Add ICU message format support

[chromium-blink-merge.git] / chrome / common / safe_browsing / csd.proto

// 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.
//
// Client side phishing and malware detection request and response
// protocol buffers.  Those protocol messages should be kept in sync
// with the server implementation.
//
// If you want to change this protocol definition or you have questions
// regarding its format please contact chrome-anti-phishing@googlegroups.com.

syntax = "proto2";

option optimize_for = LITE_RUNTIME;

package safe_browsing;

// Protocol buffer describing the Chrome user population of the user reporting
// data.
message ChromeUserPopulation {
  enum UserPopulation {
    UNKNOWN_USER_POPULATION = 0;
    SAFE_BROWSING = 1;
    EXTENDED_REPORTING = 2;
  }
  optional UserPopulation user_population = 1;
}


message ClientPhishingRequest {
  // URL that the client visited.  The CGI parameters are stripped by the
  // client.
  optional string url = 1;

  // A 5-byte SHA-256 hash prefix of the URL.  Before hashing the URL is
  // canonicalized, converted to a suffix-prefix expression and broadened
  // (www prefix is removed and everything past the last '/' is stripped).
  //
  // Marked OBSOLETE because the URL is sent for all users, making the hash
  // prefix unnecessary.
  optional bytes OBSOLETE_hash_prefix = 10;

  // Score that was computed on the client.  Value is between 0.0 and 1.0.
  // The larger the value the more likely the url is phishing.
  required float client_score = 2;

  // Note: we're skipping tag 3 because it was previously used.

  // Is true if the features for this URL were classified as phishing.
  // Currently, this will always be true for all client-phishing requests
  // that are sent to the server.
  optional bool is_phishing = 4;

  message Feature {
    // Feature name.  E.g., 'PageHasForms'.
    required string name = 1;

    // Feature value is always in the range [0.0, 1.0].  Boolean features
    // have value 1.0.
    required double value = 2;
  }

  // List of features that were extracted.  Those are the features that were
  // sent to the scorer and which resulted in client_score being computed.
  repeated Feature feature_map = 5;

  // The version number of the model that was used to compute the client-score.
  // Copied from ClientSideModel.version().
  optional int32 model_version = 6;

  // Field 7 is only used on the server.

  // List of features that are extracted in the client but are not used in the
  // machine learning model.
  repeated Feature non_model_feature_map = 8;

  // The referrer URL.  This field might not be set, for example, in the case
  // where the referrer uses HTTPs.
  // OBSOLETE: Use feature 'Referrer=<referrer>' instead.
  optional string OBSOLETE_referrer_url = 9;

  // Field 11 is only used on the server.

  // List of shingle hashes we extracted.
  repeated uint32 shingle_hashes = 12 [packed = true];

  // The model filename (basename) that was used by the client.
  optional string model_filename = 13;

  // Population that the reporting user is part of.
  optional ChromeUserPopulation population = 14;
}

message ClientPhishingResponse {
  required bool phishy = 1;

  // A list of SafeBrowsing host-suffix / path-prefix expressions that
  // are whitelisted.  The client must match the current top-level URL
  // against these whitelisted expressions and only apply a positive
  // phishing verdict above if the URL does not match any expression
  // on this whitelist.  The client must not cache these whitelisted
  // expressions.  This whitelist will be empty for the vast majority
  // of the responses but might contain up to 100 entries in emergency
  // situations.
  //
  // Marked OBSOLETE because the URL is sent for all users, so the server
  // can do whitelist matching.
  repeated string OBSOLETE_whitelist_expression = 2;
}

message ClientMalwareRequest {
  // URL that the client visited.  The CGI parameters are stripped by the
  // client.
  required string url = 1;

  // Field 2 is deleted and no longer in use.

  // Field 3 is only used on the server.

  // The referrer URL.  This field might not be set, for example, in the case
  // where the referrer uses HTTPS.
  optional string referrer_url = 4;

  // Field 5 and 6 are only used on the server.

  message UrlInfo {
    required string ip = 1;
    required string url = 2;
    optional string method = 3;
    optional string referrer = 4;
    // Resource type, the int value is a direct cast from the Type enum
    // of ResourceType class defined in //src/webkit/commom/resource_type.h
    optional int32 resource_type = 5;
  }

  // List of resource urls that match the malware IP list.
  repeated UrlInfo bad_ip_url_info = 7;

  // Population that the reporting user is part of.
  optional ChromeUserPopulation population = 9;
}

message ClientMalwareResponse {
  required bool blacklist = 1;
  // The confirmed blacklisted bad IP and its url, which will be shown in
  // malware warning, if the blacklist verdict is true.
  // This IP string could be either in IPv4 or IPv6 format, which is the same
  // as the ones client sent to server.
  optional string bad_ip = 2;
  optional string bad_url = 3;
}

message ClientDownloadRequest {
  // The final URL of the download (after all redirects).
  required string url = 1;

  // This message contains various binary digests of the download payload.
  message Digests {
    optional bytes sha256 = 1;
    optional bytes sha1 = 2;
    optional bytes md5 = 3;
  }
  required Digests digests = 2;

  // This is the length in bytes of the download payload.
  required int64 length = 3;

  // Type of the resources stored below.
  enum ResourceType {
    // The final URL of the download payload.  The resource URL should
    // correspond to the URL field above.
    DOWNLOAD_URL = 0;
    // A redirect URL that was fetched before hitting the final DOWNLOAD_URL.
    DOWNLOAD_REDIRECT = 1;
    // The final top-level URL of the tab that triggered the download.
    TAB_URL = 2;
    // A redirect URL thas was fetched before hitting the final TAB_URL.
    TAB_REDIRECT = 3;
  }

  message Resource {
    required string url = 1;
    required ResourceType type = 2;
    optional bytes remote_ip = 3;
    // This will only be set if the referrer is available and if the
    // resource type is either TAB_URL or DOWNLOAD_URL.
    optional string referrer = 4;

    // TODO(noelutz): add the transition type?
  }

  // This repeated field will store all the redirects as well as the
  // final URLs for the top-level tab URL (i.e., the URL that
  // triggered the download) as well as for the download URL itself.
  repeated Resource resources = 4;

  // A trust chain of certificates.  Each chain begins with the signing
  // certificate of the binary, and ends with a self-signed certificate,
  // typically from a trusted root CA.  This structure is analogous to
  // CERT_CHAIN_CONTEXT on Windows.
  message CertificateChain {
    // A single link in the chain.
    message Element {
      // DER-encoded X.509 representation of the certificate.
      optional bytes certificate = 1;
      // Fields 2 - 7 are only used on the server.
    }
    repeated Element element = 1;
  }

  message SignatureInfo {
    // All certificate chains for each of the binary's signers.  Multiple chains
    // may be present if the binary or any certificate has multiple signers.
    // Absence of certificate chains does not imply that the binary is not
    // signed (in that case, SignedData blobs extracted from the binary may be
    // preset), but does mean that trust has not been verified.
    repeated CertificateChain certificate_chain = 1;

    // True if the signature was trusted on the client.
    optional bool trusted = 2;

    // PKCS#7 SignedData blobs extracted from a portable executable image's
    // attribute certificate table. The presence of these does not imply that
    // the signatures were deemed trusted by the client.
    repeated bytes signed_data = 3;
  }

  // This field will only be set if the binary is signed.
  optional SignatureInfo signature = 5;

  // True if the download was user initiated.
  optional bool user_initiated = 6;

  // Fields 7 and 8 are only used on the server.

  // Name of the file where the download would be stored if the
  // download completes.  E.g., "bla.exe".
  optional string file_basename = 9;

  // Starting with Chrome M19 we're also sending back pings for Chrome
  // extensions that get downloaded by users.
  enum DownloadType {
    WIN_EXECUTABLE = 0;    // Currently all .exe, .cab and .msi files.
    CHROME_EXTENSION = 1;  // .crx files.
    ANDROID_APK = 2;       // .apk files.
    // .zip files containing one of the other executable types.
    ZIPPED_EXECUTABLE = 3;
    MAC_EXECUTABLE = 4;  // .dmg, .pkg, etc.
    ZIPPED_ARCHIVE = 5;  // .zip file containing another archive.
  }
  optional DownloadType download_type = 10 [default = WIN_EXECUTABLE];

  // Locale of the device, eg en, en_US.
  optional string locale = 11;

  message PEImageHeaders {
    // IMAGE_DOS_HEADER.
    optional bytes dos_header = 1;
    // IMAGE_FILE_HEADER.
    optional bytes file_header = 2;
    // IMAGE_OPTIONAL_HEADER32. Present only for 32-bit PE images.
    optional bytes optional_headers32 = 3;
    // IMAGE_OPTIONAL_HEADER64. Present only for 64-bit PE images.
    optional bytes optional_headers64 = 4;
    // IMAGE_SECTION_HEADER.
    repeated bytes section_header = 5;
    // Contents of the .edata section.
    optional bytes export_section_data = 6;

    message DebugData {
      // IMAGE_DEBUG_DIRECTORY.
      optional bytes directory_entry = 1;
      optional bytes raw_data = 2;
    }

    repeated DebugData debug_data = 7;
  }

  message MachOHeaders {
    // The mach_header or mach_header_64 struct.
    required bytes mach_header = 1;

    message LoadCommand {
      // |command_id| is the first uint32 of |command| as well, but is
      // extracted for easier processing.
      required uint32 command_id = 1;
      // The entire data stream of the load command.
      required bytes command = 2;
    }

    // All the load commands of the Mach-O file.
    repeated LoadCommand load_commands = 2;
  }

  message ImageHeaders {
    // Windows Portable Executable image headers.
    optional PEImageHeaders pe_headers = 1;

    // OS X Mach-O image headers.
    optional MachOHeaders mach_o_headers = 2;
  };

  // Fields 12-17 are reserved for server-side use and are never sent by the
  // client.

  optional ImageHeaders image_headers = 18;

  // Fields 19-21 are reserved for server-side use and are never sent by the
  // client.

  // A binary contained in an archive (e.g., a .zip archive).
  message ArchivedBinary {
    optional string file_basename = 1;
    optional DownloadType download_type = 2;
    optional Digests digests = 3;
    optional int64 length = 4;
    optional SignatureInfo signature = 5;
    optional ImageHeaders image_headers = 6;
  }

  repeated ArchivedBinary archived_binary = 22;

  // Population that the reporting user is part of.
  optional ChromeUserPopulation population = 24;
}

message ClientDownloadResponse {
  enum Verdict {
    // Download is considered safe.
    SAFE = 0;
    // Download is considered dangerous.  Chrome should show a warning to the
    // user.
    DANGEROUS = 1;
    // Download is unknown.  Chrome should display a less severe warning.
    UNCOMMON = 2;
    // The download is potentially unwanted.
    POTENTIALLY_UNWANTED = 3;
    // The download is from a dangerous host.
    DANGEROUS_HOST = 4;
  }
  required Verdict verdict = 1;

  message MoreInfo {
    // A human-readable string describing the nature of the warning.
    // Only if verdict != SAFE. Localized based on request.locale.
    optional string description = 1;

    // A URL to get more information about this warning, if available.
    optional string url = 2;
  }
  optional MoreInfo more_info = 2;

  // An arbitrary token that should be sent along for further server requests.
  optional bytes token = 3;
}

// The following protocol buffer holds the feedback report gathered
// from the user regarding the download.
message ClientDownloadReport {
  // The information of user who provided the feedback.
  // This is going to be useful for handling appeals.
  message UserInformation {
    optional string email = 1;
  }

  enum Reason {
    SHARE = 0;
    FALSE_POSITIVE = 1;
    APPEAL = 2;
  }

  // The type of feedback for this report.
  optional Reason reason = 1;

  // The original download ping
  optional ClientDownloadRequest download_request = 2;

  // Stores the information of the user who provided the feedback.
  optional UserInformation user_information = 3;

  // Unstructed comments provided by the user.
  optional bytes comment = 4;

  // The original download response sent from the verdict server.
  optional ClientDownloadResponse download_response = 5;
}

// This is used to send back upload status to the client after upload completion
message ClientUploadResponse {
  enum UploadStatus {
    // The upload was successful and a complete response can be expected
    SUCCESS = 0;

    // The upload was unsuccessful and the response is incomplete.
    UPLOAD_FAILURE = 1;
  }

  // Holds the upload status
  optional UploadStatus status = 1;

  // Holds the permalink where the results of scanning the binary are available
  optional string permalink = 2;
}

message ClientIncidentReport {
  message IncidentData {
    message TrackedPreferenceIncident {
      enum ValueState {
        UNKNOWN = 0;
        CLEARED = 1;
        WEAK_LEGACY_OBSOLETE = 2;
        CHANGED = 3;
        UNTRUSTED_UNKNOWN_VALUE = 4;
      }

      optional string path = 1;
      optional string atomic_value = 2;
      repeated string split_key = 3;
      optional ValueState value_state = 4;
    }
    message BinaryIntegrityIncident {
      optional string file_basename = 1;
      optional ClientDownloadRequest.SignatureInfo signature = 2;
    }
    message BlacklistLoadIncident {
      optional string path = 1;
      optional ClientDownloadRequest.Digests digest = 2;
      optional string version = 3;
      optional bool blacklist_initialized = 4;
      optional ClientDownloadRequest.SignatureInfo signature = 5;
      optional ClientDownloadRequest.ImageHeaders image_headers = 6;
    }
    message VariationsSeedSignatureIncident {
      optional string variations_seed_signature = 1;
    }
    message ResourceRequestIncident {
      enum Type {
        UNKNOWN = 0;
        TYPE_SCRIPT = 1;
        TYPE_DOMAIN = 2;
      }
      optional bytes digest = 1;
      optional string origin = 2;
      optional Type type = 3 [default = UNKNOWN];
    }
    optional int64 incident_time_msec = 1;
    optional TrackedPreferenceIncident tracked_preference = 2;
    optional BinaryIntegrityIncident binary_integrity = 3;
    optional BlacklistLoadIncident blacklist_load = 4;
    // Note: skip tag 5 because it was previously used.
    optional VariationsSeedSignatureIncident variations_seed_signature = 6;
    optional ResourceRequestIncident resource_request = 7;
  }

  repeated IncidentData incident = 1;

  message DownloadDetails {
    optional bytes token = 1;
    optional ClientDownloadRequest download = 2;
    optional int64 download_time_msec = 3;
    optional int64 open_time_msec = 4;
  }

  optional DownloadDetails download = 2;

  message EnvironmentData {
    message OS {
      optional string os_name = 1;
      optional string os_version = 2;
    }
    optional OS os = 1;
    message Machine {
      optional string cpu_architecture = 1;
      optional string cpu_vendor = 2;
      optional uint32 cpuid = 3;
    }
    optional Machine machine = 2;
    message Process {
      optional string version = 1;
      repeated string OBSOLETE_dlls = 2;
      message Patch {
        optional string function = 1;
        optional string target_dll = 2;
      }
      repeated Patch patches = 3;
      message NetworkProvider {}
      repeated NetworkProvider network_providers = 4;
      enum Channel {
        CHANNEL_UNKNOWN = 0;
        CHANNEL_CANARY = 1;
        CHANNEL_DEV = 2;
        CHANNEL_BETA = 3;
        CHANNEL_STABLE = 4;
      }
      optional Channel chrome_update_channel = 5;
      optional int64 uptime_msec = 6;
      optional bool metrics_consent = 7;
      optional bool extended_consent = 8;
      message Dll {
        enum Feature {
          UNKNOWN = 0;
          LSP = 1;
        }
        optional string path = 1;
        optional uint64 base_address = 2;
        optional uint32 length = 3;
        repeated Feature feature = 4;
        optional ClientDownloadRequest.ImageHeaders image_headers = 5;
      }
      repeated Dll dll = 9;
      repeated string blacklisted_dll = 10;
      message ModuleState {
        enum ModifiedState {
          UNKNOWN = 0;
          MODULE_STATE_UNKNOWN = 1;
          MODULE_STATE_UNMODIFIED = 2;
          MODULE_STATE_MODIFIED = 3;
        }
        optional string name = 1;
        optional ModifiedState modified_state = 2;
        repeated string OBSOLETE_modified_export = 3;

        message Modification {
          optional uint32 file_offset = 1;
          optional int32 byte_count = 2;
          optional bytes modified_bytes = 3;
          optional string export_name = 4;
        }
        repeated Modification modification = 4;
      }
      repeated ModuleState module_state = 11;
    }
    optional Process process = 3;
  }

  optional EnvironmentData environment = 3;

  // Population that the reporting user is part of.
  optional ChromeUserPopulation population = 7;
}

message ClientIncidentResponse {
  optional bytes token = 1;
  optional bool download_requested = 2;

  message EnvironmentRequest { optional int32 dll_index = 1; }

  repeated EnvironmentRequest environment_requests = 3;
}

message DownloadMetadata {
  optional uint32 download_id = 1;

  optional ClientIncidentReport.DownloadDetails download = 2;
}