[llvm-commits] CVS: llvm/lib/Bytecode/Reader/Reader.cpp

Reid Spencer reid at x10sys.com
Sun Nov 7 12:21:06 CST 2004



Changes in directory llvm/lib/Bytecode/Reader:

Reader.cpp updated: 1.136 -> 1.137
---
Log message:

* Comments & cleanup per CL code review.
* Accept 'llvc' as signature for compressed bytecode


---
Diffs of the changes:  (+55 -21)

Index: llvm/lib/Bytecode/Reader/Reader.cpp
diff -u llvm/lib/Bytecode/Reader/Reader.cpp:1.136 llvm/lib/Bytecode/Reader/Reader.cpp:1.137
--- llvm/lib/Bytecode/Reader/Reader.cpp:1.136	Sat Nov  6 17:17:23 2004
+++ llvm/lib/Bytecode/Reader/Reader.cpp	Sun Nov  7 12:20:55 2004
@@ -2153,19 +2153,40 @@
     error("Function declared, but bytecode stream ended before definition");
 }
 
-static unsigned GetUncompressionBuffer(char*&buff, unsigned& sz, void* ctxt){
+/// This function handles allocation of the buffer used for decompression of
+/// compressed bytecode files. It is called by Compressor::decompress which is
+/// called by BytecodeReader::ParseBytecode. 
+static unsigned GetDecompressionBuffer(char*&buff, unsigned& sz, void* ctxt){
+  // Case the context variable to our BufferInfo
   BytecodeReader::BufferInfo* bi = 
     reinterpret_cast<BytecodeReader::BufferInfo*>(ctxt);
+
+  // Compute the new, doubled, size of the block
   unsigned new_size = bi->size * 2;
+
+  // Extend or allocate the block (realloc(0,n) == malloc(n))
+  char* new_buff = (char*) ::realloc(bi->buff, new_size);
+
+  // Figure out what to return to the Compressor. If this is the first call,
+  // then bi->buff will be null. In this case we want to return the entire
+  // buffer because there was no previous allocation.  Otherwise, when the
+  // buffer is reallocated, we save the new base pointer in the BufferInfo.buff
+  // field but return the address of only the extension, mid-way through the
+  // buffer (since its size was doubled). Furthermore, the sz result must be
+  // 1/2 the total size of the buffer.
   if (bi->buff == 0 ) {
-    buff = bi->buff = (char*) malloc(new_size);
+    buff = bi->buff = new_buff;
     sz = new_size;
   } else {
-    bi->buff = (char*) ::realloc(bi->buff, new_size);
-    buff = bi->buff + bi->size;
+    bi->buff = new_buff;
+    buff = new_buff + bi->size;
     sz = bi->size;
   }
+
+  // Retain the size of the allocated block
   bi->size = new_size;
+
+  // Make sure we fail (return 1) if we didn't get any memory.
   return (bi->buff == 0 ? 1 : 0);
 }
 
@@ -2183,26 +2204,39 @@
 
     if (Handler) Handler->handleStart(TheModule, Length);
 
-    // Read and check signature...
-    bool compressed = 
-      (Buf[0] == 0xEC && Buf[1] == 0xEC && Buf[2] == 0xF6 && Buf[3] == 0xED);
-
-    if (compressed) {
-      bi.size = Length * 2;;
-      // Bytecode is compressed, have to decompress it first.
-      unsigned uncompressedLength = Compressor::decompress((char*)Buf+4,Length-4,
-        GetUncompressionBuffer, (void*) &bi);
+    // Read the four bytes of the signature.
+    unsigned Sig = read_uint();
+
+    // If this is a compressed file
+    if (Sig == ('l' | ('l' << 8) | ('v' << 16) | ('c' << 24))) {
 
+      // Compute the initial length of the uncompression buffer. Note that this
+      // is twice the length of the compressed buffer and will be doubled again
+      // in GetDecompressionBuffer for an initial allocation of 4xLength.  This 
+      // calculation is based on the typical compression ratio of bzip2 on LLVM 
+      // bytecode files which typically ranges in the 50%-75% range.   Since we 
+      // tyipcally get at least 50%, doubling is insufficient. By using a 4x 
+      // multiplier on the first allocation, we minimize the impact of having to
+      // copy the buffer on reallocation.
+      bi.size = Length * 2;
+
+      // Invoke the decompression of the bytecode. Note that we have to skip the
+      // file's magic number which is not part of the compressed block. Hence,
+      // the Buf+4 and Length-4.
+      unsigned decompressedLength = Compressor::decompress((char*)Buf+4,Length-4,
+        GetDecompressionBuffer, (void*) &bi);
+
+      // We must adjust the buffer pointers used by the bytecode reader to point
+      // into the new decompressed block. After decompression, the BufferInfo
+      // structure (member bi), will point to a contiguous memory area that has
+      // the decompressed data.
       At = MemStart = BlockStart = Buf = (BufPtr) bi.buff;
-      MemEnd = BlockEnd = Buf + uncompressedLength;
+      MemEnd = BlockEnd = Buf + decompressedLength;
 
-    } else {
-      if (!(Buf[0] == 'l' && Buf[1] == 'l' && Buf[2] == 'v' && Buf[3] == 'm'))
-        error("Invalid bytecode signature: " + 
-            utohexstr(Buf[0]) + utohexstr(Buf[1]) + utohexstr(Buf[2]) +
-            utohexstr(Buf[3]));
-      else
-        At += 4; // skip the bytes
+    // else if this isn't a regular (uncompressed) bytecode file, then its
+    // and error, generate that now.
+    } else if (Sig != ('l' | ('l' << 8) | ('v' << 16) | ('m' << 24))) {
+      error("Invalid bytecode signature: " + utohexstr(Sig));
     }
 
     // Tell the handler we're starting a module






More information about the llvm-commits mailing list