[SafeBrowsing] Send pings for Zip files that contain other archives.

chrome/browser/safe_browsing/download_protection_service.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 "chrome/browser/safe_browsing/download_protection_service.h"
 
 #include "base/bind.h"
 #include "base/compiler_specific.h"
 #include "base/format_macros.h"
 #include "base/memory/scoped_ptr.h"
@@ skipping 102 matching lines @@
     FILE_PATH_LITERAL(".mshxml"),
     FILE_PATH_LITERAL(".ps1"),
     FILE_PATH_LITERAL(".ps1xml"),
     FILE_PATH_LITERAL(".ps2"),
     FILE_PATH_LITERAL(".ps2xml"),
     FILE_PATH_LITERAL(".psc1"),
     FILE_PATH_LITERAL(".psc2"),
     FILE_PATH_LITERAL(".scf"),
     FILE_PATH_LITERAL(".sct"),
     FILE_PATH_LITERAL(".wsf"),
+    FILE_PATH_LITERAL(".7z"),
+    FILE_PATH_LITERAL(".xz"),
+    FILE_PATH_LITERAL(".gz"),
+    FILE_PATH_LITERAL(".tgz"),
+    FILE_PATH_LITERAL(".bz2"),
+    FILE_PATH_LITERAL(".tar"),
+    FILE_PATH_LITERAL(".arj"),
+    FILE_PATH_LITERAL(".lzh"),
+    FILE_PATH_LITERAL(".lha"),
+    FILE_PATH_LITERAL(".wim"),
+    FILE_PATH_LITERAL(".z"),
+    FILE_PATH_LITERAL(".lzma"),
+    FILE_PATH_LITERAL(".cpio"),
 };
 
 // UMA enumeration value for unrecognized file types. This is the array index of
 // the "Other" bucket in kDangerousFileTypes.
 const int EXTENSION_OTHER = 18;
 
-void RecordFileExtensionType(const base::FilePath& file) {
+// Maximum extension ID returned by GetExtensionTypeForUMA() + 1.
+const int EXTENSION_MAX = arraysize(kDangerousFileTypes);
+
+int GetExtensionTypeForUMA(const base::FilePath::StringType& extension) {
   DCHECK_EQ(static_cast<base::FilePath::CharType*>(nullptr),
             kDangerousFileTypes[EXTENSION_OTHER]);
+  DCHECK(extension.find(base::FilePath::kExtensionSeparator) == 0 ||
+         extension.empty());
+  DCHECK_EQ(extension, base::FilePath(extension).FinalExtension());
 
-  int extension_type = EXTENSION_OTHER;
-  for (const auto& extension : kDangerousFileTypes) {
-    if (extension && file.MatchesExtension(extension)) {
-      extension_type = &extension - kDangerousFileTypes;
-      break;
-    }
+  for (const auto& dangerous_extension : kDangerousFileTypes) {
+    if (dangerous_extension &&
+        base::FilePath::CompareEqualIgnoreCase(dangerous_extension, extension))
+      return &dangerous_extension - kDangerousFileTypes;
   }
+  return EXTENSION_OTHER;
+}
 
+void RecordFileExtensionType(const base::FilePath& file) {
   UMA_HISTOGRAM_ENUMERATION("SBClientDownload.DownloadExtensions",
-                            extension_type, arraysize(kDangerousFileTypes));
+                            GetExtensionTypeForUMA(file.FinalExtension()),
+                            EXTENSION_MAX);
+}
+
+void RecordArchivedArchiveFileExtensionType(
+    const base::FilePath::StringType& extension) {
+  UMA_HISTOGRAM_ENUMERATION("SBClientDownload.ArchivedArchiveExtensions",
+                            GetExtensionTypeForUMA(extension), EXTENSION_MAX);
 }
 
 // Enumerate for histogramming purposes.
 // DO NOT CHANGE THE ORDERING OF THESE VALUES (different histogram data will
 // be mixed together based on their values).
 enum SBStatsType {
   DOWNLOAD_URL_CHECKS_TOTAL,
   DOWNLOAD_URL_CHECKS_CANCELED,
   DOWNLOAD_URL_CHECKS_MALWARE,
 
@@ skipping 342 matching lines @@
                                   ClientDownloadRequest::DownloadType* type) {
     if (item.GetUrlChain().empty()) {
       *reason = REASON_EMPTY_URL_CHAIN;
       return false;
     }
     const GURL& final_url = item.GetUrlChain().back();
     if (!final_url.is_valid() || final_url.is_empty()) {
       *reason = REASON_INVALID_URL;
       return false;
     }
-    if (!download_protection_util::IsBinaryFile(target_path)) {
+    if (!download_protection_util::IsSupportedBinaryFile(target_path)) {
       *reason = REASON_NOT_BINARY_FILE;
       return false;
     }
     if ((!final_url.IsStandard() && !final_url.SchemeIsBlob() &&
          !final_url.SchemeIs(url::kDataScheme)) ||
         final_url.SchemeIsFile()) {
       *reason = REASON_UNSUPPORTED_URL_SCHEME;
       return false;
     }
     *type = download_protection_util::GetDownloadType(target_path);
@@ skipping 78 matching lines @@
     // analyzer is refcounted, it might outlive the request.
     analyzer_ = new SandboxedZipAnalyzer(
         item_->GetFullPath(),
         base::Bind(&CheckClientDownloadRequest::OnZipAnalysisFinished,
                    weakptr_factory_.GetWeakPtr()));
     analyzer_->Start();
   }
 
   void OnZipAnalysisFinished(const zip_analyzer::Results& results) {
     DCHECK_CURRENTLY_ON(BrowserThread::UI);
+    DCHECK_EQ(ClientDownloadRequest::ZIPPED_EXECUTABLE, type_);
     if (!service_)
       return;
     if (results.success) {
       zipped_executable_ = results.has_executable;
       archived_binary_.CopyFrom(results.archived_binary);
       DVLOG(1) << "Zip analysis finished for " << item_->GetFullPath().value()
                << ", has_executable=" << results.has_executable
                << " has_archive=" << results.has_archive;
     } else {
       DVLOG(1) << "Zip analysis failed for " << item_->GetFullPath().value();
     }
     UMA_HISTOGRAM_BOOLEAN("SBClientDownload.ZipFileHasExecutable",
                           zipped_executable_);
     UMA_HISTOGRAM_BOOLEAN("SBClientDownload.ZipFileHasArchiveButNoExecutable",
                           results.has_archive && !zipped_executable_);
     UMA_HISTOGRAM_TIMES("SBClientDownload.ExtractZipFeaturesTime",
                         base::TimeTicks::Now() - zip_analysis_start_time_);
+    for (const auto& file_extension : results.archived_archive_filetypes)
+      RecordArchivedArchiveFileExtensionType(file_extension);
 
-    if (!zipped_executable_) {
+    if (!zipped_executable_ && !results.has_archive) {
       PostFinishTask(UNKNOWN, REASON_ARCHIVE_WITHOUT_BINARIES);
       return;
     }
+
+    if (!zipped_executable_ && results.has_archive)
+      type_ = ClientDownloadRequest::ZIPPED_ARCHIVE;
     OnFileFeatureExtractionDone();
   }
 
   static void RecordCountOfSignedOrWhitelistedDownload() {
     UMA_HISTOGRAM_COUNTS("SBClientDownload.SignedOrWhitelistedDownload", 1);
   }
 
   void CheckWhitelists() {
     DCHECK_CURRENTLY_ON(BrowserThread::IO);
 
@@ skipping 500 matching lines @@
 GURL DownloadProtectionService::GetDownloadRequestUrl() {
   GURL url(kDownloadRequestUrl);
   std::string api_key = google_apis::GetAPIKey();
   if (!api_key.empty())
     url = url.Resolve("?key=" + net::EscapeQueryParamValue(api_key, true));
 
   return url;
 }
 
 }  // namespace safe_browsing