From 156466b10c4ffc896c1fd3c0e97a18259ac11d67 Mon Sep 17 00:00:00 2001
From: Caroline Tice <cmtice@google.com>
Date: Fri, 18 Dec 2020 00:59:16 -0800
Subject: [PATCH 12/14] gold: dwp: implement read DW_FORM_strx and co

Apply additional patch by David Blaikie to update the gold dwp tool
to handle DWARF v5.  This patch teaches dwp to read DW_FORM_strx (and
its variations), so it can read and use the str_offsets_base value to
read the .dwo file names out of the compilation unit DIEs.

(see https://critique-ng.corp.google.com/cl/348045503 for original patch)
(also see https://critique-ng.corp.google.com/cl/350363411)

This patch created for Chrome OS by Caroline Tice.
Date: 17-Dec-2020
[Adrian Ratiu: rebased from v2.27 to v2.36.1]
[Adrian Ratiu: regenerated as proper git format-patch.]

Change-Id: I3bf5924bcd650123914297443323a9eb140f9272
---
 elfcpp/elfcpp_swap.h |  47 +++++++++++++++++++
 gold/dwarf_reader.cc | 107 ++++++++++++++++++++++++++++++++++++++++++-
 gold/dwarf_reader.h  |  53 +++++++++++++++++++--
 3 files changed, 203 insertions(+), 4 deletions(-)

diff --git a/elfcpp/elfcpp_swap.h b/elfcpp/elfcpp_swap.h
index 4ccb48e2899..1c301f9ba0f 100644
--- a/elfcpp/elfcpp_swap.h
+++ b/elfcpp/elfcpp_swap.h
@@ -131,6 +131,13 @@ struct Valtype_base<16>
   typedef int16_t Signed_valtype;
 };
 
+template<>
+struct Valtype_base<24>
+{
+  typedef uint32_t Valtype;
+  typedef int32_t Signed_valtype;
+};
+
 template<>
 struct Valtype_base<32>
 {
@@ -330,6 +337,46 @@ struct Swap_unaligned<16, true>
   }
 };
 
+template<>
+struct Swap_unaligned<24, false>
+{
+  typedef Valtype_base<24>::Valtype Valtype;
+
+  static inline Valtype
+  readval(const unsigned char* wv)
+  {
+    return (wv[2] << 16) | (wv[1] << 8) | wv[0];
+  }
+
+  static inline void
+  writeval(unsigned char* wv, Valtype v)
+  {
+    wv[0] = v >> 16;
+    wv[1] = v >> 8;
+    wv[2] = v;
+  }
+};
+
+template<>
+struct Swap_unaligned<24, true>
+{
+  typedef Valtype_base<24>::Valtype Valtype;
+
+  static inline Valtype
+  readval(const unsigned char* wv)
+  {
+    return (wv[0] << 16) | (wv[1] << 8) | wv[2];
+  }
+
+  static inline void
+  writeval(unsigned char* wv, Valtype v)
+  {
+    wv[0] = v >> 16;
+    wv[1] = v >> 8;
+    wv[2] = v;
+  }
+};
+
 template<>
 struct Swap_unaligned<32, false>
 {
diff --git a/gold/dwarf_reader.cc b/gold/dwarf_reader.cc
index 04d69353efa..2b6f04518c1 100644
--- a/gold/dwarf_reader.cc
+++ b/gold/dwarf_reader.cc
@@ -33,6 +33,7 @@
 #include "dwarf_reader.h"
 #include "int_encoding.h"
 #include "compressed_output.h"
+#include "../elfcpp/dwarf.h"
 
 namespace gold {
 
@@ -651,7 +652,7 @@ Dwarf_die::Dwarf_die(
     child_offset_(0), sibling_offset_(0), abbrev_code_(NULL), attributes_(),
     attributes_read_(false), name_(NULL), name_off_(-1), linkage_name_(NULL),
     linkage_name_off_(-1), string_shndx_(0), specification_(0),
-    abstract_origin_(0)
+    abstract_origin_(0), string_offsets_base_(0)
 {
   size_t len;
   const unsigned char* pdie = dwinfo->buffer_at_offset(die_offset);
@@ -789,6 +790,8 @@ Dwarf_die::read_attributes()
 	    break;
 	  case elfcpp::DW_FORM_data1:
 	  case elfcpp::DW_FORM_flag:
+	  case elfcpp::DW_FORM_strx1:
+	  case elfcpp::DW_FORM_addrx1:
 	    attr_value.val.intval = *pattr++;
 	    break;
 	  case elfcpp::DW_FORM_ref1:
@@ -796,6 +799,8 @@ Dwarf_die::read_attributes()
 	    ref_form = true;
 	    break;
 	  case elfcpp::DW_FORM_data2:
+	  case elfcpp::DW_FORM_strx2:
+	  case elfcpp::DW_FORM_addrx2:
 	    attr_value.val.intval =
 		this->dwinfo_->read_from_pointer<16>(&pattr);
 	    break;
@@ -804,7 +809,14 @@ Dwarf_die::read_attributes()
 		this->dwinfo_->read_from_pointer<16>(&pattr);
 	    ref_form = true;
 	    break;
+	  case elfcpp::DW_FORM_strx3:
+	  case elfcpp::DW_FORM_addrx3:
+              attr_value.val.intval =
+                  this->dwinfo_->read_from_pointer<24>(&pattr);
+              break;
 	  case elfcpp::DW_FORM_data4:
+	  case elfcpp::DW_FORM_strx4:
+	  case elfcpp::DW_FORM_addrx4:
 	    {
 	      off_t sec_off;
 	      sec_off = this->dwinfo_->read_from_pointer<32>(&pattr);
@@ -858,6 +870,9 @@ Dwarf_die::read_attributes()
 	  case elfcpp::DW_FORM_udata:
 	  case elfcpp::DW_FORM_GNU_addr_index:
 	  case elfcpp::DW_FORM_GNU_str_index:
+	  case elfcpp::DW_FORM_strx:
+	  case elfcpp::DW_FORM_addrx:
+	  case elfcpp::DW_FORM_rnglistx:
 	    attr_value.val.uintval = read_unsigned_LEB_128(pattr, &len);
 	    pattr += len;
 	    break;
@@ -908,6 +923,10 @@ Dwarf_die::read_attributes()
 	    if (ref_form)
 	      this->abstract_origin_ = attr_value.val.refval;
 	    break;
+	  case elfcpp::DW_AT_str_offsets_base:
+	    if (ref_form)
+	      this->string_offsets_base_ = attr_value.val.refval;
+	    break;
 	  case elfcpp::DW_AT_sibling:
 	    if (ref_form && attr_value.aux.shndx == 0)
 	      this->sibling_offset_ = attr_value.val.refval;
@@ -992,15 +1011,25 @@ Dwarf_die::skip_attributes()
 	    }
 	  case elfcpp::DW_FORM_data1:
 	  case elfcpp::DW_FORM_ref1:
+	  case elfcpp::DW_FORM_strx1:
+	  case elfcpp::DW_FORM_addrx1:
 	  case elfcpp::DW_FORM_flag:
 	    pattr += 1;
 	    break;
 	  case elfcpp::DW_FORM_data2:
 	  case elfcpp::DW_FORM_ref2:
+	  case elfcpp::DW_FORM_strx2:
+	  case elfcpp::DW_FORM_addrx2:
 	    pattr += 2;
 	    break;
+	  case elfcpp::DW_FORM_strx3:
+	  case elfcpp::DW_FORM_addrx3:
+            pattr += 3;
+            break;
 	  case elfcpp::DW_FORM_data4:
 	  case elfcpp::DW_FORM_ref4:
+	  case elfcpp::DW_FORM_strx4:
+	  case elfcpp::DW_FORM_addrx4:
 	    pattr += 4;
 	    break;
 	  case elfcpp::DW_FORM_data8:
@@ -1012,6 +1041,8 @@ Dwarf_die::skip_attributes()
 	  case elfcpp::DW_FORM_udata:
 	  case elfcpp::DW_FORM_GNU_addr_index:
 	  case elfcpp::DW_FORM_GNU_str_index:
+	  case elfcpp::DW_FORM_strx:
+	  case elfcpp::DW_FORM_addrx:
 	    read_unsigned_LEB_128(pattr, &len);
 	    pattr += len;
 	    break;
@@ -1078,6 +1109,16 @@ Dwarf_die::string_attribute(unsigned int attr)
     return NULL;
   switch (attr_val->form)
     {
+      case elfcpp::DW_FORM_strx:
+      case elfcpp::DW_FORM_strx1:
+      case elfcpp::DW_FORM_strx2:
+      case elfcpp::DW_FORM_strx3:
+      case elfcpp::DW_FORM_strx4:
+        {
+          unsigned int index = attr_val->val.uintval;
+          off_t offset = this->dwinfo_->get_string_offset(index, 0);
+          return this->dwinfo_->get_string(offset, 0);
+        }
       case elfcpp::DW_FORM_string:
         return attr_val->val.stringval;
       case elfcpp::DW_FORM_strp:
@@ -1391,6 +1432,7 @@ Dwarf_info_reader::do_parse()
 			 NULL);
       if (root_die.tag() != 0)
 	{
+          this->string_offsets_base_ = root_die.string_offsets_base();
 	  // Visit the CU or TU.
 	  if (this->is_type_unit_)
 	    this->visit_type_unit(section_offset + this->cu_offset_,
@@ -1457,6 +1499,48 @@ Dwarf_info_reader::do_read_string_table(unsigned int string_shndx)
   return true;
 }
 
+
+// Read the string offsets table
+bool
+Dwarf_info_reader::do_read_string_offsets_table(unsigned int string_offsets_shndx)
+{
+  Relobj* object = this->object_;
+
+  if (string_offsets_shndx == 0)
+    {
+      for (unsigned int i = 1; i < this->object_->shnum(); ++i)
+	{
+	  std::string name = object->section_name(i);
+	  if (name == ".debug_str_offsets" || name == ".zdebug_str_offsets")
+	    {
+	      string_offsets_shndx = i;
+	      this->string_offsets_output_section_offset_ =
+		  object->output_section_offset(i);
+	      break;
+	    }
+	}
+      if (string_offsets_shndx == 0)
+	return false;
+    }
+
+  if (this->owns_string_offsets_buffer_ && this->string_offsets_buffer_ != NULL)
+    {
+      delete[] this->string_offsets_buffer_;
+      this->owns_string_offsets_buffer_ = false;
+    }
+
+  // Get the secton contents and decompress if necessary.
+  section_size_type buffer_size;
+  const unsigned char* buffer =
+      object->decompressed_section_contents(string_offsets_shndx,
+					    &buffer_size,
+					    &this->owns_string_offsets_buffer_);
+  this->string_offsets_buffer_ = buffer;
+  this->string_offsets_buffer_end_ = this->string_offsets_buffer_ + buffer_size;
+  this->string_offsets_shndx_ = string_offsets_shndx;
+  return true;
+}
+
 // Read a possibly unaligned integer of SIZE.
 template <int valsize>
 inline typename elfcpp::Valtype_base<valsize>::Valtype
@@ -1523,6 +1607,27 @@ Dwarf_info_reader::get_string(off_t str_off, unsigned int string_shndx)
   return p;
 }
 
+
+off_t //
+Dwarf_info_reader::get_string_offset(unsigned int index,
+    unsigned int string_offsets_shndx)
+{
+  if (!this->read_string_offsets_table(string_offsets_shndx))
+    return -1;
+
+  unsigned int offset_size = this->offset_size();
+
+  const unsigned char* offsets_start = this->string_offsets_buffer_ + this->string_offsets_base_;
+
+  if ((index + 1) * offset_size > (string_offsets_buffer_end_ - offsets_start))
+    return -1;
+
+  const unsigned char* offset = offsets_start + index * offset_size;
+
+  return offset_size == 4 ? this->read_from_pointer<32>(offset)
+                          : this->read_from_pointer<64>(offset);
+}
+
 // The following are default, do-nothing, implementations of the
 // hook methods normally provided by a derived class.  We provide
 // default implementations rather than no implementation so that
diff --git a/gold/dwarf_reader.h b/gold/dwarf_reader.h
index e0efabd9004..c8f6c2eb3bc 100644
--- a/gold/dwarf_reader.h
+++ b/gold/dwarf_reader.h
@@ -568,6 +568,15 @@ class Dwarf_die
     return this->abstract_origin_;
   }
 
+  // Return the value of the DW_AT_str_offsets_base attribute.
+  off_t
+  string_offsets_base()
+  {
+    if (!this->string_offsets_base_)
+      this->read_attributes();
+    return this->string_offsets_base_;
+  }
+
   // Return the value of attribute ATTR as a string.
   const char*
   string_attribute(unsigned int attr);
@@ -675,6 +684,8 @@ class Dwarf_die
   off_t specification_;
   // The value of a DW_AT_abstract_origin attribute.
   off_t abstract_origin_;
+  // The value of DW_AT_str_offsets_base attribute.
+  off_t string_offsets_base_;
 };
 
 // This class is used to read the debug info from the .debug_info
@@ -698,11 +709,13 @@ class Dwarf_info_reader
     : is_type_unit_(is_type_unit), object_(object), symtab_(symtab),
       symtab_size_(symtab_size), shndx_(shndx), reloc_shndx_(reloc_shndx),
       reloc_type_(reloc_type), abbrev_shndx_(0), string_shndx_(0),
-      buffer_(NULL), buffer_end_(NULL), cu_offset_(0), cu_length_(0),
-      offset_size_(0), address_size_(0), cu_version_(0),
+      string_offsets_shndx_(0), buffer_(NULL), buffer_end_(NULL), cu_offset_(0),
+      cu_length_(0), offset_size_(0), address_size_(0), cu_version_(0),
       abbrev_table_(), ranges_table_(this),
       reloc_mapper_(NULL), string_buffer_(NULL), string_buffer_end_(NULL),
-      owns_string_buffer_(false), string_output_section_offset_(0)
+      owns_string_buffer_(false), string_output_section_offset_(0),
+      string_offsets_buffer_(NULL), string_offsets_buffer_end_(NULL),
+      owns_string_offsets_buffer_(false), string_offsets_output_section_offset_(0)
   { }
 
   virtual
@@ -712,6 +725,8 @@ class Dwarf_info_reader
       delete this->reloc_mapper_;
     if (this->owns_string_buffer_ && this->string_buffer_ != NULL)
       delete[] this->string_buffer_;
+    if (this->owns_string_offsets_buffer_ && this->string_offsets_buffer_ != NULL)
+      delete[] this->string_offsets_buffer_;
   }
 
   // Begin parsing the debug info.  This calls visit_compilation_unit()
@@ -754,6 +769,10 @@ class Dwarf_info_reader
   const char*
   get_string(off_t str_off, unsigned int string_shndx);
 
+  // Return a string index from the DWARF string offsets table..
+  off_t
+  get_string_offset(unsigned int index, unsigned int string_offsets_shndx);
+
   // Return the size of a DWARF offset.
   unsigned int
   offset_size() const
@@ -878,6 +897,22 @@ class Dwarf_info_reader
   bool
   do_read_string_table(unsigned int string_shndx);
 
+  // Read the DWARF string table.
+  bool
+  read_string_offsets_table(unsigned int string_offsets_shndx)
+  {
+    // If we've already read this string table, return immediately.
+    if (this->string_offsets_shndx_ > 0 &&
+	this->string_offsets_shndx_ == string_offsets_shndx)
+      return true;
+    if (string_offsets_shndx == 0 && this->string_offsets_shndx_ > 0)
+      return true;
+    return this->do_read_string_offsets_table(string_offsets_shndx);
+  }
+
+  bool
+  do_read_string_offsets_table(unsigned int string_shndx);
+  
   // True if this is a type unit; false for a compilation unit.
   bool is_type_unit_;
   // The object containing the .debug_info or .debug_types input section.
@@ -896,6 +931,8 @@ class Dwarf_info_reader
   unsigned int abbrev_shndx_;
   // Index of the .debug_str section.
   unsigned int string_shndx_;
+ // Index of the .debug_str_offsets section.
+  unsigned int string_offsets_shndx_;
   // The buffer for the debug info.
   const unsigned char* buffer_;
   const unsigned char* buffer_end_;
@@ -925,6 +962,16 @@ class Dwarf_info_reader
   // from relocated data will be relative to the output section, and need
   // to be corrected before reading data from the input section.
   uint64_t string_output_section_offset_;
+  const unsigned char* string_offsets_buffer_;
+  const unsigned char* string_offsets_buffer_end_;
+  unsigned int string_offsets_base_;
+  // True if this object owns the buffer and needs to delete it.
+  bool owns_string_offsets_buffer_;
+  // For incremental update links, this will hold the offset of the
+  // input .debug_str section within the output section.  Offsets read
+  // from relocated data will be relative to the output section, and need
+  // to be corrected before reading data from the input section.
+  uint64_t string_offsets_output_section_offset_;
 };
 
 // We can't do better than to keep the offsets in a sorted vector.
-- 
2.32.0