tsan: store ThreadRegistry in Context by value

[dvyukov]

Summary

It's unclear why we allocate ThreadRegistry separately,
I assume it's some historical leftover.
Embed ThreadRegistry into Context.

[melver]

compiler-rt/lib/tsan/rtl/tsan_rtl.cpp

80

Some strange alignment requirements perhaps?

Is
https://reviews.llvm.org/rGcadbb9241627eefc9f589ae4376fd9ed3e272ecc
still relevant? opaque_storage is gone though.

I see the same "placeholder" idiom used for Context itself. git blame doesn't lead me very far, so I hope you or Vitaly remember the historical background.

[dvyukov]

compiler-rt/lib/tsan/rtl/tsan_rtl.cpp

80

Re https://reviews.llvm.org/rGcadbb9241627eefc9f589ae4376fd9ed3e272ecc
Context itself is 64-byte aligned, so we now should satisfy any alignment requirement for all Context members including ThreadRegistry and any Mutexes within ThreadRegistry.

I see the same "placeholder" idiom used for Context itself

Because that's the only way we can allocate anything. We can't declare and use typed global vars.

I hope you or Vitaly remember the historical background.

No, I don't. It was like 10 years ago. I remember initially ThreadRegistry was defined within tsan runtime only, then moved to sanitizer_common and generalized. Lots of things changed at that point.

[melver]

In that case, this makes sense.

compiler-rt/lib/tsan/rtl/tsan_debugging.cpp

if (IsMetaMem(addr)) {
 
     if (b != 0) {
       region_address = (uptr)allocator()->GetBlockBegin((void *)addr);
       region_size = b->siz;
       region_kind = "heap";
     } else {
       // TODO(kuba.brecka): We should not lock. This is supposed to be called
       // from within the debugger when other threads are stopped.
-      ctx->thread_registry->Lock();
+      ctx->thread_registry.Lock();
       ThreadContext *tctx = IsThreadStackOrTls(addr, &is_stack);
-      ctx->thread_registry->Unlock();
+      ctx->thread_registry.Unlock();
       if (tctx) {
         region_kind = is_stack ? "stack" : "tls";
       } else {
         region_kind = "global";
         DataInfo info;
         if (Symbolizer::GetOrInit()->SymbolizeData(addr, &info)) {
           internal_strncpy(name, info.name, name_size);
           region_address = info.start;
if (a->PointerIsMine((void *)addr)) {
     void *block_begin = a->GetBlockBegin((void *)addr);
     if (block_begin) b = ctx->metamap.GetBlock((uptr)block_begin);
   }
   if (b == 0) return 0;
 
   *thread_id = b->tid;
   // No locking.  This is supposed to be called from within the debugger when
   // other threads are stopped.
-  ThreadContextBase *tctx = ctx->thread_registry->GetThreadLocked(b->tid);
+  ThreadContextBase *tctx = ctx->thread_registry.GetThreadLocked(b->tid);
   *os_id = tctx->os_id;
 
   StackTrace stack = StackDepotGet(b->stk);
   size = Min(size, (uptr)stack.size);
   for (uptr i = 0; i < size; i++) trace[i] = stack.trace[stack.size - i - 1];
   return size;
 }

compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp

static void CallUserSignalHandler(ThreadState *thr, bool sync, bool acquire,
   // from rtl_generic_sighandler) we have not yet received the reraised
   // signal; and it looks too fragile to intercept all ways to reraise a signal.
   if (ShouldReport(thr, ReportTypeErrnoInSignal) && !sync && sig != SIGTERM &&
       errno != 99) {
     VarSizeStackTrace stack;
     // StackTrace::GetNestInstructionPc(pc) is used because return address is
     // expected, OutputReport() will undo this.
     ObtainCurrentStack(thr, StackTrace::GetNextInstructionPc(pc), &stack);
-    ThreadRegistryLock l(ctx->thread_registry);
+    ThreadRegistryLock l(&ctx->thread_registry);
     ScopedReport rep(ReportTypeErrnoInSignal);
     if (!IsFiredSuppression(ctx, ReportTypeErrnoInSignal, stack)) {
       rep.AddStack(stack, true);
       OutputReport(thr, rep);
     }
   }
   errno = saved_errno;
 }

 
 #define COMMON_INTERCEPTOR_FD_SOCKET_ACCEPT(ctx, fd, newfd) \
   FdSocketAccept(((TsanInterceptorContext *) ctx)->thr, pc, fd, newfd)
 
 #define COMMON_INTERCEPTOR_SET_THREAD_NAME(ctx, name) \
   ThreadSetName(((TsanInterceptorContext *) ctx)->thr, name)
 
 #define COMMON_INTERCEPTOR_SET_PTHREAD_NAME(ctx, thread, name) \
-  __tsan::ctx->thread_registry->SetThreadNameByUserId(thread, name)
+  __tsan::ctx->thread_registry.SetThreadNameByUserId(thread, name)
 
 #define COMMON_INTERCEPTOR_BLOCK_REAL(name) BLOCK_REAL(name)
 
 #define COMMON_INTERCEPTOR_ON_EXIT(ctx) \
   OnExit(((TsanInterceptorContext *) ctx)->thr)
 
 #define COMMON_INTERCEPTOR_MUTEX_PRE_LOCK(ctx, m) \
   MutexPreLock(((TsanInterceptorContext *)ctx)->thr, \

compiler-rt/lib/tsan/rtl/tsan_mman.cpp

 static void SignalUnsafeCall(ThreadState *thr, uptr pc) {
   if (atomic_load_relaxed(&thr->in_signal_handler) == 0 ||
       !ShouldReport(thr, ReportTypeSignalUnsafe))
     return;
   VarSizeStackTrace stack;
   ObtainCurrentStack(thr, pc, &stack);
   if (IsFiredSuppression(ctx, ReportTypeSignalUnsafe, stack))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   ScopedReport rep(ReportTypeSignalUnsafe);
   rep.AddStack(stack, true);
   OutputReport(thr, rep);
 }
 
 
 void *user_alloc_internal(ThreadState *thr, uptr pc, uptr sz, uptr align,
                           bool signal) {

compiler-rt/lib/tsan/rtl/tsan_rtl.h

#endif
   Mutex report_mtx;
   int nreported;
   int nmissed_expected;
   atomic_uint64_t last_symbolize_time_ns;
 
   void *background_thread;
   atomic_uint32_t stop_background_thread;
 
-  ThreadRegistry *thread_registry;
+  ThreadRegistry thread_registry;
 
   Mutex racy_mtx;
   Vector<RacyStacks> racy_stacks;
   Vector<RacyAddress> racy_addresses;
   // Number of fired suppressions may be large enough.
   Mutex fired_suppressions_mtx;
   InternalMmapVector<FiredSuppression> fired_suppressions;
   DDetector *dd;

compiler-rt/lib/tsan/rtl/tsan_rtl.cpp

 #if !SANITIZER_GO
   if (auto *ptr = dlsym(RTLD_DEFAULT, "__tsan_on_initialize")) {
     return reinterpret_cast<decltype(&__tsan_on_initialize)>(ptr)();
   }
 #endif
 }
 #endif
 
-static ALIGNED(64) char thread_registry_placeholder[sizeof(ThreadRegistry)];
-
 static ThreadContextBase *CreateThreadContext(u32 tid) {
   // Map thread trace when context is created.
   char name[50];
   internal_snprintf(name, sizeof(name), "trace %u", tid);
   MapThreadTrace(GetThreadTrace(tid), TraceSize() * sizeof(Event), name);
   const uptr hdr = GetThreadTraceHeader(tid);
   internal_snprintf(name, sizeof(name), "trace header %u", tid);
   MapThreadTrace(hdr, sizeof(Trace), name);

 static const u32 kThreadQuarantineSize = 64;
 #endif
 
 Context::Context()
     : initialized(),
       report_mtx(MutexTypeReport),
       nreported(),
       nmissed_expected(),
-      thread_registry(new (thread_registry_placeholder) ThreadRegistry(
-          CreateThreadContext, kMaxTid, kThreadQuarantineSize, kMaxTidReuse)),
+      thread_registry(CreateThreadContext, kMaxTid, kThreadQuarantineSize,
+                      kMaxTidReuse),
       racy_mtx(MutexTypeRacy),
       racy_stacks(),
       racy_addresses(),
       fired_suppressions_mtx(MutexTypeFired),
       clock_alloc(LINKER_INITIALIZED, "clock allocator") {
   fired_suppressions.reserve(8);
 }
 

 #endif
 {
 }
 
 #if !SANITIZER_GO
 static void MemoryProfiler(Context *ctx, fd_t fd, int i) {
   uptr n_threads;
   uptr n_running_threads;
-  ctx->thread_registry->GetNumberOfThreads(&n_threads, &n_running_threads);
+  ctx->thread_registry.GetNumberOfThreads(&n_threads, &n_running_threads);
   InternalMmapVector<char> buf(4096);
   WriteMemoryProfile(buf.data(), buf.size(), n_threads, n_running_threads);
   WriteToFile(fd, buf.data(), internal_strlen(buf.data()));
 }
 
 static void *BackgroundThread(void *arg) {
   // This is a non-initialized non-user thread, nothing to see here.
   // We don't use ScopedIgnoreInterceptors, because we want ignores to be
#endif
 
   failed = OnFinalize(failed);
 
   return failed ? common_flags()->exitcode : 0;
 }
 
 #if !SANITIZER_GO
 void ForkBefore(ThreadState *thr, uptr pc) NO_THREAD_SAFETY_ANALYSIS {
-  ctx->thread_registry->Lock();
+  ctx->thread_registry.Lock();
   ctx->report_mtx.Lock();
   ScopedErrorReportLock::Lock();
   // Suppress all reports in the pthread_atfork callbacks.
   // Reports will deadlock on the report_mtx.
   // We could ignore sync operations as well,
   // but so far it's unclear if it will do more good or harm.
   // Unnecessarily ignoring things can lead to false positives later.
   thr->suppress_reports++;
   // On OS X, REAL(fork) can call intercepted functions (OSSpinLockLock), and
   // we'll assert in CheckNoLocks() unless we ignore interceptors.
   thr->ignore_interceptors++;
 }
 
 void ForkParentAfter(ThreadState *thr, uptr pc) NO_THREAD_SAFETY_ANALYSIS {
   thr->suppress_reports--;  // Enabled in ForkBefore.
   thr->ignore_interceptors--;
   ScopedErrorReportLock::Unlock();
   ctx->report_mtx.Unlock();
-  ctx->thread_registry->Unlock();
+  ctx->thread_registry.Unlock();
 }
 
 void ForkChildAfter(ThreadState *thr, uptr pc) NO_THREAD_SAFETY_ANALYSIS {
   thr->suppress_reports--;  // Enabled in ForkBefore.
   thr->ignore_interceptors--;
   ScopedErrorReportLock::Unlock();
   ctx->report_mtx.Unlock();
-  ctx->thread_registry->Unlock();
+  ctx->thread_registry.Unlock();
 
   uptr nthread = 0;
-  ctx->thread_registry->GetNumberOfThreads(0, 0, &nthread /* alive threads */);
+  ctx->thread_registry.GetNumberOfThreads(0, 0, &nthread /* alive threads */);
   VPrintf(1, "ThreadSanitizer: forked new process with pid %d,"
       " parent had %d threads\n", (int)internal_getpid(), (int)nthread);
   if (nthread == 1) {
     StartBackgroundThread();
   } else {
     // We've just forked a multi-threaded process. We cannot reasonably function
     // after that (some mutexes may be locked before fork). So just enable
     // ignores for everything in the hope that we will exec soon.

compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cpp

 static void ReportMutexMisuse(ThreadState *thr, uptr pc, ReportType typ,
     uptr addr, u64 mid) {
   // In Go, these misuses are either impossible, or detected by std lib,
   // or false positives (e.g. unlock in a different thread).
   if (SANITIZER_GO)
     return;
   if (!ShouldReport(thr, typ))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   ScopedReport rep(typ);
   rep.AddMutex(mid);
   VarSizeStackTrace trace;
   ObtainCurrentStack(thr, pc, &trace);
   rep.AddStack(trace, true);
   rep.AddLocation(addr, 1);
   OutputReport(thr, rep);
 }
if (flags()->report_destroy_locked && s->owner_tid != kInvalidTid &&
     unlock_locked = true;
   }
   u64 mid = s->GetId();
   u64 last_lock = s->last_lock;
   if (!unlock_locked)
     s->Reset(thr->proc());  // must not reset it before the report is printed
   s->mtx.Unlock();
   if (unlock_locked && ShouldReport(thr, ReportTypeMutexDestroyLocked)) {
-    ThreadRegistryLock l(ctx->thread_registry);
+    ThreadRegistryLock l(&ctx->thread_registry);
     ScopedReport rep(ReportTypeMutexDestroyLocked);
     rep.AddMutex(mid);
     VarSizeStackTrace trace;
     ObtainCurrentStack(thr, pc, &trace);
     rep.AddStack(trace, true);
     FastState last(last_lock);
     RestoreStack(last.tid(), last.epoch(), &trace, 0);
     rep.AddStack(trace, true);
static void UpdateClockCallback(ThreadContextBase *tctx_base, void *arg) {
   }
   thr->clock.set(&thr->proc()->clock_cache, tctx->tid, epoch);
 }
 
 void AcquireGlobal(ThreadState *thr) {
   DPrintf("#%d: AcquireGlobal\n", thr->tid);
   if (thr->ignore_sync)
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
-  ctx->thread_registry->RunCallbackForEachThreadLocked(
-      UpdateClockCallback, thr);
+  ThreadRegistryLock l(&ctx->thread_registry);
+  ctx->thread_registry.RunCallbackForEachThreadLocked(UpdateClockCallback, thr);
 }
 
 void ReleaseStoreAcquire(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
   DPrintf("#%d: ReleaseStoreAcquire %zx\n", thr->tid, addr);
   if (thr->ignore_sync)
     return;
   SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
   thr->fast_state.IncrementEpoch();
static void UpdateSleepClockCallback(ThreadContextBase *tctx_base, void *arg) {
   thr->last_sleep_clock.set(&thr->proc()->clock_cache, tctx->tid, epoch);
 }
 
 void AfterSleep(ThreadState *thr, uptr pc) {
   DPrintf("#%d: AfterSleep %zx\n", thr->tid);
   if (thr->ignore_sync)
     return;
   thr->last_sleep_stack_id = CurrentStackId(thr, pc);
-  ThreadRegistryLock l(ctx->thread_registry);
-  ctx->thread_registry->RunCallbackForEachThreadLocked(
-      UpdateSleepClockCallback, thr);
+  ThreadRegistryLock l(&ctx->thread_registry);
+  ctx->thread_registry.RunCallbackForEachThreadLocked(UpdateSleepClockCallback,
+                                                      thr);
 }
 #endif
 
 void AcquireImpl(ThreadState *thr, uptr pc, SyncClock *c) {
   if (thr->ignore_sync)
     return;
   thr->clock.set(thr->fast_state.epoch());
   thr->clock.acquire(&thr->proc()->clock_cache, c);
void AcquireReleaseImpl(ThreadState *thr, uptr pc, SyncClock *c) {
   thr->clock.set(thr->fast_state.epoch());
   thr->fast_synch_epoch = thr->fast_state.epoch();
   thr->clock.acq_rel(&thr->proc()->clock_cache, c);
 }
 
 void ReportDeadlock(ThreadState *thr, uptr pc, DDReport *r) {
   if (r == 0 || !ShouldReport(thr, ReportTypeDeadlock))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   ScopedReport rep(ReportTypeDeadlock);
   for (int i = 0; i < r->n; i++) {
     rep.AddMutex(r->loop[i].mtx_ctx0);
     rep.AddUniqueTid((int)r->loop[i].thr_ctx);
     rep.AddThread((int)r->loop[i].thr_ctx);
   }
   uptr dummy_pc = 0x42;
   for (int i = 0; i < r->n; i++) {

compiler-rt/lib/tsan/rtl/tsan_rtl_report.cpp

 
 bool ShouldReport(ThreadState *thr, ReportType typ) {
   // We set thr->suppress_reports in the fork context.
   // Taking any locking in the fork context can lead to deadlocks.
   // If any locks are already taken, it's too late to do this check.
   CheckedMutex::CheckNoLocks();
   // For the same reason check we didn't lock thread_registry yet.
   if (SANITIZER_DEBUG)
-    ThreadRegistryLock l(ctx->thread_registry);
+    ThreadRegistryLock l(&ctx->thread_registry);
   if (!flags()->report_bugs || thr->suppress_reports)
     return false;
   switch (typ) {
     case ReportTypeSignalUnsafe:
       return flags()->report_signal_unsafe;
     case ReportTypeThreadLeak:
 #if !SANITIZER_GO
       // It's impossible to join phantom threads
       // in the child after fork.
       if (ctx->after_multithreaded_fork)
         return false;
 #endif
       return flags()->report_thread_leaks;
     case ReportTypeMutexDestroyLocked:
       return flags()->report_destroy_locked;
     default:
       return true;
   }
 }
 
 ScopedReportBase::ScopedReportBase(ReportType typ, uptr tag) {
-  ctx->thread_registry->CheckLocked();
+  ctx->thread_registry.CheckLocked();
   void *mem = internal_alloc(sizeof(ReportDesc));
   rep_ = new(mem) ReportDesc;
   rep_->typ = typ;
   rep_->tag = tag;
   ctx->report_mtx.Lock();
 }
 
 ScopedReportBase::~ScopedReportBase() {

 
 #if !SANITIZER_GO
 static bool FindThreadByUidLockedCallback(ThreadContextBase *tctx, void *arg) {
   int unique_id = *(int *)arg;
   return tctx->unique_id == (u32)unique_id;
 }
 
 static ThreadContext *FindThreadByUidLocked(int unique_id) {
-  ctx->thread_registry->CheckLocked();
+  ctx->thread_registry.CheckLocked();
   return static_cast<ThreadContext *>(
-      ctx->thread_registry->FindThreadContextLocked(
+      ctx->thread_registry.FindThreadContextLocked(
           FindThreadByUidLockedCallback, &unique_id));
 }
 
 static ThreadContext *FindThreadByTidLocked(int tid) {
-  ctx->thread_registry->CheckLocked();
-  return static_cast<ThreadContext*>(
-      ctx->thread_registry->GetThreadLocked(tid));
+  ctx->thread_registry.CheckLocked();
+  return static_cast<ThreadContext *>(
+      ctx->thread_registry.GetThreadLocked(tid));
 }
 
 static bool IsInStackOrTls(ThreadContextBase *tctx_base, void *arg) {
   uptr addr = (uptr)arg;
   ThreadContext *tctx = static_cast<ThreadContext*>(tctx_base);
   if (tctx->status != ThreadStatusRunning)
     return false;
   ThreadState *thr = tctx->thr;
   CHECK(thr);
   return ((addr >= thr->stk_addr && addr < thr->stk_addr + thr->stk_size) ||
           (addr >= thr->tls_addr && addr < thr->tls_addr + thr->tls_size));
 }
 
 ThreadContext *IsThreadStackOrTls(uptr addr, bool *is_stack) {
-  ctx->thread_registry->CheckLocked();
-  ThreadContext *tctx = static_cast<ThreadContext*>(
-      ctx->thread_registry->FindThreadContextLocked(IsInStackOrTls,
-                                                    (void*)addr));
+  ctx->thread_registry.CheckLocked();
+  ThreadContext *tctx =
+      static_cast<ThreadContext *>(ctx->thread_registry.FindThreadContextLocked(
+          IsInStackOrTls, (void *)addr));
   if (!tctx)
     return 0;
   ThreadState *thr = tctx->thr;
   CHECK(thr);
   *is_stack = (addr >= thr->stk_addr && addr < thr->stk_addr + thr->stk_size);
   return tctx;
 }
 #endif
void ReportRace(ThreadState *thr) {
   for (uptr i = 0; i < kMop; i++) {
     if (tags[i] != kExternalTagNone) {
       typ = ReportTypeExternalRace;
       tag = tags[i];
       break;
     }
   }
 
-  ThreadRegistryLock l0(ctx->thread_registry);
+  ThreadRegistryLock l0(&ctx->thread_registry);
   ScopedReport rep(typ, tag);
   for (uptr i = 0; i < kMop; i++) {
     Shadow s(thr->racy_state[i]);
     rep.AddMemoryAccess(addr, tags[i], s, traces[i],
                         i == 0 ? &thr->mset : mset2);
   }
 
   for (uptr i = 0; i < kMop; i++) {
     FastState s(thr->racy_state[i]);
-    ThreadContext *tctx = static_cast<ThreadContext*>(
-        ctx->thread_registry->GetThreadLocked(s.tid()));
+    ThreadContext *tctx = static_cast<ThreadContext *>(
+        ctx->thread_registry.GetThreadLocked(s.tid()));
     if (s.epoch() < tctx->epoch0 || s.epoch() > tctx->epoch1)
       continue;
     rep.AddThread(tctx);
   }
 
   rep.AddLocation(addr_min, addr_max - addr_min);
 
 #if !SANITIZER_GO

compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp

 static void ThreadCheckIgnore(ThreadState *thr) {}
 #endif
 
 void ThreadFinalize(ThreadState *thr) {
   ThreadCheckIgnore(thr);
 #if !SANITIZER_GO
   if (!ShouldReport(thr, ReportTypeThreadLeak))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   Vector<ThreadLeak> leaks;
-  ctx->thread_registry->RunCallbackForEachThreadLocked(
-      MaybeReportThreadLeak, &leaks);
+  ctx->thread_registry.RunCallbackForEachThreadLocked(MaybeReportThreadLeak,
+                                                      &leaks);
   for (uptr i = 0; i < leaks.Size(); i++) {
     ScopedReport rep(ReportTypeThreadLeak);
     rep.AddThread(leaks[i].tctx, true);
     rep.SetCount(leaks[i].count);
     OutputReport(thr, rep);
   }
 #endif
 }
 
 int ThreadCount(ThreadState *thr) {
   uptr result;
-  ctx->thread_registry->GetNumberOfThreads(0, 0, &result);
+  ctx->thread_registry.GetNumberOfThreads(0, 0, &result);
   return (int)result;
 }
 
 int ThreadCreate(ThreadState *thr, uptr pc, uptr uid, bool detached) {
   OnCreatedArgs args = { thr, pc };
   u32 parent_tid = thr ? thr->tid : kInvalidTid;  // No parent for GCD workers.
-  int tid =
-      ctx->thread_registry->CreateThread(uid, detached, parent_tid, &args);
+  int tid = ctx->thread_registry.CreateThread(uid, detached, parent_tid, &args);
   DPrintf("#%d: ThreadCreate tid=%d uid=%zu\n", parent_tid, tid, uid);
   return tid;
 }
 
 void ThreadStart(ThreadState *thr, int tid, tid_t os_id,
                  ThreadType thread_type) {
   uptr stk_addr = 0;
   uptr stk_size = 0;
   uptr tls_addr = 0;
   uptr tls_size = 0;
 #if !SANITIZER_GO
   if (thread_type != ThreadType::Fiber)
     GetThreadStackAndTls(tid == kMainTid, &stk_addr, &stk_size, &tls_addr,
                          &tls_size);
 
   if (tid != kMainTid) {
     if (stk_addr && stk_size)
       MemoryRangeImitateWrite(thr, /*pc=*/ 1, stk_addr, stk_size);
 
     if (tls_addr && tls_size) ImitateTlsWrite(thr, tls_addr, tls_size);
   }
 #endif
 
-  ThreadRegistry *tr = ctx->thread_registry;
+  ThreadRegistry *tr = &ctx->thread_registry;
   OnStartedArgs args = { thr, stk_addr, stk_size, tls_addr, tls_size };
   tr->StartThread(tid, os_id, thread_type, &args);
 
   tr->Lock();
   thr->tctx = (ThreadContext*)tr->GetThreadLocked(tid);
   tr->Unlock();
 
 #if !SANITIZER_GO
   if (ctx->after_multithreaded_fork) {
     thr->ignore_interceptors++;
     ThreadIgnoreBegin(thr, 0);
     ThreadIgnoreSyncBegin(thr, 0);
   }
 #endif
 }
 
 void ThreadFinish(ThreadState *thr) {
   ThreadCheckIgnore(thr);
   if (thr->stk_addr && thr->stk_size)
     DontNeedShadowFor(thr->stk_addr, thr->stk_size);
   if (thr->tls_addr && thr->tls_size)
     DontNeedShadowFor(thr->tls_addr, thr->tls_size);
   thr->is_dead = true;
-  ctx->thread_registry->FinishThread(thr->tid);
+  ctx->thread_registry.FinishThread(thr->tid);
 }
 
 struct ConsumeThreadContext {
   uptr uid;
   ThreadContextBase *tctx;
 };
 
 static bool ConsumeThreadByUid(ThreadContextBase *tctx, void *arg) {
if (tctx->user_id == findCtx->uid && tctx->status != ThreadStatusInvalid) {
     findCtx->tctx = tctx;
     tctx->user_id = 0;
   }
   return false;
 }
 
 int ThreadConsumeTid(ThreadState *thr, uptr pc, uptr uid) {
   ConsumeThreadContext findCtx = {uid, nullptr};
-  ctx->thread_registry->FindThread(ConsumeThreadByUid, &findCtx);
+  ctx->thread_registry.FindThread(ConsumeThreadByUid, &findCtx);
   int tid = findCtx.tctx ? findCtx.tctx->tid : kInvalidTid;
   DPrintf("#%d: ThreadTid uid=%zu tid=%d\n", thr->tid, uid, tid);
   return tid;
 }
 
 void ThreadJoin(ThreadState *thr, uptr pc, int tid) {
   CHECK_GT(tid, 0);
   CHECK_LT(tid, kMaxTid);
   DPrintf("#%d: ThreadJoin tid=%d\n", thr->tid, tid);
-  ctx->thread_registry->JoinThread(tid, thr);
+  ctx->thread_registry.JoinThread(tid, thr);
 }
 
 void ThreadDetach(ThreadState *thr, uptr pc, int tid) {
   CHECK_GT(tid, 0);
   CHECK_LT(tid, kMaxTid);
-  ctx->thread_registry->DetachThread(tid, thr);
+  ctx->thread_registry.DetachThread(tid, thr);
 }
 
 void ThreadNotJoined(ThreadState *thr, uptr pc, int tid, uptr uid) {
   CHECK_GT(tid, 0);
   CHECK_LT(tid, kMaxTid);
-  ctx->thread_registry->SetThreadUserId(tid, uid);
+  ctx->thread_registry.SetThreadUserId(tid, uid);
 }
 
 void ThreadSetName(ThreadState *thr, const char *name) {
-  ctx->thread_registry->SetThreadName(thr->tid, name);
+  ctx->thread_registry.SetThreadName(thr->tid, name);
 }
 
 void MemoryAccessRange(ThreadState *thr, uptr pc, uptr addr,
                        uptr size, bool is_write) {
   if (size == 0)
     return;
 
   u64 *shadow_mem = (u64*)MemToShadow(addr);