[LSan] Detect dynamic loader by its base address.

[alekseyshl]

Summary

Relanding D33859, which was reverted because it has "broken LOTS of
ARM/AArch64 bots for two days".

If it breaks something again, please provide some pointers to broken
bots, not just revert it, otherwise it's very hard to reason what's
wrong with this commit.

Whenever possible (Linux + glibc 2.16+), detect dynamic loader module by
its base address, not by the module name matching. The current name
matching approach fails on some configurations.

lib/lsan/lsan_common.cc

if (m.allocated() && m.tag() != kReachable && m.tag() != kIgnored) {
     if (caller_pc == 0 || (param->skip_linker_allocations &&
                            GetLinker()->containsAddress(caller_pc))) {
       m.set_tag(kReachable);
       param->frontier->push_back(chunk);
     }
   }
 }
 
-// On Linux, handles dynamically allocated TLS blocks by treating all chunks
-// allocated from ld-linux.so as reachable.
+// On Linux, treats all chunks allocated from ld-linux.so as reachable, which
+// covers dynamically allocated TLS blocks, internal dynamic loader's loaded
+// modules accounting etc.
 // Dynamic TLS blocks contain the TLS variables of dynamically loaded modules.
 // They are allocated with a __libc_memalign() call in allocate_and_init()
 // (elf/dl-tls.c). Glibc won't tell us the address ranges occupied by those
 // blocks, but we can make sure they come from our own allocator by intercepting
 // __libc_memalign(). On top of that, there is no easy way to reach them. Their
 // addresses are stored in a dynamically allocated array (the DTV) which is
 // referenced from the static TLS. Unfortunately, we can't just rely on the DTV
 // being reachable from the static TLS, and the dynamic TLS being reachable from

lib/lsan/lsan_common_linux.cc

 #if CAN_SANITIZE_LEAKS && SANITIZER_LINUX
 #include <link.h>
 
 #include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_linux.h"
 #include "sanitizer_common/sanitizer_stackdepot.h"
 
+#if SANITIZER_USE_GETAUXVAL
+#include <sys/auxv.h>
+#endif  // SANITIZER_USE_GETAUXVAL
+
 namespace __lsan {
 
 static const char kLinkerName[] = "ld";
 
 static char linker_placeholder[sizeof(LoadedModule)] ALIGNED(64);
 static LoadedModule *linker = nullptr;
 
-static bool IsLinker(const char* full_name) {
-  return LibraryNameIs(full_name, kLinkerName);
+static bool IsLinker(const LoadedModule& module) {
+#if SANITIZER_USE_GETAUXVAL
+  return module.base_address() == getauxval(AT_BASE);
+#else
+  return LibraryNameIs(module.full_name(), kLinkerName);
+#endif  // SANITIZER_USE_GETAUXVAL
 }
 
 __attribute__((tls_model("initial-exec")))
 THREADLOCAL int disable_counter;
 bool DisabledInThisThread() { return disable_counter > 0; }
 void DisableInThisThread() { disable_counter++; }
 void EnableInThisThread() {
   if (disable_counter == 0) {
     DisableCounterUnderflow();
   }
   disable_counter--;
 }
 
 void InitializePlatformSpecificModules() {
   ListOfModules modules;
   modules.init();
   for (LoadedModule &module : modules) {
-    if (!IsLinker(module.full_name())) continue;
+    if (!IsLinker(module))
+      continue;
     if (linker == nullptr) {
       linker = reinterpret_cast<LoadedModule *>(linker_placeholder);
       *linker = module;
       module = LoadedModule();
     } else {
       VReport(1, "LeakSanitizer: Multiple modules match \"%s\". "
-              "TLS will not be handled correctly.\n", kLinkerName);
+              "TLS and other allocations originating from linker might be "
+              "falsely reported as leaks.\n", kLinkerName);
       linker->clear();
       linker = nullptr;
       return;
     }
   }
   if (linker == nullptr) {
-    VReport(1, "LeakSanitizer: Dynamic linker not found. "
-               "TLS will not be handled correctly.\n");
+    VReport(1, "LeakSanitizer: Dynamic linker not found. TLS and other "
+               "allocations originating from linker might be falsely reported "
+                "as leaks.\n");
   }
 }
 
 static int ProcessGlobalRegionsCallback(struct dl_phdr_info *info, size_t size,
                                         void *data) {
   Frontier *frontier = reinterpret_cast<Frontier *>(data);
   for (uptr j = 0; j < info->dlpi_phnum; j++) {
     const ElfW(Phdr) *phdr = &(info->dlpi_phdr[j]);

lib/sanitizer_common/sanitizer_linux.cc

 #include <sys/sysctl.h>
 extern char **environ;  // provided by crt1
 #endif                  // SANITIZER_NETBSD
 
 #if !SANITIZER_ANDROID
 #include <sys/signal.h>
 #endif
 
-#ifndef __GLIBC_PREREQ
-#define __GLIBC_PREREQ(x, y) 0
-#endif
-
-#if SANITIZER_LINUX && __GLIBC_PREREQ(2, 16)
-# define SANITIZER_USE_GETAUXVAL 1
-#else
-# define SANITIZER_USE_GETAUXVAL 0
-#endif
-
 #if SANITIZER_USE_GETAUXVAL
 #include <sys/auxv.h>
 #endif
 
 #if SANITIZER_LINUX
 // <linux/time.h>
 struct kernel_timeval {
   long tv_sec;

lib/sanitizer_common/sanitizer_platform.h

 // dlopen mallocs "libgcc_s.so" string which confuses LSan, it fails to realize
 // that this allocation happens in dynamic linker and should be ignored.
 #if SANITIZER_PPC || defined(__thumb__)
 # define SANITIZER_SUPPRESS_LEAK_ON_PTHREAD_EXIT 1
 #else
 # define SANITIZER_SUPPRESS_LEAK_ON_PTHREAD_EXIT 0
 #endif
 
+#ifndef __GLIBC_PREREQ
+#define __GLIBC_PREREQ(x, y) 0
+#endif
+
+#if SANITIZER_LINUX && __GLIBC_PREREQ(2, 16)
+# define SANITIZER_USE_GETAUXVAL 1
+#else
+# define SANITIZER_USE_GETAUXVAL 0
+#endif
 
 #endif // SANITIZER_PLATFORM_H