Add --keep-all-times and default to only keeping mtime

doc/mkdwarfs.md

@@ -119,6 +119,11 @@ Most other options are concerned with compression tuning:
     reduce the size of the file system. You can pass either a unix time stamp
     or `now`.
 
+  * `--keep-all-times`:
+    As of release 0.3.0, by default, `mkdwarfs` will only save the contents of
+    the `mtime` field in order to save metadata space. If you want to save
+    `atime` and `ctime` as well, use this option.
+
   * `--order=none`|`path`|`similarity`|`nilsimsa`|`script`:
     The order in which files will be written to the filesystem. Choosing `none`,
     the files will be stored in the order in which they are discovered. With

include/dwarfs/global_entry_data.h

@@ -33,6 +33,8 @@ struct scanner_options;
 
 class global_entry_data {
  public:
+  enum class timestamp_type { ATIME, MTIME, CTIME };
+
   global_entry_data(scanner_options const& options)
       : options_(options) {}
 
@@ -41,11 +43,9 @@ class global_entry_data {
 
   void add_mode(uint16_t mode) { add(mode, modes_, next_mode_index_); }
 
-  void add_time(uint64_t time) {
-    if (time < timestamp_base_) {
-      timestamp_base_ = time;
-    }
-  }
+  void add_mtime(uint64_t time);
+  void add_atime(uint64_t time);
+  void add_ctime(uint64_t time);
 
   void add_name(std::string const& name) { names_.emplace(name, 0); }
   void add_link(std::string const& link) { links_.emplace(link, 0); }
@@ -67,7 +67,9 @@ class global_entry_data {
     return links_.at(link);
   }
 
-  uint64_t get_time_offset(uint64_t time) const;
+  uint64_t get_mtime_offset(uint64_t time) const;
+  uint64_t get_atime_offset(uint64_t time) const;
+  uint64_t get_ctime_offset(uint64_t time) const;
 
   std::vector<uint16_t> get_uids() const;
   std::vector<uint16_t> get_gids() const;

include/dwarfs/options.h

@@ -60,6 +60,7 @@ struct scanner_options {
   std::optional<uint16_t> uid;
   std::optional<uint16_t> gid;
   std::optional<uint64_t> timestamp;
+  bool keep_all_times{false};
   bool remove_empty_dirs{false};
   inode_options inode;
 };

src/dwarfs/entry.cpp

@@ -97,9 +97,9 @@ void entry::update(global_entry_data& data) const {
   data.add_uid(stat_.st_uid);
   data.add_gid(stat_.st_gid);
   data.add_mode(stat_.st_mode & 0xFFFF);
-  data.add_time(stat_.st_atime);
-  data.add_time(stat_.st_mtime);
-  data.add_time(stat_.st_ctime);
+  data.add_atime(stat_.st_atime);
+  data.add_mtime(stat_.st_mtime);
+  data.add_ctime(stat_.st_ctime);
 }
 
 void entry::pack(thrift::metadata::entry& entry_v2,
@@ -108,9 +108,9 @@ void entry::pack(thrift::metadata::entry& entry_v2,
   entry_v2.mode_index = data.get_mode_index(stat_.st_mode & 0xFFFF);
   entry_v2.owner_index = data.get_uid_index(stat_.st_uid);
   entry_v2.group_index = data.get_gid_index(stat_.st_gid);
-  entry_v2.atime_offset = data.get_time_offset(stat_.st_atime);
-  entry_v2.mtime_offset = data.get_time_offset(stat_.st_mtime);
-  entry_v2.ctime_offset = data.get_time_offset(stat_.st_ctime);
+  entry_v2.atime_offset = data.get_atime_offset(stat_.st_atime);
+  entry_v2.mtime_offset = data.get_mtime_offset(stat_.st_mtime);
+  entry_v2.ctime_offset = data.get_ctime_offset(stat_.st_ctime);
   entry_v2.inode = inode_num();
 }
 

src/dwarfs/global_entry_data.cpp

@@ -61,8 +61,18 @@ void global_entry_data::index(std::unordered_map<std::string, uint32_t>& map) {
   from(map) | get<0>() | order | [&](std::string const& s) { map[s] = ix++; };
 }
 
-uint64_t global_entry_data::get_time_offset(uint64_t time) const {
-  return options_.timestamp ? 0 : time - timestamp_base_;
+uint64_t global_entry_data::get_mtime_offset(uint64_t time) const {
+  return !options_.timestamp ? time - timestamp_base_ : UINT64_C(0);
+}
+
+uint64_t global_entry_data::get_atime_offset(uint64_t time) const {
+  return !options_.timestamp && options_.keep_all_times ? time - timestamp_base_
+                                                        : UINT64_C(0);
+}
+
+uint64_t global_entry_data::get_ctime_offset(uint64_t time) const {
+  return !options_.timestamp && options_.keep_all_times ? time - timestamp_base_
+                                                        : UINT64_C(0);
 }
 
 uint64_t global_entry_data::get_timestamp_base() const {
@@ -89,4 +99,22 @@ void global_entry_data::add_gid(uint16_t gid) {
   }
 }
 
+void global_entry_data::add_mtime(uint64_t time) {
+  if (time < timestamp_base_) {
+    timestamp_base_ = time;
+  }
+}
+
+void global_entry_data::add_atime(uint64_t time) {
+  if (options_.keep_all_times) {
+    add_mtime(time);
+  }
+}
+
+void global_entry_data::add_ctime(uint64_t time) {
+  if (options_.keep_all_times) {
+    add_mtime(time);
+  }
+}
+
 } // namespace dwarfs

src/dwarfs/metadata_v2.cpp

@@ -624,6 +624,7 @@ int metadata_<LoggerPolicy>::getattr(entry_view entry,
   auto mode = entry.mode();
   auto timebase = meta_.timestamp_base();
   auto inode = entry.inode();
+  bool mtime_only = meta_.options() && meta_.options()->mtime_only();
 
   stbuf->st_mode = mode;
 
@@ -633,9 +634,11 @@ int metadata_<LoggerPolicy>::getattr(entry_view entry,
   stbuf->st_blocks = (stbuf->st_size + 511) / 512;
   stbuf->st_uid = entry.getuid();
   stbuf->st_gid = entry.getgid();
-  stbuf->st_atime = timebase + entry.atime_offset();
   stbuf->st_mtime = timebase + entry.mtime_offset();
-  stbuf->st_ctime = timebase + entry.ctime_offset();
+  stbuf->st_atime =
+      mtime_only ? stbuf->st_mtime : timebase + entry.atime_offset();
+  stbuf->st_ctime =
+      mtime_only ? stbuf->st_mtime : timebase + entry.ctime_offset();
   stbuf->st_nlink = options_.enable_nlink && S_ISREG(mode)
                         ? nlinks_.at(inode - chunk_index_offset_)
                         : 1;

src/dwarfs/scanner.cpp

@@ -560,6 +560,9 @@ void scanner_<LoggerPolicy>::scan(filesystem_writer& fsw,
   root->accept(sdv);
   sdv.pack(mv2, ge_data);
 
+  thrift::metadata::fs_options fsopts;
+  fsopts.mtime_only = !options_.keep_all_times;
+
   mv2.uids = ge_data.get_uids();
   mv2.gids = ge_data.get_gids();
   mv2.modes = ge_data.get_modes();
@@ -568,6 +571,7 @@ void scanner_<LoggerPolicy>::scan(filesystem_writer& fsw,
   mv2.timestamp_base = ge_data.get_timestamp_base();
   mv2.block_size = UINT32_C(1) << cfg_.block_size_bits;
   mv2.total_fs_size = prog.original_size;
+  mv2.options_ref() = fsopts;
 
   auto [schema, data] = metadata_v2::freeze(mv2);
 

src/mkdwarfs.cpp

@@ -337,6 +337,9 @@ int mkdwarfs(int argc, char** argv) {
     ("set-time",
         po::value<std::string>(&timestamp),
         "set timestamp for whole file system (unixtime or 'now')")
+    ("keep-all-times",
+        po::value<bool>(&options.keep_all_times)->zero_tokens(),
+        "save atime and ctime in addition to mtime")
     ("order",
         po::value<file_order_mode>(&options.file_order)
             ->default_value(file_order_mode::SIMILARITY, "similarity"),

thrift/metadata.thrift

@@ -86,6 +86,11 @@ struct entry {
    8: required UInt64 ctime_offset,
 }
 
+struct fs_options {
+   // file system contains only mtime time stamps
+   1: required bool   mtime_only,
+}
+
 struct metadata {
    /**
     * Ranges of chunks that make up regular files. Identical
@@ -169,6 +174,13 @@ struct metadata {
    // total file system size
   16: required UInt64          total_fs_size,
 
+  //=========================================================//
+  // fields added with dwarfs-0.3.0, file system version 2.1 //
+  //=========================================================//
+
    // device ids, for lookup by (inode - device_index_offset)
   17: optional list<UInt64>    devices,
+
+   // file system options
+  18: optional fs_options      options,
 }