vdso_support: support DT_GNU_HASH

This library provides `LookupSymbol` and `LookupSymbolByAddress`. The latter
needs `GetNumSymbols` support. An object file needs either .hash (DT_HASH) or
.gnu.hash (DT_GNU_HASH). This patch adds DT_GNU_HASH support.

Note: glibc has been supporting DT_GNU_HASH since 2006 and .hash has been quite
obsoleted in the Linux communities.
PiperOrigin-RevId: 648459622
Change-Id: I3aa1274cd4617990844258175715e3be2343afd2

2 files changed, 51 insertions, 23 deletions

diff --git a/absl/debugging/internal/elf_mem_image.cc b/absl/debugging/internal/elf_mem_image.cc
index 42dcd3cd..2c168309 100644
--- a/absl/debugging/internal/elf_mem_image.cc
+++ b/absl/debugging/internal/elf_mem_image.cc
@@ -20,8 +20,11 @@
 #ifdef ABSL_HAVE_ELF_MEM_IMAGE  // defined in elf_mem_image.h
 
 #include <string.h>
+
 #include <cassert>
 #include <cstddef>
+#include <cstdint>
+
 #include "absl/base/config.h"
 #include "absl/base/internal/raw_logging.h"
 
@@ -86,20 +89,14 @@ ElfMemImage::ElfMemImage(const void *base) {
   Init(base);
 }
 
-int ElfMemImage::GetNumSymbols() const {
-  if (!hash_) {
-    return 0;
-  }
-  // See http://www.caldera.com/developers/gabi/latest/ch5.dynamic.html#hash
-  return static_cast<int>(hash_[1]);
-}
+uint32_t ElfMemImage::GetNumSymbols() const { return num_syms_; }
 
-const ElfW(Sym) *ElfMemImage::GetDynsym(int index) const {
+const ElfW(Sym) * ElfMemImage::GetDynsym(uint32_t index) const {
   ABSL_RAW_CHECK(index < GetNumSymbols(), "index out of range");
   return dynsym_ + index;
 }
 
-const ElfW(Versym) *ElfMemImage::GetVersym(int index) const {
+const ElfW(Versym) *ElfMemImage::GetVersym(uint32_t index) const {
   ABSL_RAW_CHECK(index < GetNumSymbols(), "index out of range");
   return versym_ + index;
 }
@@ -154,7 +151,7 @@ void ElfMemImage::Init(const void *base) {
   dynstr_    = nullptr;
   versym_    = nullptr;
   verdef_    = nullptr;
-  hash_      = nullptr;
+  num_syms_ = 0;
   strsize_   = 0;
   verdefnum_ = 0;
   // Sentinel: PT_LOAD .p_vaddr can't possibly be this.
@@ -219,12 +216,17 @@ void ElfMemImage::Init(const void *base) {
       base_as_char - reinterpret_cast<const char *>(link_base_);
   ElfW(Dyn)* dynamic_entry = reinterpret_cast<ElfW(Dyn)*>(
       static_cast<intptr_t>(dynamic_program_header->p_vaddr) + relocation);
+  uint32_t *sysv_hash = nullptr;
+  uint32_t *gnu_hash = nullptr;
   for (; dynamic_entry->d_tag != DT_NULL; ++dynamic_entry) {
     const auto value =
         static_cast<intptr_t>(dynamic_entry->d_un.d_val) + relocation;
     switch (dynamic_entry->d_tag) {
       case DT_HASH:
-        hash_ = reinterpret_cast<ElfW(Word) *>(value);
+        sysv_hash = reinterpret_cast<uint32_t *>(value);
+        break;
+      case DT_GNU_HASH:
+        gnu_hash = reinterpret_cast<uint32_t *>(value);
         break;
       case DT_SYMTAB:
         dynsym_ = reinterpret_cast<ElfW(Sym) *>(value);
@@ -249,13 +251,38 @@ void ElfMemImage::Init(const void *base) {
         break;
     }
   }
-  if (!hash_ || !dynsym_ || !dynstr_ || !versym_ ||
+  if ((!sysv_hash && !gnu_hash) || !dynsym_ || !dynstr_ || !versym_ ||
       !verdef_ || !verdefnum_ || !strsize_) {
     assert(false);  // invalid VDSO
     // Mark this image as not present. Can not recur infinitely.
     Init(nullptr);
     return;
   }
+  if (sysv_hash) {
+    num_syms_ = sysv_hash[1];
+  } else {
+    assert(gnu_hash);
+    // Compute the number of symbols for DT_GNU_HASH, which is specified by
+    // https://sourceware.org/gnu-gabi/program-loading-and-dynamic-linking.txt
+    uint32_t nbuckets = gnu_hash[0];
+    // The buckets array is located after the header (4 uint32) and the bloom
+    // filter (size_t array of gnu_hash[2] elements).
+    uint32_t *buckets = gnu_hash + 4 + sizeof(size_t) / 4 * gnu_hash[2];
+    // Find the chain of the last non-empty bucket.
+    uint32_t idx = 0;
+    for (uint32_t i = nbuckets; i > 0;) {
+      idx = buckets[--i];
+      if (idx != 0) break;
+    }
+    if (idx != 0) {
+      // Find the last element of the chain, which has an odd value.
+      // Add one to get the number of symbols.
+      uint32_t *chain = buckets + nbuckets - gnu_hash[1];
+      while (chain[idx++] % 2 == 0) {
+      }
+    }
+    num_syms_ = idx;
+  }
 }
 
 bool ElfMemImage::LookupSymbol(const char *name,
@@ -300,9 +327,9 @@ bool ElfMemImage::LookupSymbolByAddress(const void *address,
   return false;
 }
 
-ElfMemImage::SymbolIterator::SymbolIterator(const void *const image, int index)
-    : index_(index), image_(image) {
-}
+ElfMemImage::SymbolIterator::SymbolIterator(const void *const image,
+                                            uint32_t index)
+    : index_(index), image_(image) {}
 
 const ElfMemImage::SymbolInfo *ElfMemImage::SymbolIterator::operator->() const {
   return &info_;
@@ -335,7 +362,7 @@ ElfMemImage::SymbolIterator ElfMemImage::end() const {
   return SymbolIterator(this, GetNumSymbols());
 }
 
-void ElfMemImage::SymbolIterator::Update(int increment) {
+void ElfMemImage::SymbolIterator::Update(uint32_t increment) {
   const ElfMemImage *image = reinterpret_cast<const ElfMemImage *>(image_);
   ABSL_RAW_CHECK(image->IsPresent() || increment == 0, "");
   if (!image->IsPresent()) {
diff --git a/absl/debugging/internal/elf_mem_image.h b/absl/debugging/internal/elf_mem_image.h
index e7fe6ab0..19c4952e 100644
--- a/absl/debugging/internal/elf_mem_image.h
+++ b/absl/debugging/internal/elf_mem_image.h
@@ -22,6 +22,7 @@
 // Including this will define the __GLIBC__ macro if glibc is being
 // used.
 #include <climits>
+#include <cstdint>
 
 #include "absl/base/config.h"
 
@@ -82,10 +83,10 @@ class ElfMemImage {
     bool operator!=(const SymbolIterator &rhs) const;
     bool operator==(const SymbolIterator &rhs) const;
    private:
-    SymbolIterator(const void *const image, int index);
-    void Update(int incr);
+    SymbolIterator(const void *const image, uint32_t index);
+    void Update(uint32_t incr);
     SymbolInfo info_;
-    int index_;
+    uint32_t index_;
     const void *const image_;
   };
 
@@ -94,14 +95,14 @@ class ElfMemImage {
   void                 Init(const void *base);
   bool                 IsPresent() const { return ehdr_ != nullptr; }
   const ElfW(Phdr)*    GetPhdr(int index) const;
-  const ElfW(Sym)*     GetDynsym(int index) const;
-  const ElfW(Versym)*  GetVersym(int index) const;
+  const ElfW(Sym) * GetDynsym(uint32_t index) const;
+  const ElfW(Versym)*  GetVersym(uint32_t index) const;
   const ElfW(Verdef)*  GetVerdef(int index) const;
   const ElfW(Verdaux)* GetVerdefAux(const ElfW(Verdef) *verdef) const;
   const char*          GetDynstr(ElfW(Word) offset) const;
   const void*          GetSymAddr(const ElfW(Sym) *sym) const;
   const char*          GetVerstr(ElfW(Word) offset) const;
-  int                  GetNumSymbols() const;
+  uint32_t GetNumSymbols() const;
 
   SymbolIterator begin() const;
   SymbolIterator end() const;
@@ -124,8 +125,8 @@ class ElfMemImage {
   const ElfW(Sym) *dynsym_;
   const ElfW(Versym) *versym_;
   const ElfW(Verdef) *verdef_;
-  const ElfW(Word) *hash_;
   const char *dynstr_;
+  uint32_t num_syms_;
   size_t strsize_;
   size_t verdefnum_;
   ElfW(Addr) link_base_;     // Link-time base (p_vaddr of first PT_LOAD).