1 /* Copyright (c) 2015-2023. The SimGrid Team. All rights reserved. */
3 /* This program is free software; you can redistribute it and/or modify it
4 * under the terms of the license (GNU LGPL) which comes with this package. */
6 #include <simgrid/plugins/file_system.h>
7 #include <simgrid/s4u/Comm.hpp>
8 #include <simgrid/s4u/Disk.hpp>
9 #include <simgrid/s4u/Engine.hpp>
10 #include <simgrid/s4u/Host.hpp>
11 #include <simgrid/simix.hpp>
12 #include <xbt/asserts.h>
13 #include <xbt/config.hpp>
14 #include <xbt/file.hpp>
16 #include <xbt/parse_units.hpp>
18 #include <boost/algorithm/string.hpp>
19 #include <boost/algorithm/string/split.hpp>
23 XBT_LOG_NEW_DEFAULT_SUBCATEGORY(s4u_file, s4u, "S4U files");
25 /** @defgroup plugin_filesystem Plugin FileSystem
27 * This adds the notion of Files on top of the storage notion that provided by the core of SimGrid.
28 * Activate this plugin at will.
33 template class xbt::Extendable<s4u::File>;
36 simgrid::xbt::Extension<Disk, FileSystemDiskExt> FileSystemDiskExt::EXTENSION_ID;
37 simgrid::xbt::Extension<Host, FileDescriptorHostExt> FileDescriptorHostExt::EXTENSION_ID;
38 int FileDescriptorHostExt::max_file_descriptors;
40 const Disk* File::find_local_disk_on(const Host* host)
42 const Disk* d = nullptr;
43 size_t longest_prefix_length = 0;
44 for (auto const& disk : host->get_disks()) {
45 std::string current_mount;
46 if (disk->get_host() != host)
47 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point(disk->get_host());
49 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
50 mount_point_ = fullpath_.substr(0, current_mount.length());
51 if (mount_point_ == current_mount && current_mount.length() > longest_prefix_length) {
52 /* The current mount name is found in the full path and is bigger than the previous*/
53 longest_prefix_length = current_mount.length();
56 xbt_assert(longest_prefix_length > 0, "Can't find mount point for '%s' on '%s'", fullpath_.c_str(),
58 /* Mount point found, split fullpath_ into mount_name and path+filename*/
59 mount_point_ = fullpath_.substr(0, longest_prefix_length);
60 if (mount_point_ == "/")
63 path_ = fullpath_.substr(longest_prefix_length, fullpath_.length());
64 XBT_DEBUG("%s + %s", mount_point_.c_str(), path_.c_str());
69 File::File(const std::string& fullpath, void* userdata) : File(fullpath, Host::current(), userdata) {}
71 File::File(const std::string& fullpath, const_sg_host_t host, void* userdata) : fullpath_(fullpath)
73 kernel::actor::simcall_answered([this, &host, userdata] {
74 this->set_data(userdata);
75 // this cannot fail because we get a xbt_die if the mountpoint does not exist
76 local_disk_ = find_local_disk_on(host);
78 // assign a file descriptor id to the newly opened File
79 auto* ext = host->extension<FileDescriptorHostExt>();
80 if (ext->file_descriptor_table == nullptr) {
81 ext->file_descriptor_table = std::make_unique<std::vector<int>>(FileDescriptorHostExt::max_file_descriptors);
82 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
84 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
85 desc_id = ext->file_descriptor_table->back();
86 ext->file_descriptor_table->pop_back();
88 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
89 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
91 // if file does not exist create an empty file
93 auto sz = content->find(path_);
94 if (sz != content->end()) {
96 XBT_DEBUG("\tOpen file '%s', size %llu", path_.c_str(), size_);
99 content->insert({path_, size_});
100 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
106 File::~File() = default;
108 File* File::open(const std::string& fullpath, void* userdata)
110 return new File(fullpath, userdata);
113 File* File::open(const std::string& fullpath, const_sg_host_t host, void* userdata)
115 return new File(fullpath, host, userdata);
120 std::vector<int>* desc_table = Host::current()->extension<FileDescriptorHostExt>()->file_descriptor_table.get();
121 kernel::actor::simcall_answered([this, desc_table] { desc_table->push_back(this->desc_id); });
125 void File::dump() const
127 XBT_INFO("File Descriptor information:\n"
128 "\t\tFull path: '%s'\n"
130 "\t\tMount point: '%s'\n"
131 "\t\tDisk Id: '%s'\n"
132 "\t\tHost Id: '%s'\n"
133 "\t\tFile Descriptor Id: %d",
134 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
138 sg_size_t File::read(sg_size_t size)
140 if (size_ == 0) /* Nothing to read, return */
142 Host* host = nullptr;
143 // if the current position is close to the end of the file, we may not be able to read the requested size
144 sg_size_t to_read = std::min(size, size_ - current_position_);
145 sg_size_t read_size = 0;
147 /* Find the host where the file is physically located and read it */
148 host = local_disk_->get_host();
149 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
150 read_size = local_disk_->read(to_read);
152 current_position_ += read_size;
154 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
155 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
156 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
157 Comm::sendto(host, Host::current(), read_size);
163 /** @brief Write into a file (local or remote)
164 * @ingroup plugin_filesystem
166 * @param size of the file to write
167 * @param write_inside
168 * @return the number of bytes successfully write or -1 if an error occurred
170 sg_size_t File::write(sg_size_t size, bool write_inside)
172 if (size == 0) /* Nothing to write, return */
175 sg_size_t write_size = 0;
176 /* Find the host where the file is physically located (remote or local)*/
177 if (Host* host = local_disk_->get_host(); host && host->get_name() != Host::current()->get_name()) {
178 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
179 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
180 Comm::sendto(Host::current(), host, size);
182 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
183 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
184 // If the disk is full before even starting to write
185 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
187 if (not write_inside)
188 /* Subtract the part of the file that might disappear from the used sized on the storage element */
189 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
190 write_size = local_disk_->write(size);
191 update_position(current_position_ + write_size);
196 sg_size_t File::size() const
201 void File::seek(sg_offset_t offset)
203 current_position_ = offset;
206 void File::seek(sg_offset_t offset, int origin)
210 update_position(offset);
213 update_position(current_position_ + offset);
216 update_position(size_ + offset);
223 void File::update_position(sg_offset_t position)
225 xbt_assert(position >= 0, "Error in seek, cannot seek before file %s", get_path());
226 current_position_ = position;
227 if(current_position_>size_){
228 XBT_DEBUG("Updating size of file %s from %llu to %lld", path_.c_str(), size_, position);
229 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(current_position_-size_);
230 size_ = current_position_;
232 kernel::actor::simcall_answered([this] {
233 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
234 content->erase(path_);
235 content->insert({path_, size_});
240 sg_size_t File::tell() const
242 return current_position_;
245 void File::move(const std::string& fullpath) const
247 /* Check if the new full path is on the same mount point */
248 if (fullpath.rfind(mount_point_, 0) == 0) {
249 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
250 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
252 auto sz = content->find(path_);
253 if (sz != content->end()) { // src file exists
254 sg_size_t new_size = sz->second;
255 content->erase(path_);
256 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
257 content->insert({path.c_str(), new_size});
258 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
260 XBT_WARN("File %s doesn't exist", path_.c_str());
264 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
268 int File::unlink() const
270 /* Check if the file is on local storage */
271 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
272 const char* name = local_disk_->get_cname();
274 if (not content || content->find(path_) == content->end()) {
275 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
278 XBT_DEBUG("UNLINK %s of size %llu on disk '%s'", path_.c_str(), size_, name);
280 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_);
281 // Remove the file from storage
282 content->erase(path_);
288 int File::remote_copy(sg_host_t host, const std::string& fullpath)
290 /* Find the host where the file is physically located and read it */
291 Host* src_host = nullptr;
292 sg_size_t read_size = 0;
294 Host* dst_host = host;
295 size_t longest_prefix_length = 0;
299 src_host = local_disk_->get_host();
300 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
301 read_size = local_disk_->read(size_);
302 current_position_ += read_size;
304 const Disk* dst_disk = nullptr;
306 for (auto const& disk : host->get_disks()) {
307 std::string current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
308 std::string mount_point = fullpath.substr(0, current_mount.length());
309 if (mount_point == current_mount && current_mount.length() > longest_prefix_length) {
310 /* The current mount name is found in the full path and is bigger than the previous*/
311 longest_prefix_length = current_mount.length();
316 if (dst_disk == nullptr) {
317 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath.c_str(), host->get_cname());
322 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
323 dst_host->get_cname());
324 Comm::sendto(src_host, dst_host, read_size);
327 /* Create file on remote host, write it and close it */
328 auto* fd = File::open(fullpath, dst_host, nullptr);
329 fd->write(read_size);
334 int File::remote_move(sg_host_t host, const std::string& fullpath)
336 int res = remote_copy(host, fullpath);
341 FileSystemDiskExt::FileSystemDiskExt(const Disk* ptr)
343 if (const char* size_str = ptr->get_property("size")) {
344 std::string dummyfile;
345 size_ = xbt_parse_get_size(dummyfile, -1, size_str, "disk size " + ptr->get_name());
348 if (const char* current_mount_str = ptr->get_property("mount"))
349 mount_point_ = current_mount_str;
353 if (const char* content_str = ptr->get_property("content"))
354 content_.reset(parse_content(content_str));
357 std::map<std::string, sg_size_t, std::less<>>* FileSystemDiskExt::parse_content(const std::string& filename)
359 if (filename.empty())
362 auto* parse_content = new std::map<std::string, sg_size_t, std::less<>>();
364 auto fs = std::unique_ptr<std::ifstream>(simgrid::xbt::path_ifsopen(filename));
365 xbt_assert(not fs->fail(), "Cannot open file '%s' (path=%s)", filename.c_str(),
366 simgrid::xbt::path_to_string().c_str());
369 std::vector<std::string> tokens;
371 std::getline(*fs, line);
373 if (line.length() > 0) {
374 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
375 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
376 sg_size_t size = std::stoull(tokens.at(1));
379 parse_content->insert({tokens.front(), size});
381 } while (not fs->eof());
382 return parse_content;
385 void FileSystemDiskExt::add_remote_mount(Host* host, const std::string& mount_point)
387 remote_mount_points_.try_emplace(host, mount_point);
390 void FileSystemDiskExt::decr_used_size(sg_size_t size)
392 simgrid::kernel::actor::simcall_answered([this, size] { used_size_ -= size; });
395 void FileSystemDiskExt::incr_used_size(sg_size_t size)
397 simgrid::kernel::actor::simcall_answered([this, size] { used_size_ += size; });
402 using simgrid::s4u::FileDescriptorHostExt;
403 using simgrid::s4u::FileSystemDiskExt;
405 static void on_disk_creation(simgrid::s4u::Disk& d)
407 d.extension_set(new FileSystemDiskExt(&d));
410 static void on_host_creation(simgrid::s4u::Host& host)
412 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
415 static void on_platform_created()
417 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
418 const char* remote_disk_str = host->get_property("remote_disk");
419 if (not remote_disk_str)
421 std::vector<std::string> tokens;
422 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
423 std::string mount_point = tokens[0];
424 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
425 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
427 const simgrid::s4u::Disk* disk = nullptr;
428 for (auto const& d : remote_host->get_disks())
429 if (d->get_name() == tokens[1]) {
434 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
435 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
439 /* **************************** Public interface *************************** */
440 /** @brief Initialize the file system plugin.
441 @ingroup plugin_filesystem
444 See the examples in :ref:`s4u_ex_disk_io`.
447 void sg_storage_file_system_init()
449 FileDescriptorHostExt::max_file_descriptors = 1024;
450 simgrid::config::bind_flag(FileDescriptorHostExt::max_file_descriptors, "storage/max_file_descriptors",
451 "Maximum number of concurrently opened files per host. Default is 1024");
453 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
454 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
455 simgrid::s4u::Disk::on_creation_cb(&on_disk_creation);
458 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
459 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
460 simgrid::s4u::Host::on_creation_cb(&on_host_creation);
462 simgrid::s4u::Engine::on_platform_created_cb(&on_platform_created);
465 sg_file_t sg_file_open(const char* fullpath, void* data)
467 return simgrid::s4u::File::open(fullpath, data);
470 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
472 return fd->read(size);
475 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
477 return fd->write(size);
480 void sg_file_close(sg_file_t fd)
485 /** Retrieves the path to the file
486 * @ingroup plugin_filesystem
488 const char* sg_file_get_name(const_sg_file_t fd)
490 xbt_assert((fd != nullptr), "Invalid file descriptor");
491 return fd->get_path();
494 /** Retrieves the size of the file
495 * @ingroup plugin_filesystem
497 sg_size_t sg_file_get_size(const_sg_file_t fd)
502 void sg_file_dump(const_sg_file_t fd)
507 /** Retrieves the user data associated with the file
508 * @ingroup plugin_filesystem
510 void* sg_file_get_data(const_sg_file_t fd)
512 return fd->get_data<void>();
515 /** Changes the user data associated with the file
516 * @ingroup plugin_filesystem
518 void sg_file_set_data(sg_file_t fd, void* data)
524 * @brief Set the file position indicator in the sg_file_t by adding offset bytes to the position specified by origin (either SEEK_SET, SEEK_CUR, or SEEK_END).
525 * @ingroup plugin_filesystem
527 * @param fd : file object that identifies the stream
528 * @param offset : number of bytes to offset from origin
529 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
530 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
531 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
533 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
535 fd->seek(offset, origin);
538 sg_size_t sg_file_tell(const_sg_file_t fd)
543 void sg_file_move(const_sg_file_t fd, const char* fullpath)
548 void sg_file_unlink(sg_file_t fd)
555 * @brief Copy a file to another location on a remote host.
556 * @ingroup plugin_filesystem
558 * @param file : the file to move
559 * @param host : the remote host where the file has to be copied
560 * @param fullpath : the complete path destination on the remote host
561 * @return If successful, the function returns 0. Otherwise, it returns -1.
563 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
565 return file->remote_copy(host, fullpath);
569 * @brief Move a file to another location on a remote host.
570 * @ingroup plugin_filesystem
572 * @param file : the file to move
573 * @param host : the remote host where the file has to be moved
574 * @param fullpath : the complete path destination on the remote host
575 * @return If successful, the function returns 0. Otherwise, it returns -1.
577 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
579 return file->remote_move(host, fullpath);
582 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
584 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
587 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
589 return d->extension<FileSystemDiskExt>()->get_used_size();
592 sg_size_t sg_disk_get_size(const_sg_disk_t d)
594 return d->extension<FileSystemDiskExt>()->get_size();
597 const char* sg_disk_get_mount_point(const_sg_disk_t d)
599 return d->extension<FileSystemDiskExt>()->get_mount_point();