grpc 1.50.0-x86_64-linux → 1.51.0-x86_64-linux

data/src/core/lib/event_engine/posix_engine/ev_epoll1_linux.cc

@@ -0,0 +1,618 @@
+// Copyright 2022 The gRPC Authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include <grpc/support/port_platform.h>
+
+#include "src/core/lib/event_engine/posix_engine/ev_epoll1_linux.h"
+
+#include <stdint.h>
+
+#include <atomic>
+#include <memory>
+
+#include "absl/status/status.h"
+#include "absl/status/statusor.h"
+
+#include <grpc/event_engine/event_engine.h>
+#include <grpc/support/log.h>
+#include <grpc/support/sync.h>
+
+#include "src/core/lib/event_engine/poller.h"
+#include "src/core/lib/event_engine/time_util.h"
+#include "src/core/lib/iomgr/port.h"
+
+// This polling engine is only relevant on linux kernels supporting epoll
+// epoll_create() or epoll_create1()
+#ifdef GRPC_LINUX_EPOLL
+#include <errno.h>
+#include <limits.h>
+#include <sys/epoll.h>
+#include <sys/socket.h>
+#include <unistd.h>
+
+#include "src/core/lib/event_engine/posix_engine/event_poller.h"
+#include "src/core/lib/event_engine/posix_engine/lockfree_event.h"
+#include "src/core/lib/event_engine/posix_engine/posix_engine_closure.h"
+#include "src/core/lib/event_engine/posix_engine/wakeup_fd_posix.h"
+#include "src/core/lib/event_engine/posix_engine/wakeup_fd_posix_default.h"
+#include "src/core/lib/gprpp/fork.h"
+#include "src/core/lib/gprpp/strerror.h"
+#include "src/core/lib/gprpp/sync.h"
+
+using ::grpc_event_engine::posix_engine::LockfreeEvent;
+using ::grpc_event_engine::posix_engine::WakeupFd;
+
+#define MAX_EPOLL_EVENTS_HANDLED_PER_ITERATION 1
+
+namespace grpc_event_engine {
+namespace posix_engine {
+
+using ::grpc_event_engine::experimental::EventEngine;
+using ::grpc_event_engine::experimental::Poller;
+using ::grpc_event_engine::posix_engine::LockfreeEvent;
+using ::grpc_event_engine::posix_engine::WakeupFd;
+
+class Epoll1EventHandle : public EventHandle {
+ public:
+  Epoll1EventHandle(int fd, Epoll1Poller* poller)
+      : fd_(fd),
+        list_(this),
+        poller_(poller),
+        read_closure_(std::make_unique<LockfreeEvent>(poller->GetScheduler())),
+        write_closure_(std::make_unique<LockfreeEvent>(poller->GetScheduler())),
+        error_closure_(
+            std::make_unique<LockfreeEvent>(poller->GetScheduler())) {
+    read_closure_->InitEvent();
+    write_closure_->InitEvent();
+    error_closure_->InitEvent();
+    pending_read_.store(false, std::memory_order_relaxed);
+    pending_write_.store(false, std::memory_order_relaxed);
+    pending_error_.store(false, std::memory_order_relaxed);
+  }
+  void ReInit(int fd) {
+    fd_ = fd;
+    read_closure_->InitEvent();
+    write_closure_->InitEvent();
+    error_closure_->InitEvent();
+    pending_read_.store(false, std::memory_order_relaxed);
+    pending_write_.store(false, std::memory_order_relaxed);
+    pending_error_.store(false, std::memory_order_relaxed);
+  }
+  Epoll1Poller* Poller() override { return poller_; }
+  bool SetPendingActions(bool pending_read, bool pending_write,
+                         bool pending_error) {
+    // Another thread may be executing ExecutePendingActions() at this point
+    // This is possible for instance, if one instantiation of Work(..) sets
+    // an fd to be readable while the next instantiation of Work(...) may
+    // set the fd to be writable. While the second instantiation is running,
+    // ExecutePendingActions() of the first instantiation may execute in
+    // parallel and read the pending_<***>_ variables. So we need to use
+    // atomics to manipulate pending_<***>_ variables.
+
+    if (pending_read) {
+      pending_read_.store(true, std::memory_order_release);
+    }
+
+    if (pending_write) {
+      pending_write_.store(true, std::memory_order_release);
+    }
+
+    if (pending_error) {
+      pending_error_.store(true, std::memory_order_release);
+    }
+
+    return pending_read || pending_write || pending_error;
+  }
+  int WrappedFd() override { return fd_; }
+  void OrphanHandle(PosixEngineClosure* on_done, int* release_fd,
+                    absl::string_view reason) override;
+  void ShutdownHandle(absl::Status why) override;
+  void NotifyOnRead(PosixEngineClosure* on_read) override;
+  void NotifyOnWrite(PosixEngineClosure* on_write) override;
+  void NotifyOnError(PosixEngineClosure* on_error) override;
+  void SetReadable() override;
+  void SetWritable() override;
+  void SetHasError() override;
+  bool IsHandleShutdown() override;
+  inline void ExecutePendingActions() {
+    // These may execute in Parallel with ShutdownHandle. Thats not an issue
+    // because the lockfree event implementation should be able to handle it.
+    if (pending_read_.exchange(false, std::memory_order_acq_rel)) {
+      read_closure_->SetReady();
+    }
+    if (pending_write_.exchange(false, std::memory_order_acq_rel)) {
+      write_closure_->SetReady();
+    }
+    if (pending_error_.exchange(false, std::memory_order_acq_rel)) {
+      error_closure_->SetReady();
+    }
+  }
+  grpc_core::Mutex* mu() { return &mu_; }
+  LockfreeEvent* ReadClosure() { return read_closure_.get(); }
+  LockfreeEvent* WriteClosure() { return write_closure_.get(); }
+  LockfreeEvent* ErrorClosure() { return error_closure_.get(); }
+  Epoll1Poller::HandlesList& ForkFdListPos() { return list_; }
+  ~Epoll1EventHandle() override = default;
+
+ private:
+  void HandleShutdownInternal(absl::Status why, bool releasing_fd);
+  // See Epoll1Poller::ShutdownHandle for explanation on why a mutex is
+  // required.
+  grpc_core::Mutex mu_;
+  int fd_;
+  // See Epoll1Poller::SetPendingActions for explanation on why pending_<***>_
+  // need to be atomic.
+  std::atomic<bool> pending_read_{false};
+  std::atomic<bool> pending_write_{false};
+  std::atomic<bool> pending_error_{false};
+  Epoll1Poller::HandlesList list_;
+  Epoll1Poller* poller_;
+  std::unique_ptr<LockfreeEvent> read_closure_;
+  std::unique_ptr<LockfreeEvent> write_closure_;
+  std::unique_ptr<LockfreeEvent> error_closure_;
+};
+
+namespace {
+
+int EpollCreateAndCloexec() {
+#ifdef GRPC_LINUX_EPOLL_CREATE1
+  int fd = epoll_create1(EPOLL_CLOEXEC);
+  if (fd < 0) {
+    gpr_log(GPR_ERROR, "epoll_create1 unavailable");
+  }
+#else
+  int fd = epoll_create(MAX_EPOLL_EVENTS);
+  if (fd < 0) {
+    gpr_log(GPR_ERROR, "epoll_create unavailable");
+  } else if (fcntl(fd, F_SETFD, FD_CLOEXEC) != 0) {
+    gpr_log(GPR_ERROR, "fcntl following epoll_create failed");
+    return -1;
+  }
+#endif
+  return fd;
+}
+
+// Only used when GRPC_ENABLE_FORK_SUPPORT=1
+std::list<Epoll1Poller*> fork_poller_list;
+
+// Only used when GRPC_ENABLE_FORK_SUPPORT=1
+Epoll1EventHandle* fork_fd_list_head = nullptr;
+gpr_mu fork_fd_list_mu;
+
+void ForkFdListAddHandle(Epoll1EventHandle* handle) {
+  if (grpc_core::Fork::Enabled()) {
+    gpr_mu_lock(&fork_fd_list_mu);
+    handle->ForkFdListPos().next = fork_fd_list_head;
+    handle->ForkFdListPos().prev = nullptr;
+    if (fork_fd_list_head != nullptr) {
+      fork_fd_list_head->ForkFdListPos().prev = handle;
+    }
+    fork_fd_list_head = handle;
+    gpr_mu_unlock(&fork_fd_list_mu);
+  }
+}
+
+void ForkFdListRemoveHandle(Epoll1EventHandle* handle) {
+  if (grpc_core::Fork::Enabled()) {
+    gpr_mu_lock(&fork_fd_list_mu);
+    if (fork_fd_list_head == handle) {
+      fork_fd_list_head = handle->ForkFdListPos().next;
+    }
+    if (handle->ForkFdListPos().prev != nullptr) {
+      handle->ForkFdListPos().prev->ForkFdListPos().next =
+          handle->ForkFdListPos().next;
+    }
+    if (handle->ForkFdListPos().next != nullptr) {
+      handle->ForkFdListPos().next->ForkFdListPos().prev =
+          handle->ForkFdListPos().prev;
+    }
+    gpr_mu_unlock(&fork_fd_list_mu);
+  }
+}
+
+void ForkPollerListAddPoller(Epoll1Poller* poller) {
+  if (grpc_core::Fork::Enabled()) {
+    gpr_mu_lock(&fork_fd_list_mu);
+    fork_poller_list.push_back(poller);
+    gpr_mu_unlock(&fork_fd_list_mu);
+  }
+}
+
+void ForkPollerListRemovePoller(Epoll1Poller* poller) {
+  if (grpc_core::Fork::Enabled()) {
+    gpr_mu_lock(&fork_fd_list_mu);
+    fork_poller_list.remove(poller);
+    gpr_mu_unlock(&fork_fd_list_mu);
+  }
+}
+
+bool InitEpoll1PollerLinux();
+
+// Called by the child process's post-fork handler to close open fds,
+// including the global epoll fd of each poller. This allows gRPC to shutdown in
+// the child process without interfering with connections or RPCs ongoing in the
+// parent.
+void ResetEventManagerOnFork() {
+  // Delete all pending Epoll1EventHandles.
+  gpr_mu_lock(&fork_fd_list_mu);
+  while (fork_fd_list_head != nullptr) {
+    close(fork_fd_list_head->WrappedFd());
+    Epoll1EventHandle* next = fork_fd_list_head->ForkFdListPos().next;
+    delete fork_fd_list_head;
+    fork_fd_list_head = next;
+  }
+  // Delete all registered pollers. This also closes all open epoll_sets
+  while (!fork_poller_list.empty()) {
+    Epoll1Poller* poller = fork_poller_list.front();
+    fork_poller_list.pop_front();
+    delete poller;
+  }
+  gpr_mu_unlock(&fork_fd_list_mu);
+  if (grpc_core::Fork::Enabled()) {
+    gpr_mu_destroy(&fork_fd_list_mu);
+    grpc_core::Fork::SetResetChildPollingEngineFunc(nullptr);
+  }
+  InitEpoll1PollerLinux();
+}
+
+// It is possible that GLIBC has epoll but the underlying kernel doesn't.
+// Create epoll_fd to make sure epoll support is available
+bool InitEpoll1PollerLinux() {
+  if (!grpc_event_engine::posix_engine::SupportsWakeupFd()) {
+    return false;
+  }
+  int fd = EpollCreateAndCloexec();
+  if (fd <= 0) {
+    return false;
+  }
+  if (grpc_core::Fork::Enabled()) {
+    gpr_mu_init(&fork_fd_list_mu);
+    grpc_core::Fork::SetResetChildPollingEngineFunc(ResetEventManagerOnFork);
+  }
+  close(fd);
+  return true;
+}
+
+}  // namespace
+
+void Epoll1EventHandle::OrphanHandle(PosixEngineClosure* on_done,
+                                     int* release_fd,
+                                     absl::string_view reason) {
+  bool is_release_fd = (release_fd != nullptr);
+  if (!read_closure_->IsShutdown()) {
+    HandleShutdownInternal(absl::Status(absl::StatusCode::kUnknown, reason),
+                           is_release_fd);
+  }
+
+  // If release_fd is not NULL, we should be relinquishing control of the file
+  // descriptor fd->fd (but we still own the grpc_fd structure).
+  if (is_release_fd) {
+    *release_fd = fd_;
+  } else {
+    close(fd_);
+  }
+
+  ForkFdListRemoveHandle(this);
+  {
+    // See Epoll1Poller::ShutdownHandle for explanation on why a mutex is
+    // required here.
+    grpc_core::MutexLock lock(&mu_);
+    read_closure_->DestroyEvent();
+    write_closure_->DestroyEvent();
+    error_closure_->DestroyEvent();
+  }
+  pending_read_.store(false, std::memory_order_release);
+  pending_write_.store(false, std::memory_order_release);
+  pending_error_.store(false, std::memory_order_release);
+  {
+    grpc_core::MutexLock lock(&poller_->mu_);
+    poller_->free_epoll1_handles_list_.push_back(this);
+  }
+  if (on_done != nullptr) {
+    on_done->SetStatus(absl::OkStatus());
+    poller_->GetScheduler()->Run(on_done);
+  }
+}
+
+// if 'releasing_fd' is true, it means that we are going to detach the internal
+// fd from grpc_fd structure (i.e which means we should not be calling
+// shutdown() syscall on that fd)
+void Epoll1EventHandle::HandleShutdownInternal(absl::Status why,
+                                               bool releasing_fd) {
+  if (read_closure_->SetShutdown(why)) {
+    if (!releasing_fd) {
+      shutdown(fd_, SHUT_RDWR);
+    } else {
+      epoll_event phony_event;
+      if (epoll_ctl(poller_->g_epoll_set_.epfd, EPOLL_CTL_DEL, fd_,
+                    &phony_event) != 0) {
+        gpr_log(GPR_ERROR, "epoll_ctl failed: %s",
+                grpc_core::StrError(errno).c_str());
+      }
+    }
+    write_closure_->SetShutdown(why);
+    write_closure_->SetShutdown(why);
+  }
+}
+
+Epoll1Poller::Epoll1Poller(Scheduler* scheduler)
+    : scheduler_(scheduler), was_kicked_(false) {
+  g_epoll_set_.epfd = EpollCreateAndCloexec();
+  wakeup_fd_ = *CreateWakeupFd();
+  GPR_ASSERT(wakeup_fd_ != nullptr);
+  GPR_ASSERT(g_epoll_set_.epfd >= 0);
+  gpr_log(GPR_INFO, "grpc epoll fd: %d", g_epoll_set_.epfd);
+  struct epoll_event ev;
+  ev.events = static_cast<uint32_t>(EPOLLIN | EPOLLET);
+  ev.data.ptr = wakeup_fd_.get();
+  GPR_ASSERT(epoll_ctl(g_epoll_set_.epfd, EPOLL_CTL_ADD, wakeup_fd_->ReadFd(),
+                       &ev) == 0);
+  g_epoll_set_.num_events = 0;
+  g_epoll_set_.cursor = 0;
+  ForkPollerListAddPoller(this);
+}
+
+void Epoll1Poller::Shutdown() {
+  ForkPollerListRemovePoller(this);
+  delete this;
+}
+
+Epoll1Poller::~Epoll1Poller() {
+  if (g_epoll_set_.epfd >= 0) {
+    close(g_epoll_set_.epfd);
+    g_epoll_set_.epfd = -1;
+  }
+  {
+    grpc_core::MutexLock lock(&mu_);
+    while (!free_epoll1_handles_list_.empty()) {
+      Epoll1EventHandle* handle = reinterpret_cast<Epoll1EventHandle*>(
+          free_epoll1_handles_list_.front());
+      free_epoll1_handles_list_.pop_front();
+      delete handle;
+    }
+  }
+}
+
+EventHandle* Epoll1Poller::CreateHandle(int fd, absl::string_view /*name*/,
+                                        bool track_err) {
+  Epoll1EventHandle* new_handle = nullptr;
+  {
+    grpc_core::MutexLock lock(&mu_);
+    if (free_epoll1_handles_list_.empty()) {
+      new_handle = new Epoll1EventHandle(fd, this);
+    } else {
+      new_handle = reinterpret_cast<Epoll1EventHandle*>(
+          free_epoll1_handles_list_.front());
+      free_epoll1_handles_list_.pop_front();
+      new_handle->ReInit(fd);
+    }
+  }
+  ForkFdListAddHandle(new_handle);
+  struct epoll_event ev;
+  ev.events = static_cast<uint32_t>(EPOLLIN | EPOLLOUT | EPOLLET);
+  // Use the least significant bit of ev.data.ptr to store track_err. We expect
+  // the addresses to be word aligned. We need to store track_err to avoid
+  // synchronization issues when accessing it after receiving an event.
+  // Accessing fd would be a data race there because the fd might have been
+  // returned to the free list at that point.
+  ev.data.ptr = reinterpret_cast<void*>(reinterpret_cast<intptr_t>(new_handle) |
+                                        (track_err ? 1 : 0));
+  if (epoll_ctl(g_epoll_set_.epfd, EPOLL_CTL_ADD, fd, &ev) != 0) {
+    gpr_log(GPR_ERROR, "epoll_ctl failed: %s",
+            grpc_core::StrError(errno).c_str());
+  }
+
+  return new_handle;
+}
+
+// Process the epoll events found by DoEpollWait() function.
+// - g_epoll_set.cursor points to the index of the first event to be processed
+// - This function then processes up-to max_epoll_events_to_handle and
+//   updates the g_epoll_set.cursor.
+// It returns true, it there was a Kick that forced invocation of this
+// function. It also returns the list of closures to run to take action
+// on file descriptors that became readable/writable.
+bool Epoll1Poller::ProcessEpollEvents(int max_epoll_events_to_handle,
+                                      Events& pending_events) {
+  int64_t num_events = g_epoll_set_.num_events;
+  int64_t cursor = g_epoll_set_.cursor;
+  bool was_kicked = false;
+  for (int idx = 0; (idx < max_epoll_events_to_handle) && cursor != num_events;
+       idx++) {
+    int64_t c = cursor++;
+    struct epoll_event* ev = &g_epoll_set_.events[c];
+    void* data_ptr = ev->data.ptr;
+    if (data_ptr == wakeup_fd_.get()) {
+      GPR_ASSERT(wakeup_fd_->ConsumeWakeup().ok());
+      was_kicked = true;
+    } else {
+      Epoll1EventHandle* handle = reinterpret_cast<Epoll1EventHandle*>(
+          reinterpret_cast<intptr_t>(data_ptr) & ~static_cast<intptr_t>(1));
+      bool track_err =
+          reinterpret_cast<intptr_t>(data_ptr) & static_cast<intptr_t>(1);
+      bool cancel = (ev->events & EPOLLHUP) != 0;
+      bool error = (ev->events & EPOLLERR) != 0;
+      bool read_ev = (ev->events & (EPOLLIN | EPOLLPRI)) != 0;
+      bool write_ev = (ev->events & EPOLLOUT) != 0;
+      bool err_fallback = error && !track_err;
+      if (handle->SetPendingActions(read_ev || cancel || err_fallback,
+                                    write_ev || cancel || err_fallback,
+                                    error && !err_fallback)) {
+        pending_events.push_back(handle);
+      }
+    }
+  }
+  g_epoll_set_.cursor = cursor;
+  return was_kicked;
+}
+
+//  Do epoll_wait and store the events in g_epoll_set.events field. This does
+//  not "process" any of the events yet; that is done in ProcessEpollEvents().
+//  See ProcessEpollEvents() function for more details. It returns the number
+// of events generated by epoll_wait.
+int Epoll1Poller::DoEpollWait(EventEngine::Duration timeout) {
+  int r;
+  do {
+    r = epoll_wait(g_epoll_set_.epfd, g_epoll_set_.events, MAX_EPOLL_EVENTS,
+                   static_cast<int>(
+                       grpc_event_engine::experimental::Milliseconds(timeout)));
+  } while (r < 0 && errno == EINTR);
+  if (r < 0) {
+    gpr_log(GPR_ERROR,
+            "(event_engine) Epoll1Poller:%p encountered epoll_wait error: %s",
+            this, grpc_core::StrError(errno).c_str());
+    GPR_ASSERT(false);
+  }
+  g_epoll_set_.num_events = r;
+  g_epoll_set_.cursor = 0;
+  return r;
+}
+
+// Might be called multiple times
+void Epoll1EventHandle::ShutdownHandle(absl::Status why) {
+  // A mutex is required here because, the SetShutdown method of the
+  // lockfree event may schedule a closure if it is already ready and that
+  // closure may call OrphanHandle. Execution of ShutdownHandle and OrphanHandle
+  // in parallel is not safe because some of the lockfree event types e.g, read,
+  // write, error may-not have called SetShutdown when DestroyEvent gets
+  // called in the OrphanHandle method.
+  grpc_core::MutexLock lock(&mu_);
+  HandleShutdownInternal(why, false);
+}
+
+bool Epoll1EventHandle::IsHandleShutdown() {
+  return read_closure_->IsShutdown();
+}
+
+void Epoll1EventHandle::NotifyOnRead(PosixEngineClosure* on_read) {
+  read_closure_->NotifyOn(on_read);
+}
+
+void Epoll1EventHandle::NotifyOnWrite(PosixEngineClosure* on_write) {
+  write_closure_->NotifyOn(on_write);
+}
+
+void Epoll1EventHandle::NotifyOnError(PosixEngineClosure* on_error) {
+  error_closure_->NotifyOn(on_error);
+}
+
+void Epoll1EventHandle::SetReadable() { read_closure_->SetReady(); }
+
+void Epoll1EventHandle::SetWritable() { write_closure_->SetReady(); }
+
+void Epoll1EventHandle::SetHasError() { error_closure_->SetReady(); }
+
+// Polls the registered Fds for events until timeout is reached or there is a
+// Kick(). If there is a Kick(), it collects and processes any previously
+// un-processed events. If there are no un-processed events, it returns
+// Poller::WorkResult::Kicked{}
+Poller::WorkResult Epoll1Poller::Work(
+    EventEngine::Duration timeout,
+    absl::FunctionRef<void()> schedule_poll_again) {
+  Events pending_events;
+  bool was_kicked_ext = false;
+  if (g_epoll_set_.cursor == g_epoll_set_.num_events) {
+    if (DoEpollWait(timeout) == 0) {
+      return Poller::WorkResult::kDeadlineExceeded;
+    }
+  }
+  {
+    grpc_core::MutexLock lock(&mu_);
+    // If was_kicked_ is true, collect all pending events in this iteration.
+    if (ProcessEpollEvents(
+            was_kicked_ ? INT_MAX : MAX_EPOLL_EVENTS_HANDLED_PER_ITERATION,
+            pending_events)) {
+      was_kicked_ = false;
+      was_kicked_ext = true;
+    }
+    if (pending_events.empty()) {
+      return Poller::WorkResult::kKicked;
+    }
+  }
+  // Run the provided callback.
+  schedule_poll_again();
+  // Process all pending events inline.
+  for (auto& it : pending_events) {
+    it->ExecutePendingActions();
+  }
+  return was_kicked_ext ? Poller::WorkResult::kKicked : Poller::WorkResult::kOk;
+}
+
+void Epoll1Poller::Kick() {
+  grpc_core::MutexLock lock(&mu_);
+  if (was_kicked_) {
+    return;
+  }
+  was_kicked_ = true;
+  GPR_ASSERT(wakeup_fd_->Wakeup().ok());
+}
+
+Epoll1Poller* MakeEpoll1Poller(Scheduler* scheduler) {
+  static bool kEpoll1PollerSupported = InitEpoll1PollerLinux();
+  if (kEpoll1PollerSupported) {
+    return new Epoll1Poller(scheduler);
+  }
+  return nullptr;
+}
+
+}  // namespace posix_engine
+}  // namespace grpc_event_engine
+
+#else /* defined(GRPC_LINUX_EPOLL) */
+#if defined(GRPC_POSIX_SOCKET_EV_EPOLL1)
+
+namespace grpc_event_engine {
+namespace posix_engine {
+
+using ::grpc_event_engine::experimental::EventEngine;
+using ::grpc_event_engine::experimental::Poller;
+
+Epoll1Poller::Epoll1Poller(Scheduler* /* engine */) {
+  GPR_ASSERT(false && "unimplemented");
+}
+
+void Epoll1Poller::Shutdown() { GPR_ASSERT(false && "unimplemented"); }
+
+Epoll1Poller::~Epoll1Poller() { GPR_ASSERT(false && "unimplemented"); }
+
+EventHandle* Epoll1Poller::CreateHandle(int /*fd*/, absl::string_view /*name*/,
+                                        bool /*track_err*/) {
+  GPR_ASSERT(false && "unimplemented");
+}
+
+bool Epoll1Poller::ProcessEpollEvents(int /*max_epoll_events_to_handle*/,
+                                      Events& /*pending_events*/) {
+  GPR_ASSERT(false && "unimplemented");
+}
+
+int Epoll1Poller::DoEpollWait(EventEngine::Duration /*timeout*/) {
+  GPR_ASSERT(false && "unimplemented");
+}
+
+Poller::WorkResult Epoll1Poller::Work(
+    EventEngine::Duration /*timeout*/,
+    absl::FunctionRef<void()> /*schedule_poll_again*/) {
+  GPR_ASSERT(false && "unimplemented");
+}
+
+void Epoll1Poller::Kick() { GPR_ASSERT(false && "unimplemented"); }
+
+// If GRPC_LINUX_EPOLL is not defined, it means epoll is not available. Return
+// nullptr.
+Epoll1Poller* MakeEpoll1Poller(Scheduler* /*scheduler*/) { return nullptr; }
+
+}  // namespace posix_engine
+}  // namespace grpc_event_engine
+
+#endif /* defined(GRPC_POSIX_SOCKET_EV_EPOLL1) */
+#endif /* !defined(GRPC_LINUX_EPOLL) */