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/Comm.hpp"
9 #include "simgrid/s4u/Engine.hpp"
10 #include "src/surf/HostImpl.hpp"
11 #include "src/surf/xml/platf_private.hpp"
12 #include "xbt/config.hpp"
13 #include "xbt/parse_units.hpp"
16 #include <boost/algorithm/string.hpp>
17 #include <boost/algorithm/string/join.hpp>
18 #include <boost/algorithm/string/split.hpp>
23 XBT_LOG_NEW_DEFAULT_SUBCATEGORY(s4u_file, s4u, "S4U files");
24 int sg_storage_max_file_descriptors = 1024;
26 /** @defgroup plugin_filesystem Plugin FileSystem
28 * This adds the notion of Files on top of the storage notion that provided by the core of SimGrid.
29 * Activate this plugin at will.
34 template class xbt::Extendable<s4u::File>;
37 simgrid::xbt::Extension<Disk, FileSystemDiskExt> FileSystemDiskExt::EXTENSION_ID;
38 simgrid::xbt::Extension<Host, FileDescriptorHostExt> FileDescriptorHostExt::EXTENSION_ID;
40 const Disk* File::find_local_disk_on(const Host* host)
42 const Disk* d = nullptr;
43 size_t longest_prefix_length = 0;
44 for (auto const& disk : host->get_disks()) {
45 std::string current_mount;
46 if (disk->get_host() != host)
47 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point(disk->get_host());
49 current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
50 mount_point_ = fullpath_.substr(0, current_mount.length());
51 if (mount_point_ == current_mount && current_mount.length() > longest_prefix_length) {
52 /* The current mount name is found in the full path and is bigger than the previous*/
53 longest_prefix_length = current_mount.length();
56 xbt_assert(longest_prefix_length > 0, "Can't find mount point for '%s' on '%s'", fullpath_.c_str(),
58 /* Mount point found, split fullpath_ into mount_name and path+filename*/
59 mount_point_ = fullpath_.substr(0, longest_prefix_length);
60 if (mount_point_ == std::string("/"))
63 path_ = fullpath_.substr(longest_prefix_length, fullpath_.length());
64 XBT_DEBUG("%s + %s", mount_point_.c_str(), path_.c_str());
69 File::File(const std::string& fullpath, void* userdata) : File(fullpath, Host::current(), userdata) {}
71 File::File(const std::string& fullpath, const_sg_host_t host, void* userdata) : fullpath_(fullpath)
73 kernel::actor::simcall([this, &host, userdata] {
74 this->set_data(userdata);
75 // this cannot fail because we get a xbt_die if the mountpoint does not exist
76 local_disk_ = find_local_disk_on(host);
78 // assign a file descriptor id to the newly opened File
79 auto* ext = host->extension<simgrid::s4u::FileDescriptorHostExt>();
80 if (ext->file_descriptor_table == nullptr) {
81 ext->file_descriptor_table = std::make_unique<std::vector<int>>(sg_storage_max_file_descriptors);
82 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
84 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
85 desc_id = ext->file_descriptor_table->back();
86 ext->file_descriptor_table->pop_back();
88 XBT_DEBUG("\tOpen file '%s'", path_.c_str());
89 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
90 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
92 // if file does not exist create an empty file
94 auto sz = content->find(path_);
95 if (sz != content->end()) {
99 content->insert({path_, size_});
100 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
108 std::vector<int>* desc_table =
109 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table.get();
110 kernel::actor::simcall([this, desc_table] { desc_table->push_back(this->desc_id); });
113 void File::dump() const
115 XBT_INFO("File Descriptor information:\n"
116 "\t\tFull path: '%s'\n"
118 "\t\tMount point: '%s'\n"
119 "\t\tDisk Id: '%s'\n"
120 "\t\tHost Id: '%s'\n"
121 "\t\tFile Descriptor Id: %d",
122 get_path(), size_, mount_point_.c_str(), local_disk_->get_cname(), local_disk_->get_host()->get_cname(),
126 sg_size_t File::read(sg_size_t size)
128 if (size_ == 0) /* Nothing to read, return */
130 Host* host = nullptr;
131 // if the current position is close to the end of the file, we may not be able to read the requested size
132 sg_size_t to_read = std::min(size, size_ - current_position_);
133 sg_size_t read_size = 0;
135 /* Find the host where the file is physically located and read it */
136 host = local_disk_->get_host();
137 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
138 read_size = local_disk_->read(to_read);
140 current_position_ += read_size;
142 if (host && host->get_name() != Host::current()->get_name() && read_size > 0) {
143 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
144 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
145 Comm::sendto(host, Host::current(), read_size);
151 /** @brief Write into a file (local or remote)
152 * @ingroup plugin_filesystem
154 * @param size of the file to write
155 * @return the number of bytes successfully write or -1 if an error occurred
157 sg_size_t File::write(sg_size_t size, bool write_inside)
159 if (size == 0) /* Nothing to write, return */
162 sg_size_t write_size = 0;
163 /* Find the host where the file is physically located (remote or local)*/
164 Host* host = local_disk_->get_host();
166 if (host && host->get_name() != Host::current()->get_name()) {
167 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
168 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
169 Comm::sendto(Host::current(), host, size);
171 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_disk_->get_cname(), size, size_,
172 sg_disk_get_size_used(local_disk_), sg_disk_get_size(local_disk_));
173 // If the disk is full before even starting to write
174 if (sg_disk_get_size_used(local_disk_) >= sg_disk_get_size(local_disk_))
176 if (not write_inside) {
177 /* Subtract the part of the file that might disappear from the used sized on the storage element */
178 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_ - current_position_);
179 write_size = local_disk_->write(size);
180 local_disk_->extension<FileSystemDiskExt>()->incr_used_size(write_size);
181 current_position_ += write_size;
182 size_ = current_position_;
184 write_size = local_disk_->write(size);
185 current_position_ += write_size;
186 if (current_position_ > size_)
187 size_ = current_position_;
189 kernel::actor::simcall([this] {
190 std::map<std::string, sg_size_t, std::less<>>* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
192 content->erase(path_);
193 content->insert({path_, 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 current_position_ = offset;
216 current_position_ += offset;
219 current_position_ = size_ + offset;
226 sg_size_t File::tell() const
228 return current_position_;
231 void File::move(const std::string& fullpath) const
233 /* Check if the new full path is on the same mount point */
234 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
235 std::map<std::string, sg_size_t, std::less<>>* content = nullptr;
236 content = local_disk_->extension<FileSystemDiskExt>()->get_content();
238 auto sz = content->find(path_);
239 if (sz != content->end()) { // src file exists
240 sg_size_t new_size = sz->second;
241 content->erase(path_);
242 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
243 content->insert({path.c_str(), new_size});
244 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
246 XBT_WARN("File %s doesn't exist", path_.c_str());
250 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
254 int File::unlink() const
256 /* Check if the file is on local storage */
257 auto* content = local_disk_->extension<FileSystemDiskExt>()->get_content();
258 const char* name = local_disk_->get_cname();
260 if (not content || content->find(path_) == content->end()) {
261 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), name);
264 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), name);
266 local_disk_->extension<FileSystemDiskExt>()->decr_used_size(size_);
268 // Remove the file from storage
269 content->erase(path_);
275 int File::remote_copy(sg_host_t host, const std::string& fullpath)
277 /* Find the host where the file is physically located and read it */
278 Host* src_host = nullptr;
279 sg_size_t read_size = 0;
281 Host* dst_host = host;
282 size_t longest_prefix_length = 0;
286 src_host = local_disk_->get_host();
287 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_disk_->get_cname());
288 read_size = local_disk_->read(size_);
289 current_position_ += read_size;
291 const Disk* dst_disk = nullptr;
293 for (auto const& disk : host->get_disks()) {
294 std::string current_mount = disk->extension<FileSystemDiskExt>()->get_mount_point();
295 std::string mount_point = std::string(fullpath).substr(0, current_mount.length());
296 if (mount_point == current_mount && current_mount.length() > longest_prefix_length) {
297 /* The current mount name is found in the full path and is bigger than the previous*/
298 longest_prefix_length = current_mount.length();
303 if (dst_disk == nullptr) {
304 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath.c_str(), host->get_cname());
309 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
310 dst_host->get_cname());
311 Comm::sendto(src_host, dst_host, read_size);
314 /* Create file on remote host, write it and close it */
315 File fd(fullpath, dst_host, nullptr);
320 int File::remote_move(sg_host_t host, const std::string& fullpath)
322 int res = remote_copy(host, fullpath);
327 FileSystemDiskExt::FileSystemDiskExt(const Disk* ptr)
329 const char* size_str = ptr->get_property("size");
330 std::string dummyfile;
332 size_ = xbt_parse_get_size(dummyfile, -1, size_str, "disk size " + ptr->get_name());
334 const char* current_mount_str = ptr->get_property("mount");
335 if (current_mount_str)
336 mount_point_ = std::string(current_mount_str);
338 mount_point_ = std::string("/");
340 const char* content_str = ptr->get_property("content");
342 content_.reset(parse_content(content_str));
345 std::map<std::string, sg_size_t, std::less<>>* FileSystemDiskExt::parse_content(const std::string& filename)
347 if (filename.empty())
350 auto* parse_content = new std::map<std::string, sg_size_t, std::less<>>();
352 auto fs = std::unique_ptr<std::ifstream>(surf_ifsopen(filename));
353 xbt_assert(not fs->fail(), "Cannot open file '%s' (path=%s)", filename.c_str(),
354 (boost::join(surf_path, ":")).c_str());
357 std::vector<std::string> tokens;
359 std::getline(*fs, line);
361 if (line.length() > 0) {
362 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
363 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
364 sg_size_t size = std::stoull(tokens.at(1));
367 parse_content->insert({tokens.front(), size});
369 } while (not fs->eof());
370 return parse_content;
373 void FileSystemDiskExt::decr_used_size(sg_size_t size)
375 simgrid::kernel::actor::simcall([this, size] { used_size_ -= size; });
378 void FileSystemDiskExt::incr_used_size(sg_size_t size)
380 simgrid::kernel::actor::simcall([this, size] { used_size_ += size; });
385 using simgrid::s4u::FileDescriptorHostExt;
386 using simgrid::s4u::FileSystemDiskExt;
388 static void on_disk_creation(simgrid::s4u::Disk& d)
390 d.extension_set(new FileSystemDiskExt(&d));
393 static void on_host_creation(simgrid::s4u::Host& host)
395 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
398 static void on_platform_created()
400 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
401 const char* remote_disk_str = host->get_property("remote_disk");
402 if (remote_disk_str) {
403 std::vector<std::string> tokens;
404 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
405 std::string mount_point = tokens[0];
406 simgrid::s4u::Host* remote_host = simgrid::s4u::Host::by_name_or_null(tokens[2]);
407 xbt_assert(remote_host, "You're trying to access a host that does not exist. Please check your platform file");
409 const simgrid::s4u::Disk* disk = nullptr;
410 for (auto const& d : remote_host->get_disks())
411 if (d->get_name() == tokens[1]) {
416 xbt_assert(disk, "You're trying to mount a disk that does not exist. Please check your platform file");
417 disk->extension<FileSystemDiskExt>()->add_remote_mount(remote_host, mount_point);
418 host->add_disk(disk);
420 XBT_DEBUG("Host '%s' wants to mount a remote disk: %s of %s mounted on %s", host->get_cname(), disk->get_cname(),
421 remote_host->get_cname(), mount_point.c_str());
422 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
427 static void on_simulation_end()
429 XBT_DEBUG("Simulation is over, time to unregister remote disks if any");
430 for (auto const& host : simgrid::s4u::Engine::get_instance()->get_all_hosts()) {
431 const char* remote_disk_str = host->get_property("remote_disk");
432 if (remote_disk_str) {
433 std::vector<std::string> tokens;
434 boost::split(tokens, remote_disk_str, boost::is_any_of(":"));
435 XBT_DEBUG("Host '%s' wants to unmount a remote disk: %s of %s mounted on %s", host->get_cname(),
436 tokens[1].c_str(), tokens[2].c_str(), tokens[0].c_str());
437 host->remove_disk(tokens[1]);
438 XBT_DEBUG("Host '%s' now has %zu disks", host->get_cname(), host->get_disks().size());
443 /* **************************** Public interface *************************** */
444 /** @brief Initialize the file system plugin.
445 @ingroup plugin_filesystem
448 See the examples in :ref:`s4u_ex_disk_io`.
451 void sg_storage_file_system_init()
453 sg_storage_max_file_descriptors = 1024;
454 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
455 "Maximum number of concurrently opened files per host. Default is 1024");
457 if (not FileSystemDiskExt::EXTENSION_ID.valid()) {
458 FileSystemDiskExt::EXTENSION_ID = simgrid::s4u::Disk::extension_create<FileSystemDiskExt>();
459 simgrid::s4u::Disk::on_creation.connect(&on_disk_creation);
462 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
463 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
464 simgrid::s4u::Host::on_creation.connect(&on_host_creation);
466 simgrid::s4u::Engine::on_platform_created.connect(&on_platform_created);
467 simgrid::s4u::Engine::on_simulation_end.connect(&on_simulation_end);
470 sg_file_t sg_file_open(const char* fullpath, void* data)
472 return new simgrid::s4u::File(fullpath, data);
475 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
477 return fd->read(size);
480 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
482 return fd->write(size);
485 void sg_file_close(const_sg_file_t fd)
490 /** Retrieves the path to the file
491 * @ingroup plugin_filesystem
493 const char* sg_file_get_name(const_sg_file_t fd)
495 xbt_assert((fd != nullptr), "Invalid file descriptor");
496 return fd->get_path();
499 /** Retrieves the size of the file
500 * @ingroup plugin_filesystem
502 sg_size_t sg_file_get_size(const_sg_file_t fd)
507 void sg_file_dump(const_sg_file_t fd)
512 /** Retrieves the user data associated with the file
513 * @ingroup plugin_filesystem
515 void* sg_file_get_data(const_sg_file_t fd)
517 return fd->get_data();
520 /** Changes the user data associated with the file
521 * @ingroup plugin_filesystem
523 void sg_file_set_data(sg_file_t fd, void* data)
529 * @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).
530 * @ingroup plugin_filesystem
532 * @param fd : file object that identifies the stream
533 * @param offset : number of bytes to offset from origin
534 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
535 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
536 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
538 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
540 fd->seek(offset, origin);
543 sg_size_t sg_file_tell(const_sg_file_t fd)
548 void sg_file_move(const_sg_file_t fd, const char* fullpath)
553 void sg_file_unlink(sg_file_t fd)
560 * @brief Copy a file to another location on a remote host.
561 * @ingroup plugin_filesystem
563 * @param file : the file to move
564 * @param host : the remote host where the file has to be copied
565 * @param fullpath : the complete path destination on the remote host
566 * @return If successful, the function returns 0. Otherwise, it returns -1.
568 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
570 return file->remote_copy(host, fullpath);
574 * @brief Move a file to another location on a remote host.
575 * @ingroup plugin_filesystem
577 * @param file : the file to move
578 * @param host : the remote host where the file has to be moved
579 * @param fullpath : the complete path destination on the remote host
580 * @return If successful, the function returns 0. Otherwise, it returns -1.
582 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
584 return file->remote_move(host, fullpath);
587 sg_size_t sg_disk_get_size_free(const_sg_disk_t d)
589 return d->extension<FileSystemDiskExt>()->get_size() - d->extension<FileSystemDiskExt>()->get_used_size();
592 sg_size_t sg_disk_get_size_used(const_sg_disk_t d)
594 return d->extension<FileSystemDiskExt>()->get_used_size();
597 sg_size_t sg_disk_get_size(const_sg_disk_t d)
599 return d->extension<FileSystemDiskExt>()->get_size();
602 const char* sg_disk_get_mount_point(const_sg_disk_t d)
604 return d->extension<FileSystemDiskExt>()->get_mount_point();