.next(); 48 ... 49 il.delete(match[1], match[5]); 50 ... 51 } 52 </pre> 53 * @author <A HREF="http://www.berlin.de/~markus.dahm/">M. Dahm</A> 54 * @see Instruction 55 * @see InstructionList 56 */ 57 public class InstructionFinder { 58 private static final int OFFSET = 32767; // char + OFFSET is outside of LATIN-1 59 private static final int NO_OPCODES = 256; // Potential number, some are not used 60 61 private static final HashMap map = new HashMap(); // Map<String,Pattern> 62 63 private InstructionList il; 64 private String il_string; // instruction list as string 65 private InstructionHandle[] handles; // map instruction list to array 66 67 /** 68 * @param il instruction list to search for given patterns 69 */ InstructionFinder(InstructionList il)70 public InstructionFinder(InstructionList il) { 71 this.il = il; 72 reread(); 73 } 74 75 /** 76 * Reread the instruction list, e.g., after you've altered the list upon a match. 77 */ reread()78 public final void reread() { 79 int size = il.getLength(); 80 char[] buf = new char[size]; // Create a string with length equal to il length 81 handles = il.getInstructionHandles(); 82 83 // Map opcodes to characters 84 for(int i=0; i < size; i++) 85 buf[i] = makeChar(handles[i].getInstruction().getOpcode()); 86 87 il_string = new String(buf); 88 } 89 90 /** 91 * Map symbolic instruction names like "getfield" to a single character. 92 * 93 * @param pattern instruction pattern in lower case 94 * @return encoded string for a pattern such as "BranchInstruction". 95 */ mapName(String pattern)96 private static final String mapName(String pattern) { 97 String result = (String)map.get(pattern); 98 99 if(result != null) 100 return result; 101 102 for(short i=0; i < NO_OPCODES; i++) 103 if(pattern.equals(Constants.OPCODE_NAMES[i])) 104 return "" + makeChar(i); 105 106 throw new RuntimeException("Instruction unknown: " + pattern); 107 } 108 109 /** 110 * Replace symbolic names of instructions with the appropiate character and remove 111 * all white space from string. Meta characters such as +, * are ignored. 112 * 113 * @param pattern The pattern to compile 114 * @return translated regular expression string 115 */ compilePattern(String pattern)116 private static final String compilePattern(String pattern) { 117 String lower = pattern.toLowerCase(); 118 StringBuffer buf = new StringBuffer(); 119 int size = pattern.length(); 120 121 for(int i=0; i < size; i++) { 122 char ch = lower.charAt(i); 123 124 if(Character.isLetterOrDigit(ch)) { 125 StringBuffer name = new StringBuffer(); 126 127 while((Character.isLetterOrDigit(ch) || ch == '_') && i < size) { 128 name.append(ch); 129 130 if(++i < size) 131 ch = lower.charAt(i); 132 else 133 break; 134 } 135 136 i--; 137 138 buf.append(mapName(name.toString())); 139 } else if(!Character.isWhitespace(ch)) 140 buf.append(ch); 141 } 142 143 return buf.toString(); 144 } 145 146 /** 147 * @return the matched piece of code as an array of instruction (handles) 148 */ getMatch(int matched_from, int match_length)149 private InstructionHandle[] getMatch(int matched_from, int match_length) { 150 InstructionHandle[] match = new InstructionHandle[match_length]; 151 System.arraycopy(handles, matched_from, match, 0, match_length); 152 153 return match; 154 } 155 156 /** 157 * Search for the given pattern in the instruction list. You can search for any valid 158 * opcode via its symbolic name, e.g. "istore". You can also use a super class or 159 * an interface name to match a whole set of instructions, e.g. "BranchInstruction" or 160 * "LoadInstruction". "istore" is also an alias for all "istore_x" instructions. Additional 161 * aliases are "if" for "ifxx", "if_icmp" for "if_icmpxx", "if_acmp" for "if_acmpxx". 162 * 163 * Consecutive instruction names must be separated by white space which will be removed 164 * during the compilation of the pattern. 165 * 166 * For the rest the usual pattern matching rules for regular expressions apply.<P> 167 * Example pattern: 168 * <pre> 169 search("BranchInstruction NOP ((IfInstruction|GOTO)+ ISTORE Instruction)*"); 170 * </pre> 171 * 172 * <p>If you alter the instruction list upon a match such that other 173 * matching areas are affected, you should call reread() to update 174 * the finder and call search() again, because the matches are cached. 175 * 176 * @param pattern the instruction pattern to search for, where case is ignored 177 * @param from where to start the search in the instruction list 178 * @param constraint optional CodeConstraint to check the found code pattern for 179 * user-defined constraints 180 * @return iterator of matches where e.nextElement() returns an array of instruction handles 181 * describing the matched area 182 */ search(String pattern, InstructionHandle from, CodeConstraint constraint)183 public final Iterator search(String pattern, InstructionHandle from, 184 CodeConstraint constraint) 185 { 186 String search = compilePattern(pattern); 187 int start = -1; 188 189 for(int i=0; i < handles.length; i++) { 190 if(handles[i] == from) { 191 start = i; // Where to start search from (index) 192 break; 193 } 194 } 195 196 if(start == -1) 197 throw new ClassGenException("Instruction handle " + from + 198 " not found in instruction list."); 199 200 Pattern regex = Pattern.compile(search); 201 List<InstructionHandle[]> matches = new ArrayList<>(); 202 Matcher matcher = regex.matcher(il_string); 203 while(start < il_string.length() && matcher.find(start)) { 204 int startExpr = matcher.start(); 205 int endExpr = matcher.end(); 206 int lenExpr = endExpr - startExpr; 207 InstructionHandle[] match = getMatch(startExpr, lenExpr); 208 209 if((constraint == null) || constraint.checkCode(match)) 210 matches.add(match); 211 start = endExpr; 212 } 213 214 return matches.iterator(); 215 } 216 217 /** 218 * Start search beginning from the start of the given instruction list. 219 * 220 * @param pattern the instruction pattern to search for, where case is ignored 221 * @return iterator of matches where e.nextElement() 222 * returns an array of instruction handles describing the matched 223 * area 224 */ search(String pattern)225 public final Iterator search(String pattern) { 226 return search(pattern, il.getStart(), null); 227 } 228 229 /** 230 * Start search beginning from `from'. 231 * 232 * @param pattern the instruction pattern to search for, where case is ignored 233 * @param from where to start the search in the instruction list 234 * @return iterator of matches where e.nextElement() returns an array of instruction handles 235 * describing the matched area 236 */ search(String pattern, InstructionHandle from)237 public final Iterator search(String pattern, InstructionHandle from) { 238 return search(pattern, from, null); 239 } 240 241 /** 242 * Start search beginning from the start of the given instruction list. 243 * Check found matches with the constraint object. 244 * 245 * @param pattern the instruction pattern to search for, case is ignored 246 * @param constraint constraints to be checked on matching code 247 * @return instruction handle or `null' if the match failed 248 */ search(String pattern, CodeConstraint constraint)249 public final Iterator search(String pattern, CodeConstraint constraint) { 250 return search(pattern, il.getStart(), constraint); 251 } 252 253 /** 254 * Convert opcode number to char. 255 */ makeChar(short opcode)256 private static final char makeChar(short opcode) { 257 return (char)(opcode + OFFSET); 258 } 259 260 /** 261 * @return the inquired instruction list 262 */ getInstructionList()263 public final InstructionList getInstructionList() { return il; } 264 265 /** 266 * Code patterns found may be checked using an additional 267 * user-defined constraint object whether they really match the needed criterion. 268 * I.e., check constraints that can not expressed with regular expressions. 269 * 270 */ 271 public interface CodeConstraint { 272 /** 273 * @param match array of instructions matching the requested pattern 274 * @return true if the matched area is really useful 275 */ checkCode(InstructionHandle[] match)276 public boolean checkCode(InstructionHandle[] match); 277 } 278 279 // Initialize pattern map 280 281 static { 282 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)"); 283 map.put("invokeinstruction", "(invokevirtual|invokeinterface|invokestatic|invokespecial)"); 284 map.put("arrayinstruction", "(baload|aastore|saload|caload|fastore|lastore|iaload|castore|iastore|aaload|bastore|sastore|faload|laload|daload|dastore)"); 285 map.put("gotoinstruction", "(goto|goto_w)"); 286 map.put("conversioninstruction", "(d2l|l2d|i2s|d2i|l2i|i2b|l2f|d2f|f2i|i2d|i2l|f2d|i2c|f2l|i2f)"); 287 map.put("localvariableinstruction", "(fstore|iinc|lload|dstore|dload|iload|aload|astore|istore|fload|lstore)"); 288 map.put("loadinstruction", "(fload|dload|lload|iload|aload)"); 289 map.put("fieldinstruction", "(getfield|putstatic|getstatic|putfield)"); 290 map.put("cpinstruction", "(ldc2_w|invokeinterface|multianewarray|putstatic|instanceof|getstatic|checkcast|getfield|invokespecial|ldc_w|invokestatic|invokevirtual|putfield|ldc|new|anewarray)"); 291 map.put("stackinstruction", "(dup2|swap|dup2_x2|pop|pop2|dup|dup2_x1|dup_x2|dup_x1)"); 292 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)"); 293 map.put("returninstruction", "(lreturn|ireturn|freturn|dreturn|areturn|return)"); 294 map.put("storeinstruction", "(istore|fstore|dstore|astore|lstore)"); 295 map.put("select", "(tableswitch|lookupswitch)"); 296 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)"); 297 map.put("jsrinstruction", "(jsr|jsr_w)"); 298 map.put("variablelengthinstruction", "(tableswitch|jsr|goto|lookupswitch)"); 299 map.put("unconditionalbranch", "(goto|jsr|jsr_w|athrow|goto_w)"); 300 map.put("constantpushinstruction", "(dconst|bipush|sipush|fconst|iconst|lconst)"); 301 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)"); 302 map.put("popinstruction", "(fstore|dstore|pop|pop2|astore|putstatic|istore|lstore)"); 303 map.put("allocationinstruction", "(multianewarray|new|anewarray|newarray)"); 304 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)"); 305 map.put("pushinstruction", "(dup|lload|dup2|bipush|fload|ldc2_w|sipush|lconst|fconst|dload|getstatic|ldc_w|aconst_null|dconst|iload|ldc|iconst|aload)"); 306 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)"); 307 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)"); 308 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)"); 309 map.put("loadclass", "(multianewarray|invokeinterface|instanceof|invokespecial|putfield|checkcast|putstatic|invokevirtual|new|getstatic|invokestatic|getfield|anewarray)"); 310 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)"); 311 312 // Some aliases 313 map.put("if_icmp", "(if_icmpne|if_icmpeq|if_icmple|if_icmpge|if_icmplt|if_icmpgt)"); 314 map.put("if_acmp", "(if_acmpeq|if_acmpne)"); 315 map.put("if", "(ifeq|ifne|iflt|ifge|ifgt|ifle)"); 316 317 // Precompile some aliases first 318 map.put("iconst", precompile(Constants.ICONST_0, Constants.ICONST_5, Constants.ICONST_M1)); 319 map.put("lconst", new String(new char[] { '(', makeChar(Constants.LCONST_0), '|', 320 makeChar(Constants.LCONST_1), ')' })); 321 map.put("dconst", new String(new char[] { '(', makeChar(Constants.DCONST_0), '|', 322 makeChar(Constants.DCONST_1), ')' })); 323 map.put("fconst", new String(new char[] { '(', makeChar(Constants.FCONST_0), '|', 324 makeChar(Constants.FCONST_1), ')' })); 325 326 map.put("iload", precompile(Constants.ILOAD_0, Constants.ILOAD_3, Constants.ILOAD)); 327 map.put("dload", precompile(Constants.DLOAD_0, Constants.DLOAD_3, Constants.DLOAD)); 328 map.put("fload", precompile(Constants.FLOAD_0, Constants.FLOAD_3, Constants.FLOAD)); 329 map.put("aload", precompile(Constants.ALOAD_0, Constants.ALOAD_3, Constants.ALOAD)); 330 331 map.put("istore", precompile(Constants.ISTORE_0, Constants.ISTORE_3, Constants.ISTORE)); 332 map.put("dstore", precompile(Constants.DSTORE_0, Constants.DSTORE_3, Constants.DSTORE)); 333 map.put("fstore", precompile(Constants.FSTORE_0, Constants.FSTORE_3, Constants.FSTORE)); 334 map.put("astore", precompile(Constants.ASTORE_0, Constants.ASTORE_3, Constants.ASTORE)); 335 336 // Compile strings 337 338 for(Iterator i = map.keySet().iterator(); i.hasNext(); ) { 339 String key = (String)i.next(); 340 String value = (String)map.get(key); 341 342 char ch = value.charAt(1); // Omit already precompiled patterns 343 if(ch < OFFSET) { map.put(key, compilePattern(value))344 map.put(key, compilePattern(value)); // precompile all patterns 345 } 346 } 347 348 // Add instruction alias to match anything 349 350 StringBuffer buf = new StringBuffer("("); 351 352 for(short i=0; i < NO_OPCODES; i++) { 353 if(Constants.NO_OF_OPERANDS[i] != Constants.UNDEFINED) { // Not an invalid opcode makeChar(i)354 buf.append(makeChar(i)); 355 356 if(i < NO_OPCODES - 1) 357 buf.append('|'); 358 } 359 } 360 buf.append(')'); 361 362 map.put("instruction", buf.toString()); 363 } 364 precompile(short from, short to, short extra)365 private static String precompile(short from, short to, short extra) { 366 StringBuffer buf = new StringBuffer("("); 367 368 for(short i=from; i <= to; i++) { 369 buf.append(makeChar(i)); 370 buf.append('|'); 371 } 372 373 buf.append(makeChar(extra)); 374 buf.append(")"); 375 return buf.toString(); 376 } 377 378 /* 379 * Internal debugging routines. 380 */ pattern2string(String pattern)381 private static final String pattern2string(String pattern) { 382 return pattern2string(pattern, true); 383 } 384 pattern2string(String pattern, boolean make_string)385 private static final String pattern2string(String pattern, boolean make_string) { 386 StringBuffer buf = new StringBuffer(); 387 388 for(int i=0; i < pattern.length(); i++) { 389 char ch = pattern.charAt(i); 390 391 if(ch >= OFFSET) { 392 if(make_string) 393 buf.append(Constants.OPCODE_NAMES[ch - OFFSET]); 394 else 395 buf.append((int)(ch - OFFSET)); 396 } else 397 buf.append(ch); 398 } 399 400 return buf.toString(); 401 } 402 } 403