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/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([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([this, desc_table] { desc_table->push_back(this->desc_id); });
114 void File::dump() const
116 XBT_INFO("File Descriptor information:\n"
117 "\t\tFull path: '%s'\n"
119 "\t\tMount point: '%s'\n"
120 "\t\tDisk Id: '%s'\n"
121 "\t\tHost Id: '%s'\n"
122 "\t\tFile Descriptor Id: %d",
123 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
127 sg_size_t File::read(sg_size_t size)
129 if (size_ == 0) /* Nothing to read, return */
131 Host* host = nullptr;
132 // if the current position is close to the end of the file, we may not be able to read the requested size
133 sg_size_t to_read = std::min(size, size_ - current_position_);
134 sg_size_t read_size = 0;
136 /* Find the host where the file is physically located and read it */
137 host = local_disk_->get_host();
138 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
139 read_size = local_disk_->read(to_read);
141 current_position_ += read_size;
143 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
144 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
145 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
146 Comm::sendto(host, Host::current(), read_size);
152 /** @brief Write into a file (local or remote)
153 * @ingroup plugin_filesystem
155 * @param size of the file to write
156 * @param write_inside
157 * @return the number of bytes successfully write or -1 if an error occurred
159 sg_size_t File::write(sg_size_t size, bool write_inside)
161 if (size == 0) /* Nothing to write, return */
164 sg_size_t write_size = 0;
165 /* Find the host where the file is physically located (remote or local)*/
166 Host* host = local_disk_->get_host();
168 if (host && host->get_name() != Host::current()->get_name()) {
169 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
170 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
171 Comm::sendto(Host::current(), host, size);
173 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
174 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
175 // If the disk is full before even starting to write
176 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
178 if (not write_inside) {
179 /* Subtract the part of the file that might disappear from the used sized on the storage element */
180 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
181 write_size = local_disk_->write(size);
182 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(write_size);
183 current_position_ += write_size;
184 size_ = current_position_;
186 write_size = local_disk_->write(size);
187 current_position_ += write_size;
188 if (current_position_ > size_)
189 size_ = current_position_;
191 kernel::actor::simcall([this] {
192 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
194 content->erase(path_);
195 content->insert({path_, size_});
201 sg_size_t File::size() const
206 void File::seek(sg_offset_t offset)
208 current_position_ = offset;
211 void File::seek(sg_offset_t offset, int origin)
215 current_position_ = offset;
218 current_position_ += offset;
221 current_position_ = size_ + offset;
228 sg_size_t File::tell() const
230 return current_position_;
233 void File::move(const std::string& fullpath) const
235 /* Check if the new full path is on the same mount point */
236 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
237 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
238 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
240 auto sz = content->find(path_);
241 if (sz != content->end()) { // src file exists
242 sg_size_t new_size = sz->second;
243 content->erase(path_);
244 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
245 content->insert({path.c_str(), new_size});
246 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
248 XBT_WARN("File %s doesn't exist", path_.c_str());
252 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
256 int File::unlink() const
258 /* Check if the file is on local storage */
259 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
260 const char* name = local_disk_->get_cname();
262 if (not content || content->find(path_) == content->end()) {
263 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
266 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), name);
268 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_);
270 // Remove the file from storage
271 content->erase(path_);
277 int File::remote_copy(sg_host_t host, const std::string& fullpath)
279 /* Find the host where the file is physically located and read it */
280 Host* src_host = nullptr;
281 sg_size_t read_size = 0;
283 Host* dst_host = host;
284 size_t longest_prefix_length = 0;
288 src_host = local_disk_->get_host();
289 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
290 read_size = local_disk_->read(size_);
291 current_position_ += read_size;
293 const Disk* dst_disk = nullptr;
295 for (auto const& disk : host->get_disks()) {
296 std::string current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
297 std::string mount_point = std::string(fullpath).substr(0, current_mount.length());
298 if (mount_point == current_mount && current_mount.length() > longest_prefix_length) {
299 /* The current mount name is found in the full path and is bigger than the previous*/
300 longest_prefix_length = current_mount.length();
305 if (dst_disk == nullptr) {
306 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath.c_str(), host->get_cname());
311 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
312 dst_host->get_cname());
313 Comm::sendto(src_host, dst_host, read_size);
316 /* Create file on remote host, write it and close it */
317 File fd(fullpath, dst_host, nullptr);
322 int File::remote_move(sg_host_t host, const std::string& fullpath)
324 int res = remote_copy(host, fullpath);
329 FileSystemDiskExt::FileSystemDiskExt(const Disk* ptr)
331 const char* size_str = ptr->get_property("size");
332 std::string dummyfile;
334 size_ = xbt_parse_get_size(dummyfile, -1, size_str, "disk size " + ptr->get_name());
336 const char* current_mount_str = ptr->get_property("mount");
337 if (current_mount_str)
338 mount_point_ = std::string(current_mount_str);
340 mount_point_ = std::string("/");
342 const char* content_str = ptr->get_property("content");
344 content_.reset(parse_content(content_str));
347 std::map<std::string, sg_size_t, std::less<>>* FileSystemDiskExt::parse_content(const std::string& filename)
349 if (filename.empty())
352 auto* parse_content = new std::map<std::string, sg_size_t, std::less<>>();
354 auto fs = std::unique_ptr<std::ifstream>(surf_ifsopen(filename));
355 xbt_assert(not fs->fail(), "Cannot open file '%s' (path=%s)", filename.c_str(),
356 (boost::join(surf_path, ":")).c_str());
359 std::vector<std::string> tokens;
361 std::getline(*fs, line);
363 if (line.length() > 0) {
364 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
365 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
366 sg_size_t size = std::stoull(tokens.at(1));
369 parse_content->insert({tokens.front(), size});
371 } while (not fs->eof());
372 return parse_content;
375 void FileSystemDiskExt::decr_used_size(sg_size_t size)
377 simgrid::kernel::actor::simcall([this, size] { used_size_ -= size; });
380 void FileSystemDiskExt::incr_used_size(sg_size_t size)
382 simgrid::kernel::actor::simcall([this, size] { used_size_ += size; });
387 using simgrid::s4u::FileDescriptorHostExt;
388 using simgrid::s4u::FileSystemDiskExt;
390 static void on_disk_creation(simgrid::s4u::Disk& d)
392 d.extension_set(new FileSystemDiskExt(&d));
395 static void on_host_creation(simgrid::s4u::Host& host)
397 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
400 static void on_platform_created()
402 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
403 const char* remote_disk_str = host->get_property("remote_disk");
404 if (remote_disk_str) {
405 std::vector<std::string> tokens;
406 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
407 std::string mount_point = tokens[0];
408 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
409 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
411 const simgrid::s4u::Disk* disk = nullptr;
412 for (auto const& d : remote_host->get_disks())
413 if (d->get_name() == tokens[1]) {
418 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
419 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
420 host->add_disk(disk);
422 XBT_DEBUG("Host '%s' wants to mount a remote disk: %s of %s mounted on %s", host->get_cname(), disk->get_cname(),
423 remote_host->get_cname(), mount_point.c_str());
424 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
429 static void on_simulation_end()
431 XBT_DEBUG("Simulation is over, time to unregister remote disks if any");
432 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
433 const char* remote_disk_str = host->get_property("remote_disk");
434 if (remote_disk_str) {
435 std::vector<std::string> tokens;
436 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
437 XBT_DEBUG("Host '%s' wants to unmount a remote disk: %s of %s mounted on %s", host->get_cname(),
438 tokens[1].c_str(), tokens[2].c_str(), tokens[0].c_str());
439 host->remove_disk(tokens[1]);
440 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
445 /* **************************** Public interface *************************** */
446 /** @brief Initialize the file system plugin.
447 @ingroup plugin_filesystem
450 See the examples in :ref:`s4u_ex_disk_io`.
453 void sg_storage_file_system_init()
455 sg_storage_max_file_descriptors = 1024;
456 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
457 "Maximum number of concurrently opened files per host. Default is 1024");
459 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
460 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
461 simgrid::s4u::Disk::on_creation.connect(&on_disk_creation);
464 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
465 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
466 simgrid::s4u::Host::on_creation.connect(&on_host_creation);
468 simgrid::s4u::Engine::on_platform_created.connect(&on_platform_created);
469 simgrid::s4u::Engine::on_simulation_end.connect(&on_simulation_end);
472 sg_file_t sg_file_open(const char* fullpath, void* data)
474 return new simgrid::s4u::File(fullpath, data);
477 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
479 return fd->read(size);
482 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
484 return fd->write(size);
487 void sg_file_close(const_sg_file_t fd)
492 /** Retrieves the path to the file
493 * @ingroup plugin_filesystem
495 const char* sg_file_get_name(const_sg_file_t fd)
497 xbt_assert((fd != nullptr), "Invalid file descriptor");
498 return fd->get_path();
501 /** Retrieves the size of the file
502 * @ingroup plugin_filesystem
504 sg_size_t sg_file_get_size(const_sg_file_t fd)
509 void sg_file_dump(const_sg_file_t fd)
514 /** Retrieves the user data associated with the file
515 * @ingroup plugin_filesystem
517 void* sg_file_get_data(const_sg_file_t fd)
519 return fd->get_data();
522 /** Changes the user data associated with the file
523 * @ingroup plugin_filesystem
525 void sg_file_set_data(sg_file_t fd, void* data)
531 * @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).
532 * @ingroup plugin_filesystem
534 * @param fd : file object that identifies the stream
535 * @param offset : number of bytes to offset from origin
536 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
537 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
538 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
540 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
542 fd->seek(offset, origin);
545 sg_size_t sg_file_tell(const_sg_file_t fd)
550 void sg_file_move(const_sg_file_t fd, const char* fullpath)
555 void sg_file_unlink(sg_file_t fd)
562 * @brief Copy a file to another location on a remote host.
563 * @ingroup plugin_filesystem
565 * @param file : the file to move
566 * @param host : the remote host where the file has to be copied
567 * @param fullpath : the complete path destination on the remote host
568 * @return If successful, the function returns 0. Otherwise, it returns -1.
570 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
572 return file->remote_copy(host, fullpath);
576 * @brief Move a file to another location on a remote host.
577 * @ingroup plugin_filesystem
579 * @param file : the file to move
580 * @param host : the remote host where the file has to be moved
581 * @param fullpath : the complete path destination on the remote host
582 * @return If successful, the function returns 0. Otherwise, it returns -1.
584 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
586 return file->remote_move(host, fullpath);
589 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
591 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
594 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
596 return d->extension<FileSystemDiskExt>()->get_used_size();
599 sg_size_t sg_disk_get_size(const_sg_disk_t d)
601 return d->extension<FileSystemDiskExt>()->get_size();
604 const char* sg_disk_get_mount_point(const_sg_disk_t d)
606 return d->extension<FileSystemDiskExt>()->get_mount_point();