1 /* Copyright (c) 2015-2022. 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>
15 #include <xbt/parse_units.hpp>
17 #include "src/surf/surf_interface.hpp"
19 #include <boost/algorithm/string.hpp>
20 #include <boost/algorithm/string/split.hpp>
24 XBT_LOG_NEW_DEFAULT_SUBCATEGORY(s4u_file, s4u, "S4U files");
25 int sg_storage_max_file_descriptors = 1024;
27 /** @defgroup plugin_filesystem Plugin FileSystem
29 * This adds the notion of Files on top of the storage notion that provided by the core of SimGrid.
30 * Activate this plugin at will.
35 template class xbt::Extendable<s4u::File>;
38 simgrid::xbt::Extension<Disk, FileSystemDiskExt> FileSystemDiskExt::EXTENSION_ID;
39 simgrid::xbt::Extension<Host, FileDescriptorHostExt> FileDescriptorHostExt::EXTENSION_ID;
41 const Disk* File::find_local_disk_on(const Host* host)
43 const Disk* d = nullptr;
44 size_t longest_prefix_length = 0;
45 for (auto const& disk : host->get_disks()) {
46 std::string current_mount;
47 if (disk->get_host() != host)
48 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point(disk->get_host());
50 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
51 mount_point_ = fullpath_.substr(0, current_mount.length());
52 if (mount_point_ == current_mount && current_mount.length() > longest_prefix_length) {
53 /* The current mount name is found in the full path and is bigger than the previous*/
54 longest_prefix_length = current_mount.length();
57 xbt_assert(longest_prefix_length > 0, "Can't find mount point for '%s' on '%s'", fullpath_.c_str(),
59 /* Mount point found, split fullpath_ into mount_name and path+filename*/
60 mount_point_ = fullpath_.substr(0, longest_prefix_length);
61 if (mount_point_ == std::string("/"))
64 path_ = fullpath_.substr(longest_prefix_length, fullpath_.length());
65 XBT_DEBUG("%s + %s", mount_point_.c_str(), path_.c_str());
70 File::File(const std::string& fullpath, void* userdata) : File(fullpath, Host::current(), userdata) {}
72 File::File(const std::string& fullpath, const_sg_host_t host, void* userdata) : fullpath_(fullpath)
74 kernel::actor::simcall_answered([this, &host, userdata] {
75 this->set_data(userdata);
76 // this cannot fail because we get a xbt_die if the mountpoint does not exist
77 local_disk_ = find_local_disk_on(host);
79 // assign a file descriptor id to the newly opened File
80 auto* ext = host->extension<simgrid::s4u::FileDescriptorHostExt>();
81 if (ext->file_descriptor_table == nullptr) {
82 ext->file_descriptor_table = std::make_unique<std::vector<int>>(sg_storage_max_file_descriptors);
83 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
85 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
86 desc_id = ext->file_descriptor_table->back();
87 ext->file_descriptor_table->pop_back();
89 XBT_DEBUG("\tOpen file '%s'", path_.c_str());
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()) {
100 content->insert({path_, size_});
101 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
109 std::vector<int>* desc_table =
110 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table.get();
111 kernel::actor::simcall_answered([this, desc_table] { desc_table->push_back(this->desc_id); });
114 File* File::open(const std::string& fullpath, void* userdata)
116 return new File(fullpath, userdata);
119 File* File::open(const std::string& fullpath, const_sg_host_t host, void* userdata)
121 return new File(fullpath, host, userdata);
124 void File::dump() const
126 XBT_INFO("File Descriptor information:\n"
127 "\t\tFull path: '%s'\n"
129 "\t\tMount point: '%s'\n"
130 "\t\tDisk Id: '%s'\n"
131 "\t\tHost Id: '%s'\n"
132 "\t\tFile Descriptor Id: %d",
133 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
137 sg_size_t File::read(sg_size_t size)
139 if (size_ == 0) /* Nothing to read, return */
141 Host* host = nullptr;
142 // if the current position is close to the end of the file, we may not be able to read the requested size
143 sg_size_t to_read = std::min(size, size_ - current_position_);
144 sg_size_t read_size = 0;
146 /* Find the host where the file is physically located and read it */
147 host = local_disk_->get_host();
148 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
149 read_size = local_disk_->read(to_read);
151 current_position_ += read_size;
153 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
154 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
155 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
156 Comm::sendto(host, Host::current(), read_size);
162 /** @brief Write into a file (local or remote)
163 * @ingroup plugin_filesystem
165 * @param size of the file to write
166 * @param write_inside
167 * @return the number of bytes successfully write or -1 if an error occurred
169 sg_size_t File::write(sg_size_t size, bool write_inside)
171 if (size == 0) /* Nothing to write, return */
174 sg_size_t write_size = 0;
175 /* Find the host where the file is physically located (remote or local)*/
176 Host* host = local_disk_->get_host();
178 if (host && host->get_name() != Host::current()->get_name()) {
179 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
180 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
181 Comm::sendto(Host::current(), host, size);
183 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
184 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
185 // If the disk is full before even starting to write
186 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
188 if (not write_inside) {
189 /* Subtract the part of the file that might disappear from the used sized on the storage element */
190 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
191 write_size = local_disk_->write(size);
192 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(write_size);
193 current_position_ += write_size;
194 size_ = current_position_;
196 write_size = local_disk_->write(size);
197 current_position_ += write_size;
198 if (current_position_ > size_)
199 size_ = current_position_;
201 kernel::actor::simcall_answered([this] {
202 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
204 content->erase(path_);
205 content->insert({path_, size_});
211 sg_size_t File::size() const
216 void File::seek(sg_offset_t offset)
218 current_position_ = offset;
221 void File::seek(sg_offset_t offset, int origin)
225 current_position_ = offset;
228 current_position_ += offset;
231 current_position_ = size_ + offset;
238 sg_size_t File::tell() const
240 return current_position_;
243 void File::move(const std::string& fullpath) const
245 /* Check if the new full path is on the same mount point */
246 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
247 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
248 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
250 auto sz = content->find(path_);
251 if (sz != content->end()) { // src file exists
252 sg_size_t new_size = sz->second;
253 content->erase(path_);
254 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
255 content->insert({path.c_str(), new_size});
256 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
258 XBT_WARN("File %s doesn't exist", path_.c_str());
262 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
266 int File::unlink() const
268 /* Check if the file is on local storage */
269 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
270 const char* name = local_disk_->get_cname();
272 if (not content || content->find(path_) == content->end()) {
273 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
276 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), name);
278 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_);
280 // Remove the file from storage
281 content->erase(path_);
287 int File::remote_copy(sg_host_t host, const std::string& fullpath)
289 /* Find the host where the file is physically located and read it */
290 Host* src_host = nullptr;
291 sg_size_t read_size = 0;
293 Host* dst_host = host;
294 size_t longest_prefix_length = 0;
298 src_host = local_disk_->get_host();
299 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
300 read_size = local_disk_->read(size_);
301 current_position_ += read_size;
303 const Disk* dst_disk = nullptr;
305 for (auto const& disk : host->get_disks()) {
306 std::string current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
307 std::string mount_point = std::string(fullpath).substr(0, current_mount.length());
308 if (mount_point == current_mount && current_mount.length() > longest_prefix_length) {
309 /* The current mount name is found in the full path and is bigger than the previous*/
310 longest_prefix_length = current_mount.length();
315 if (dst_disk == nullptr) {
316 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath.c_str(), host->get_cname());
321 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
322 dst_host->get_cname());
323 Comm::sendto(src_host, dst_host, read_size);
326 /* Create file on remote host, write it and close it */
327 File fd(fullpath, dst_host, nullptr);
332 int File::remote_move(sg_host_t host, const std::string& fullpath)
334 int res = remote_copy(host, fullpath);
339 FileSystemDiskExt::FileSystemDiskExt(const Disk* ptr)
341 const char* size_str = ptr->get_property("size");
342 std::string dummyfile;
344 size_ = xbt_parse_get_size(dummyfile, -1, size_str, "disk size " + ptr->get_name());
346 const char* current_mount_str = ptr->get_property("mount");
347 if (current_mount_str)
348 mount_point_ = std::string(current_mount_str);
350 mount_point_ = std::string("/");
352 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>(surf_ifsopen(filename));
365 xbt_assert(not fs->fail(), "Cannot open file '%s' (path=%s)", filename.c_str(),
366 (boost::join(surf_path, ":")).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::decr_used_size(sg_size_t size)
387 simgrid::kernel::actor::simcall_answered([this, size] { used_size_ -= size; });
390 void FileSystemDiskExt::incr_used_size(sg_size_t size)
392 simgrid::kernel::actor::simcall_answered([this, size] { used_size_ += size; });
397 using simgrid::s4u::FileDescriptorHostExt;
398 using simgrid::s4u::FileSystemDiskExt;
400 static void on_disk_creation(simgrid::s4u::Disk& d)
402 d.extension_set(new FileSystemDiskExt(&d));
405 static void on_host_creation(simgrid::s4u::Host& host)
407 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
410 static void on_platform_created()
412 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
413 const char* remote_disk_str = host->get_property("remote_disk");
414 if (remote_disk_str) {
415 std::vector<std::string> tokens;
416 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
417 std::string mount_point = tokens[0];
418 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
419 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
421 const simgrid::s4u::Disk* disk = nullptr;
422 for (auto const& d : remote_host->get_disks())
423 if (d->get_name() == tokens[1]) {
428 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
429 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
430 host->add_disk(disk);
432 XBT_DEBUG("Host '%s' wants to mount a remote disk: %s of %s mounted on %s", host->get_cname(), disk->get_cname(),
433 remote_host->get_cname(), mount_point.c_str());
434 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
439 static void on_simulation_end()
441 XBT_DEBUG("Simulation is over, time to unregister remote disks if any");
442 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
443 const char* remote_disk_str = host->get_property("remote_disk");
444 if (remote_disk_str) {
445 std::vector<std::string> tokens;
446 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
447 XBT_DEBUG("Host '%s' wants to unmount a remote disk: %s of %s mounted on %s", host->get_cname(),
448 tokens[1].c_str(), tokens[2].c_str(), tokens[0].c_str());
449 host->remove_disk(tokens[1]);
450 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
455 /* **************************** Public interface *************************** */
456 /** @brief Initialize the file system plugin.
457 @ingroup plugin_filesystem
460 See the examples in :ref:`s4u_ex_disk_io`.
463 void sg_storage_file_system_init()
465 sg_storage_max_file_descriptors = 1024;
466 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
467 "Maximum number of concurrently opened files per host. Default is 1024");
469 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
470 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
471 simgrid::s4u::Disk::on_creation_cb(&on_disk_creation);
474 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
475 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
476 simgrid::s4u::Host::on_creation_cb(&on_host_creation);
478 simgrid::s4u::Engine::on_platform_created_cb(&on_platform_created);
479 simgrid::s4u::Engine::on_simulation_end_cb(&on_simulation_end);
482 sg_file_t sg_file_open(const char* fullpath, void* data)
484 return simgrid::s4u::File::open(fullpath, data);
487 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
489 return fd->read(size);
492 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
494 return fd->write(size);
497 void sg_file_close(sg_file_t fd)
502 /** Retrieves the path to the file
503 * @ingroup plugin_filesystem
505 const char* sg_file_get_name(const_sg_file_t fd)
507 xbt_assert((fd != nullptr), "Invalid file descriptor");
508 return fd->get_path();
511 /** Retrieves the size of the file
512 * @ingroup plugin_filesystem
514 sg_size_t sg_file_get_size(const_sg_file_t fd)
519 void sg_file_dump(const_sg_file_t fd)
524 /** Retrieves the user data associated with the file
525 * @ingroup plugin_filesystem
527 void* sg_file_get_data(const_sg_file_t fd)
529 return fd->get_data<void>();
532 /** Changes the user data associated with the file
533 * @ingroup plugin_filesystem
535 void sg_file_set_data(sg_file_t fd, void* data)
541 * @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).
542 * @ingroup plugin_filesystem
544 * @param fd : file object that identifies the stream
545 * @param offset : number of bytes to offset from origin
546 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
547 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
548 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
550 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
552 fd->seek(offset, origin);
555 sg_size_t sg_file_tell(const_sg_file_t fd)
560 void sg_file_move(const_sg_file_t fd, const char* fullpath)
565 void sg_file_unlink(sg_file_t fd)
572 * @brief Copy a file to another location on a remote host.
573 * @ingroup plugin_filesystem
575 * @param file : the file to move
576 * @param host : the remote host where the file has to be copied
577 * @param fullpath : the complete path destination on the remote host
578 * @return If successful, the function returns 0. Otherwise, it returns -1.
580 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
582 return file->remote_copy(host, fullpath);
586 * @brief Move a file to another location on a remote host.
587 * @ingroup plugin_filesystem
589 * @param file : the file to move
590 * @param host : the remote host where the file has to be moved
591 * @param fullpath : the complete path destination on the remote host
592 * @return If successful, the function returns 0. Otherwise, it returns -1.
594 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
596 return file->remote_move(host, fullpath);
599 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
601 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
604 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
606 return d->extension<FileSystemDiskExt>()->get_used_size();
609 sg_size_t sg_disk_get_size(const_sg_disk_t d)
611 return d->extension<FileSystemDiskExt>()->get_size();
614 const char* sg_disk_get_mount_point(const_sg_disk_t d)
616 return d->extension<FileSystemDiskExt>()->get_mount_point();