1 /* Copyright (c) 2015-2018. 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 "src/surf/HostImpl.hpp"
9 #include "xbt/config.hpp"
12 #include <boost/algorithm/string.hpp>
13 #include <boost/algorithm/string/join.hpp>
14 #include <boost/algorithm/string/split.hpp>
18 XBT_LOG_NEW_DEFAULT_CATEGORY(s4u_file, "S4U files");
19 int sg_storage_max_file_descriptors = 1024;
23 simgrid::xbt::Extension<Storage, FileSystemStorageExt> FileSystemStorageExt::EXTENSION_ID;
24 simgrid::xbt::Extension<Host, FileDescriptorHostExt> FileDescriptorHostExt::EXTENSION_ID;
26 File::File(std::string fullpath, void* userdata) : File(fullpath, Host::current(), userdata){};
28 File::File(std::string fullpath, sg_host_t host, void* userdata) : fullpath_(fullpath), userdata_(userdata)
30 // this cannot fail because we get a xbt_die if the mountpoint does not exist
31 Storage* st = nullptr;
32 size_t longest_prefix_length = 0;
33 XBT_DEBUG("Search for storage name for '%s' on '%s'", fullpath.c_str(), host->get_cname());
35 for (auto const& mnt : host->get_mounted_storages()) {
36 XBT_DEBUG("See '%s'", mnt.first.c_str());
37 mount_point_ = fullpath.substr(0, mnt.first.length());
39 if (mount_point_ == mnt.first && mnt.first.length() > longest_prefix_length) {
40 /* The current mount name is found in the full path and is bigger than the previous*/
41 longest_prefix_length = mnt.first.length();
45 if (longest_prefix_length > 0) { /* Mount point found, split fullpath into mount_name and path+filename*/
46 mount_point_ = fullpath.substr(0, longest_prefix_length);
47 path_ = fullpath.substr(longest_prefix_length, fullpath.length());
49 xbt_die("Can't find mount point for '%s' on '%s'", fullpath.c_str(), host->get_cname());
53 // assign a file descriptor id to the newly opened File
54 FileDescriptorHostExt* ext = host->extension<simgrid::s4u::FileDescriptorHostExt>();
55 if (ext->file_descriptor_table == nullptr) {
56 ext->file_descriptor_table = new std::vector<int>(sg_storage_max_file_descriptors);
57 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
59 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
60 desc_id = ext->file_descriptor_table->back();
61 ext->file_descriptor_table->pop_back();
63 XBT_DEBUG("\tOpen file '%s'", path_.c_str());
64 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
65 // if file does not exist create an empty file
66 auto sz = content->find(path_);
67 if (sz != content->end()) {
71 content->insert({path_, size_});
72 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
78 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table->push_back(desc_id);
83 XBT_INFO("File Descriptor information:\n"
84 "\t\tFull path: '%s'\n"
86 "\t\tMount point: '%s'\n"
87 "\t\tStorage Id: '%s'\n"
88 "\t\tStorage Type: '%s'\n"
89 "\t\tFile Descriptor Id: %d",
90 get_path(), size_, mount_point_.c_str(), local_storage_->get_cname(), local_storage_->get_type(), desc_id);
93 sg_size_t File::read(sg_size_t size)
95 if (size_ == 0) /* Nothing to read, return */
98 /* Find the host where the file is physically located and read it */
99 Host* host = local_storage_->get_host();
100 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_storage_->get_cname());
101 // if the current position is close to the end of the file, we may not be able to read the requested size
102 sg_size_t read_size = local_storage_->read(std::min(size, size_ - current_position_));
103 current_position_ += read_size;
105 if (strcmp(host->get_cname(), Host::current()->get_cname())) {
106 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
107 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
108 Host* m_host_list[] = {Host::current(), host};
109 double* flops_amount = new double[2]{0, 0};
110 double* bytes_amount = new double[4]{0, 0, static_cast<double>(read_size), 0};
112 this_actor::parallel_execute(2, m_host_list, flops_amount, bytes_amount);
118 /** @brief Write into a file (local or remote)
120 * @param size of the file to write
121 * @return the number of bytes successfully write or -1 if an error occurred
123 sg_size_t File::write(sg_size_t size)
125 if (size == 0) /* Nothing to write, return */
128 /* Find the host where the file is physically located (remote or local)*/
129 Host* host = local_storage_->get_host();
131 if (strcmp(host->get_cname(), Host::current()->get_cname())) {
132 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
133 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
134 Host* m_host_list[] = {Host::current(), host};
135 double* flops_amount = new double[2]{0, 0};
136 double* bytes_amount = new double[4]{0, static_cast<double>(size), 0, 0};
138 this_actor::parallel_execute(2, m_host_list, flops_amount, bytes_amount);
141 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu'", get_path(), local_storage_->get_cname(), size, size_);
142 // If the storage is full before even starting to write
143 if (sg_storage_get_size_used(local_storage_) >= sg_storage_get_size(local_storage_))
145 /* Substract the part of the file that might disappear from the used sized on the storage element */
146 local_storage_->extension<FileSystemStorageExt>()->decr_used_size(size_ - current_position_);
148 sg_size_t write_size = local_storage_->write(size);
149 local_storage_->extension<FileSystemStorageExt>()->incr_used_size(write_size);
151 current_position_ += write_size;
152 size_ = current_position_;
153 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
155 content->erase(path_);
156 content->insert({path_, size_});
161 sg_size_t File::size()
166 void File::seek(sg_offset_t offset)
168 current_position_ = offset;
171 void File::seek(sg_offset_t offset, int origin)
175 current_position_ = offset;
178 current_position_ += offset;
181 current_position_ = size_ + offset;
188 sg_size_t File::tell()
190 return current_position_;
193 void File::move(std::string fullpath)
195 /* Check if the new full path is on the same mount point */
196 if (not strncmp(mount_point_.c_str(), fullpath.c_str(), mount_point_.length())) {
197 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
198 auto sz = content->find(path_);
199 if (sz != content->end()) { // src file exists
200 sg_size_t new_size = sz->second;
201 content->erase(path_);
202 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
203 content->insert({path.c_str(), new_size});
204 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
206 XBT_WARN("File %s doesn't exist", path_.c_str());
209 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
215 /* Check if the file is on local storage */
216 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
218 if (content->find(path_) == content->end()) {
219 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), local_storage_->get_cname());
222 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), local_storage_->get_cname());
223 local_storage_->extension<FileSystemStorageExt>()->decr_used_size(size_);
225 // Remove the file from storage
226 content->erase(fullpath_);
232 int File::remote_copy(sg_host_t host, const char* fullpath)
234 /* Find the host where the file is physically located and read it */
235 Storage* storage_src = local_storage_;
236 Host* src_host = storage_src->get_host();
238 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_storage_->get_cname());
239 // if the current position is close to the end of the file, we may not be able to read the requested size
240 sg_size_t read_size = local_storage_->read(size_);
241 current_position_ += read_size;
243 /* Find the host that owns the storage where the file has to be copied */
244 Storage* storage_dest = nullptr;
246 size_t longest_prefix_length = 0;
248 for (auto const& elm : host->get_mounted_storages()) {
249 std::string mount_point = std::string(fullpath).substr(0, elm.first.size());
250 if (mount_point == elm.first && elm.first.length() > longest_prefix_length) {
251 /* The current mount name is found in the full path and is bigger than the previous*/
252 longest_prefix_length = elm.first.length();
253 storage_dest = elm.second;
257 if (storage_dest != nullptr) {
258 /* Mount point found, retrieve the host the storage is attached to */
259 dst_host = storage_dest->get_host();
261 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath, host->get_cname());
265 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
266 storage_dest->get_host()->get_cname());
267 Host* m_host_list[] = {src_host, dst_host};
268 double* flops_amount = new double[2]{0, 0};
269 double* bytes_amount = new double[4]{0, static_cast<double>(read_size), 0, 0};
271 this_actor::parallel_execute(2, m_host_list, flops_amount, bytes_amount);
273 /* Create file on remote host, write it and close it */
274 File* fd = new File(fullpath, dst_host, nullptr);
275 sg_size_t write_size = fd->local_storage_->write(read_size);
276 fd->local_storage_->extension<FileSystemStorageExt>()->incr_used_size(write_size);
277 (*(fd->local_storage_->extension<FileSystemStorageExt>()->get_content()))[path_] = size_;
282 int File::remote_move(sg_host_t host, const char* fullpath)
284 int res = remote_copy(host, fullpath);
289 FileSystemStorageExt::FileSystemStorageExt(simgrid::s4u::Storage* ptr)
291 content_ = parse_content(ptr->get_impl()->content_name);
292 size_ = ptr->get_impl()->size_;
295 FileSystemStorageExt::~FileSystemStorageExt()
300 std::map<std::string, sg_size_t>* FileSystemStorageExt::parse_content(std::string filename)
302 if (filename.empty())
305 std::map<std::string, sg_size_t>* parse_content = new std::map<std::string, sg_size_t>();
307 std::ifstream* fs = surf_ifsopen(filename);
310 std::vector<std::string> tokens;
312 std::getline(*fs, line);
314 if (line.length() > 0) {
315 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
316 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
317 sg_size_t size = std::stoull(tokens.at(1));
320 parse_content->insert({tokens.front(), size});
322 } while (not fs->eof());
324 return parse_content;
329 using simgrid::s4u::FileSystemStorageExt;
330 using simgrid::s4u::FileDescriptorHostExt;
332 static void on_storage_creation(simgrid::s4u::Storage& st)
334 st.extension_set(new FileSystemStorageExt(&st));
337 static void on_host_creation(simgrid::s4u::Host& host)
339 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
342 /* **************************** Public interface *************************** */
343 void sg_storage_file_system_init()
345 sg_storage_max_file_descriptors = 1024;
346 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
347 "Maximum number of concurrently opened files per host. Default is 1024");
349 if (not FileSystemStorageExt::EXTENSION_ID.valid()) {
350 FileSystemStorageExt::EXTENSION_ID = simgrid::s4u::Storage::extension_create<FileSystemStorageExt>();
351 simgrid::s4u::Storage::on_creation.connect(&on_storage_creation);
354 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
355 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
356 simgrid::s4u::Host::on_creation.connect(&on_host_creation);
360 sg_file_t sg_file_open(const char* fullpath, void* data)
362 return new simgrid::s4u::File(fullpath, data);
365 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
367 return fd->read(size);
370 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
372 return fd->write(size);
375 void sg_file_close(sg_file_t fd)
380 const char* sg_file_get_name(sg_file_t fd)
382 xbt_assert((fd != nullptr), "Invalid file descriptor");
383 return fd->get_path();
386 sg_size_t sg_file_get_size(sg_file_t fd)
391 void sg_file_dump(sg_file_t fd)
396 void* sg_file_get_data(sg_file_t fd)
398 return fd->get_userdata();
401 void sg_file_set_data(sg_file_t fd, void* data)
403 fd->set_userdata(data);
407 * @brief Set the file position indicator in the sg_file_t by adding offset bytes
408 * to the position specified by origin (either SEEK_SET, SEEK_CUR, or SEEK_END).
410 * @param fd : file object that identifies the stream
411 * @param offset : number of bytes to offset from origin
412 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
413 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
414 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
416 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
418 fd->seek(offset, origin);
421 sg_size_t sg_file_tell(sg_file_t fd)
426 void sg_file_move(sg_file_t fd, const char* fullpath)
431 void sg_file_unlink(sg_file_t fd)
438 * @brief Copy a file to another location on a remote host.
439 * @param file : the file to move
440 * @param host : the remote host where the file has to be copied
441 * @param fullpath : the complete path destination on the remote host
442 * @return If successful, the function returns 0. Otherwise, it returns -1.
444 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
446 return file->remote_copy(host, fullpath);
450 * @brief Move a file to another location on a remote host.
451 * @param file : the file to move
452 * @param host : the remote host where the file has to be moved
453 * @param fullpath : the complete path destination on the remote host
454 * @return If successful, the function returns 0. Otherwise, it returns -1.
456 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
458 return file->remote_move(host, fullpath);
461 sg_size_t sg_storage_get_size_free(sg_storage_t st)
463 return st->extension<FileSystemStorageExt>()->get_size() - st->extension<FileSystemStorageExt>()->get_used_size();
466 sg_size_t sg_storage_get_size_used(sg_storage_t st)
468 return st->extension<FileSystemStorageExt>()->get_used_size();
471 sg_size_t sg_storage_get_size(sg_storage_t st)
473 return st->extension<FileSystemStorageExt>()->get_size();
476 xbt_dict_t sg_storage_get_content(sg_storage_t storage)
478 std::map<std::string, sg_size_t>* content = storage->extension<simgrid::s4u::FileSystemStorageExt>()->get_content();
479 // Note: ::operator delete is ok here (no destructor called) since the dict elements are of POD type sg_size_t.
480 xbt_dict_t content_as_dict = xbt_dict_new_homogeneous(::operator delete);
482 for (auto const& entry : *content) {
483 sg_size_t* psize = new sg_size_t;
484 *psize = entry.second;
485 xbt_dict_set(content_as_dict, entry.first.c_str(), psize, nullptr);
487 return content_as_dict;
490 xbt_dict_t sg_host_get_storage_content(sg_host_t host)
492 xbt_assert((host != nullptr), "Invalid parameters");
493 xbt_dict_t contents = xbt_dict_new_homogeneous(nullptr);
494 for (auto const& elm : host->get_mounted_storages())
495 xbt_dict_set(contents, elm.first.c_str(), sg_storage_get_content(elm.second), nullptr);