1 /* Copyright (c) 2015-2019. 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(const std::string& fullpath, void* userdata) : File(fullpath, Host::current(), userdata){};
28 File::File(const 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.reset(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 (host->get_name() != Host::current()->get_name() && read_size > 0) {
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->send_to(Host::current(), read_size);
114 /** @brief Write into a file (local or remote)
116 * @param size of the file to write
117 * @return the number of bytes successfully write or -1 if an error occurred
119 sg_size_t File::write(sg_size_t size, int write_inside)
121 if (size == 0) /* Nothing to write, return */
124 /* Find the host where the file is physically located (remote or local)*/
125 Host* host = local_storage_->get_host();
127 if (host->get_name() != Host::current()->get_name()) {
128 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
129 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
130 Host::current()->send_to(host, size);
133 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu' '%llu:%llu'", get_path(), local_storage_->get_cname(), size, size_, sg_storage_get_size_used(local_storage_), sg_storage_get_size(local_storage_));
134 // If the storage is full before even starting to write
135 if (sg_storage_get_size_used(local_storage_) >= sg_storage_get_size(local_storage_))
137 sg_size_t write_size=0;
139 /* Substract the part of the file that might disappear from the used sized on the storage element */
140 local_storage_->extension<FileSystemStorageExt>()->decr_used_size(size_ - current_position_);
141 write_size = local_storage_->write(size);
142 local_storage_->extension<FileSystemStorageExt>()->incr_used_size(write_size);
143 current_position_ += write_size;
144 size_ = current_position_;
146 write_size = local_storage_->write(size);
147 current_position_ += write_size;
148 if(current_position_>size_)
149 size_ = current_position_;
151 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
153 content->erase(path_);
154 content->insert({path_, size_});
159 sg_size_t File::size()
164 void File::seek(sg_offset_t offset)
166 current_position_ = offset;
169 void File::seek(sg_offset_t offset, int origin)
173 current_position_ = offset;
176 current_position_ += offset;
179 current_position_ = size_ + offset;
186 sg_size_t File::tell()
188 return current_position_;
191 void File::move(const std::string& fullpath)
193 /* Check if the new full path is on the same mount point */
194 if (fullpath.compare(0, mount_point_.length(), mount_point_) == 0) {
195 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
196 auto sz = content->find(path_);
197 if (sz != content->end()) { // src file exists
198 sg_size_t new_size = sz->second;
199 content->erase(path_);
200 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
201 content->insert({path.c_str(), new_size});
202 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
204 XBT_WARN("File %s doesn't exist", path_.c_str());
207 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
213 /* Check if the file is on local storage */
214 std::map<std::string, sg_size_t>* content = local_storage_->extension<FileSystemStorageExt>()->get_content();
216 if (content->find(path_) == content->end()) {
217 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), local_storage_->get_cname());
220 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), local_storage_->get_cname());
221 local_storage_->extension<FileSystemStorageExt>()->decr_used_size(size_);
223 // Remove the file from storage
224 content->erase(fullpath_);
230 int File::remote_copy(sg_host_t host, const char* fullpath)
232 /* Find the host where the file is physically located and read it */
233 Storage* storage_src = local_storage_;
234 Host* src_host = storage_src->get_host();
236 XBT_DEBUG("READ %s on disk '%s'", get_path(), local_storage_->get_cname());
237 // if the current position is close to the end of the file, we may not be able to read the requested size
238 sg_size_t read_size = local_storage_->read(size_);
239 current_position_ += read_size;
241 /* Find the host that owns the storage where the file has to be copied */
242 Storage* storage_dest = nullptr;
244 size_t longest_prefix_length = 0;
246 for (auto const& elm : host->get_mounted_storages()) {
247 std::string mount_point = std::string(fullpath).substr(0, elm.first.size());
248 if (mount_point == elm.first && elm.first.length() > longest_prefix_length) {
249 /* The current mount name is found in the full path and is bigger than the previous*/
250 longest_prefix_length = elm.first.length();
251 storage_dest = elm.second;
255 if (storage_dest != nullptr) {
256 /* Mount point found, retrieve the host the storage is attached to */
257 dst_host = storage_dest->get_host();
259 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath, host->get_cname());
263 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
264 storage_dest->get_host()->get_cname());
265 src_host->send_to(dst_host, read_size);
267 /* Create file on remote host, write it and close it */
268 File* fd = new File(fullpath, dst_host, nullptr);
269 sg_size_t write_size = fd->local_storage_->write(read_size);
270 fd->local_storage_->extension<FileSystemStorageExt>()->incr_used_size(write_size);
271 (*(fd->local_storage_->extension<FileSystemStorageExt>()->get_content()))[path_] = size_;
276 int File::remote_move(sg_host_t host, const char* fullpath)
278 int res = remote_copy(host, fullpath);
283 FileSystemStorageExt::FileSystemStorageExt(simgrid::s4u::Storage* ptr)
285 content_.reset(parse_content(ptr->get_impl()->content_name_));
286 size_ = ptr->get_impl()->size_;
289 std::map<std::string, sg_size_t>* FileSystemStorageExt::parse_content(const std::string& filename)
291 if (filename.empty())
294 std::map<std::string, sg_size_t>* parse_content = new std::map<std::string, sg_size_t>();
296 std::ifstream* fs = surf_ifsopen(filename);
299 std::vector<std::string> tokens;
301 std::getline(*fs, line);
303 if (line.length() > 0) {
304 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
305 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
306 sg_size_t size = std::stoull(tokens.at(1));
309 parse_content->insert({tokens.front(), size});
311 } while (not fs->eof());
313 return parse_content;
318 using simgrid::s4u::FileSystemStorageExt;
319 using simgrid::s4u::FileDescriptorHostExt;
321 static void on_storage_creation(simgrid::s4u::Storage& st)
323 st.extension_set(new FileSystemStorageExt(&st));
326 static void on_host_creation(simgrid::s4u::Host& host)
328 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
331 /* **************************** Public interface *************************** */
332 void sg_storage_file_system_init()
334 sg_storage_max_file_descriptors = 1024;
335 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
336 "Maximum number of concurrently opened files per host. Default is 1024");
338 if (not FileSystemStorageExt::EXTENSION_ID.valid()) {
339 FileSystemStorageExt::EXTENSION_ID = simgrid::s4u::Storage::extension_create<FileSystemStorageExt>();
340 simgrid::s4u::Storage::on_creation.connect(&on_storage_creation);
343 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
344 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
345 simgrid::s4u::Host::on_creation.connect(&on_host_creation);
349 sg_file_t sg_file_open(const char* fullpath, void* data)
351 return new simgrid::s4u::File(fullpath, data);
354 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
356 return fd->read(size);
359 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
361 return fd->write(size);
364 void sg_file_close(sg_file_t fd)
369 const char* sg_file_get_name(sg_file_t fd)
371 xbt_assert((fd != nullptr), "Invalid file descriptor");
372 return fd->get_path();
375 sg_size_t sg_file_get_size(sg_file_t fd)
380 void sg_file_dump(sg_file_t fd)
385 void* sg_file_get_data(sg_file_t fd)
387 return fd->get_userdata();
390 void sg_file_set_data(sg_file_t fd, void* data)
392 fd->set_userdata(data);
396 * @brief Set the file position indicator in the sg_file_t by adding offset bytes
397 * to the position specified by origin (either SEEK_SET, SEEK_CUR, or SEEK_END).
399 * @param fd : file object that identifies the stream
400 * @param offset : number of bytes to offset from origin
401 * @param origin : Position used as reference for the offset. It is specified by one of the following constants defined
402 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
403 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
405 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
407 fd->seek(offset, origin);
410 sg_size_t sg_file_tell(sg_file_t fd)
415 void sg_file_move(sg_file_t fd, const char* fullpath)
420 void sg_file_unlink(sg_file_t fd)
427 * @brief Copy a file to another location on a remote host.
428 * @param file : the file to move
429 * @param host : the remote host where the file has to be copied
430 * @param fullpath : the complete path destination on the remote host
431 * @return If successful, the function returns 0. Otherwise, it returns -1.
433 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
435 return file->remote_copy(host, fullpath);
439 * @brief Move a file to another location on a remote host.
440 * @param file : the file to move
441 * @param host : the remote host where the file has to be moved
442 * @param fullpath : the complete path destination on the remote host
443 * @return If successful, the function returns 0. Otherwise, it returns -1.
445 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
447 return file->remote_move(host, fullpath);
450 sg_size_t sg_storage_get_size_free(sg_storage_t st)
452 return st->extension<FileSystemStorageExt>()->get_size() - st->extension<FileSystemStorageExt>()->get_used_size();
455 sg_size_t sg_storage_get_size_used(sg_storage_t st)
457 return st->extension<FileSystemStorageExt>()->get_used_size();
460 sg_size_t sg_storage_get_size(sg_storage_t st)
462 return st->extension<FileSystemStorageExt>()->get_size();
465 xbt_dict_t sg_storage_get_content(sg_storage_t storage)
467 std::map<std::string, sg_size_t>* content = storage->extension<simgrid::s4u::FileSystemStorageExt>()->get_content();
468 // Note: ::operator delete is ok here (no destructor called) since the dict elements are of POD type sg_size_t.
469 xbt_dict_t content_as_dict = xbt_dict_new_homogeneous(::operator delete);
471 for (auto const& entry : *content) {
472 sg_size_t* psize = new sg_size_t;
473 *psize = entry.second;
474 xbt_dict_set(content_as_dict, entry.first.c_str(), psize, nullptr);
476 return content_as_dict;
479 xbt_dict_t sg_host_get_storage_content(sg_host_t host)
481 xbt_assert((host != nullptr), "Invalid parameters");
482 xbt_dict_t contents = xbt_dict_new_homogeneous(nullptr);
483 for (auto const& elm : host->get_mounted_storages())
484 xbt_dict_set(contents, elm.first.c_str(), sg_storage_get_content(elm.second), nullptr);