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 "src/surf/surf_interface.hpp"
20 #include <boost/algorithm/string.hpp>
21 #include <boost/algorithm/string/split.hpp>
25 XBT_LOG_NEW_DEFAULT_SUBCATEGORY(s4u_file, s4u, "S4U files");
26 int sg_storage_max_file_descriptors = 1024;
28 /** @defgroup plugin_filesystem Plugin FileSystem
30 * This adds the notion of Files on top of the storage notion that provided by the core of SimGrid.
31 * Activate this plugin at will.
36 template class xbt::Extendable<s4u::File>;
39 simgrid::xbt::Extension<Disk, FileSystemDiskExt> FileSystemDiskExt::EXTENSION_ID;
40 simgrid::xbt::Extension<Host, FileDescriptorHostExt> FileDescriptorHostExt::EXTENSION_ID;
42 const Disk* File::find_local_disk_on(const Host* host)
44 const Disk* d = nullptr;
45 size_t longest_prefix_length = 0;
46 for (auto const& disk : host->get_disks()) {
47 std::string current_mount;
48 if (disk->get_host() != host)
49 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point(disk->get_host());
51 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
52 mount_point_ = fullpath_.substr(0, current_mount.length());
53 if (mount_point_ == current_mount && current_mount.length() > longest_prefix_length) {
54 /* The current mount name is found in the full path and is bigger than the previous*/
55 longest_prefix_length = current_mount.length();
58 xbt_assert(longest_prefix_length > 0, "Can't find mount point for '%s' on '%s'", fullpath_.c_str(),
60 /* Mount point found, split fullpath_ into mount_name and path+filename*/
61 mount_point_ = fullpath_.substr(0, longest_prefix_length);
62 if (mount_point_ == "/")
65 path_ = fullpath_.substr(longest_prefix_length, fullpath_.length());
66 XBT_DEBUG("%s + %s", mount_point_.c_str(), path_.c_str());
71 File::File(const std::string& fullpath, void* userdata) : File(fullpath, Host::current(), userdata) {}
73 File::File(const std::string& fullpath, const_sg_host_t host, void* userdata) : fullpath_(fullpath)
75 kernel::actor::simcall_answered([this, &host, userdata] {
76 this->set_data(userdata);
77 // this cannot fail because we get a xbt_die if the mountpoint does not exist
78 local_disk_ = find_local_disk_on(host);
80 // assign a file descriptor id to the newly opened File
81 auto* ext = host->extension<simgrid::s4u::FileDescriptorHostExt>();
82 if (ext->file_descriptor_table == nullptr) {
83 ext->file_descriptor_table = std::make_unique<std::vector<int>>(sg_storage_max_file_descriptors);
84 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
86 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
87 desc_id = ext->file_descriptor_table->back();
88 ext->file_descriptor_table->pop_back();
90 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
91 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
93 // if file does not exist create an empty file
95 auto sz = content->find(path_);
96 if (sz != content->end()) {
98 XBT_DEBUG("\tOpen file '%s', size %llu", path_.c_str(), size_);
101 content->insert({path_, size_});
102 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
108 File::~File() = default;
110 File* File::open(const std::string& fullpath, void* userdata)
112 return new File(fullpath, userdata);
115 File* File::open(const std::string& fullpath, const_sg_host_t host, void* userdata)
117 return new File(fullpath, host, userdata);
122 std::vector<int>* desc_table =
123 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table.get();
124 kernel::actor::simcall_answered([this, desc_table] { desc_table->push_back(this->desc_id); });
128 void File::dump() const
130 XBT_INFO("File Descriptor information:\n"
131 "\t\tFull path: '%s'\n"
133 "\t\tMount point: '%s'\n"
134 "\t\tDisk Id: '%s'\n"
135 "\t\tHost Id: '%s'\n"
136 "\t\tFile Descriptor Id: %d",
137 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
141 sg_size_t File::read(sg_size_t size)
143 if (size_ == 0) /* Nothing to read, return */
145 Host* host = nullptr;
146 // if the current position is close to the end of the file, we may not be able to read the requested size
147 sg_size_t to_read = std::min(size, size_ - current_position_);
148 sg_size_t read_size = 0;
150 /* Find the host where the file is physically located and read it */
151 host = local_disk_->get_host();
152 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
153 read_size = local_disk_->read(to_read);
155 current_position_ += read_size;
157 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
158 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
159 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
160 Comm::sendto(host, Host::current(), read_size);
166 /** @brief Write into a file (local or remote)
167 * @ingroup plugin_filesystem
169 * @param size of the file to write
170 * @param write_inside
171 * @return the number of bytes successfully write or -1 if an error occurred
173 sg_size_t File::write(sg_size_t size, bool write_inside)
175 if (size == 0) /* Nothing to write, return */
178 sg_size_t write_size = 0;
179 /* Find the host where the file is physically located (remote or local)*/
180 if (Host* host = local_disk_->get_host(); host && host->get_name() != Host::current()->get_name()) {
181 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
182 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
183 Comm::sendto(Host::current(), host, size);
185 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
186 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
187 // If the disk is full before even starting to write
188 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
190 if (not write_inside)
191 /* Subtract the part of the file that might disappear from the used sized on the storage element */
192 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
193 write_size = local_disk_->write(size);
194 update_position(current_position_ + write_size);
199 sg_size_t File::size() const
204 void File::seek(sg_offset_t offset)
206 current_position_ = offset;
209 void File::seek(sg_offset_t offset, int origin)
213 update_position(offset);
216 update_position(current_position_ + offset);
219 update_position(size_ + offset);
226 void File::update_position(sg_offset_t position)
228 xbt_assert(position >= 0, "Error in seek, cannot seek before file %s", get_path());
229 current_position_ = position;
230 if(current_position_>size_){
231 XBT_DEBUG("Updating size of file %s from %llu to %lld", path_.c_str(), size_, position);
232 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(current_position_-size_);
233 size_ = current_position_;
235 kernel::actor::simcall_answered([this] {
236 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
237 content->erase(path_);
238 content->insert({path_, size_});
243 sg_size_t File::tell() const
245 return current_position_;
248 void File::move(const std::string& fullpath) const
250 /* Check if the new full path is on the same mount point */
251 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
252 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
253 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
255 auto sz = content->find(path_);
256 if (sz != content->end()) { // src file exists
257 sg_size_t new_size = sz->second;
258 content->erase(path_);
259 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
260 content->insert({path.c_str(), new_size});
261 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
263 XBT_WARN("File %s doesn't exist", path_.c_str());
267 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
271 int File::unlink() const
273 /* Check if the file is on local storage */
274 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
275 const char* name = local_disk_->get_cname();
277 if (not content || content->find(path_) == content->end()) {
278 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
281 XBT_DEBUG("UNLINK %s of size %llu on disk '%s'", path_.c_str(), size_, name);
283 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_);
284 // Remove the file from storage
285 content->erase(path_);
291 int File::remote_copy(sg_host_t host, const std::string& fullpath)
293 /* Find the host where the file is physically located and read it */
294 Host* src_host = nullptr;
295 sg_size_t read_size = 0;
297 Host* dst_host = host;
298 size_t longest_prefix_length = 0;
302 src_host = local_disk_->get_host();
303 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
304 read_size = local_disk_->read(size_);
305 current_position_ += read_size;
307 const Disk* dst_disk = nullptr;
309 for (auto const& disk : host->get_disks()) {
310 std::string current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
311 std::string mount_point = fullpath.substr(0, current_mount.length());
312 if (mount_point == current_mount && current_mount.length() > longest_prefix_length) {
313 /* The current mount name is found in the full path and is bigger than the previous*/
314 longest_prefix_length = current_mount.length();
319 if (dst_disk == nullptr) {
320 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath.c_str(), host->get_cname());
325 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
326 dst_host->get_cname());
327 Comm::sendto(src_host, dst_host, read_size);
330 /* Create file on remote host, write it and close it */
331 auto* fd = File::open(fullpath, dst_host, nullptr);
332 fd->write(read_size);
337 int File::remote_move(sg_host_t host, const std::string& fullpath)
339 int res = remote_copy(host, fullpath);
344 FileSystemDiskExt::FileSystemDiskExt(const Disk* ptr)
346 if (const char* size_str = ptr->get_property("size")) {
347 std::string dummyfile;
348 size_ = xbt_parse_get_size(dummyfile, -1, size_str, "disk size " + ptr->get_name());
351 if (const char* current_mount_str = ptr->get_property("mount"))
352 mount_point_ = current_mount_str;
356 if (const char* content_str = ptr->get_property("content"))
357 content_.reset(parse_content(content_str));
360 std::map<std::string, sg_size_t, std::less<>>* FileSystemDiskExt::parse_content(const std::string& filename)
362 if (filename.empty())
365 auto* parse_content = new std::map<std::string, sg_size_t, std::less<>>();
367 auto fs = std::unique_ptr<std::ifstream>(simgrid::xbt::path_ifsopen(filename));
368 xbt_assert(not fs->fail(), "Cannot open file '%s' (path=%s)", filename.c_str(),
369 simgrid::xbt::path_to_string().c_str());
372 std::vector<std::string> tokens;
374 std::getline(*fs, line);
376 if (line.length() > 0) {
377 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
378 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
379 sg_size_t size = std::stoull(tokens.at(1));
382 parse_content->insert({tokens.front(), size});
384 } while (not fs->eof());
385 return parse_content;
388 void FileSystemDiskExt::add_remote_mount(Host* host, const std::string& mount_point)
390 remote_mount_points_.try_emplace(host, mount_point);
393 void FileSystemDiskExt::decr_used_size(sg_size_t size)
395 simgrid::kernel::actor::simcall_answered([this, size] { used_size_ -= size; });
398 void FileSystemDiskExt::incr_used_size(sg_size_t size)
400 simgrid::kernel::actor::simcall_answered([this, size] { used_size_ += size; });
405 using simgrid::s4u::FileDescriptorHostExt;
406 using simgrid::s4u::FileSystemDiskExt;
408 static void on_disk_creation(simgrid::s4u::Disk& d)
410 d.extension_set(new FileSystemDiskExt(&d));
413 static void on_host_creation(simgrid::s4u::Host& host)
415 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
418 static void on_platform_created()
420 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
421 const char* remote_disk_str = host->get_property("remote_disk");
422 if (not remote_disk_str)
424 std::vector<std::string> tokens;
425 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
426 std::string mount_point = tokens[0];
427 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
428 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
430 const simgrid::s4u::Disk* disk = nullptr;
431 for (auto const& d : remote_host->get_disks())
432 if (d->get_name() == tokens[1]) {
437 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
438 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
439 host->add_disk(disk);
441 XBT_DEBUG("Host '%s' wants to mount a remote disk: %s of %s mounted on %s", host->get_cname(), disk->get_cname(),
442 remote_host->get_cname(), mount_point.c_str());
443 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
447 static void on_simulation_end()
449 XBT_DEBUG("Simulation is over, time to unregister remote disks if any");
450 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
451 const char* remote_disk_str = host->get_property("remote_disk");
452 if (remote_disk_str) {
453 std::vector<std::string> tokens;
454 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
455 XBT_DEBUG("Host '%s' wants to unmount a remote disk: %s of %s mounted on %s", host->get_cname(),
456 tokens[1].c_str(), tokens[2].c_str(), tokens[0].c_str());
457 host->remove_disk(tokens[1]);
458 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
463 /* **************************** Public interface *************************** */
464 /** @brief Initialize the file system plugin.
465 @ingroup plugin_filesystem
468 See the examples in :ref:`s4u_ex_disk_io`.
471 void sg_storage_file_system_init()
473 sg_storage_max_file_descriptors = 1024;
474 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
475 "Maximum number of concurrently opened files per host. Default is 1024");
477 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
478 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
479 simgrid::s4u::Disk::on_creation_cb(&on_disk_creation);
482 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
483 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
484 simgrid::s4u::Host::on_creation_cb(&on_host_creation);
486 simgrid::s4u::Engine::on_platform_created_cb(&on_platform_created);
487 simgrid::s4u::Engine::on_simulation_end_cb(&on_simulation_end);
490 sg_file_t sg_file_open(const char* fullpath, void* data)
492 return simgrid::s4u::File::open(fullpath, data);
495 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
497 return fd->read(size);
500 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
502 return fd->write(size);
505 void sg_file_close(sg_file_t fd)
510 /** Retrieves the path to the file
511 * @ingroup plugin_filesystem
513 const char* sg_file_get_name(const_sg_file_t fd)
515 xbt_assert((fd != nullptr), "Invalid file descriptor");
516 return fd->get_path();
519 /** Retrieves the size of the file
520 * @ingroup plugin_filesystem
522 sg_size_t sg_file_get_size(const_sg_file_t fd)
527 void sg_file_dump(const_sg_file_t fd)
532 /** Retrieves the user data associated with the file
533 * @ingroup plugin_filesystem
535 void* sg_file_get_data(const_sg_file_t fd)
537 return fd->get_data<void>();
540 /** Changes the user data associated with the file
541 * @ingroup plugin_filesystem
543 void sg_file_set_data(sg_file_t fd, void* data)
549 * @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).
550 * @ingroup plugin_filesystem
552 * @param fd : file object that identifies the stream
553 * @param offset : number of bytes to offset from origin
554 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
555 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
556 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
558 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
560 fd->seek(offset, origin);
563 sg_size_t sg_file_tell(const_sg_file_t fd)
568 void sg_file_move(const_sg_file_t fd, const char* fullpath)
573 void sg_file_unlink(sg_file_t fd)
580 * @brief Copy a file to another location on a remote host.
581 * @ingroup plugin_filesystem
583 * @param file : the file to move
584 * @param host : the remote host where the file has to be copied
585 * @param fullpath : the complete path destination on the remote host
586 * @return If successful, the function returns 0. Otherwise, it returns -1.
588 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
590 return file->remote_copy(host, fullpath);
594 * @brief Move a file to another location on a remote host.
595 * @ingroup plugin_filesystem
597 * @param file : the file to move
598 * @param host : the remote host where the file has to be moved
599 * @param fullpath : the complete path destination on the remote host
600 * @return If successful, the function returns 0. Otherwise, it returns -1.
602 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
604 return file->remote_move(host, fullpath);
607 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
609 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
612 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
614 return d->extension<FileSystemDiskExt>()->get_used_size();
617 sg_size_t sg_disk_get_size(const_sg_disk_t d)
619 return d->extension<FileSystemDiskExt>()->get_size();
622 const char* sg_disk_get_mount_point(const_sg_disk_t d)
624 return d->extension<FileSystemDiskExt>()->get_mount_point();