[libcxx-commits] [libcxx] [libcxx] Caches file attributes during directory iteration. (PR #93316)
Eduard Satdarov via libcxx-commits
libcxx-commits at lists.llvm.org
Fri May 24 09:13:39 PDT 2024
https://github.com/ed-sat created https://github.com/llvm/llvm-project/pull/93316
Added caching file attributes during directory iteration for OS Windows. Allows improving performance working with files in directory.
>From 08ca1dcb567ee9ec6ed14d5f296e15e20cb1334a Mon Sep 17 00:00:00 2001
From: Eduard Satdarov <sath at yandex-team.ru>
Date: Fri, 24 May 2024 16:41:35 +0300
Subject: [PATCH] [libcxx] Cache file attributes during directory iteration.
---
libcxx/include/__filesystem/directory_entry.h | 38 +++++++++++++++++++
libcxx/src/filesystem/directory_iterator.cpp | 12 +++---
libcxx/src/filesystem/file_descriptor.h | 16 +++++---
3 files changed, 55 insertions(+), 11 deletions(-)
diff --git a/libcxx/include/__filesystem/directory_entry.h b/libcxx/include/__filesystem/directory_entry.h
index 016ad94a853dc..6d35d92ebf483 100644
--- a/libcxx/include/__filesystem/directory_entry.h
+++ b/libcxx/include/__filesystem/directory_entry.h
@@ -200,6 +200,8 @@ class directory_entry {
_Empty,
_IterSymlink,
_IterNonSymlink,
+ _IterCachedSymlink,
+ _IterCachedNonSymlink,
_RefreshSymlink,
_RefreshSymlinkUnresolved,
_RefreshNonSymlink
@@ -241,6 +243,28 @@ class directory_entry {
return __data;
}
+ _LIBCPP_HIDE_FROM_ABI static __cached_data __create_iter_cached_result(file_type __ft, uintmax_t __size, perms __perm, file_time_type __write_time) {
+ __cached_data __data;
+ __data.__type_ = __ft;
+ __data.__size_ = __size;
+ __data.__write_time_ = __write_time;
+ if (__ft == file_type::symlink)
+ __data.__sym_perms_ = __perm;
+ else
+ __data.__non_sym_perms_ = __perm;
+ __data.__cache_type_ = [&]() {
+ switch (__ft) {
+ case file_type::none:
+ return _Empty;
+ case file_type::symlink:
+ return _IterCachedSymlink;
+ default:
+ return _IterCachedNonSymlink;
+ }
+ }();
+ return __data;
+ }
+
_LIBCPP_HIDE_FROM_ABI void __assign_iter_entry(_Path&& __p, __cached_data __dt) {
__p_ = std::move(__p);
__data_ = __dt;
@@ -282,12 +306,14 @@ class directory_entry {
case _Empty:
return __symlink_status(__p_, __ec).type();
case _IterSymlink:
+ case _IterCachedSymlink:
case _RefreshSymlink:
case _RefreshSymlinkUnresolved:
if (__ec)
__ec->clear();
return file_type::symlink;
case _IterNonSymlink:
+ case _IterCachedNonSymlink:
case _RefreshNonSymlink:
file_status __st(__data_.__type_);
if (__ec && !filesystem::exists(__st))
@@ -303,9 +329,11 @@ class directory_entry {
switch (__data_.__cache_type_) {
case _Empty:
case _IterSymlink:
+ case _IterCachedSymlink:
case _RefreshSymlinkUnresolved:
return __status(__p_, __ec).type();
case _IterNonSymlink:
+ case _IterCachedNonSymlink:
case _RefreshNonSymlink:
case _RefreshSymlink: {
file_status __st(__data_.__type_);
@@ -324,8 +352,10 @@ class directory_entry {
case _Empty:
case _IterNonSymlink:
case _IterSymlink:
+ case _IterCachedSymlink:
case _RefreshSymlinkUnresolved:
return __status(__p_, __ec);
+ case _IterCachedNonSymlink:
case _RefreshNonSymlink:
case _RefreshSymlink:
return file_status(__get_ft(__ec), __data_.__non_sym_perms_);
@@ -339,8 +369,10 @@ class directory_entry {
case _IterNonSymlink:
case _IterSymlink:
return __symlink_status(__p_, __ec);
+ case _IterCachedNonSymlink:
case _RefreshNonSymlink:
return file_status(__get_sym_ft(__ec), __data_.__non_sym_perms_);
+ case _IterCachedSymlink:
case _RefreshSymlink:
case _RefreshSymlinkUnresolved:
return file_status(__get_sym_ft(__ec), __data_.__sym_perms_);
@@ -353,8 +385,10 @@ class directory_entry {
case _Empty:
case _IterNonSymlink:
case _IterSymlink:
+ case _IterCachedSymlink:
case _RefreshSymlinkUnresolved:
return filesystem::__file_size(__p_, __ec);
+ case _IterCachedNonSymlink:
case _RefreshSymlink:
case _RefreshNonSymlink: {
error_code __m_ec;
@@ -375,6 +409,8 @@ class directory_entry {
case _Empty:
case _IterNonSymlink:
case _IterSymlink:
+ case _IterCachedNonSymlink:
+ case _IterCachedSymlink:
case _RefreshSymlinkUnresolved:
return filesystem::__hard_link_count(__p_, __ec);
case _RefreshSymlink:
@@ -395,6 +431,8 @@ class directory_entry {
case _IterSymlink:
case _RefreshSymlinkUnresolved:
return filesystem::__last_write_time(__p_, __ec);
+ case _IterCachedNonSymlink:
+ case _IterCachedSymlink:
case _RefreshSymlink:
case _RefreshNonSymlink: {
error_code __m_ec;
diff --git a/libcxx/src/filesystem/directory_iterator.cpp b/libcxx/src/filesystem/directory_iterator.cpp
index dceb3486279f8..d7ed9a358f559 100644
--- a/libcxx/src/filesystem/directory_iterator.cpp
+++ b/libcxx/src/filesystem/directory_iterator.cpp
@@ -77,13 +77,13 @@ class __dir_stream {
bool assign() {
if (!wcscmp(__data_.cFileName, L".") || !wcscmp(__data_.cFileName, L".."))
return false;
- // FIXME: Cache more of this
- // directory_entry::__cached_data cdata;
- // cdata.__type_ = get_file_type(__data_);
- // cdata.__size_ = get_file_size(__data_);
- // cdata.__write_time_ = get_write_time(__data_);
__entry_.__assign_iter_entry(
- __root_ / __data_.cFileName, directory_entry::__create_iter_result(detail::get_file_type(__data_)));
+ __root_ / __data_.cFileName,
+ directory_entry::__create_iter_cached_result(
+ detail::get_file_type(__data_),
+ detail::get_file_size(__data_),
+ detail::get_file_perm(__data_),
+ detail::get_write_time(__data_)));
return true;
}
diff --git a/libcxx/src/filesystem/file_descriptor.h b/libcxx/src/filesystem/file_descriptor.h
index 50178ff84e03f..44a072b72e766 100644
--- a/libcxx/src/filesystem/file_descriptor.h
+++ b/libcxx/src/filesystem/file_descriptor.h
@@ -97,11 +97,17 @@ inline uintmax_t get_file_size(const WIN32_FIND_DATAW& data) {
return (static_cast<uint64_t>(data.nFileSizeHigh) << 32) + data.nFileSizeLow;
}
inline file_time_type get_write_time(const WIN32_FIND_DATAW& data) {
- ULARGE_INTEGER tmp;
- const FILETIME& time = data.ftLastWriteTime;
- tmp.u.LowPart = time.dwLowDateTime;
- tmp.u.HighPart = time.dwHighDateTime;
- return file_time_type(file_time_type::duration(tmp.QuadPart));
+ using detail::fs_time;
+ auto ts = filetime_to_timespec(time);
+ if (!fs_time::is_representable(ts))
+ return file_time_type::min();
+ return fs_time::convert_from_timespec(ts);
+}
+inline perms get_file_perm(const WIN32_FIND_DATAW& data) {
+ unsigned st_mode = 0555; // Read-only
+ if (!(data.dwFileAttributes & FILE_ATTRIBUTE_READONLY))
+ st_mode |= 0222; // Write
+ return static_cast<perms>(st_mode) & perms::mask;
}
#endif // !_LIBCPP_WIN32API
More information about the libcxx-commits
mailing list