1 /* Copyright (c) 2015-2021. 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/Actor.hpp"
8 #include "simgrid/s4u/Engine.hpp"
9 #include "src/surf/HostImpl.hpp"
10 #include "src/surf/xml/platf_private.hpp"
11 #include "xbt/config.hpp"
12 #include "xbt/parse_units.hpp"
15 #include <boost/algorithm/string.hpp>
16 #include <boost/algorithm/string/join.hpp>
17 #include <boost/algorithm/string/split.hpp>
22 XBT_LOG_NEW_DEFAULT_SUBCATEGORY(s4u_file, s4u, "S4U files");
23 int sg_storage_max_file_descriptors = 1024;
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;
39 Disk* File::find_local_disk_on(const Host* host)
42 size_t longest_prefix_length = 0;
43 for (auto const& disk : host->get_disks()) {
44 std::string current_mount;
45 if (disk->get_host() != host)
46 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point(disk->get_host());
48 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
49 mount_point_ = fullpath_.substr(0, current_mount.length());
50 if (mount_point_ == current_mount && current_mount.length() > longest_prefix_length) {
51 /* The current mount name is found in the full path and is bigger than the previous*/
52 longest_prefix_length = current_mount.length();
55 if (longest_prefix_length > 0) { /* Mount point found, split fullpath_ into mount_name and path+filename*/
56 mount_point_ = fullpath_.substr(0, longest_prefix_length);
57 if (mount_point_ == std::string("/"))
60 path_ = fullpath_.substr(longest_prefix_length, fullpath_.length());
61 XBT_DEBUG("%s + %s", mount_point_.c_str(), path_.c_str());
63 xbt_die("Can't find mount point for '%s' on '%s'", fullpath_.c_str(), host->get_cname());
68 File::File(const std::string& fullpath, void* userdata) : File(fullpath, Host::current(), userdata) {}
70 File::File(const std::string& fullpath, sg_host_t host, void* userdata) : fullpath_(fullpath)
72 kernel::actor::simcall([this, &host, userdata] {
73 this->set_data(userdata);
74 // this cannot fail because we get a xbt_die if the mountpoint does not exist
75 local_disk_ = find_local_disk_on(host);
77 // assign a file descriptor id to the newly opened File
78 auto* ext = host->extension<simgrid::s4u::FileDescriptorHostExt>();
79 if (ext->file_descriptor_table == nullptr) {
80 ext->file_descriptor_table = std::make_unique<std::vector<int>>(sg_storage_max_file_descriptors);
81 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
83 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
84 desc_id = ext->file_descriptor_table->back();
85 ext->file_descriptor_table->pop_back();
87 XBT_DEBUG("\tOpen file '%s'", path_.c_str());
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()) {
98 content->insert({path_, size_});
99 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
107 std::vector<int>* desc_table =
108 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table.get();
109 kernel::actor::simcall([this, desc_table] { desc_table->push_back(this->desc_id); });
112 void File::dump() const
114 XBT_INFO("File Descriptor information:\n"
115 "\t\tFull path: '%s'\n"
117 "\t\tMount point: '%s'\n"
118 "\t\tDisk Id: '%s'\n"
119 "\t\tHost Id: '%s'\n"
120 "\t\tFile Descriptor Id: %d",
121 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
125 sg_size_t File::read(sg_size_t size)
127 if (size_ == 0) /* Nothing to read, return */
129 Host* host = nullptr;
130 // if the current position is close to the end of the file, we may not be able to read the requested size
131 sg_size_t to_read = std::min(size, size_ - current_position_);
132 sg_size_t read_size = 0;
134 /* Find the host where the file is physically located and read it */
135 host = local_disk_->get_host();
136 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
137 read_size = local_disk_->read(to_read);
139 current_position_ += read_size;
141 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
142 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
143 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
144 host->sendto(Host::current(), read_size);
150 /** @brief Write into a file (local or remote)
151 * @ingroup plugin_filesystem
153 * @param size of the file to write
154 * @return the number of bytes successfully write or -1 if an error occurred
156 sg_size_t File::write(sg_size_t size, bool write_inside)
158 if (size == 0) /* Nothing to write, return */
161 sg_size_t write_size = 0;
162 /* Find the host where the file is physically located (remote or local)*/
163 Host* host = local_disk_->get_host();
165 if (host && host->get_name() != Host::current()->get_name()) {
166 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
167 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
168 Host::current()->sendto(host, size);
170 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
171 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
172 // If the disk is full before even starting to write
173 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
175 if (not write_inside) {
176 /* Subtract the part of the file that might disappear from the used sized on the storage element */
177 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
178 write_size = local_disk_->write(size);
179 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(write_size);
180 current_position_ += write_size;
181 size_ = current_position_;
183 write_size = local_disk_->write(size);
184 current_position_ += write_size;
185 if (current_position_ > size_)
186 size_ = current_position_;
188 kernel::actor::simcall([this] {
189 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
191 content->erase(path_);
192 content->insert({path_, size_});
198 sg_size_t File::size() const
203 void File::seek(sg_offset_t offset)
205 current_position_ = offset;
208 void File::seek(sg_offset_t offset, int origin)
212 current_position_ = offset;
215 current_position_ += offset;
218 current_position_ = size_ + offset;
225 sg_size_t File::tell() const
227 return current_position_;
230 void File::move(const std::string& fullpath) const
232 /* Check if the new full path is on the same mount point */
233 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
234 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
235 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
237 auto sz = content->find(path_);
238 if (sz != content->end()) { // src file exists
239 sg_size_t new_size = sz->second;
240 content->erase(path_);
241 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
242 content->insert({path.c_str(), new_size});
243 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
245 XBT_WARN("File %s doesn't exist", path_.c_str());
249 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
253 int File::unlink() const
255 /* Check if the file is on local storage */
256 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
257 const char* name = local_disk_->get_cname();
259 if (not content || content->find(path_) == content->end()) {
260 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
263 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), name);
265 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_);
267 // Remove the file from storage
268 content->erase(path_);
274 int File::remote_copy(sg_host_t host, const std::string& fullpath)
276 /* Find the host where the file is physically located and read it */
277 Host* src_host = nullptr;
278 sg_size_t read_size = 0;
280 Host* dst_host = host;
281 size_t longest_prefix_length = 0;
285 src_host = local_disk_->get_host();
286 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
287 read_size = local_disk_->read(size_);
288 current_position_ += read_size;
290 const Disk* dst_disk = nullptr;
292 for (auto const& disk : host->get_disks()) {
293 std::string current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
294 std::string mount_point = std::string(fullpath).substr(0, current_mount.length());
295 if (mount_point == current_mount && current_mount.length() > longest_prefix_length) {
296 /* The current mount name is found in the full path and is bigger than the previous*/
297 longest_prefix_length = current_mount.length();
302 if (dst_disk == nullptr) {
303 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath.c_str(), host->get_cname());
308 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
309 dst_host->get_cname());
310 src_host->sendto(dst_host, read_size);
313 /* Create file on remote host, write it and close it */
314 File fd(fullpath, dst_host, nullptr);
319 int File::remote_move(sg_host_t host, const std::string& fullpath)
321 int res = remote_copy(host, fullpath);
326 FileSystemDiskExt::FileSystemDiskExt(const Disk* ptr)
328 const char* size_str = ptr->get_property("size");
329 std::string dummyfile;
331 size_ = surf_parse_get_size(dummyfile, -1, size_str, "disk size", ptr->get_name());
333 const char* current_mount_str = ptr->get_property("mount");
334 if (current_mount_str)
335 mount_point_ = std::string(current_mount_str);
337 mount_point_ = std::string("/");
339 const char* content_str = ptr->get_property("content");
341 content_.reset(parse_content(content_str));
344 std::map<std::string, sg_size_t, std::less<>>* FileSystemDiskExt::parse_content(const std::string& filename)
346 if (filename.empty())
349 auto* parse_content = new std::map<std::string, sg_size_t, std::less<>>();
351 auto fs = std::unique_ptr<std::ifstream>(surf_ifsopen(filename));
352 xbt_assert(not fs->fail(), "Cannot open file '%s' (path=%s)", filename.c_str(),
353 (boost::join(surf_path, ":")).c_str());
356 std::vector<std::string> tokens;
358 std::getline(*fs, line);
360 if (line.length() > 0) {
361 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
362 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
363 sg_size_t size = std::stoull(tokens.at(1));
366 parse_content->insert({tokens.front(), size});
368 } while (not fs->eof());
369 return parse_content;
372 void FileSystemDiskExt::decr_used_size(sg_size_t size)
374 simgrid::kernel::actor::simcall([this, size] { used_size_ -= size; });
377 void FileSystemDiskExt::incr_used_size(sg_size_t size)
379 simgrid::kernel::actor::simcall([this, size] { used_size_ += size; });
384 using simgrid::s4u::FileDescriptorHostExt;
385 using simgrid::s4u::FileSystemDiskExt;
387 static void on_disk_creation(simgrid::s4u::Disk& d)
389 d.extension_set(new FileSystemDiskExt(&d));
392 static void on_host_creation(simgrid::s4u::Host& host)
394 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
397 static void on_platform_created()
399 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
400 const char* remote_disk_str = host->get_property("remote_disk");
401 if (remote_disk_str) {
402 std::vector<std::string> tokens;
403 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
404 std::string mount_point = tokens[0];
405 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
406 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
408 const simgrid::s4u::Disk* disk = nullptr;
409 for (auto const& d : remote_host->get_disks())
410 if (d->get_name() == tokens[1]) {
415 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
416 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
417 host->add_disk(disk);
419 XBT_DEBUG("Host '%s' wants to mount a remote disk: %s of %s mounted on %s", host->get_cname(), disk->get_cname(),
420 remote_host->get_cname(), mount_point.c_str());
421 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
426 static void on_simulation_end()
428 XBT_DEBUG("Simulation is over, time to unregister remote disks if any");
429 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
430 const char* remote_disk_str = host->get_property("remote_disk");
431 if (remote_disk_str) {
432 std::vector<std::string> tokens;
433 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
434 XBT_DEBUG("Host '%s' wants to unmount a remote disk: %s of %s mounted on %s", host->get_cname(),
435 tokens[1].c_str(), tokens[2].c_str(), tokens[0].c_str());
436 host->remove_disk(tokens[1]);
437 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
442 /* **************************** Public interface *************************** */
443 /** @brief Initialize the file system plugin.
444 @ingroup plugin_filesystem
447 See the examples in :ref:`s4u_ex_disk_io`.
450 void sg_storage_file_system_init()
452 sg_storage_max_file_descriptors = 1024;
453 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
454 "Maximum number of concurrently opened files per host. Default is 1024");
456 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
457 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
458 simgrid::s4u::Disk::on_creation.connect(&on_disk_creation);
461 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
462 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
463 simgrid::s4u::Host::on_creation.connect(&on_host_creation);
465 simgrid::s4u::Engine::on_platform_created.connect(&on_platform_created);
466 simgrid::s4u::Engine::on_simulation_end.connect(&on_simulation_end);
469 sg_file_t sg_file_open(const char* fullpath, void* data)
471 return new simgrid::s4u::File(fullpath, data);
474 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
476 return fd->read(size);
479 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
481 return fd->write(size);
484 void sg_file_close(const_sg_file_t fd)
489 /** Retrieves the path to the file
490 * @ingroup plugin_filesystem
492 const char* sg_file_get_name(const_sg_file_t fd)
494 xbt_assert((fd != nullptr), "Invalid file descriptor");
495 return fd->get_path();
498 /** Retrieves the size of the file
499 * @ingroup plugin_filesystem
501 sg_size_t sg_file_get_size(const_sg_file_t fd)
506 void sg_file_dump(const_sg_file_t fd)
511 /** Retrieves the user data associated with the file
512 * @ingroup plugin_filesystem
514 void* sg_file_get_data(const_sg_file_t fd)
516 return fd->get_data();
519 /** Changes the user data associated with the file
520 * @ingroup plugin_filesystem
522 void sg_file_set_data(sg_file_t fd, void* data)
528 * @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).
529 * @ingroup plugin_filesystem
531 * @param fd : file object that identifies the stream
532 * @param offset : number of bytes to offset from origin
533 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
534 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
535 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
537 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
539 fd->seek(offset, origin);
542 sg_size_t sg_file_tell(const_sg_file_t fd)
547 void sg_file_move(const_sg_file_t fd, const char* fullpath)
552 void sg_file_unlink(sg_file_t fd)
559 * @brief Copy a file to another location on a remote host.
560 * @ingroup plugin_filesystem
562 * @param file : the file to move
563 * @param host : the remote host where the file has to be copied
564 * @param fullpath : the complete path destination on the remote host
565 * @return If successful, the function returns 0. Otherwise, it returns -1.
567 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
569 return file->remote_copy(host, fullpath);
573 * @brief Move a file to another location on a remote host.
574 * @ingroup plugin_filesystem
576 * @param file : the file to move
577 * @param host : the remote host where the file has to be moved
578 * @param fullpath : the complete path destination on the remote host
579 * @return If successful, the function returns 0. Otherwise, it returns -1.
581 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
583 return file->remote_move(host, fullpath);
586 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
588 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
591 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
593 return d->extension<FileSystemDiskExt>()->get_used_size();
596 sg_size_t sg_disk_get_size(const_sg_disk_t d)
598 return d->extension<FileSystemDiskExt>()->get_size();
601 const char* sg_disk_get_mount_point(const_sg_disk_t d)
603 return d->extension<FileSystemDiskExt>()->get_mount_point();