Refactor rel32 searching process for x64 to make it more similar to x86.

courgette/rel32_finder_x64.cc

Index: courgette/rel32_finder_x64.cc
diff --git a/courgette/disassembler_win32_x64.cc b/courgette/rel32_finder_x64.cc
similarity index 52%
copy from courgette/disassembler_win32_x64.cc
copy to courgette/rel32_finder_x64.cc
index c4d89e452e62875fe402af773f60110710691236..080d68f59d1f02867b0b6167b7c8a834a73a0f96 100644
--- a/courgette/disassembler_win32_x64.cc
+++ b/courgette/rel32_finder_x64.cc
@@ -1,72 +1,37 @@
-// Copyright 2013 The Chromium Authors. All rights reserved.
+// Copyright 2016 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
-#include "courgette/disassembler_win32_x64.h"
-
-#include <algorithm>
-
-#include "base/logging.h"
-#include "base/numerics/safe_conversions.h"
-#include "courgette/assembly_program.h"
-#include "courgette/courgette.h"
-
-#if COURGETTE_HISTOGRAM_TARGETS
-#include <iostream>
-#endif
+#include "courgette/rel32_finder_x64.h"
 
 namespace courgette {
 
-DisassemblerWin32X64::DisassemblerWin32X64(const void* start, size_t length)
-    : DisassemblerWin32(start, length) {}
-
-RVA DisassemblerWin32X64::PointerToTargetRVA(const uint8_t* p) const {
-  return Address64ToRVA(Read64LittleEndian(p));
-}
-
-RVA DisassemblerWin32X64::Address64ToRVA(uint64_t address) const {
-  if (address < image_base() || address >= image_base() + size_of_image_)
-    return kNoRVA;
-  return base::checked_cast<RVA>(address - image_base());
-}
-
-CheckBool DisassemblerWin32X64::EmitAbs(Label* label,
-                                        AssemblyProgram* program) {
-  return program->EmitAbs64(label);
-}
-
-void DisassemblerWin32X64::ParseRel32RelocsFromSection(const Section* section) {
-  // TODO(sra): use characteristic.
-  bool isCode = strcmp(section->name, ".text") == 0;
-  if (!isCode)
-    return;
-
-  FileOffset start_file_offset = section->file_offset_of_raw_data;
-  FileOffset end_file_offset = start_file_offset + section->size_of_raw_data;
-  RVA relocs_start_rva = base_relocation_table().address_;
-
-  const uint8_t* start_pointer = FileOffsetToPointer(start_file_offset);
-  const uint8_t* end_pointer = FileOffsetToPointer(end_file_offset);
-
-  RVA start_rva = FileOffsetToRVA(start_file_offset);
-  RVA end_rva = start_rva + section->virtual_size;
+Rel32FinderX64::Rel32FinderX64(RVA relocs_start_rva, RVA relocs_end_rva)
+    : Rel32Finder(relocs_start_rva, relocs_end_rva) {}
 
+void Rel32FinderX64::Find(const uint8_t* start_pointer,
+                          const uint8_t* end_pointer,
+                          RVA start_rva,
+                          RVA end_rva,
+                          uint32_t size_of_image,
+                          const std::vector<RVA>& abs32_locations) {
   // Quick way to convert from Pointer to RVA within a single Section is to
-  // subtract |pointer_to_rva|.
+  // subtract 'pointer_to_rva'.

[huangs, 2016/05/26 22:11:17]

NIT: |adjust_pointer_to_rva|.

[etiennep, 2016/05/27 18:12:34]

Done.

   const uint8_t* const adjust_pointer_to_rva = start_pointer - start_rva;
 
-  std::vector<RVA>::iterator abs32_pos = abs32_locations_.begin();
+  std::vector<RVA>::const_iterator abs32_pos = abs32_locations.begin();
 
   // Find the rel32 relocations.
   const uint8_t* p = start_pointer;
   while (p < end_pointer) {
     RVA current_rva = static_cast<RVA>(p - adjust_pointer_to_rva);
-    if (current_rva == relocs_start_rva) {
-      uint32_t relocs_size = base_relocation_table().size_;
-      if (relocs_size) {
-        p += relocs_size;
-        continue;
-      }
+
+    // Skip the base reloation table if we encounter it.
+    // Note: We're not bothering to handle the edge case where a Rel32 pointer
+    // collides with |relocs_start_rva_| by being {1, 2, 3}-bytes before it.
+    if (current_rva >= relocs_start_rva_ && current_rva < relocs_end_rva_) {
+      p += relocs_end_rva_ - current_rva;
+      continue;
     }
 
     // Heuristic discovery of rel32 locations in instruction stream: are the
@@ -81,7 +46,7 @@ void DisassemblerWin32X64::ParseRel32RelocsFromSection(const Section* section) {
     }
     if (p + 6 <= end_pointer) {
       if (*p == 0x0F && (*(p + 1) & 0xF0) == 0x80) {  // Jcc long form
-        if (p[1] != 0x8A && p[1] != 0x8B)  // JPE/JPO unlikely
+        if (p[1] != 0x8A && p[1] != 0x8B)             // JPE/JPO unlikely
           rel32 = p + 2;
       } else if (*p == 0xFF && (*(p + 1) == 0x15 || *(p + 1) == 0x25)) {
         // rip relative call/jmp
@@ -107,24 +72,23 @@ void DisassemblerWin32X64::ParseRel32RelocsFromSection(const Section* section) {
       RVA rel32_rva = static_cast<RVA>(rel32 - adjust_pointer_to_rva);
 
       // Is there an abs32 reloc overlapping the candidate?
-      while (abs32_pos != abs32_locations_.end() && *abs32_pos < rel32_rva - 3)
+      while (abs32_pos != abs32_locations.end() && *abs32_pos < rel32_rva - 3)
         ++abs32_pos;
       // Now: (*abs32_pos > rel32_rva - 4) i.e. the lowest addressed 4-byte
       // region that could overlap rel32_rva.
-      if (abs32_pos != abs32_locations_.end()) {
+      if (abs32_pos != abs32_locations.end()) {
         if (*abs32_pos < rel32_rva + 4) {
           // Beginning of abs32 reloc is before end of rel32 reloc so they
-          // overlap. Skip four bytes past the abs32 reloc.
+          // overlap.  Skip four bytes past the abs32 reloc.

[huangs, 2016/05/26 22:11:17]

NIT: Remove space before "Skip".

[etiennep, 2016/05/27 18:12:34]

Done.

           p += (*abs32_pos + 4) - current_rva;
           continue;
         }
       }
 

[huangs, 2016/05/26 22:40:55]

Why did the comment disappear?

[etiennep, 2016/05/27 18:12:34]

Idk, it's back now!

-      // + 4 since offset is relative to start of next instruction.
       RVA target_rva = rel32_rva + 4 + Read32LittleEndian(rel32);
       // To be valid, rel32 target must be within image, and within this
       // section.
-      if (target_rva < size_of_image_ &&  // Subsumes rva != kUnassignedRVA.
+      if (target_rva < size_of_image &&  // Subsumes rva != kUnassignedRVA.

[huangs, 2016/05/26 22:40:55]

target_rva < size_of_image

seems kinda bogus and we should revisit this later.  Leave it alone in this
refactoring CL.  But one thing we can do is to remove |size_of_image| from
Rel32FinderX64::Find() interface, and save as a member variable of
Rel32FinderX64, to be initialized at construction.

[etiennep, 2016/05/27 18:12:34]

I moved size_of_image as an member variable. It would be tempting to put
abs32_locations as a member variable as well, since it's only of use for this
implementation (full disassemble wouldn't need to check collisions)

           (is_rip_relative ||
            (start_rva <= target_rva && target_rva < end_rva))) {
         rel32_locations_.push_back(rel32_rva);