Mypal/toolkit/components/url-classifier/nsUrlClassifierStreamUpdater.cpp
2019-03-11 13:26:37 +03:00

698 lines
22 KiB
C++

//* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "nsCRT.h"
#include "nsIHttpChannel.h"
#include "nsIObserverService.h"
#include "nsIStringStream.h"
#include "nsIUploadChannel.h"
#include "nsIURI.h"
#include "nsIUrlClassifierDBService.h"
#include "nsNetUtil.h"
#include "nsStreamUtils.h"
#include "nsStringStream.h"
#include "nsToolkitCompsCID.h"
#include "nsUrlClassifierStreamUpdater.h"
#include "mozilla/BasePrincipal.h"
#include "mozilla/ErrorNames.h"
#include "mozilla/Logging.h"
#include "nsIInterfaceRequestor.h"
#include "mozilla/LoadContext.h"
#include "nsContentUtils.h"
#include "nsIURLFormatter.h"
using mozilla::DocShellOriginAttributes;
static const char* gQuitApplicationMessage = "quit-application";
// Limit the list file size to 32mb
const uint32_t MAX_FILE_SIZE = (32 * 1024 * 1024);
#undef LOG
// MOZ_LOG=UrlClassifierStreamUpdater:5
static mozilla::LazyLogModule gUrlClassifierStreamUpdaterLog("UrlClassifierStreamUpdater");
#define LOG(args) TrimAndLog args
// Calls nsIURLFormatter::TrimSensitiveURLs to remove sensitive
// info from the logging message.
static void TrimAndLog(const char* aFmt, ...)
{
nsString raw;
va_list ap;
va_start(ap, aFmt);
raw.AppendPrintf(aFmt, ap);
va_end(ap);
nsCOMPtr<nsIURLFormatter> urlFormatter =
do_GetService("@mozilla.org/toolkit/URLFormatterService;1");
nsString trimmed;
nsresult rv = urlFormatter->TrimSensitiveURLs(raw, trimmed);
if (NS_FAILED(rv)) {
trimmed = EmptyString();
}
MOZ_LOG(gUrlClassifierStreamUpdaterLog,
mozilla::LogLevel::Debug,
(NS_ConvertUTF16toUTF8(trimmed).get()));
}
// This class does absolutely nothing, except pass requests onto the DBService.
///////////////////////////////////////////////////////////////////////////////
// nsIUrlClassiferStreamUpdater implementation
// Handles creating/running the stream listener
nsUrlClassifierStreamUpdater::nsUrlClassifierStreamUpdater()
: mIsUpdating(false), mInitialized(false), mDownloadError(false),
mBeganStream(false), mChannel(nullptr)
{
LOG(("nsUrlClassifierStreamUpdater init [this=%p]", this));
}
NS_IMPL_ISUPPORTS(nsUrlClassifierStreamUpdater,
nsIUrlClassifierStreamUpdater,
nsIUrlClassifierUpdateObserver,
nsIRequestObserver,
nsIStreamListener,
nsIObserver,
nsIInterfaceRequestor,
nsITimerCallback)
/**
* Clear out the update.
*/
void
nsUrlClassifierStreamUpdater::DownloadDone()
{
LOG(("nsUrlClassifierStreamUpdater::DownloadDone [this=%p]", this));
mIsUpdating = false;
mPendingUpdates.Clear();
mDownloadError = false;
mSuccessCallback = nullptr;
mUpdateErrorCallback = nullptr;
mDownloadErrorCallback = nullptr;
}
///////////////////////////////////////////////////////////////////////////////
// nsIUrlClassifierStreamUpdater implementation
nsresult
nsUrlClassifierStreamUpdater::FetchUpdate(nsIURI *aUpdateUrl,
const nsACString & aRequestPayload,
bool aIsPostRequest,
const nsACString & aStreamTable)
{
#ifdef DEBUG
LOG(("Fetching update %s from %s",
aRequestPayload.Data(), aUpdateUrl->GetSpecOrDefault().get()));
#endif
nsresult rv;
uint32_t loadFlags = nsIChannel::INHIBIT_CACHING |
nsIChannel::LOAD_BYPASS_CACHE;
rv = NS_NewChannel(getter_AddRefs(mChannel),
aUpdateUrl,
nsContentUtils::GetSystemPrincipal(),
nsILoadInfo::SEC_ALLOW_CROSS_ORIGIN_DATA_IS_NULL,
nsIContentPolicy::TYPE_OTHER,
nullptr, // aLoadGroup
this, // aInterfaceRequestor
loadFlags);
NS_ENSURE_SUCCESS(rv, rv);
nsCOMPtr<nsILoadInfo> loadInfo = mChannel->GetLoadInfo();
loadInfo->SetOriginAttributes(mozilla::NeckoOriginAttributes(NECKO_SAFEBROWSING_APP_ID, false));
mBeganStream = false;
if (!aIsPostRequest) {
// We use POST method to send our request in v2. In v4, the request
// needs to be embedded to the URL and use GET method to send.
// However, from the Chromium source code, a extended HTTP header has
// to be sent along with the request to make the request succeed.
// The following description is from Chromium source code:
//
// "The following header informs the envelope server (which sits in
// front of Google's stubby server) that the received GET request should be
// interpreted as a POST."
//
nsCOMPtr<nsIHttpChannel> httpChannel = do_QueryInterface(mChannel, &rv);
NS_ENSURE_SUCCESS(rv, rv);
rv = httpChannel->SetRequestHeader(NS_LITERAL_CSTRING("X-HTTP-Method-Override"),
NS_LITERAL_CSTRING("POST"),
false);
NS_ENSURE_SUCCESS(rv, rv);
} else if (!aRequestPayload.IsEmpty()) {
rv = AddRequestBody(aRequestPayload);
NS_ENSURE_SUCCESS(rv, rv);
}
// Set the appropriate content type for file/data URIs, for unit testing
// purposes.
// This is only used for testing and should be deleted.
bool match;
if ((NS_SUCCEEDED(aUpdateUrl->SchemeIs("file", &match)) && match) ||
(NS_SUCCEEDED(aUpdateUrl->SchemeIs("data", &match)) && match)) {
mChannel->SetContentType(NS_LITERAL_CSTRING("application/vnd.google.safebrowsing-update"));
} else {
// We assume everything else is an HTTP request.
// Disable keepalive.
nsCOMPtr<nsIHttpChannel> httpChannel = do_QueryInterface(mChannel, &rv);
NS_ENSURE_SUCCESS(rv, rv);
rv = httpChannel->SetRequestHeader(NS_LITERAL_CSTRING("Connection"), NS_LITERAL_CSTRING("close"), false);
NS_ENSURE_SUCCESS(rv, rv);
}
// Create a custom LoadContext for SafeBrowsing, so we can use callbacks on
// the channel to query the appId which allows separation of safebrowsing
// cookies in a separate jar.
DocShellOriginAttributes attrs;
attrs.mAppId = NECKO_SAFEBROWSING_APP_ID;
nsCOMPtr<nsIInterfaceRequestor> sbContext = new mozilla::LoadContext(attrs);
rv = mChannel->SetNotificationCallbacks(sbContext);
NS_ENSURE_SUCCESS(rv, rv);
// Make the request.
rv = mChannel->AsyncOpen2(this);
NS_ENSURE_SUCCESS(rv, rv);
mStreamTable = aStreamTable;
return NS_OK;
}
nsresult
nsUrlClassifierStreamUpdater::FetchUpdate(const nsACString & aUpdateUrl,
const nsACString & aRequestPayload,
bool aIsPostRequest,
const nsACString & aStreamTable)
{
LOG(("(pre) Fetching update from %s\n", PromiseFlatCString(aUpdateUrl).get()));
nsCString updateUrl(aUpdateUrl);
if (!aIsPostRequest) {
updateUrl.AppendPrintf("&$req=%s", nsCString(aRequestPayload).get());
}
nsCOMPtr<nsIURI> uri;
nsresult rv = NS_NewURI(getter_AddRefs(uri), updateUrl);
NS_ENSURE_SUCCESS(rv, rv);
nsAutoCString urlSpec;
uri->GetAsciiSpec(urlSpec);
LOG(("(post) Fetching update from %s\n", urlSpec.get()));
return FetchUpdate(uri, aRequestPayload, aIsPostRequest, aStreamTable);
}
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::DownloadUpdates(
const nsACString &aRequestTables,
const nsACString &aRequestPayload,
bool aIsPostRequest,
const nsACString &aUpdateUrl,
nsIUrlClassifierCallback *aSuccessCallback,
nsIUrlClassifierCallback *aUpdateErrorCallback,
nsIUrlClassifierCallback *aDownloadErrorCallback,
bool *_retval)
{
NS_ENSURE_ARG(aSuccessCallback);
NS_ENSURE_ARG(aUpdateErrorCallback);
NS_ENSURE_ARG(aDownloadErrorCallback);
if (mIsUpdating) {
LOG(("Already updating, queueing update %s from %s", aRequestPayload.Data(),
aUpdateUrl.Data()));
*_retval = false;
PendingRequest *request = mPendingRequests.AppendElement();
request->mTables = aRequestTables;
request->mRequestPayload = aRequestPayload;
request->mIsPostRequest = aIsPostRequest;
request->mUrl = aUpdateUrl;
request->mSuccessCallback = aSuccessCallback;
request->mUpdateErrorCallback = aUpdateErrorCallback;
request->mDownloadErrorCallback = aDownloadErrorCallback;
return NS_OK;
}
if (aUpdateUrl.IsEmpty()) {
NS_ERROR("updateUrl not set");
return NS_ERROR_NOT_INITIALIZED;
}
nsresult rv;
if (!mInitialized) {
// Add an observer for shutdown so we can cancel any pending list
// downloads. quit-application is the same event that the download
// manager listens for and uses to cancel pending downloads.
nsCOMPtr<nsIObserverService> observerService =
mozilla::services::GetObserverService();
if (!observerService)
return NS_ERROR_FAILURE;
observerService->AddObserver(this, gQuitApplicationMessage, false);
mDBService = do_GetService(NS_URLCLASSIFIERDBSERVICE_CONTRACTID, &rv);
NS_ENSURE_SUCCESS(rv, rv);
mInitialized = true;
}
rv = mDBService->BeginUpdate(this, aRequestTables);
if (rv == NS_ERROR_NOT_AVAILABLE) {
LOG(("Service busy, already updating, queuing update %s from %s",
aRequestPayload.Data(), aUpdateUrl.Data()));
*_retval = false;
PendingRequest *request = mPendingRequests.AppendElement();
request->mTables = aRequestTables;
request->mRequestPayload = aRequestPayload;
request->mIsPostRequest = aIsPostRequest;
request->mUrl = aUpdateUrl;
request->mSuccessCallback = aSuccessCallback;
request->mUpdateErrorCallback = aUpdateErrorCallback;
request->mDownloadErrorCallback = aDownloadErrorCallback;
return NS_OK;
}
if (NS_FAILED(rv)) {
return rv;
}
mSuccessCallback = aSuccessCallback;
mUpdateErrorCallback = aUpdateErrorCallback;
mDownloadErrorCallback = aDownloadErrorCallback;
mIsUpdating = true;
*_retval = true;
LOG(("FetchUpdate: %s", aUpdateUrl.Data()));
return FetchUpdate(aUpdateUrl, aRequestPayload, aIsPostRequest, EmptyCString());
}
///////////////////////////////////////////////////////////////////////////////
// nsIUrlClassifierUpdateObserver implementation
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::UpdateUrlRequested(const nsACString &aUrl,
const nsACString &aTable)
{
LOG(("Queuing requested update from %s\n", PromiseFlatCString(aUrl).get()));
PendingUpdate *update = mPendingUpdates.AppendElement();
if (!update)
return NS_ERROR_OUT_OF_MEMORY;
// Allow data: and file: urls for unit testing purposes, otherwise assume http
if (StringBeginsWith(aUrl, NS_LITERAL_CSTRING("data:")) ||
StringBeginsWith(aUrl, NS_LITERAL_CSTRING("file:"))) {
update->mUrl = aUrl;
} else {
// For unittesting update urls to localhost should use http, not https
// (otherwise the connection will fail silently, since there will be no
// cert available).
if (!StringBeginsWith(aUrl, NS_LITERAL_CSTRING("localhost"))) {
update->mUrl = NS_LITERAL_CSTRING("https://") + aUrl;
} else {
update->mUrl = NS_LITERAL_CSTRING("http://") + aUrl;
}
}
update->mTable = aTable;
return NS_OK;
}
nsresult
nsUrlClassifierStreamUpdater::FetchNext()
{
if (mPendingUpdates.Length() == 0) {
return NS_OK;
}
PendingUpdate &update = mPendingUpdates[0];
LOG(("Fetching update url: %s\n", update.mUrl.get()));
nsresult rv = FetchUpdate(update.mUrl,
EmptyCString(),
true, // This method is for v2 and v2 is always a POST.
update.mTable);
if (NS_FAILED(rv)) {
LOG(("Error fetching update url: %s\n", update.mUrl.get()));
// We can commit the urls that we've applied so far. This is
// probably a transient server problem, so trigger backoff.
mDownloadErrorCallback->HandleEvent(EmptyCString());
mDownloadError = true;
mDBService->FinishUpdate();
return rv;
}
mPendingUpdates.RemoveElementAt(0);
return NS_OK;
}
nsresult
nsUrlClassifierStreamUpdater::FetchNextRequest()
{
if (mPendingRequests.Length() == 0) {
LOG(("No more requests, returning"));
return NS_OK;
}
PendingRequest &request = mPendingRequests[0];
LOG(("Stream updater: fetching next request: %s, %s",
request.mTables.get(), request.mUrl.get()));
bool dummy;
DownloadUpdates(
request.mTables,
request.mRequestPayload,
request.mIsPostRequest,
request.mUrl,
request.mSuccessCallback,
request.mUpdateErrorCallback,
request.mDownloadErrorCallback,
&dummy);
request.mSuccessCallback = nullptr;
request.mUpdateErrorCallback = nullptr;
request.mDownloadErrorCallback = nullptr;
mPendingRequests.RemoveElementAt(0);
return NS_OK;
}
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::StreamFinished(nsresult status,
uint32_t requestedDelay)
{
// We are a service and may not be reset with Init between calls, so reset
// mBeganStream manually.
mBeganStream = false;
LOG(("nsUrlClassifierStreamUpdater::StreamFinished [%x, %d]", status, requestedDelay));
if (NS_FAILED(status) || mPendingUpdates.Length() == 0) {
// We're done.
LOG(("nsUrlClassifierStreamUpdater::Done [this=%p]", this));
mDBService->FinishUpdate();
return NS_OK;
}
// This timer is for fetching indirect updates ("forwards") from any "u:" lines
// that we encountered while processing the server response. It is NOT for
// scheduling the next time we pull the list from the server. That's a different
// timer in listmanager.js (see bug 1110891).
nsresult rv;
mTimer = do_CreateInstance("@mozilla.org/timer;1", &rv);
if (NS_SUCCEEDED(rv)) {
rv = mTimer->InitWithCallback(this, requestedDelay,
nsITimer::TYPE_ONE_SHOT);
}
if (NS_FAILED(rv)) {
NS_WARNING("Unable to initialize timer, fetching next safebrowsing item immediately");
return FetchNext();
}
return NS_OK;
}
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::UpdateSuccess(uint32_t requestedTimeout)
{
LOG(("nsUrlClassifierStreamUpdater::UpdateSuccess [this=%p]", this));
if (mPendingUpdates.Length() != 0) {
NS_WARNING("Didn't fetch all safebrowsing update redirects");
}
// DownloadDone() clears mSuccessCallback, so we save it off here.
nsCOMPtr<nsIUrlClassifierCallback> successCallback = mDownloadError ? nullptr : mSuccessCallback.get();
DownloadDone();
nsAutoCString strTimeout;
strTimeout.AppendInt(requestedTimeout);
if (successCallback) {
LOG(("nsUrlClassifierStreamUpdater::UpdateSuccess callback [this=%p]",
this));
successCallback->HandleEvent(strTimeout);
} else {
LOG(("nsUrlClassifierStreamUpdater::UpdateSuccess skipping callback [this=%p]",
this));
}
// Now fetch the next request
LOG(("stream updater: calling into fetch next request"));
FetchNextRequest();
return NS_OK;
}
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::UpdateError(nsresult result)
{
LOG(("nsUrlClassifierStreamUpdater::UpdateError [this=%p]", this));
// DownloadDone() clears mUpdateErrorCallback, so we save it off here.
nsCOMPtr<nsIUrlClassifierCallback> errorCallback = mDownloadError ? nullptr : mUpdateErrorCallback.get();
DownloadDone();
nsAutoCString strResult;
strResult.AppendInt(static_cast<uint32_t>(result));
if (errorCallback) {
errorCallback->HandleEvent(strResult);
}
return NS_OK;
}
nsresult
nsUrlClassifierStreamUpdater::AddRequestBody(const nsACString &aRequestBody)
{
nsresult rv;
nsCOMPtr<nsIStringInputStream> strStream =
do_CreateInstance(NS_STRINGINPUTSTREAM_CONTRACTID, &rv);
NS_ENSURE_SUCCESS(rv, rv);
rv = strStream->SetData(aRequestBody.BeginReading(),
aRequestBody.Length());
NS_ENSURE_SUCCESS(rv, rv);
nsCOMPtr<nsIUploadChannel> uploadChannel = do_QueryInterface(mChannel, &rv);
NS_ENSURE_SUCCESS(rv, rv);
rv = uploadChannel->SetUploadStream(strStream,
NS_LITERAL_CSTRING("text/plain"),
-1);
NS_ENSURE_SUCCESS(rv, rv);
nsCOMPtr<nsIHttpChannel> httpChannel = do_QueryInterface(mChannel, &rv);
NS_ENSURE_SUCCESS(rv, rv);
rv = httpChannel->SetRequestMethod(NS_LITERAL_CSTRING("POST"));
NS_ENSURE_SUCCESS(rv, rv);
return NS_OK;
}
///////////////////////////////////////////////////////////////////////////////
// nsIStreamListenerObserver implementation
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::OnStartRequest(nsIRequest *request,
nsISupports* context)
{
nsresult rv;
bool downloadError = false;
nsAutoCString strStatus;
nsresult status = NS_OK;
// Only update if we got http success header
nsCOMPtr<nsIHttpChannel> httpChannel = do_QueryInterface(request);
if (httpChannel) {
rv = httpChannel->GetStatus(&status);
NS_ENSURE_SUCCESS(rv, rv);
if (MOZ_LOG_TEST(gUrlClassifierStreamUpdaterLog, mozilla::LogLevel::Debug)) {
nsAutoCString errorName, spec;
mozilla::GetErrorName(status, errorName);
nsCOMPtr<nsIURI> uri;
rv = httpChannel->GetURI(getter_AddRefs(uri));
if (NS_SUCCEEDED(rv) && uri) {
uri->GetAsciiSpec(spec);
}
LOG(("nsUrlClassifierStreamUpdater::OnStartRequest "
"(status=%s, uri=%s, this=%p)", errorName.get(),
spec.get(), this));
}
if (NS_FAILED(status)) {
// Assume we're overloading the server and trigger backoff.
downloadError = true;
} else {
bool succeeded = false;
rv = httpChannel->GetRequestSucceeded(&succeeded);
NS_ENSURE_SUCCESS(rv, rv);
uint32_t requestStatus;
rv = httpChannel->GetResponseStatus(&requestStatus);
NS_ENSURE_SUCCESS(rv, rv);
LOG(("nsUrlClassifierStreamUpdater::OnStartRequest %s (%d)", succeeded ?
"succeeded" : "failed", requestStatus));
if (!succeeded) {
// 404 or other error, pass error status back
strStatus.AppendInt(requestStatus);
downloadError = true;
}
}
}
if (downloadError) {
LOG(("nsUrlClassifierStreamUpdater::Download error [this=%p]", this));
// It's possible for mDownloadErrorCallback to be null on shutdown.
if (mDownloadErrorCallback) {
mDownloadErrorCallback->HandleEvent(strStatus);
}
mDownloadError = true;
status = NS_ERROR_ABORT;
} else if (NS_SUCCEEDED(status)) {
MOZ_ASSERT(mDownloadErrorCallback);
mBeganStream = true;
LOG(("nsUrlClassifierStreamUpdater::Beginning stream [this=%p]", this));
rv = mDBService->BeginStream(mStreamTable);
NS_ENSURE_SUCCESS(rv, rv);
}
mStreamTable.Truncate();
return status;
}
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::OnDataAvailable(nsIRequest *request,
nsISupports* context,
nsIInputStream *aIStream,
uint64_t aSourceOffset,
uint32_t aLength)
{
if (!mDBService)
return NS_ERROR_NOT_INITIALIZED;
LOG(("OnDataAvailable (%d bytes)", aLength));
if (aSourceOffset > MAX_FILE_SIZE) {
LOG(("OnDataAvailable::Abort because exceeded the maximum file size(%lld)", aSourceOffset));
return NS_ERROR_FILE_TOO_BIG;
}
nsresult rv;
// Copy the data into a nsCString
nsCString chunk;
rv = NS_ConsumeStream(aIStream, aLength, chunk);
NS_ENSURE_SUCCESS(rv, rv);
//LOG(("Chunk (%d): %s\n\n", chunk.Length(), chunk.get()));
rv = mDBService->UpdateStream(chunk);
NS_ENSURE_SUCCESS(rv, rv);
return NS_OK;
}
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::OnStopRequest(nsIRequest *request, nsISupports* context,
nsresult aStatus)
{
if (!mDBService)
return NS_ERROR_NOT_INITIALIZED;
LOG(("OnStopRequest (status %x, beganStream %s, this=%p)", aStatus,
mBeganStream ? "true" : "false", this));
nsresult rv;
if (NS_SUCCEEDED(aStatus)) {
// Success, finish this stream and move on to the next.
rv = mDBService->FinishStream();
} else if (mBeganStream) {
LOG(("OnStopRequest::Canceling update [this=%p]", this));
// We began this stream and couldn't finish it. We have to cancel the
// update, it's not in a consistent state.
rv = mDBService->CancelUpdate();
} else {
LOG(("OnStopRequest::Finishing update [this=%p]", this));
// The fetch failed, but we didn't start the stream (probably a
// server or connection error). We can commit what we've applied
// so far, and request again later.
rv = mDBService->FinishUpdate();
}
mChannel = nullptr;
// If the fetch failed, return the network status rather than NS_OK, the
// result of finishing a possibly-empty update
if (NS_SUCCEEDED(aStatus)) {
return rv;
}
return aStatus;
}
///////////////////////////////////////////////////////////////////////////////
// nsIObserver implementation
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::Observe(nsISupports *aSubject, const char *aTopic,
const char16_t *aData)
{
if (nsCRT::strcmp(aTopic, gQuitApplicationMessage) == 0) {
if (mIsUpdating && mChannel) {
LOG(("Cancel download"));
nsresult rv;
rv = mChannel->Cancel(NS_ERROR_ABORT);
NS_ENSURE_SUCCESS(rv, rv);
mIsUpdating = false;
mChannel = nullptr;
}
if (mTimer) {
mTimer->Cancel();
mTimer = nullptr;
}
}
return NS_OK;
}
///////////////////////////////////////////////////////////////////////////////
// nsIInterfaceRequestor implementation
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::GetInterface(const nsIID & eventSinkIID, void* *_retval)
{
return QueryInterface(eventSinkIID, _retval);
}
///////////////////////////////////////////////////////////////////////////////
// nsITimerCallback implementation
NS_IMETHODIMP
nsUrlClassifierStreamUpdater::Notify(nsITimer *timer)
{
LOG(("nsUrlClassifierStreamUpdater::Notify [%p]", this));
mTimer = nullptr;
// Start the update process up again.
FetchNext();
return NS_OK;
}