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());
107 File::~File() = default;
109 File* File::open(const std::string& fullpath, void* userdata)
111 return new File(fullpath, userdata);
114 File* File::open(const std::string& fullpath, const_sg_host_t host, void* userdata)
116 return new File(fullpath, host, userdata);
121 std::vector<int>* desc_table =
122 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table.get();
123 kernel::actor::simcall_answered([this, desc_table] { desc_table->push_back(this->desc_id); });
127 void File::dump() const
129 XBT_INFO("File Descriptor information:\n"
130 "\t\tFull path: '%s'\n"
132 "\t\tMount point: '%s'\n"
133 "\t\tDisk Id: '%s'\n"
134 "\t\tHost Id: '%s'\n"
135 "\t\tFile Descriptor Id: %d",
136 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
140 sg_size_t File::read(sg_size_t size)
142 if (size_ == 0) /* Nothing to read, return */
144 Host* host = nullptr;
145 // if the current position is close to the end of the file, we may not be able to read the requested size
146 sg_size_t to_read = std::min(size, size_ - current_position_);
147 sg_size_t read_size = 0;
149 /* Find the host where the file is physically located and read it */
150 host = local_disk_->get_host();
151 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
152 read_size = local_disk_->read(to_read);
154 current_position_ += read_size;
156 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
157 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
158 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
159 Comm::sendto(host, Host::current(), read_size);
165 /** @brief Write into a file (local or remote)
166 * @ingroup plugin_filesystem
168 * @param size of the file to write
169 * @param write_inside
170 * @return the number of bytes successfully write or -1 if an error occurred
172 sg_size_t File::write(sg_size_t size, bool write_inside)
174 if (size == 0) /* Nothing to write, return */
177 sg_size_t write_size = 0;
178 /* Find the host where the file is physically located (remote or local)*/
179 if (Host* host = local_disk_->get_host(); host && host->get_name() != Host::current()->get_name()) {
180 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
181 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
182 Comm::sendto(Host::current(), host, size);
184 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
185 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
186 // If the disk is full before even starting to write
187 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
189 if (not write_inside) {
190 /* Subtract the part of the file that might disappear from the used sized on the storage element */
191 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
192 write_size = local_disk_->write(size);
193 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(write_size);
194 current_position_ += write_size;
195 size_ = current_position_;
197 write_size = local_disk_->write(size);
198 current_position_ += write_size;
199 if (current_position_ > size_)
200 size_ = current_position_;
202 kernel::actor::simcall_answered([this] {
203 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
205 content->erase(path_);
206 content->insert({path_, size_});
212 sg_size_t File::size() const
217 void File::seek(sg_offset_t offset)
219 current_position_ = offset;
222 void File::seek(sg_offset_t offset, int origin)
226 current_position_ = offset;
229 current_position_ += offset;
232 current_position_ = size_ + offset;
239 sg_size_t File::tell() const
241 return current_position_;
244 void File::move(const std::string& fullpath) const
246 /* Check if the new full path is on the same mount point */
247 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
248 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
249 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
251 auto sz = content->find(path_);
252 if (sz != content->end()) { // src file exists
253 sg_size_t new_size = sz->second;
254 content->erase(path_);
255 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
256 content->insert({path.c_str(), new_size});
257 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
259 XBT_WARN("File %s doesn't exist", path_.c_str());
263 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
267 int File::unlink() const
269 /* Check if the file is on local storage */
270 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
271 const char* name = local_disk_->get_cname();
273 if (not content || content->find(path_) == content->end()) {
274 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
277 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), name);
279 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 = std::string(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_ = std::string(current_mount_str);
351 mount_point_ = std::string("/");
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>(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::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 (remote_disk_str) {
420 std::vector<std::string> tokens;
421 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
422 std::string mount_point = tokens[0];
423 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
424 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
426 const simgrid::s4u::Disk* disk = nullptr;
427 for (auto const& d : remote_host->get_disks())
428 if (d->get_name() == tokens[1]) {
433 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
434 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
435 host->add_disk(disk);
437 XBT_DEBUG("Host '%s' wants to mount a remote disk: %s of %s mounted on %s", host->get_cname(), disk->get_cname(),
438 remote_host->get_cname(), mount_point.c_str());
439 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
444 static void on_simulation_end()
446 XBT_DEBUG("Simulation is over, time to unregister remote disks if any");
447 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
448 const char* remote_disk_str = host->get_property("remote_disk");
449 if (remote_disk_str) {
450 std::vector<std::string> tokens;
451 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
452 XBT_DEBUG("Host '%s' wants to unmount a remote disk: %s of %s mounted on %s", host->get_cname(),
453 tokens[1].c_str(), tokens[2].c_str(), tokens[0].c_str());
454 host->remove_disk(tokens[1]);
455 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
460 /* **************************** Public interface *************************** */
461 /** @brief Initialize the file system plugin.
462 @ingroup plugin_filesystem
465 See the examples in :ref:`s4u_ex_disk_io`.
468 void sg_storage_file_system_init()
470 sg_storage_max_file_descriptors = 1024;
471 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
472 "Maximum number of concurrently opened files per host. Default is 1024");
474 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
475 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
476 simgrid::s4u::Disk::on_creation_cb(&on_disk_creation);
479 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
480 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
481 simgrid::s4u::Host::on_creation_cb(&on_host_creation);
483 simgrid::s4u::Engine::on_platform_created_cb(&on_platform_created);
484 simgrid::s4u::Engine::on_simulation_end_cb(&on_simulation_end);
487 sg_file_t sg_file_open(const char* fullpath, void* data)
489 return simgrid::s4u::File::open(fullpath, data);
492 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
494 return fd->read(size);
497 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
499 return fd->write(size);
502 void sg_file_close(sg_file_t fd)
507 /** Retrieves the path to the file
508 * @ingroup plugin_filesystem
510 const char* sg_file_get_name(const_sg_file_t fd)
512 xbt_assert((fd != nullptr), "Invalid file descriptor");
513 return fd->get_path();
516 /** Retrieves the size of the file
517 * @ingroup plugin_filesystem
519 sg_size_t sg_file_get_size(const_sg_file_t fd)
524 void sg_file_dump(const_sg_file_t fd)
529 /** Retrieves the user data associated with the file
530 * @ingroup plugin_filesystem
532 void* sg_file_get_data(const_sg_file_t fd)
534 return fd->get_data<void>();
537 /** Changes the user data associated with the file
538 * @ingroup plugin_filesystem
540 void sg_file_set_data(sg_file_t fd, void* data)
546 * @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).
547 * @ingroup plugin_filesystem
549 * @param fd : file object that identifies the stream
550 * @param offset : number of bytes to offset from origin
551 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
552 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
553 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
555 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
557 fd->seek(offset, origin);
560 sg_size_t sg_file_tell(const_sg_file_t fd)
565 void sg_file_move(const_sg_file_t fd, const char* fullpath)
570 void sg_file_unlink(sg_file_t fd)
577 * @brief Copy a file to another location on a remote host.
578 * @ingroup plugin_filesystem
580 * @param file : the file to move
581 * @param host : the remote host where the file has to be copied
582 * @param fullpath : the complete path destination on the remote host
583 * @return If successful, the function returns 0. Otherwise, it returns -1.
585 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
587 return file->remote_copy(host, fullpath);
591 * @brief Move a file to another location on a remote host.
592 * @ingroup plugin_filesystem
594 * @param file : the file to move
595 * @param host : the remote host where the file has to be moved
596 * @param fullpath : the complete path destination on the remote host
597 * @return If successful, the function returns 0. Otherwise, it returns -1.
599 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
601 return file->remote_move(host, fullpath);
604 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
606 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
609 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
611 return d->extension<FileSystemDiskExt>()->get_used_size();
614 sg_size_t sg_disk_get_size(const_sg_disk_t d)
616 return d->extension<FileSystemDiskExt>()->get_size();
619 const char* sg_disk_get_mount_point(const_sg_disk_t d)
621 return d->extension<FileSystemDiskExt>()->get_mount_point();