[MemoryInfra] Support dump providers running on SequencedTaskRunner

base/trace_event/memory_dump_manager.cc

 // Copyright 2015 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 "base/trace_event/memory_dump_manager.h"
 
 #include <algorithm>
 #include <utility>
 
 #include "base/atomic_sequence_num.h"
@@ skipping 166 matching lines @@
   TRACE_EVENT0(kTraceCategory, "init");  // Add to trace-viewer category list.
   TraceLog::GetInstance()->AddEnabledStateObserver(this);
   if (is_tracing_already_enabled)
     OnTraceLogEnabled();
 }
 
 void MemoryDumpManager::RegisterDumpProvider(
     MemoryDumpProvider* mdp,
     const char* name,
     const scoped_refptr<SingleThreadTaskRunner>& task_runner,
+    MemoryDumpProvider::Options options) {
+  options.dumps_on_single_thread_task_runner = true;
+  RegisterDumpProviderInternal(mdp, name, task_runner, options);
+}
+
+void MemoryDumpManager::RegisterDumpProvider(
+    MemoryDumpProvider* mdp,
+    const char* name,
+    const scoped_refptr<SingleThreadTaskRunner>& task_runner) {
+  // Set |dumps_on_single_thread_task_runner| to true because all providers
+  // without task runner are run on dump thread.
+  MemoryDumpProvider::Options options;
+  options.dumps_on_single_thread_task_runner = true;
+  RegisterDumpProviderInternal(mdp, name, task_runner, options);
+}
+
+void MemoryDumpManager::RegisterDumpProviderWithSequencedTaskRunner(
+    MemoryDumpProvider* mdp,
+    const char* name,
+    const scoped_refptr<SequencedTaskRunner>& task_runner,
+    MemoryDumpProvider::Options options) {
+  DCHECK(task_runner);
+  options.dumps_on_single_thread_task_runner = false;
+  RegisterDumpProviderInternal(mdp, name, task_runner, options);
+}
+
+void MemoryDumpManager::RegisterDumpProviderInternal(
+    MemoryDumpProvider* mdp,
+    const char* name,
+    const scoped_refptr<SequencedTaskRunner>& task_runner,
     const MemoryDumpProvider::Options& options) {
   if (dumper_registrations_ignored_for_testing_)
     return;
 
   scoped_refptr<MemoryDumpProviderInfo> mdpinfo =
       new MemoryDumpProviderInfo(mdp, name, task_runner, options);
 
   {
     AutoLock lock(lock_);
     bool already_registered = !dump_providers_.insert(mdpinfo).second;
     // This actually happens in some tests which don't have a clean tear-down
     // path for RenderThreadImpl::Init().
     if (already_registered)
       return;
   }
 
   if (heap_profiling_enabled_)
     mdp->OnHeapProfilingEnabled(true);
 }
 
-void MemoryDumpManager::RegisterDumpProvider(
-    MemoryDumpProvider* mdp,
-    const char* name,
-    const scoped_refptr<SingleThreadTaskRunner>& task_runner) {
-  RegisterDumpProvider(mdp, name, task_runner, MemoryDumpProvider::Options());
-}
-
 void MemoryDumpManager::UnregisterDumpProvider(MemoryDumpProvider* mdp) {
   UnregisterDumpProviderInternal(mdp, false /* delete_async */);
 }
 
 void MemoryDumpManager::UnregisterAndDeleteDumpProviderSoon(
     scoped_ptr<MemoryDumpProvider> mdp) {
   UnregisterDumpProviderInternal(mdp.release(), true /* delete_async */);
 }
 
 void MemoryDumpManager::UnregisterDumpProviderInternal(
@@ skipping 10 matching lines @@
     if ((*mdp_iter)->dump_provider == mdp)
       break;
   }
 
   if (mdp_iter == dump_providers_.end())
     return;  // Not registered / already unregistered.
 
   if (take_mdp_ownership_and_delete_async) {
     // The MDP will be deleted whenever the MDPInfo struct will, that is either:
     // - At the end of this function, if no dump is in progress.
-    // - In the prologue of the ContinueAsyncProcessDump().
+    // - In FinalizeCurrentDump() after the current dump finishes.
     DCHECK(!(*mdp_iter)->owned_dump_provider);
     (*mdp_iter)->owned_dump_provider = std::move(owned_mdp);
   } else if (subtle::NoBarrier_Load(&memory_tracing_enabled_)) {
     // If you hit this DCHECK, your dump provider has a bug.
     // Unregistration of a MemoryDumpProvider is safe only if:
-    // - The MDP has specified a thread affinity (via task_runner()) AND
-    //   the unregistration happens on the same thread (so the MDP cannot
+    // - The MDP has specified a sequenced task runner affinity AND the
+    //   unregistration happens on the same task runner. So that the MDP cannot
     //   unregister and be in the middle of a OnMemoryDump() at the same time.
-    // - The MDP has NOT specified a thread affinity and its ownership is
+    // - The MDP has NOT specified a task runner affinity and its ownership is
     //   transferred via UnregisterAndDeleteDumpProviderSoon().
     // In all the other cases, it is not possible to guarantee that the
     // unregistration will not race with OnMemoryDump() calls.
     DCHECK((*mdp_iter)->task_runner &&
-           (*mdp_iter)->task_runner->BelongsToCurrentThread())
+           (*mdp_iter)->task_runner->RunsTasksOnCurrentThread())
         << "MemoryDumpProvider \"" << (*mdp_iter)->name << "\" attempted to "
         << "unregister itself in a racy way. Please file a crbug.";
   }
 
   // The MDPInfo instance can still be referenced by the
   // |ProcessMemoryDumpAsyncState.pending_dump_providers|. For this reason
-  // the MDPInfo is flagged as disabled. It will cause ContinueAsyncProcessDump
+  // the MDPInfo is flagged as disabled. It will cause InvokeOnMemoryDump()
   // to just skip it, without actually invoking the |mdp|, which might be
   // destroyed by the caller soon after this method returns.
   (*mdp_iter)->disabled = true;
   dump_providers_.erase(mdp_iter);
 }
 
 void MemoryDumpManager::RequestGlobalDump(
     MemoryDumpType dump_type,
     MemoryDumpLevelOfDetail level_of_detail,
     const MemoryDumpCallback& callback) {
@@ skipping 39 matching lines @@
 
 void MemoryDumpManager::CreateProcessDump(const MemoryDumpRequestArgs& args,
                                           const MemoryDumpCallback& callback) {
   TRACE_EVENT_NESTABLE_ASYNC_BEGIN0(kTraceCategory, "ProcessMemoryDump",
                                     TRACE_ID_MANGLE(args.dump_guid));
 
   scoped_ptr<ProcessMemoryDumpAsyncState> pmd_async_state;
   {
     AutoLock lock(lock_);
     // |dump_thread_| can be nullptr is tracing was disabled before reaching
-    // here. ContinueAsyncProcessDump is robust enough to tolerate it and will
+    // here. SetupNextMemoryDump is robust enough to tolerate it and will
     // NACK the dump.
     pmd_async_state.reset(new ProcessMemoryDumpAsyncState(
         args, dump_providers_, session_state_, callback,
         dump_thread_ ? dump_thread_->task_runner() : nullptr));
   }
 
   TRACE_EVENT_WITH_FLOW0(kTraceCategory, "MemoryDumpManager::CreateProcessDump",
                          TRACE_ID_MANGLE(args.dump_guid),
                          TRACE_EVENT_FLAG_FLOW_OUT);
 
-  // Start the thread hop. |dump_providers_| are kept sorted by thread, so
-  // ContinueAsyncProcessDump will hop at most once per thread (w.r.t. thread
-  // affinity specified by the MemoryDumpProvider(s) in RegisterDumpProvider()).
-  ContinueAsyncProcessDump(pmd_async_state.release());
+  // Start the process dump. This involves task runner hops as specified by the
+  // MemoryDumpProvider(s) in RegisterDumpProvider()).
+  SetupNextMemoryDump(std::move(pmd_async_state));
 }
 
-// At most one ContinueAsyncProcessDump() can be active at any time for a given
-// PMD, regardless of status of the |lock_|. |lock_| is used here purely to
-// ensure consistency w.r.t. (un)registrations of |dump_providers_|.
-// The linearization of dump providers' OnMemoryDump invocations is achieved by
-// means of subsequent PostTask(s).
-//
-// 1) Prologue:
-//   - If this was the last hop, create a trace event, add it to the trace
-//     and finalize (invoke callback).
-//   - Check if we are on the right thread. If not hop and continue there.
-//   - Check if the dump provider is disabled, if so skip the dump.
-// 2) Invoke the dump provider's OnMemoryDump() (unless skipped).
-// 3) Epilogue:
-//   - Unregister the dump provider if it failed too many times consecutively.
-//   - Pop() the MDP from the |pending_dump_providers| list, eventually
-//     destroying the MDPInfo if that was unregistered in the meantime.
-void MemoryDumpManager::ContinueAsyncProcessDump(
-    ProcessMemoryDumpAsyncState* owned_pmd_async_state) {
+// PostTask InvokeOnMemoryDump() to the dump provider's sequenced task runner. A
+// PostTask is always required for SequencedTaskRunner with no additional

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

s/SequencedTaskRunner with no additional guarantees/generic SequencedTaskRunner/

[ssid, 2016/02/11 01:45:03]

Done.

+// guarantees to ensure that no other task is running on it concurrently.
+// SetupNextMemoryDump, InvokeOnMemoryDump and FinalizeCurrentDump are called in

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

add () after function names, e.g., SetupNextMemoryDump(),
InvokeOnMemoryDump()...

[ssid, 2016/02/11 01:45:03]

Done.

+// series which linearizes the dump provider's OnMemoryDump invocations.
+// At most one of either SetupNextMemoryDump() or InvokeOnMemoryDump() can be
+// active at any time for a given PMD, regardless of status of the |lock_|.
+// |lock_| is used in these functions purely to ensure consistency w.r.t.
+// (un)registrations of |dump_providers_|.
+void MemoryDumpManager::SetupNextMemoryDump(
+    scoped_ptr<ProcessMemoryDumpAsyncState> pmd_async_state) {
+  if (pmd_async_state->pending_dump_providers.empty())
+    return FinalizeCurrentDump(std::move(pmd_async_state));
+
   // Initalizes the ThreadLocalEventBuffer to guarantee that the TRACE_EVENTs

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

I think you still need to keep this chunk first, to avoid surprises in the case
of an empty dump. Move lines 378-381 before the initial if.

[ssid, 2016/02/11 01:45:03]

Done.

   // in the PostTask below don't end up registering their own dump providers
   // (for discounting trace memory overhead) while holding the |lock_|.
   TraceLog::GetInstance()->InitializeThreadLocalEventBufferIfSupported();
 
+  // Read MemoryDumpProviderInfo thread safety considerations in
+  // memory_dump_manager.h when accessing |mdpinfo| fields.
+  MemoryDumpProviderInfo* mdpinfo =
+      pmd_async_state->pending_dump_providers.back().get();
+
+  // If the dump provider did not specify a task runner affinity, dump on
+  // |dump_thread_|. Note that |dump_thread_| might have been destroyed
+  // meanwhile.
+  SequencedTaskRunner* task_runner = mdpinfo->task_runner.get();
+  if (!task_runner) {
+    DCHECK(mdpinfo->options.dumps_on_single_thread_task_runner);
+    task_runner = pmd_async_state->dump_thread_task_runner.get();
+    if (!task_runner) {
+      // If tracing was disabled before reaching CreateProcessDump() the
+      // dump_thread_ would have been already torn down. Nack current dump and
+      // finalize.
+      pmd_async_state->dump_successful = false;
+      FinalizeCurrentDump(std::move(pmd_async_state));

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

micro-nit. for consistency with above I'd do the call and return in one line,
i.e return FinalizeCurrentDump...
this also saves one precious line :)

[ssid, 2016/02/11 01:45:03]

Done.

+      return;
+    }
+  }
+
+  if (mdpinfo->options.dumps_on_single_thread_task_runner &&
+      task_runner->RunsTasksOnCurrentThread()) {
+    // If |dumps_on_single_thread_task_runner| is true then no PostTask is
+    // required if we are on the right thread. So invoke dump on the current

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

I'd shorten this saying just:

// If |dumps_on_single_thread_task_runner| is true then no PostTask is
// required as we are already on the right thread.

[ssid, 2016/02/11 01:45:03]

Done.

+    // thread.
+    InvokeOnMemoryDump(pmd_async_state.release());
+    return;

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

ditto return Invoke...

[ssid, 2016/02/11 01:45:03]

Done.

+  }
+
+  bool did_post_task = task_runner->PostTask(
+      FROM_HERE, Bind(&MemoryDumpManager::InvokeOnMemoryDump, Unretained(this),
+                      Unretained(pmd_async_state.get())));
+
+  if (did_post_task) {
+    // Ownership is tranferred to InvokeOnMemoryDump().
+    ignore_result(pmd_async_state.release());
+    return;
+  }
+
+  // PostTask usually fails only if the process or thread is shut down. So, the
+  // dump provider is disabled here. But, don't disable unbound dump providers.
+  // The utility thread is normally shutdown when disabling the trace and
+  // getting here in this case is expected.
+  if (mdpinfo->task_runner) {
+    LOG(ERROR) << "Disabling MemoryDumpProvider \"" << mdpinfo->name
+               << "\". Failed to post task on the task runner provided.";
+
+    // A locked access is required to R/W |disabled| (for the
+    // UnregisterAndDeleteDumpProviderSoon() case).
+    AutoLock lock(lock_);
+    mdpinfo->disabled = true;
+  }
+
+  // PostTask failed. Ignore the MDP and call FinalizeCurrentDump.
+  FinalizeCurrentDump(std::move(pmd_async_state));
+}
+
+// This function is called on the right task runner for current MDP. It is
+// either the task runner specified by MDP or |dump_thread_task_runner| if the
+// MDP did not specify task runner. Invokes the dump provider's OnMemoryDump()
+// (unless disabled).
+void MemoryDumpManager::InvokeOnMemoryDump(
+    ProcessMemoryDumpAsyncState* owned_pmd_async_state) {
   // In theory |owned_pmd_async_state| should be a scoped_ptr. The only reason
-  // why it isn't is because of the corner case logic of |did_post_task| below,
-  // which needs to take back the ownership of the |pmd_async_state| when a
-  // thread goes away and consequently the PostTask() fails.
+  // why it isn't is because of the corner case logic of |did_post_task|
+  // above, which needs to take back the ownership of the |pmd_async_state| when
+  // the PostTask() fails.
   // Unfortunately, PostTask() destroys the scoped_ptr arguments upon failure
   // to prevent accidental leaks. Using a scoped_ptr would prevent us to to
   // skip the hop and move on. Hence the manual naked -> scoped ptr juggling.
   auto pmd_async_state = make_scoped_ptr(owned_pmd_async_state);
   owned_pmd_async_state = nullptr;
 
-  if (pmd_async_state->pending_dump_providers.empty())
-    return FinalizeDumpAndAddToTrace(std::move(pmd_async_state));
-
   // Read MemoryDumpProviderInfo thread safety considerations in
   // memory_dump_manager.h when accessing |mdpinfo| fields.
   MemoryDumpProviderInfo* mdpinfo =
       pmd_async_state->pending_dump_providers.back().get();
 
-  // If the dump provider did not specify a thread affinity, dump on
-  // |dump_thread_|. Note that |dump_thread_| might have been Stop()-ed at this
-  // point (if tracing was disabled in the meanwhile). In such case the
-  // PostTask() below will fail, but |task_runner| should always be non-null.
-  SingleThreadTaskRunner* task_runner = mdpinfo->task_runner.get();
-  if (!task_runner)
-    task_runner = pmd_async_state->dump_thread_task_runner.get();
+  DCHECK(!mdpinfo->task_runner ||
+         mdpinfo->task_runner->RunsTasksOnCurrentThread());
 
-  bool post_task_failed = false;
-  if (!task_runner) {
-    // If tracing was disabled before reaching CreateProcessDump() |task_runner|
-    // will be null, as the dump_thread_ would have been already torn down.
-    post_task_failed = true;
-    pmd_async_state->dump_successful = false;
-  } else if (!task_runner->BelongsToCurrentThread()) {
-    // It's time to hop onto another thread.
-    post_task_failed = !task_runner->PostTask(
-        FROM_HERE, Bind(&MemoryDumpManager::ContinueAsyncProcessDump,
-                        Unretained(this), Unretained(pmd_async_state.get())));
-    if (!post_task_failed) {
-      // Ownership is tranferred to the next ContinueAsyncProcessDump().
-      ignore_result(pmd_async_state.release());
-      return;
+  bool should_dump;
+  {
+    // A locked access is required to R/W |disabled| (for the
+    // UnregisterAndDeleteDumpProviderSoon() case).
+    AutoLock lock(lock_);
+
+    // Unregister the dump provider if it failed too many times consecutively.
+    if (!mdpinfo->disabled &&
+        mdpinfo->consecutive_failures >= kMaxConsecutiveFailuresCount) {
+      mdpinfo->disabled = true;
+      LOG(ERROR) << "Disabling MemoryDumpProvider \"" << mdpinfo->name
+                 << "\". Dump failed multiple times consecutively.";
     }
-  }
-
-  // At this point either:
-  // - The MDP has a task runner affinity and we are on the right thread.
-  // - The MDP has a task runner affinity but the underlying thread is gone,
-  //   hence the above |post_task_failed| == true.
-  // - The MDP does NOT have a task runner affinity. A locked access is required
-  //   to R/W |disabled| (for the UnregisterAndDeleteDumpProviderSoon() case).
-  bool should_dump;
-  const char* disabled_reason = nullptr;
-  {
-    AutoLock lock(lock_);
-    if (!mdpinfo->disabled) {
-      if (mdpinfo->consecutive_failures >= kMaxConsecutiveFailuresCount) {
-        mdpinfo->disabled = true;
-        disabled_reason =
-            "Dump failure, possibly related with sandboxing (crbug.com/461788)."
-            " Try --no-sandbox.";
-      } else if (post_task_failed && mdpinfo->task_runner) {
-        // Don't disable unbound dump providers. The utility thread is normally
-        // shutdown when disabling the trace and getting here in this case is
-        // expected.
-        mdpinfo->disabled = true;
-        disabled_reason = "The thread it was meant to dump onto is gone.";
-      }
-    }
-    should_dump = !mdpinfo->disabled && !post_task_failed;
+    should_dump = !mdpinfo->disabled;
   }  // AutoLock lock(lock_);
 
-  if (disabled_reason) {
-    LOG(ERROR) << "Disabling MemoryDumpProvider \"" << mdpinfo->name << "\". "
-               << disabled_reason;
-  }
-
   if (should_dump) {
     // Invoke the dump provider.
     TRACE_EVENT_WITH_FLOW1(kTraceCategory,
-                           "MemoryDumpManager::ContinueAsyncProcessDump",
+                           "MemoryDumpManager::SetupNextMemoryDump",

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

Shouldn't this be InvokeOnMemoryDump?

[ssid, 2016/02/11 01:45:03]

Done.

                            TRACE_ID_MANGLE(pmd_async_state->req_args.dump_guid),
                            TRACE_EVENT_FLAG_FLOW_IN | TRACE_EVENT_FLAG_FLOW_OUT,
                            "dump_provider.name", mdpinfo->name);
 
     // Pid of the target process being dumped. Often kNullProcessId (= current
     // process), non-zero when the coordinator process creates dumps on behalf
     // of child processes (see crbug.com/461788).
     ProcessId target_pid = mdpinfo->options.target_pid;
     ProcessMemoryDump* pmd =
         pmd_async_state->GetOrCreateMemoryDumpContainerForProcess(target_pid);
     MemoryDumpArgs args = {pmd_async_state->req_args.level_of_detail};
     bool dump_successful = mdpinfo->dump_provider->OnMemoryDump(args, pmd);
     mdpinfo->consecutive_failures =
         dump_successful ? 0 : mdpinfo->consecutive_failures + 1;
-  }  // if (!mdpinfo->disabled)
+  }
 
-  pmd_async_state->pending_dump_providers.pop_back();
-  ContinueAsyncProcessDump(pmd_async_state.release());
+  FinalizeCurrentDump(std::move(pmd_async_state));
+}
+

[Primiano Tucci (use gerrit), 2016/02/08 12:03:15]

Hmm this function:
- Has a misleading name, especially considering that now we have both
FinalizeCurrentDump and FinalizeDumpAndAddToTrace, but they now refer to two
different finalization stages.
- Doesn't seem to do anything more than the epilogue of SetupNextMemoryDump.

Why don't we make the sequence just
SetupNextMemoryDump -> InvokeOnMemoryDump -> SetupNextMemoryDump ... eventually
goes to FinalizeAndAddToTrace at the end.

 which feels a bit more easier to follow in my mind

[ssid, 2016/02/11 01:45:03]

I did not want multiple functions destroying mdpinfo. It's little hard to guess
where all MDP is destroyed. Also I had to copy these 2 lines at few places. That
is why made a function.

+void MemoryDumpManager::FinalizeCurrentDump(
+    scoped_ptr<ProcessMemoryDumpAsyncState> pmd_async_state) {
+  // Pop() the MDP from the |pending_dump_providers| list, eventually destroying
+  // the MDPInfo if that was unregistered in the meantime.
+  if (!pmd_async_state->pending_dump_providers.empty())
+    pmd_async_state->pending_dump_providers.pop_back();
+
+  // Call SetupNextMemoryDump to continue with the next dump provider in the
+  // process dump.
+  if (!pmd_async_state->pending_dump_providers.empty())
+    return SetupNextMemoryDump(std::move(pmd_async_state));
+
+  // If this was the last hop, create a trace event, add it to the trace and
+  // finalize process dump (invoke callback).
+  FinalizeDumpAndAddToTrace(std::move(pmd_async_state));
 }
 
 // static
 void MemoryDumpManager::FinalizeDumpAndAddToTrace(
     scoped_ptr<ProcessMemoryDumpAsyncState> pmd_async_state) {
   DCHECK(pmd_async_state->pending_dump_providers.empty());
   const uint64_t dump_guid = pmd_async_state->req_args.dump_guid;
   if (!pmd_async_state->callback_task_runner->BelongsToCurrentThread()) {
     scoped_refptr<SingleThreadTaskRunner> callback_task_runner =
         pmd_async_state->callback_task_runner;
@@ skipping 132 matching lines @@
   // state is always accessed by the dumping methods holding the |lock_|.
   subtle::NoBarrier_Store(&memory_tracing_enabled_, 0);
   scoped_ptr<Thread> dump_thread;
   {
     AutoLock lock(lock_);
     dump_thread = std::move(dump_thread_);
     session_state_ = nullptr;
   }
 
   // Thread stops are blocking and must be performed outside of the |lock_|
-  // or will deadlock (e.g., if ContinueAsyncProcessDump() tries to acquire it).
+  // or will deadlock (e.g., if SetupNextMemoryDump() tries to acquire it).
   periodic_dump_timer_.Stop();
   if (dump_thread)
     dump_thread->Stop();
 }
 
 uint64_t MemoryDumpManager::GetTracingProcessId() const {
   return delegate_->GetTracingProcessId();
 }
 
 MemoryDumpManager::MemoryDumpProviderInfo::MemoryDumpProviderInfo(
     MemoryDumpProvider* dump_provider,
     const char* name,
-    const scoped_refptr<SingleThreadTaskRunner>& task_runner,
+    const scoped_refptr<SequencedTaskRunner>& task_runner,
     const MemoryDumpProvider::Options& options)
     : dump_provider(dump_provider),
       name(name),
       task_runner(task_runner),
       options(options),
       consecutive_failures(0),
       disabled(false) {}
 
 MemoryDumpManager::MemoryDumpProviderInfo::~MemoryDumpProviderInfo() {}
 
@@ skipping 33 matching lines @@
   auto iter = process_dumps.find(pid);
   if (iter == process_dumps.end()) {
     scoped_ptr<ProcessMemoryDump> new_pmd(new ProcessMemoryDump(session_state));
     iter = process_dumps.insert(std::make_pair(pid, std::move(new_pmd))).first;
   }
   return iter->second.get();
 }
 
 }  // namespace trace_event
 }  // namespace base