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. */
7 #include "simgrid/plugins/file_system.h"
8 #include "simgrid/s4u/Actor.hpp"
9 #include "src/surf/HostImpl.hpp"
10 #include "xbt/config.hpp"
13 #include <boost/algorithm/string.hpp>
14 #include <boost/algorithm/string/join.hpp>
15 #include <boost/algorithm/string/split.hpp>
19 XBT_LOG_NEW_DEFAULT_CATEGORY(s4u_file, "S4U files");
20 int sg_storage_max_file_descriptors = 1024;
24 simgrid::xbt::Extension<Storage, FileSystemStorageExt> FileSystemStorageExt::EXTENSION_ID;
25 simgrid::xbt::Extension<Host, FileDescriptorHostExt> FileDescriptorHostExt::EXTENSION_ID;
27 File::File(std::string fullpath, void* userdata) : File(fullpath, Host::current(), userdata){};
29 File::File(std::string fullpath, sg_host_t host, void* userdata) : fullpath_(fullpath), userdata_(userdata)
31 // this cannot fail because we get a xbt_die if the mountpoint does not exist
32 Storage* st = nullptr;
33 size_t longest_prefix_length = 0;
34 XBT_DEBUG("Search for storage name for '%s' on '%s'", fullpath.c_str(), host->get_cname());
36 for (auto const& mnt : host->getMountedStorages()) {
37 XBT_DEBUG("See '%s'", mnt.first.c_str());
38 mount_point_ = fullpath.substr(0, mnt.first.length());
40 if (mount_point_ == mnt.first && mnt.first.length() > longest_prefix_length) {
41 /* The current mount name is found in the full path and is bigger than the previous*/
42 longest_prefix_length = mnt.first.length();
46 if (longest_prefix_length > 0) { /* Mount point found, split fullpath into mount_name and path+filename*/
47 mount_point_ = fullpath.substr(0, longest_prefix_length);
48 path_ = fullpath.substr(longest_prefix_length, fullpath.length());
50 xbt_die("Can't find mount point for '%s' on '%s'", fullpath.c_str(), host->get_cname());
54 // assign a file descriptor id to the newly opened File
55 FileDescriptorHostExt* ext = host->extension<simgrid::s4u::FileDescriptorHostExt>();
56 if (ext->file_descriptor_table == nullptr) {
57 ext->file_descriptor_table = new std::vector<int>(sg_storage_max_file_descriptors);
58 std::iota(ext->file_descriptor_table->rbegin(), ext->file_descriptor_table->rend(), 0); // Fill with ..., 1, 0.
60 xbt_assert(not ext->file_descriptor_table->empty(), "Too much files are opened! Some have to be closed.");
61 desc_id = ext->file_descriptor_table->back();
62 ext->file_descriptor_table->pop_back();
64 XBT_DEBUG("\tOpen file '%s'", path_.c_str());
65 std::map<std::string, sg_size_t>* content = localStorage->extension<FileSystemStorageExt>()->getContent();
66 // if file does not exist create an empty file
67 auto sz = content->find(path_);
68 if (sz != content->end()) {
72 content->insert({path_, size_});
73 XBT_DEBUG("File '%s' was not found, file created.", path_.c_str());
79 Host::current()->extension<simgrid::s4u::FileDescriptorHostExt>()->file_descriptor_table->push_back(desc_id);
84 XBT_INFO("File Descriptor information:\n"
85 "\t\tFull path: '%s'\n"
87 "\t\tMount point: '%s'\n"
88 "\t\tStorage Id: '%s'\n"
89 "\t\tStorage Type: '%s'\n"
90 "\t\tFile Descriptor Id: %d",
91 getPath(), size_, mount_point_.c_str(), localStorage->get_cname(), localStorage->getType(), desc_id);
94 sg_size_t File::read(sg_size_t size)
96 if (size_ == 0) /* Nothing to read, return */
99 /* Find the host where the file is physically located and read it */
100 Host* host = localStorage->getHost();
101 XBT_DEBUG("READ %s on disk '%s'", getPath(), localStorage->get_cname());
102 // if the current position is close to the end of the file, we may not be able to read the requested size
103 sg_size_t read_size = localStorage->read(std::min(size, size_ - current_position_));
104 current_position_ += read_size;
106 if (strcmp(host->get_cname(), Host::current()->get_cname())) {
107 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
108 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), read_size);
109 Host* m_host_list[] = {Host::current(), host};
110 double* flops_amount = new double[2]{0, 0};
111 double* bytes_amount = new double[4]{0, 0, static_cast<double>(read_size), 0};
113 this_actor::parallel_execute(2, m_host_list, flops_amount, bytes_amount);
119 /** \brief Write into a file (local or remote)
121 * \param size of the file to write
122 * \param fd is a the file descriptor
123 * \return the number of bytes successfully write or -1 if an error occurred
125 sg_size_t File::write(sg_size_t size)
127 if (size == 0) /* Nothing to write, return */
130 /* Find the host where the file is physically located (remote or local)*/
131 Host* host = localStorage->getHost();
133 if (strcmp(host->get_cname(), Host::current()->get_cname())) {
134 /* the file is hosted on a remote host, initiate a communication between src and dest hosts for data transfer */
135 XBT_DEBUG("File is on %s remote host, initiate data transfer of %llu bytes.", host->get_cname(), size);
136 Host* m_host_list[] = {Host::current(), host};
137 double* flops_amount = new double[2]{0, 0};
138 double* bytes_amount = new double[4]{0, static_cast<double>(size), 0, 0};
140 this_actor::parallel_execute(2, m_host_list, flops_amount, bytes_amount);
143 XBT_DEBUG("WRITE %s on disk '%s'. size '%llu/%llu'", getPath(), localStorage->get_cname(), size, size_);
144 // If the storage is full before even starting to write
145 if (sg_storage_get_size_used(localStorage) >= sg_storage_get_size(localStorage))
147 /* Substract the part of the file that might disappear from the used sized on the storage element */
148 localStorage->extension<FileSystemStorageExt>()->decrUsedSize(size_ - current_position_);
150 sg_size_t write_size = localStorage->write(size);
151 localStorage->extension<FileSystemStorageExt>()->incrUsedSize(write_size);
153 current_position_ += write_size;
154 size_ = current_position_;
155 std::map<std::string, sg_size_t>* content = localStorage->extension<FileSystemStorageExt>()->getContent();
157 content->erase(path_);
158 content->insert({path_, size_});
163 sg_size_t File::size()
168 void File::seek(sg_offset_t offset)
170 current_position_ = offset;
173 void File::seek(sg_offset_t offset, int origin)
177 current_position_ = offset;
180 current_position_ += offset;
183 current_position_ = size_ + offset;
190 sg_size_t File::tell()
192 return current_position_;
195 void File::move(std::string fullpath)
197 /* Check if the new full path is on the same mount point */
198 if (not strncmp(mount_point_.c_str(), fullpath.c_str(), mount_point_.length())) {
199 std::map<std::string, sg_size_t>* content = localStorage->extension<FileSystemStorageExt>()->getContent();
200 auto sz = content->find(path_);
201 if (sz != content->end()) { // src file exists
202 sg_size_t new_size = sz->second;
203 content->erase(path_);
204 std::string path = fullpath.substr(mount_point_.length(), fullpath.length());
205 content->insert({path.c_str(), new_size});
206 XBT_DEBUG("Move file from %s to %s, size '%llu'", path_.c_str(), fullpath.c_str(), new_size);
208 XBT_WARN("File %s doesn't exist", path_.c_str());
211 XBT_WARN("New full path %s is not on the same mount point: %s.", fullpath.c_str(), mount_point_.c_str());
217 /* Check if the file is on local storage */
218 std::map<std::string, sg_size_t>* content = localStorage->extension<FileSystemStorageExt>()->getContent();
220 if (content->find(path_) == content->end()) {
221 XBT_WARN("File %s is not on disk %s. Impossible to unlink", path_.c_str(), localStorage->get_cname());
224 XBT_DEBUG("UNLINK %s on disk '%s'", path_.c_str(), localStorage->get_cname());
225 localStorage->extension<FileSystemStorageExt>()->decrUsedSize(size_);
227 // Remove the file from storage
228 content->erase(fullpath_);
234 int File::remoteCopy(sg_host_t host, const char* fullpath)
236 /* Find the host where the file is physically located and read it */
237 Storage* storage_src = localStorage;
238 Host* src_host = storage_src->getHost();
240 XBT_DEBUG("READ %s on disk '%s'", getPath(), localStorage->get_cname());
241 // if the current position is close to the end of the file, we may not be able to read the requested size
242 sg_size_t read_size = localStorage->read(size_);
243 current_position_ += read_size;
245 /* Find the host that owns the storage where the file has to be copied */
246 Storage* storage_dest = nullptr;
248 size_t longest_prefix_length = 0;
250 for (auto const& elm : host->getMountedStorages()) {
251 std::string mount_point = std::string(fullpath).substr(0, elm.first.size());
252 if (mount_point == elm.first && elm.first.length() > longest_prefix_length) {
253 /* The current mount name is found in the full path and is bigger than the previous*/
254 longest_prefix_length = elm.first.length();
255 storage_dest = elm.second;
259 if (storage_dest != nullptr) {
260 /* Mount point found, retrieve the host the storage is attached to */
261 dst_host = storage_dest->getHost();
263 XBT_WARN("Can't find mount point for '%s' on destination host '%s'", fullpath, host->get_cname());
267 XBT_DEBUG("Initiate data transfer of %llu bytes between %s and %s.", read_size, src_host->get_cname(),
268 storage_dest->getHost()->get_cname());
269 Host* m_host_list[] = {src_host, dst_host};
270 double* flops_amount = new double[2]{0, 0};
271 double* bytes_amount = new double[4]{0, static_cast<double>(read_size), 0, 0};
273 this_actor::parallel_execute(2, m_host_list, flops_amount, bytes_amount);
275 /* Create file on remote host, write it and close it */
276 File* fd = new File(fullpath, dst_host, nullptr);
277 sg_size_t write_size = fd->localStorage->write(read_size);
278 fd->localStorage->extension<FileSystemStorageExt>()->incrUsedSize(write_size);
279 (*(fd->localStorage->extension<FileSystemStorageExt>()->getContent()))[path_] = size_;
284 int File::remoteMove(sg_host_t host, const char* fullpath)
286 int res = remoteCopy(host, fullpath);
291 FileSystemStorageExt::FileSystemStorageExt(simgrid::s4u::Storage* ptr)
293 content_ = parseContent(ptr->getImpl()->content_name);
294 size_ = ptr->getImpl()->size_;
297 FileSystemStorageExt::~FileSystemStorageExt()
302 std::map<std::string, sg_size_t>* FileSystemStorageExt::parseContent(std::string filename)
304 if (filename.empty())
307 std::map<std::string, sg_size_t>* parse_content = new std::map<std::string, sg_size_t>();
309 std::ifstream* fs = surf_ifsopen(filename);
312 std::vector<std::string> tokens;
314 std::getline(*fs, line);
316 if (line.length() > 0) {
317 boost::split(tokens, line, boost::is_any_of(" \t"), boost::token_compress_on);
318 xbt_assert(tokens.size() == 2, "Parse error in %s: %s", filename.c_str(), line.c_str());
319 sg_size_t size = std::stoull(tokens.at(1));
322 parse_content->insert({tokens.front(), size});
324 } while (not fs->eof());
326 return parse_content;
331 using simgrid::s4u::FileSystemStorageExt;
332 using simgrid::s4u::FileDescriptorHostExt;
334 static void onStorageCreation(simgrid::s4u::Storage& st)
336 st.extension_set(new FileSystemStorageExt(&st));
339 static void onStorageDestruction(simgrid::s4u::Storage& st)
341 delete st.extension<FileSystemStorageExt>();
344 static void onHostCreation(simgrid::s4u::Host& host)
346 host.extension_set<FileDescriptorHostExt>(new FileDescriptorHostExt());
349 /* **************************** Public interface *************************** */
350 void sg_storage_file_system_init()
352 sg_storage_max_file_descriptors = 1024;
353 simgrid::config::bind_flag(sg_storage_max_file_descriptors, "storage/max_file_descriptors",
354 "Maximum number of concurrently opened files per host. Default is 1024");
356 if (not FileSystemStorageExt::EXTENSION_ID.valid()) {
357 FileSystemStorageExt::EXTENSION_ID = simgrid::s4u::Storage::extension_create<FileSystemStorageExt>();
358 simgrid::s4u::Storage::onCreation.connect(&onStorageCreation);
359 simgrid::s4u::Storage::onDestruction.connect(&onStorageDestruction);
362 if (not FileDescriptorHostExt::EXTENSION_ID.valid()) {
363 FileDescriptorHostExt::EXTENSION_ID = simgrid::s4u::Host::extension_create<FileDescriptorHostExt>();
364 simgrid::s4u::Host::onCreation.connect(&onHostCreation);
368 sg_file_t sg_file_open(const char* fullpath, void* data)
370 return new simgrid::s4u::File(fullpath, data);
373 sg_size_t sg_file_read(sg_file_t fd, sg_size_t size)
375 return fd->read(size);
378 sg_size_t sg_file_write(sg_file_t fd, sg_size_t size)
380 return fd->write(size);
383 void sg_file_close(sg_file_t fd)
388 const char* sg_file_get_name(sg_file_t fd)
390 xbt_assert((fd != nullptr), "Invalid file descriptor");
391 return fd->getPath();
394 sg_size_t sg_file_get_size(sg_file_t fd)
399 void sg_file_dump(sg_file_t fd)
404 void* sg_file_get_data(sg_file_t fd)
406 return fd->getUserdata();
409 void sg_file_set_data(sg_file_t fd, void* data)
411 fd->setUserdata(data);
415 * \brief Set the file position indicator in the sg_file_t by adding offset bytes
416 * to the position specified by origin (either SEEK_SET, SEEK_CUR, or SEEK_END).
418 * \param fd : file object that identifies the stream
419 * \param offset : number of bytes to offset from origin
420 * \param origin : Position used as reference for the offset. It is specified by one of the following constants defined
421 * in \<stdio.h\> exclusively to be used as arguments for this function (SEEK_SET = beginning of file,
422 * SEEK_CUR = current position of the file pointer, SEEK_END = end of file)
424 void sg_file_seek(sg_file_t fd, sg_offset_t offset, int origin)
426 fd->seek(offset, origin);
429 sg_size_t sg_file_tell(sg_file_t fd)
434 void sg_file_move(sg_file_t fd, const char* fullpath)
439 void sg_file_unlink(sg_file_t fd)
446 * \brief Copy a file to another location on a remote host.
447 * \param file : the file to move
448 * \param host : the remote host where the file has to be copied
449 * \param fullpath : the complete path destination on the remote host
450 * \return If successful, the function returns 0. Otherwise, it returns -1.
452 int sg_file_rcopy(sg_file_t file, sg_host_t host, const char* fullpath)
454 return file->remoteCopy(host, fullpath);
458 * \brief Move a file to another location on a remote host.
459 * \param file : the file to move
460 * \param host : the remote host where the file has to be moved
461 * \param fullpath : the complete path destination on the remote host
462 * \return If successful, the function returns 0. Otherwise, it returns -1.
464 int sg_file_rmove(sg_file_t file, sg_host_t host, const char* fullpath)
466 return file->remoteMove(host, fullpath);
469 sg_size_t sg_storage_get_size_free(sg_storage_t st)
471 return st->extension<FileSystemStorageExt>()->getSize() - st->extension<FileSystemStorageExt>()->getUsedSize();
474 sg_size_t sg_storage_get_size_used(sg_storage_t st)
476 return st->extension<FileSystemStorageExt>()->getUsedSize();
479 sg_size_t sg_storage_get_size(sg_storage_t st)
481 return st->extension<FileSystemStorageExt>()->getSize();
484 xbt_dict_t sg_storage_get_content(sg_storage_t storage)
486 std::map<std::string, sg_size_t>* content = storage->extension<simgrid::s4u::FileSystemStorageExt>()->getContent();
487 // Note: ::operator delete is ok here (no destructor called) since the dict elements are of POD type sg_size_t.
488 xbt_dict_t content_as_dict = xbt_dict_new_homogeneous(::operator delete);
490 for (auto const& entry : *content) {
491 sg_size_t* psize = new sg_size_t;
492 *psize = entry.second;
493 xbt_dict_set(content_as_dict, entry.first.c_str(), psize, nullptr);
495 return content_as_dict;
498 xbt_dict_t sg_host_get_storage_content(sg_host_t host)
500 xbt_assert((host != nullptr), "Invalid parameters");
501 xbt_dict_t contents = xbt_dict_new_homogeneous(nullptr);
502 for (auto const& elm : host->getMountedStorages())
503 xbt_dict_set(contents, elm.first.c_str(), sg_storage_get_content(elm.second), nullptr);