stale-while-revalidate experimental implementation.

net/http/http_cache.cc

 // Copyright (c) 2012 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 "net/http/http_cache.h"
 
 #include <algorithm>
 
 #include "base/compiler_specific.h"
 
 #if defined(OS_POSIX)
 #include <unistd.h>
 #endif
 
 #include "base/bind.h"
 #include "base/bind_helpers.h"
 #include "base/callback.h"
 #include "base/file_util.h"
 #include "base/format_macros.h"
 #include "base/location.h"
 #include "base/memory/ref_counted.h"
 #include "base/message_loop/message_loop.h"
 #include "base/metrics/field_trial.h"
+#include "base/metrics/histogram.h"
 #include "base/pickle.h"
 #include "base/stl_util.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_util.h"
 #include "base/strings/stringprintf.h"
 #include "base/threading/worker_pool.h"
+#include "base/time/time.h"
 #include "net/base/cache_type.h"
 #include "net/base/io_buffer.h"
 #include "net/base/load_flags.h"
 #include "net/base/net_errors.h"
+#include "net/base/network_delegate.h"
 #include "net/base/upload_data_stream.h"
 #include "net/disk_cache/disk_cache.h"
 #include "net/http/disk_based_cert_cache.h"
 #include "net/http/disk_cache_based_quic_server_info.h"
 #include "net/http/http_cache_transaction.h"
 #include "net/http/http_network_layer.h"
 #include "net/http/http_network_session.h"
 #include "net/http/http_request_info.h"
 #include "net/http/http_response_headers.h"
 #include "net/http/http_response_info.h"
@@ skipping 241 matching lines @@
   virtual QuicServerInfo* GetForServer(
       const QuicServerId& server_id) OVERRIDE {
     return new DiskCacheBasedQuicServerInfo(server_id, http_cache_);
   }
 
  private:
   HttpCache* const http_cache_;
 };
 
 //-----------------------------------------------------------------------------
+
+class HttpCache::AsyncValidation {
+ public:
+  AsyncValidation(const HttpRequestInfo& original_request, HttpCache* cache)
+      : request_(original_request), cache_(cache) {}
+  ~AsyncValidation() {}
+
+  void Start(const BoundNetLog& net_log, NetworkDelegate* network_delegate);
+
+ private:
+  void OnStarted(int result);
+  void DoRead();
+  void OnRead(int result);
+
+  // Terminate this request with net error code |result|. Logs the transaction
+  // result and asks HttpCache to delete this object.
+  // If there was a client or server certificate error, it cannot be recovered
+  // asynchronously, so we need to prevent future attempts to asynchronously
+  // fetch the resource. In this case, the cache entry is doomed.
+  void Terminate(int result);
+
+  HttpRequestInfo request_;
+  scoped_refptr<IOBuffer> buf_;
+  CompletionCallback read_callback_;
+  scoped_ptr<Transaction> transaction_;
+  base::Time start_time_;
+
+  // The HttpCache object owns this object. This object is always deleted before
+  // the pointer to the cache becomes invalid.
+  HttpCache* cache_;
+
+  DISALLOW_COPY_AND_ASSIGN(AsyncValidation);
+};
+
+void HttpCache::AsyncValidation::Start(const BoundNetLog& net_log,
+                                       NetworkDelegate* network_delegate) {
+  scoped_ptr<HttpTransaction> transaction;
+  cache_->CreateTransaction(IDLE, &transaction);
+  if (network_delegate) {
+    // This code is necessary to enable async transactions to pass over the
+    // data-reduction proxy. This is a violation of the "once-and-only-once"
+    // principle, since it copies code from URLRequestHttpJob. We cannot use the
+    // original callback passed to HttpCache::Transaction by URLRequestHttpJob
+    // as it will only be valid as long as the URLRequestHttpJob object is
+    // alive, and that object will be deleted as soon as the synchronous request
+    // completes.
+    //
+    // This code is also an encapsulation violation. We are exploiting the fact
+    // that the |request| parameter to NotifyBeforeSendProxyHeaders() is never
+    // actually used for anything, and so can be NULL.
+    //
+    // TODO(ricea): Do this better.
+    transaction->SetBeforeProxyHeadersSentCallback(
+        base::Bind(&NetworkDelegate::NotifyBeforeSendProxyHeaders,
+                   base::Unretained(network_delegate),
+                   static_cast<URLRequest*>(NULL)));
+    // The above use of base::Unretained is safe because the NetworkDelegate has
+    // to live at least as long as the HttpNetworkSession which has to live as
+    // least as long as the HttpNetworkLayer which has to live at least as long
+    // this HttpCache object.
+  }
+
+  transaction_.reset(static_cast<Transaction*>(transaction.release()));
+  DCHECK_EQ(0, request_.load_flags & LOAD_ASYNC_REVALIDATION);
+  request_.load_flags |= LOAD_ASYNC_REVALIDATION;
+  start_time_ = base::Time::Now();
+  // This use of base::Unretained is safe because |transaction_| is owned by
+  // this object.
+  read_callback_ = base::Bind(&AsyncValidation::OnRead, base::Unretained(this));
+  // This use of base::Unretained is safe as above.
+  int rv = transaction_->Start(
+      &request_,
+      base::Bind(&AsyncValidation::OnStarted, base::Unretained(this)),
+      net_log);
+
+  if (rv == ERR_IO_PENDING)
+    return;
+
+  OnStarted(rv);
+}
+
+void HttpCache::AsyncValidation::OnStarted(int result) {
+  if (result != OK) {
+    DVLOG(1) << "Asynchronous transaction start failed for " << request_.url;
+    Terminate(result);
+    return;
+  }
+
+  DoRead();
+}
+
+void HttpCache::AsyncValidation::DoRead() {
+  const size_t kBufSize = 4096;
+  if (!buf_)
+    buf_ = new IOBuffer(kBufSize);
+
+  int rv = 0;
+  do {
+    rv = transaction_->Read(buf_.get(), kBufSize, read_callback_);
+  } while (rv > 0);
+
+  if (rv == ERR_IO_PENDING)
+    return;
+
+  OnRead(rv);
+}
+
+void HttpCache::AsyncValidation::OnRead(int result) {
+  if (result > 0) {
+    DoRead();
+    return;
+  }
+  Terminate(result);
+}
+
+void HttpCache::AsyncValidation::Terminate(int result) {
+  if (result == ERR_SSL_CLIENT_AUTH_CERT_NEEDED || IsCertificateError(result)) {
+    // We should not attempt to access this resource asynchronously again until
+    // the certificate problem has been resolved.
+    // TODO(ricea): Mark the entry as requiring synchronous revalidation rather
+    // than just deleting it.
+    cache_->DoomEntry(transaction_->key(), transaction_.get());
+  }
+  base::TimeDelta duration = base::Time::Now() - start_time_;
+  UMA_HISTOGRAM_TIMES("HttpCache.AsyncValidationDuration", duration);
+  transaction_->net_log().EndEventWithNetErrorCode(
+      NetLog::TYPE_ASYNC_REVALIDATION, result);
+  cache_->DeleteAsyncValidation(request_.url.possibly_invalid_spec());

[rvargas (doing something else), 2014/09/11 02:33:06]

We should rely one way or another on GenerateCacheKey

[Adam Rice, 2014/09/12 01:46:48]

Done.

+  // |this| is deleted.
+}
+
+//-----------------------------------------------------------------------------
 HttpCache::HttpCache(const net::HttpNetworkSession::Params& params,
                      BackendFactory* backend_factory)
     : net_log_(params.net_log),
       backend_factory_(backend_factory),
       building_backend_(false),
       bypass_lock_for_test_(false),
+      use_stale_while_revalidate_(params.use_stale_while_revalidate),
       mode_(NORMAL),
       network_layer_(new HttpNetworkLayer(new HttpNetworkSession(params))),
       weak_factory_(this) {
   SetupQuicServerInfoFactory(network_layer_->GetSession());
 }
 
 
 // This call doesn't change the shared |session|'s QuicServerInfoFactory because
 // |session| is shared.
 HttpCache::HttpCache(HttpNetworkSession* session,
                      BackendFactory* backend_factory)
     : net_log_(session->net_log()),
       backend_factory_(backend_factory),
       building_backend_(false),
       bypass_lock_for_test_(false),
+      use_stale_while_revalidate_(session->params().use_stale_while_revalidate),
       mode_(NORMAL),
       network_layer_(new HttpNetworkLayer(session)),
       weak_factory_(this) {
 }
 
 HttpCache::HttpCache(HttpTransactionFactory* network_layer,
                      NetLog* net_log,
                      BackendFactory* backend_factory)
     : net_log_(net_log),
       backend_factory_(backend_factory),
       building_backend_(false),
       bypass_lock_for_test_(false),
+      use_stale_while_revalidate_(false),
       mode_(NORMAL),
       network_layer_(network_layer),
       weak_factory_(this) {
   SetupQuicServerInfoFactory(network_layer_->GetSession());
+  HttpNetworkSession* session = network_layer_->GetSession();
+  if (session)
+    use_stale_while_revalidate_ = session->params().use_stale_while_revalidate;
 }
 
 HttpCache::~HttpCache() {
   // Transactions should see an invalid cache after this point; otherwise they
   // could see an inconsistent object (half destroyed).
   weak_factory_.InvalidateWeakPtrs();
 
   // If we have any active entries remaining, then we need to deactivate them.
   // We may have some pending calls to OnProcessPendingQueue, but since those
   // won't run (due to our destruction), we can simply ignore the corresponding
   // will_process_pending_queue flag.
   while (!active_entries_.empty()) {
     ActiveEntry* entry = active_entries_.begin()->second;
     entry->will_process_pending_queue = false;
     entry->pending_queue.clear();
     entry->readers.clear();
     entry->writer = NULL;
     DeactivateEntry(entry);
   }
 
   STLDeleteElements(&doomed_entries_);
+  STLDeleteValues(&async_validations_);
 
   // Before deleting pending_ops_, we have to make sure that the disk cache is
   // done with said operations, or it will attempt to use deleted data.
   cert_cache_.reset();
   disk_cache_.reset();
 
   PendingOpsMap::iterator pending_it = pending_ops_.begin();
   for (; pending_it != pending_ops_.end(); ++pending_it) {
     // We are not notifying the transactions about the cache going away, even
     // though they are waiting for a callback that will never fire.
@@ skipping 668 matching lines @@
   // not delete the entry before OnProcessPendingQueue runs.
   if (entry->will_process_pending_queue)
     return;
   entry->will_process_pending_queue = true;
 
   base::MessageLoop::current()->PostTask(
       FROM_HERE,
       base::Bind(&HttpCache::OnProcessPendingQueue, GetWeakPtr(), entry));
 }
 
+void HttpCache::PerformAsyncValidation(const HttpRequestInfo& original_request,
+                                       const BoundNetLog& net_log) {
+  DCHECK(use_stale_while_revalidate_);
+  std::string spec = original_request.url.possibly_invalid_spec();
+  if (ContainsKey(async_validations_, spec)) {

[rvargas (doing something else), 2014/09/11 02:33:06]

Avoid searching twice by using insert_ok to detect this case.

[Adam Rice, 2014/09/12 01:46:48]

Sorry, yes. I optimised for the conflict case without thinking about it.

+    DVLOG(1) << "Harmless race condition detected on URL " << spec
+             << "; discarding redundant revalidation.";
+    return;
+  }
+  AsyncValidation* async_validation =
+      new AsyncValidation(original_request, this);
+  typedef AsyncValidationMap::value_type AsyncValidationKeyValue;
+  bool insert_ok =
+      async_validations_.insert(AsyncValidationKeyValue(spec, async_validation))
+          .second;
+  DCHECK(insert_ok);
+  HttpNetworkSession* network_session = GetSession();
+  NetworkDelegate* network_delegate = NULL;
+  if (network_session)
+    network_delegate = network_session->network_delegate();
+  async_validation->Start(net_log, network_delegate);
+  // |async_validation| may have been deleted here.
+}
+
+void HttpCache::DeleteAsyncValidation(const std::string& url) {
+  AsyncValidationMap::iterator it = async_validations_.find(url);
+  CHECK(it != async_validations_.end());  // security-critical invariant
+  AsyncValidation* async_validation = it->second;
+  async_validations_.erase(it);
+  delete async_validation;  // |url| is no longer valid.

[rvargas (doing something else), 2014/09/11 02:33:06]

Is the comment really needed? If so, it needs rephrasing.

[Adam Rice, 2014/09/12 01:46:48]

Probably not. I thought it was worth noting as non-obvious behaviour, but this
code shouldn't be around long enough for someone to add a reference to url after
the delete (and AddressSanitizer would spot that straight away anyway).

+}
+
 void HttpCache::OnProcessPendingQueue(ActiveEntry* entry) {
   entry->will_process_pending_queue = false;
   DCHECK(!entry->writer);
 
   // If no one is interested in this entry, then we can deactivate it.
   if (entry->pending_queue.empty()) {
     if (entry->readers.empty())
       DestroyEntry(entry);
     return;
   }
@@ skipping 150 matching lines @@
     building_backend_ = false;
     DeletePendingOp(pending_op);
   }
 
   // The cache may be gone when we return from the callback.
   if (!item->DoCallback(result, disk_cache_.get()))
     item->NotifyTransaction(result, NULL);
 }
 
 }  // namespace net