Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 1 | #include "config.h" |
| 2 | |
Gunnar Mills | 94df8c9 | 2018-09-14 14:50:03 -0500 | [diff] [blame] | 3 | #include "occ_manager.hpp" |
| 4 | |
| 5 | #include "i2c_occ.hpp" |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 6 | #include "occ_dbus.hpp" |
Gunnar Mills | 94df8c9 | 2018-09-14 14:50:03 -0500 | [diff] [blame] | 7 | #include "utils.hpp" |
| 8 | |
George Liu | b5ca101 | 2021-09-10 12:53:11 +0800 | [diff] [blame] | 9 | #include <phosphor-logging/elog-errors.hpp> |
| 10 | #include <phosphor-logging/log.hpp> |
| 11 | #include <xyz/openbmc_project/Common/error.hpp> |
| 12 | |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 13 | #include <chrono> |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 14 | #include <cmath> |
George Liu | bcef3b4 | 2021-09-10 12:39:02 +0800 | [diff] [blame] | 15 | #include <filesystem> |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 16 | #include <fstream> |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 17 | #include <regex> |
Gunnar Mills | 94df8c9 | 2018-09-14 14:50:03 -0500 | [diff] [blame] | 18 | |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 19 | namespace open_power |
| 20 | { |
| 21 | namespace occ |
| 22 | { |
| 23 | |
Matt Spinler | 8b8abee | 2021-08-25 15:18:21 -0500 | [diff] [blame] | 24 | constexpr uint32_t fruTypeNotAvailable = 0xFF; |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 25 | constexpr auto fruTypeSuffix = "fru_type"; |
| 26 | constexpr auto faultSuffix = "fault"; |
| 27 | constexpr auto inputSuffix = "input"; |
Matt Spinler | ace67d8 | 2021-10-18 13:41:57 -0500 | [diff] [blame] | 28 | constexpr auto maxSuffix = "max"; |
Matt Spinler | 8b8abee | 2021-08-25 15:18:21 -0500 | [diff] [blame] | 29 | |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 30 | const auto HOST_ON_FILE = "/run/openbmc/host@0-on"; |
| 31 | |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 32 | using namespace phosphor::logging; |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 33 | using namespace std::literals::chrono_literals; |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 34 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 35 | template <typename T> |
| 36 | T readFile(const std::string& path) |
| 37 | { |
| 38 | std::ifstream ifs; |
| 39 | ifs.exceptions(std::ifstream::failbit | std::ifstream::badbit | |
| 40 | std::ifstream::eofbit); |
| 41 | T data; |
| 42 | |
| 43 | try |
| 44 | { |
| 45 | ifs.open(path); |
| 46 | ifs >> data; |
| 47 | ifs.close(); |
| 48 | } |
| 49 | catch (const std::exception& e) |
| 50 | { |
| 51 | auto err = errno; |
| 52 | throw std::system_error(err, std::generic_category()); |
| 53 | } |
| 54 | |
| 55 | return data; |
| 56 | } |
| 57 | |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 58 | void Manager::findAndCreateObjects() |
| 59 | { |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 60 | #ifndef POWER10 |
Deepak Kodihalli | 370f06b | 2017-10-25 04:26:07 -0500 | [diff] [blame] | 61 | for (auto id = 0; id < MAX_CPUS; ++id) |
| 62 | { |
Deepak Kodihalli | 30417a1 | 2017-12-04 00:54:01 -0600 | [diff] [blame] | 63 | // Create one occ per cpu |
| 64 | auto occ = std::string(OCC_NAME) + std::to_string(id); |
| 65 | createObjects(occ); |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 66 | } |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 67 | #else |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 68 | if (!fs::exists(HOST_ON_FILE)) |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 69 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 70 | static bool statusObjCreated = false; |
| 71 | if (!statusObjCreated) |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 72 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 73 | // Create the OCCs based on on the /dev/occX devices |
| 74 | auto occs = findOCCsInDev(); |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 75 | |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 76 | if (occs.empty() || (prevOCCSearch.size() != occs.size())) |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 77 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 78 | // Something changed or no OCCs yet, try again in 10s. |
| 79 | // Note on the first pass prevOCCSearch will be empty, |
| 80 | // so there will be at least one delay to give things |
| 81 | // a chance to settle. |
| 82 | prevOCCSearch = occs; |
| 83 | |
| 84 | log<level::INFO>( |
| 85 | fmt::format( |
| 86 | "Manager::findAndCreateObjects(): Waiting for OCCs (currently {})", |
| 87 | occs.size()) |
| 88 | .c_str()); |
| 89 | |
| 90 | discoverTimer->restartOnce(10s); |
| 91 | } |
| 92 | else |
| 93 | { |
| 94 | // All OCCs appear to be available, create status objects |
| 95 | |
| 96 | // createObjects requires OCC0 first. |
| 97 | std::sort(occs.begin(), occs.end()); |
| 98 | |
| 99 | log<level::INFO>( |
| 100 | fmt::format( |
| 101 | "Manager::findAndCreateObjects(): Creating {} OCC Status Objects", |
| 102 | occs.size()) |
| 103 | .c_str()); |
| 104 | for (auto id : occs) |
| 105 | { |
| 106 | createObjects(std::string(OCC_NAME) + std::to_string(id)); |
| 107 | } |
| 108 | statusObjCreated = true; |
| 109 | } |
| 110 | } |
| 111 | |
| 112 | if (statusObjCreated) |
| 113 | { |
| 114 | static bool tracedHostWait = false; |
| 115 | if (utils::isHostRunning()) |
| 116 | { |
| 117 | if (tracedHostWait) |
| 118 | { |
| 119 | log<level::INFO>( |
| 120 | "Manager::findAndCreateObjects(): Host is running"); |
| 121 | tracedHostWait = false; |
| 122 | } |
| 123 | waitingForAllOccActiveSensors = true; |
| 124 | checkAllActiveSensors(); |
| 125 | } |
| 126 | else |
| 127 | { |
| 128 | if (!tracedHostWait) |
| 129 | { |
| 130 | log<level::INFO>( |
| 131 | "Manager::findAndCreateObjects(): Waiting for host to start"); |
| 132 | tracedHostWait = true; |
| 133 | } |
| 134 | discoverTimer->restartOnce(30s); |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 135 | } |
| 136 | } |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 137 | } |
| 138 | else |
| 139 | { |
Chris Cain | 1718fd8 | 2022-02-16 16:39:50 -0600 | [diff] [blame] | 140 | log<level::INFO>( |
| 141 | fmt::format( |
| 142 | "Manager::findAndCreateObjects(): Waiting for {} to complete...", |
| 143 | HOST_ON_FILE) |
| 144 | .c_str()); |
| 145 | discoverTimer->restartOnce(10s); |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 146 | } |
| 147 | #endif |
| 148 | } |
| 149 | |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 150 | #ifdef POWER10 |
| 151 | // Check if all occActive sensors are available |
| 152 | void Manager::checkAllActiveSensors() |
| 153 | { |
| 154 | static bool allActiveSensorAvailable = false; |
| 155 | static bool tracedSensorWait = false; |
| 156 | |
| 157 | // Start with the assumption that all are available |
| 158 | allActiveSensorAvailable = true; |
| 159 | for (auto& obj : statusObjects) |
| 160 | { |
| 161 | // If active sensor is already true, then no need to query sensor |
| 162 | if (!obj->occActive()) |
| 163 | { |
| 164 | allActiveSensorAvailable = false; |
| 165 | if (!tracedSensorWait) |
| 166 | { |
| 167 | log<level::INFO>( |
| 168 | fmt::format( |
| 169 | "Manager::checkAllActiveSensors(): Waiting on OCC{} Active sensor", |
| 170 | obj->getOccInstanceID()) |
| 171 | .c_str()); |
| 172 | tracedSensorWait = true; |
| 173 | } |
| 174 | pldmHandle->checkActiveSensor(obj->getOccInstanceID()); |
| 175 | break; |
| 176 | } |
| 177 | } |
| 178 | |
| 179 | if (allActiveSensorAvailable) |
| 180 | { |
| 181 | // All sensors were found, disable the discovery timer |
| 182 | discoverTimer.reset(); |
| 183 | waitingForAllOccActiveSensors = false; |
| 184 | |
| 185 | log<level::INFO>( |
| 186 | "Manager::checkAllActiveSensors(): OCC Active sensors are available"); |
| 187 | tracedSensorWait = false; |
| 188 | } |
| 189 | else |
| 190 | { |
| 191 | // Not all sensors were available, so keep waiting |
| 192 | if (!tracedSensorWait) |
| 193 | { |
| 194 | log<level::INFO>( |
| 195 | "Manager::checkAllActiveSensors(): Waiting for OCC Active sensors to become available"); |
| 196 | tracedSensorWait = true; |
| 197 | } |
| 198 | discoverTimer->restartOnce(30s); |
| 199 | } |
| 200 | } |
| 201 | #endif |
| 202 | |
Matt Spinler | d267cec | 2021-09-01 14:49:19 -0500 | [diff] [blame] | 203 | std::vector<int> Manager::findOCCsInDev() |
| 204 | { |
| 205 | std::vector<int> occs; |
| 206 | std::regex expr{R"(occ(\d+)$)"}; |
| 207 | |
| 208 | for (auto& file : fs::directory_iterator("/dev")) |
| 209 | { |
| 210 | std::smatch match; |
| 211 | std::string path{file.path().string()}; |
| 212 | if (std::regex_search(path, match, expr)) |
| 213 | { |
| 214 | auto num = std::stoi(match[1].str()); |
| 215 | |
| 216 | // /dev numbering starts at 1, ours starts at 0. |
| 217 | occs.push_back(num - 1); |
| 218 | } |
| 219 | } |
| 220 | |
| 221 | return occs; |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 222 | } |
| 223 | |
| 224 | int Manager::cpuCreated(sdbusplus::message::message& msg) |
| 225 | { |
George Liu | bcef3b4 | 2021-09-10 12:39:02 +0800 | [diff] [blame] | 226 | namespace fs = std::filesystem; |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 227 | |
| 228 | sdbusplus::message::object_path o; |
| 229 | msg.read(o); |
| 230 | fs::path cpuPath(std::string(std::move(o))); |
| 231 | |
| 232 | auto name = cpuPath.filename().string(); |
| 233 | auto index = name.find(CPU_NAME); |
| 234 | name.replace(index, std::strlen(CPU_NAME), OCC_NAME); |
| 235 | |
| 236 | createObjects(name); |
| 237 | |
| 238 | return 0; |
| 239 | } |
| 240 | |
| 241 | void Manager::createObjects(const std::string& occ) |
| 242 | { |
| 243 | auto path = fs::path(OCC_CONTROL_ROOT) / occ; |
| 244 | |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 245 | #ifdef POWER10 |
| 246 | if (!pmode) |
| 247 | { |
Chris Cain | 1be4337 | 2021-12-09 19:29:37 -0600 | [diff] [blame] | 248 | // Create the power mode object |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 249 | pmode = std::make_unique<powermode::PowerMode>( |
Chris Cain | 1be4337 | 2021-12-09 19:29:37 -0600 | [diff] [blame] | 250 | *this, powermode::PMODE_PATH, powermode::PIPS_PATH); |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 251 | } |
| 252 | #endif |
| 253 | |
Gunnar Mills | 94df8c9 | 2018-09-14 14:50:03 -0500 | [diff] [blame] | 254 | statusObjects.emplace_back(std::make_unique<Status>( |
George Liu | f3b7514 | 2021-06-10 11:22:50 +0800 | [diff] [blame] | 255 | event, path.c_str(), *this, |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 256 | #ifdef POWER10 |
| 257 | pmode, |
| 258 | #endif |
Gunnar Mills | 94df8c9 | 2018-09-14 14:50:03 -0500 | [diff] [blame] | 259 | std::bind(std::mem_fn(&Manager::statusCallBack), this, |
Sheldon Bailey | 373af75 | 2022-02-21 15:14:00 -0600 | [diff] [blame] | 260 | std::placeholders::_1, std::placeholders::_2) |
Tom Joseph | 0032523 | 2020-07-29 17:51:48 +0530 | [diff] [blame] | 261 | #ifdef PLDM |
| 262 | , |
| 263 | std::bind(std::mem_fn(&pldm::Interface::resetOCC), pldmHandle.get(), |
| 264 | std::placeholders::_1) |
| 265 | #endif |
| 266 | )); |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 267 | |
Chris Cain | 40501a2 | 2022-03-14 17:33:27 -0500 | [diff] [blame] | 268 | // Create the power cap monitor object |
| 269 | if (!pcap) |
| 270 | { |
| 271 | pcap = std::make_unique<open_power::occ::powercap::PowerCap>( |
| 272 | *statusObjects.back()); |
| 273 | } |
| 274 | |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 275 | if (statusObjects.back()->isMasterOcc()) |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 276 | { |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 277 | log<level::INFO>( |
| 278 | fmt::format("Manager::createObjects(): OCC{} is the master", |
| 279 | statusObjects.back()->getOccInstanceID()) |
| 280 | .c_str()); |
| 281 | _pollTimer->setEnabled(false); |
| 282 | |
Chris Cain | 78e8601 | 2021-03-04 16:15:31 -0600 | [diff] [blame] | 283 | #ifdef POWER10 |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 284 | // Set the master OCC on the PowerMode object |
| 285 | pmode->setMasterOcc(path); |
Chris Cain | 78e8601 | 2021-03-04 16:15:31 -0600 | [diff] [blame] | 286 | #endif |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 287 | } |
| 288 | |
| 289 | passThroughObjects.emplace_back(std::make_unique<PassThrough>(path.c_str() |
| 290 | #ifdef POWER10 |
| 291 | , |
| 292 | pmode |
| 293 | #endif |
| 294 | )); |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 295 | } |
| 296 | |
Sheldon Bailey | 373af75 | 2022-02-21 15:14:00 -0600 | [diff] [blame] | 297 | void Manager::statusCallBack(instanceID instance, bool status) |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 298 | { |
Gunnar Mills | 94df8c9 | 2018-09-14 14:50:03 -0500 | [diff] [blame] | 299 | using InternalFailure = |
| 300 | sdbusplus::xyz::openbmc_project::Common::Error::InternalFailure; |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 301 | |
| 302 | // At this time, it won't happen but keeping it |
| 303 | // here just in case something changes in the future |
| 304 | if ((activeCount == 0) && (!status)) |
| 305 | { |
Sheldon Bailey | 373af75 | 2022-02-21 15:14:00 -0600 | [diff] [blame] | 306 | log<level::ERR>( |
| 307 | fmt::format("Invalid update on OCCActive with OCC{}", instance) |
| 308 | .c_str()); |
| 309 | |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 310 | elog<InternalFailure>(); |
| 311 | } |
| 312 | |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 313 | if (status == true) |
Eddie James | dae2d94 | 2017-12-20 10:50:03 -0600 | [diff] [blame] | 314 | { |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 315 | // OCC went active |
| 316 | ++activeCount; |
| 317 | |
| 318 | #ifdef POWER10 |
| 319 | if (activeCount == 1) |
Eddie James | dae2d94 | 2017-12-20 10:50:03 -0600 | [diff] [blame] | 320 | { |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 321 | // First OCC went active (allow some time for all OCCs to go active) |
| 322 | waitForAllOccsTimer->restartOnce(30s); |
Matt Spinler | 53f6814 | 2021-08-25 15:47:31 -0500 | [diff] [blame] | 323 | } |
| 324 | #endif |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 325 | |
| 326 | if (activeCount == statusObjects.size()) |
| 327 | { |
| 328 | #ifdef POWER10 |
| 329 | // All OCCs are now running |
| 330 | if (waitForAllOccsTimer->isEnabled()) |
| 331 | { |
| 332 | // stop occ wait timer |
| 333 | waitForAllOccsTimer->setEnabled(false); |
| 334 | } |
| 335 | #endif |
| 336 | |
| 337 | // Verify master OCC and start presence monitor |
| 338 | validateOccMaster(); |
| 339 | } |
| 340 | |
| 341 | // Start poll timer if not already started |
| 342 | if (!_pollTimer->isEnabled()) |
| 343 | { |
| 344 | log<level::INFO>( |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 345 | fmt::format("Manager: OCCs will be polled every {} seconds", |
| 346 | pollInterval) |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 347 | .c_str()); |
| 348 | |
| 349 | // Send poll and start OCC poll timer |
| 350 | pollerTimerExpired(); |
| 351 | } |
| 352 | } |
| 353 | else |
| 354 | { |
| 355 | // OCC went away |
| 356 | --activeCount; |
| 357 | |
| 358 | if (activeCount == 0) |
| 359 | { |
| 360 | // No OCCs are running |
| 361 | |
| 362 | // Stop OCC poll timer |
| 363 | if (_pollTimer->isEnabled()) |
| 364 | { |
| 365 | log<level::INFO>( |
| 366 | "Manager::statusCallBack(): OCCs are not running, stopping poll timer"); |
| 367 | _pollTimer->setEnabled(false); |
| 368 | } |
| 369 | |
| 370 | #ifdef POWER10 |
| 371 | // stop wait timer |
| 372 | if (waitForAllOccsTimer->isEnabled()) |
| 373 | { |
| 374 | waitForAllOccsTimer->setEnabled(false); |
| 375 | } |
| 376 | #endif |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 377 | } |
Sheldon Bailey | 373af75 | 2022-02-21 15:14:00 -0600 | [diff] [blame] | 378 | #ifdef READ_OCC_SENSORS |
| 379 | // Clear OCC sensors |
| 380 | setSensorValueToNonFunctional(instance); |
| 381 | #endif |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 382 | } |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 383 | |
| 384 | #ifdef POWER10 |
| 385 | if (waitingForAllOccActiveSensors) |
| 386 | { |
| 387 | checkAllActiveSensors(); |
| 388 | } |
| 389 | #endif |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 390 | } |
| 391 | |
| 392 | #ifdef I2C_OCC |
| 393 | void Manager::initStatusObjects() |
| 394 | { |
| 395 | // Make sure we have a valid path string |
| 396 | static_assert(sizeof(DEV_PATH) != 0); |
| 397 | |
| 398 | auto deviceNames = i2c_occ::getOccHwmonDevices(DEV_PATH); |
| 399 | for (auto& name : deviceNames) |
| 400 | { |
| 401 | i2c_occ::i2cToDbus(name); |
Lei YU | b5259a1 | 2017-09-01 16:22:40 +0800 | [diff] [blame] | 402 | name = std::string(OCC_NAME) + '_' + name; |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 403 | auto path = fs::path(OCC_CONTROL_ROOT) / name; |
| 404 | statusObjects.emplace_back( |
George Liu | f3b7514 | 2021-06-10 11:22:50 +0800 | [diff] [blame] | 405 | std::make_unique<Status>(event, path.c_str(), *this)); |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 406 | } |
Chris Cain | 40501a2 | 2022-03-14 17:33:27 -0500 | [diff] [blame] | 407 | // The first device is master occ |
| 408 | pcap = std::make_unique<open_power::occ::powercap::PowerCap>( |
| 409 | *statusObjects.front()); |
Chris Cain | 78e8601 | 2021-03-04 16:15:31 -0600 | [diff] [blame] | 410 | #ifdef POWER10 |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 411 | pmode = std::make_unique<powermode::PowerMode>(*this, powermode::PMODE_PATH, |
| 412 | powermode::PIPS_PATH); |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 413 | // Set the master OCC on the PowerMode object |
| 414 | pmode->setMasterOcc(path); |
Chris Cain | 78e8601 | 2021-03-04 16:15:31 -0600 | [diff] [blame] | 415 | #endif |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 416 | } |
| 417 | #endif |
| 418 | |
Tom Joseph | 815f9f5 | 2020-07-27 12:12:13 +0530 | [diff] [blame] | 419 | #ifdef PLDM |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 420 | void Manager::sbeTimeout(unsigned int instance) |
| 421 | { |
Eddie James | 2a751d7 | 2022-03-04 09:16:12 -0600 | [diff] [blame] | 422 | auto obj = std::find_if(statusObjects.begin(), statusObjects.end(), |
| 423 | [instance](const auto& obj) { |
| 424 | return instance == obj->getOccInstanceID(); |
| 425 | }); |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 426 | |
Eddie James | cb018da | 2022-03-05 11:49:37 -0600 | [diff] [blame] | 427 | if (obj != statusObjects.end() && (*obj)->occActive()) |
Eddie James | 2a751d7 | 2022-03-04 09:16:12 -0600 | [diff] [blame] | 428 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 429 | log<level::INFO>( |
| 430 | fmt::format("SBE timeout, requesting HRESET (OCC{})", instance) |
| 431 | .c_str()); |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 432 | |
Eddie James | 2a751d7 | 2022-03-04 09:16:12 -0600 | [diff] [blame] | 433 | setSBEState(instance, SBE_STATE_NOT_USABLE); |
| 434 | |
| 435 | pldmHandle->sendHRESET(instance); |
| 436 | } |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 437 | } |
| 438 | |
Tom Joseph | 815f9f5 | 2020-07-27 12:12:13 +0530 | [diff] [blame] | 439 | bool Manager::updateOCCActive(instanceID instance, bool status) |
| 440 | { |
Chris Cain | 7e374fb | 2022-04-07 09:47:23 -0500 | [diff] [blame] | 441 | auto obj = std::find_if(statusObjects.begin(), statusObjects.end(), |
| 442 | [instance](const auto& obj) { |
| 443 | return instance == obj->getOccInstanceID(); |
| 444 | }); |
| 445 | |
| 446 | if (obj != statusObjects.end()) |
| 447 | { |
| 448 | return (*obj)->occActive(status); |
| 449 | } |
| 450 | else |
| 451 | { |
| 452 | log<level::WARNING>( |
| 453 | fmt::format( |
| 454 | "Manager::updateOCCActive: No status object to update for OCC{} (active={})", |
| 455 | instance, status) |
| 456 | .c_str()); |
| 457 | return false; |
| 458 | } |
Tom Joseph | 815f9f5 | 2020-07-27 12:12:13 +0530 | [diff] [blame] | 459 | } |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 460 | |
| 461 | void Manager::sbeHRESETResult(instanceID instance, bool success) |
| 462 | { |
| 463 | if (success) |
| 464 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 465 | log<level::INFO>( |
| 466 | fmt::format("HRESET succeeded (OCC{})", instance).c_str()); |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 467 | |
| 468 | setSBEState(instance, SBE_STATE_BOOTED); |
| 469 | |
| 470 | return; |
| 471 | } |
| 472 | |
| 473 | setSBEState(instance, SBE_STATE_FAILED); |
| 474 | |
| 475 | if (sbeCanDump(instance)) |
| 476 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 477 | log<level::INFO>( |
| 478 | fmt::format("HRESET failed (OCC{}), triggering SBE dump", instance) |
| 479 | .c_str()); |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 480 | |
| 481 | auto& bus = utils::getBus(); |
| 482 | uint32_t src6 = instance << 16; |
| 483 | uint32_t logId = |
| 484 | FFDC::createPEL("org.open_power.Processor.Error.SbeChipOpTimeout", |
| 485 | src6, "SBE command timeout"); |
| 486 | |
| 487 | try |
| 488 | { |
George Liu | f3a4a69 | 2021-12-28 13:59:51 +0800 | [diff] [blame] | 489 | constexpr auto path = "/org/openpower/dump"; |
| 490 | constexpr auto interface = "xyz.openbmc_project.Dump.Create"; |
| 491 | constexpr auto function = "CreateDump"; |
| 492 | |
Eddie James | cbad219 | 2021-10-07 09:39:39 -0500 | [diff] [blame] | 493 | std::string service = utils::getService(path, interface); |
| 494 | auto method = |
| 495 | bus.new_method_call(service.c_str(), path, interface, function); |
| 496 | |
| 497 | std::map<std::string, std::variant<std::string, uint64_t>> |
| 498 | createParams{ |
| 499 | {"com.ibm.Dump.Create.CreateParameters.ErrorLogId", |
| 500 | uint64_t(logId)}, |
| 501 | {"com.ibm.Dump.Create.CreateParameters.DumpType", |
| 502 | "com.ibm.Dump.Create.DumpType.SBE"}, |
| 503 | {"com.ibm.Dump.Create.CreateParameters.FailingUnitId", |
| 504 | uint64_t(instance)}, |
| 505 | }; |
| 506 | |
| 507 | method.append(createParams); |
| 508 | |
| 509 | auto response = bus.call(method); |
| 510 | } |
| 511 | catch (const sdbusplus::exception::exception& e) |
| 512 | { |
| 513 | constexpr auto ERROR_DUMP_DISABLED = |
| 514 | "xyz.openbmc_project.Dump.Create.Error.Disabled"; |
| 515 | if (e.name() == ERROR_DUMP_DISABLED) |
| 516 | { |
| 517 | log<level::INFO>("Dump is disabled, skipping"); |
| 518 | } |
| 519 | else |
| 520 | { |
| 521 | log<level::ERR>("Dump failed"); |
| 522 | } |
| 523 | } |
| 524 | } |
| 525 | } |
| 526 | |
| 527 | bool Manager::sbeCanDump(unsigned int instance) |
| 528 | { |
| 529 | struct pdbg_target* proc = getPdbgTarget(instance); |
| 530 | |
| 531 | if (!proc) |
| 532 | { |
| 533 | // allow the dump in the error case |
| 534 | return true; |
| 535 | } |
| 536 | |
| 537 | try |
| 538 | { |
| 539 | if (!openpower::phal::sbe::isDumpAllowed(proc)) |
| 540 | { |
| 541 | return false; |
| 542 | } |
| 543 | |
| 544 | if (openpower::phal::pdbg::isSbeVitalAttnActive(proc)) |
| 545 | { |
| 546 | return false; |
| 547 | } |
| 548 | } |
| 549 | catch (openpower::phal::exception::SbeError& e) |
| 550 | { |
| 551 | log<level::INFO>("Failed to query SBE state"); |
| 552 | } |
| 553 | |
| 554 | // allow the dump in the error case |
| 555 | return true; |
| 556 | } |
| 557 | |
| 558 | void Manager::setSBEState(unsigned int instance, enum sbe_state state) |
| 559 | { |
| 560 | struct pdbg_target* proc = getPdbgTarget(instance); |
| 561 | |
| 562 | if (!proc) |
| 563 | { |
| 564 | return; |
| 565 | } |
| 566 | |
| 567 | try |
| 568 | { |
| 569 | openpower::phal::sbe::setState(proc, state); |
| 570 | } |
| 571 | catch (const openpower::phal::exception::SbeError& e) |
| 572 | { |
| 573 | log<level::ERR>("Failed to set SBE state"); |
| 574 | } |
| 575 | } |
| 576 | |
| 577 | struct pdbg_target* Manager::getPdbgTarget(unsigned int instance) |
| 578 | { |
| 579 | if (!pdbgInitialized) |
| 580 | { |
| 581 | try |
| 582 | { |
| 583 | openpower::phal::pdbg::init(); |
| 584 | pdbgInitialized = true; |
| 585 | } |
| 586 | catch (const openpower::phal::exception::PdbgError& e) |
| 587 | { |
| 588 | log<level::ERR>("pdbg initialization failed"); |
| 589 | return nullptr; |
| 590 | } |
| 591 | } |
| 592 | |
| 593 | struct pdbg_target* proc = nullptr; |
| 594 | pdbg_for_each_class_target("proc", proc) |
| 595 | { |
| 596 | if (pdbg_target_index(proc) == instance) |
| 597 | { |
| 598 | return proc; |
| 599 | } |
| 600 | } |
| 601 | |
| 602 | log<level::ERR>("Failed to get pdbg target"); |
| 603 | return nullptr; |
| 604 | } |
Tom Joseph | 815f9f5 | 2020-07-27 12:12:13 +0530 | [diff] [blame] | 605 | #endif |
| 606 | |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 607 | void Manager::pollerTimerExpired() |
| 608 | { |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 609 | if (!_pollTimer) |
| 610 | { |
| 611 | log<level::ERR>( |
| 612 | "Manager::pollerTimerExpired() ERROR: Timer not defined"); |
| 613 | return; |
| 614 | } |
| 615 | |
| 616 | for (auto& obj : statusObjects) |
| 617 | { |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 618 | if (!obj->occActive()) |
| 619 | { |
| 620 | // OCC is not running yet |
| 621 | #ifdef READ_OCC_SENSORS |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 622 | auto id = obj->getOccInstanceID(); |
Sheldon Bailey | 373af75 | 2022-02-21 15:14:00 -0600 | [diff] [blame] | 623 | setSensorValueToNonFunctional(id); |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 624 | #endif |
| 625 | continue; |
| 626 | } |
| 627 | |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 628 | // Read sysfs to force kernel to poll OCC |
| 629 | obj->readOccState(); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 630 | |
| 631 | #ifdef READ_OCC_SENSORS |
| 632 | // Read occ sensor values |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 633 | getSensorValues(obj); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 634 | #endif |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 635 | } |
| 636 | |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 637 | if (activeCount > 0) |
| 638 | { |
| 639 | // Restart OCC poll timer |
| 640 | _pollTimer->restartOnce(std::chrono::seconds(pollInterval)); |
| 641 | } |
| 642 | else |
| 643 | { |
| 644 | // No OCCs running, so poll timer will not be restarted |
| 645 | log<level::INFO>( |
| 646 | fmt::format( |
| 647 | "Manager::pollerTimerExpired: poll timer will not be restarted") |
| 648 | .c_str()); |
| 649 | } |
Chris Cain | a8857c5 | 2021-01-27 11:53:05 -0600 | [diff] [blame] | 650 | } |
| 651 | |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 652 | #ifdef READ_OCC_SENSORS |
| 653 | void Manager::readTempSensors(const fs::path& path, uint32_t id) |
| 654 | { |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 655 | std::regex expr{"temp\\d+_label$"}; // Example: temp5_label |
| 656 | for (auto& file : fs::directory_iterator(path)) |
| 657 | { |
| 658 | if (!std::regex_search(file.path().string(), expr)) |
| 659 | { |
| 660 | continue; |
| 661 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 662 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 663 | uint32_t labelValue{0}; |
| 664 | |
| 665 | try |
| 666 | { |
| 667 | labelValue = readFile<uint32_t>(file.path()); |
| 668 | } |
| 669 | catch (const std::system_error& e) |
| 670 | { |
| 671 | log<level::DEBUG>( |
| 672 | fmt::format("readTempSensors: Failed reading {}, errno = {}", |
| 673 | file.path().string(), e.code().value()) |
| 674 | .c_str()); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 675 | continue; |
| 676 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 677 | |
| 678 | const std::string& tempLabel = "label"; |
| 679 | const std::string filePathString = file.path().string().substr( |
| 680 | 0, file.path().string().length() - tempLabel.length()); |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 681 | |
| 682 | uint32_t fruTypeValue{0}; |
| 683 | try |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 684 | { |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 685 | fruTypeValue = readFile<uint32_t>(filePathString + fruTypeSuffix); |
| 686 | } |
| 687 | catch (const std::system_error& e) |
| 688 | { |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 689 | log<level::DEBUG>( |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 690 | fmt::format("readTempSensors: Failed reading {}, errno = {}", |
| 691 | filePathString + fruTypeSuffix, e.code().value()) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 692 | .c_str()); |
| 693 | continue; |
| 694 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 695 | |
| 696 | std::string sensorPath = |
| 697 | OCC_SENSORS_ROOT + std::string("/temperature/"); |
| 698 | |
Matt Spinler | ace67d8 | 2021-10-18 13:41:57 -0500 | [diff] [blame] | 699 | std::string dvfsTempPath; |
| 700 | |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 701 | if (fruTypeValue == VRMVdd) |
| 702 | { |
| 703 | sensorPath.append("vrm_vdd" + std::to_string(id) + "_temp"); |
| 704 | } |
Matt Spinler | ace67d8 | 2021-10-18 13:41:57 -0500 | [diff] [blame] | 705 | else if (fruTypeValue == processorIoRing) |
| 706 | { |
| 707 | sensorPath.append("proc" + std::to_string(id) + "_ioring_temp"); |
| 708 | dvfsTempPath = std::string{OCC_SENSORS_ROOT} + "/temperature/proc" + |
| 709 | std::to_string(id) + "_ioring_dvfs_temp"; |
| 710 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 711 | else |
| 712 | { |
Matt Spinler | 14d1402 | 2021-08-25 15:38:29 -0500 | [diff] [blame] | 713 | uint16_t type = (labelValue & 0xFF000000) >> 24; |
| 714 | uint16_t instanceID = labelValue & 0x0000FFFF; |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 715 | |
| 716 | if (type == OCC_DIMM_TEMP_SENSOR_TYPE) |
| 717 | { |
Matt Spinler | 8b8abee | 2021-08-25 15:18:21 -0500 | [diff] [blame] | 718 | if (fruTypeValue == fruTypeNotAvailable) |
| 719 | { |
| 720 | // Not all DIMM related temps are available to read |
| 721 | // (no _input file in this case) |
| 722 | continue; |
| 723 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 724 | auto iter = dimmTempSensorName.find(fruTypeValue); |
| 725 | if (iter == dimmTempSensorName.end()) |
| 726 | { |
George Liu | b5ca101 | 2021-09-10 12:53:11 +0800 | [diff] [blame] | 727 | log<level::ERR>( |
| 728 | fmt::format( |
| 729 | "readTempSensors: Fru type error! fruTypeValue = {}) ", |
| 730 | fruTypeValue) |
| 731 | .c_str()); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 732 | continue; |
| 733 | } |
| 734 | |
| 735 | sensorPath.append("dimm" + std::to_string(instanceID) + |
| 736 | iter->second); |
| 737 | } |
| 738 | else if (type == OCC_CPU_TEMP_SENSOR_TYPE) |
| 739 | { |
Matt Spinler | ace67d8 | 2021-10-18 13:41:57 -0500 | [diff] [blame] | 740 | if (fruTypeValue == processorCore) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 741 | { |
Matt Spinler | ace67d8 | 2021-10-18 13:41:57 -0500 | [diff] [blame] | 742 | // The OCC reports small core temps, of which there are |
| 743 | // two per big core. All current P10 systems are in big |
| 744 | // core mode, so use a big core name. |
| 745 | uint16_t coreNum = instanceID / 2; |
| 746 | uint16_t tempNum = instanceID % 2; |
| 747 | sensorPath.append("proc" + std::to_string(id) + "_core" + |
| 748 | std::to_string(coreNum) + "_" + |
| 749 | std::to_string(tempNum) + "_temp"); |
| 750 | |
| 751 | dvfsTempPath = std::string{OCC_SENSORS_ROOT} + |
| 752 | "/temperature/proc" + std::to_string(id) + |
| 753 | "_core_dvfs_temp"; |
| 754 | } |
| 755 | else |
| 756 | { |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 757 | continue; |
| 758 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 759 | } |
| 760 | else |
| 761 | { |
| 762 | continue; |
| 763 | } |
| 764 | } |
| 765 | |
Matt Spinler | ace67d8 | 2021-10-18 13:41:57 -0500 | [diff] [blame] | 766 | // The dvfs temp file only needs to be read once per chip per type. |
| 767 | if (!dvfsTempPath.empty() && |
| 768 | !dbus::OccDBusSensors::getOccDBus().hasDvfsTemp(dvfsTempPath)) |
| 769 | { |
| 770 | try |
| 771 | { |
| 772 | auto dvfsValue = readFile<double>(filePathString + maxSuffix); |
| 773 | |
| 774 | dbus::OccDBusSensors::getOccDBus().setDvfsTemp( |
| 775 | dvfsTempPath, dvfsValue * std::pow(10, -3)); |
| 776 | } |
| 777 | catch (const std::system_error& e) |
| 778 | { |
| 779 | log<level::DEBUG>( |
| 780 | fmt::format( |
| 781 | "readTempSensors: Failed reading {}, errno = {}", |
| 782 | filePathString + maxSuffix, e.code().value()) |
| 783 | .c_str()); |
| 784 | } |
| 785 | } |
| 786 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 787 | uint32_t faultValue{0}; |
| 788 | try |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 789 | { |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 790 | faultValue = readFile<uint32_t>(filePathString + faultSuffix); |
| 791 | } |
| 792 | catch (const std::system_error& e) |
| 793 | { |
| 794 | log<level::DEBUG>( |
| 795 | fmt::format("readTempSensors: Failed reading {}, errno = {}", |
| 796 | filePathString + faultSuffix, e.code().value()) |
| 797 | .c_str()); |
| 798 | continue; |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 799 | } |
| 800 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 801 | if (faultValue != 0) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 802 | { |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 803 | dbus::OccDBusSensors::getOccDBus().setValue( |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 804 | sensorPath, std::numeric_limits<double>::quiet_NaN()); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 805 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 806 | dbus::OccDBusSensors::getOccDBus().setOperationalStatus(sensorPath, |
| 807 | false); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 808 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 809 | continue; |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 810 | } |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 811 | |
| 812 | double tempValue{0}; |
| 813 | |
| 814 | try |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 815 | { |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 816 | tempValue = readFile<double>(filePathString + inputSuffix); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 817 | } |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 818 | catch (const std::system_error& e) |
| 819 | { |
| 820 | log<level::DEBUG>( |
| 821 | fmt::format("readTempSensors: Failed reading {}, errno = {}", |
| 822 | filePathString + inputSuffix, e.code().value()) |
| 823 | .c_str()); |
| 824 | continue; |
| 825 | } |
| 826 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 827 | dbus::OccDBusSensors::getOccDBus().setValue( |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 828 | sensorPath, tempValue * std::pow(10, -3)); |
| 829 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 830 | dbus::OccDBusSensors::getOccDBus().setOperationalStatus(sensorPath, |
| 831 | true); |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 832 | |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 833 | // At this point, the sensor will be created for sure. |
| 834 | if (existingSensors.find(sensorPath) == existingSensors.end()) |
| 835 | { |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 836 | dbus::OccDBusSensors::getOccDBus().setChassisAssociation( |
| 837 | sensorPath); |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 838 | } |
| 839 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 840 | existingSensors[sensorPath] = id; |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 841 | } |
| 842 | return; |
| 843 | } |
| 844 | |
| 845 | std::optional<std::string> |
| 846 | Manager::getPowerLabelFunctionID(const std::string& value) |
| 847 | { |
| 848 | // If the value is "system", then the FunctionID is "system". |
| 849 | if (value == "system") |
| 850 | { |
| 851 | return value; |
| 852 | } |
| 853 | |
| 854 | // If the value is not "system", then the label value have 3 numbers, of |
| 855 | // which we only care about the middle one: |
| 856 | // <sensor id>_<function id>_<apss channel> |
| 857 | // eg: The value is "0_10_5" , then the FunctionID is "10". |
| 858 | if (value.find("_") == std::string::npos) |
| 859 | { |
| 860 | return std::nullopt; |
| 861 | } |
| 862 | |
| 863 | auto powerLabelValue = value.substr((value.find("_") + 1)); |
| 864 | |
| 865 | if (powerLabelValue.find("_") == std::string::npos) |
| 866 | { |
| 867 | return std::nullopt; |
| 868 | } |
| 869 | |
| 870 | return powerLabelValue.substr(0, powerLabelValue.find("_")); |
| 871 | } |
| 872 | |
| 873 | void Manager::readPowerSensors(const fs::path& path, uint32_t id) |
| 874 | { |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 875 | std::regex expr{"power\\d+_label$"}; // Example: power5_label |
| 876 | for (auto& file : fs::directory_iterator(path)) |
| 877 | { |
| 878 | if (!std::regex_search(file.path().string(), expr)) |
| 879 | { |
| 880 | continue; |
| 881 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 882 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 883 | std::string labelValue; |
| 884 | try |
| 885 | { |
| 886 | labelValue = readFile<std::string>(file.path()); |
| 887 | } |
| 888 | catch (const std::system_error& e) |
| 889 | { |
| 890 | log<level::DEBUG>( |
| 891 | fmt::format("readPowerSensors: Failed reading {}, errno = {}", |
| 892 | file.path().string(), e.code().value()) |
| 893 | .c_str()); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 894 | continue; |
| 895 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 896 | |
| 897 | auto functionID = getPowerLabelFunctionID(labelValue); |
| 898 | if (functionID == std::nullopt) |
| 899 | { |
| 900 | continue; |
| 901 | } |
| 902 | |
| 903 | const std::string& tempLabel = "label"; |
| 904 | const std::string filePathString = file.path().string().substr( |
| 905 | 0, file.path().string().length() - tempLabel.length()); |
| 906 | |
| 907 | std::string sensorPath = OCC_SENSORS_ROOT + std::string("/power/"); |
| 908 | |
| 909 | auto iter = powerSensorName.find(*functionID); |
| 910 | if (iter == powerSensorName.end()) |
| 911 | { |
| 912 | continue; |
| 913 | } |
| 914 | sensorPath.append(iter->second); |
| 915 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 916 | double tempValue{0}; |
| 917 | |
| 918 | try |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 919 | { |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 920 | tempValue = readFile<double>(filePathString + inputSuffix); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 921 | } |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 922 | catch (const std::system_error& e) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 923 | { |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 924 | log<level::DEBUG>( |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 925 | fmt::format("readPowerSensors: Failed reading {}, errno = {}", |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 926 | filePathString + inputSuffix, e.code().value()) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 927 | .c_str()); |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 928 | continue; |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 929 | } |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 930 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 931 | dbus::OccDBusSensors::getOccDBus().setUnit( |
Chris Cain | d84a833 | 2022-01-13 08:58:45 -0600 | [diff] [blame] | 932 | sensorPath, "xyz.openbmc_project.Sensor.Value.Unit.Watts"); |
| 933 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 934 | dbus::OccDBusSensors::getOccDBus().setValue( |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 935 | sensorPath, tempValue * std::pow(10, -3) * std::pow(10, -3)); |
| 936 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 937 | dbus::OccDBusSensors::getOccDBus().setOperationalStatus(sensorPath, |
| 938 | true); |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 939 | |
Matt Spinler | 5901abd | 2021-09-23 13:50:03 -0500 | [diff] [blame] | 940 | if (existingSensors.find(sensorPath) == existingSensors.end()) |
| 941 | { |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 942 | dbus::OccDBusSensors::getOccDBus().setChassisAssociation( |
| 943 | sensorPath); |
Matt Spinler | 5901abd | 2021-09-23 13:50:03 -0500 | [diff] [blame] | 944 | } |
| 945 | |
Matt Spinler | a26f152 | 2021-08-25 15:50:20 -0500 | [diff] [blame] | 946 | existingSensors[sensorPath] = id; |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 947 | } |
| 948 | return; |
| 949 | } |
| 950 | |
| 951 | void Manager::setSensorValueToNaN(uint32_t id) |
| 952 | { |
| 953 | for (const auto& [sensorPath, occId] : existingSensors) |
| 954 | { |
| 955 | if (occId == id) |
| 956 | { |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 957 | dbus::OccDBusSensors::getOccDBus().setValue( |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 958 | sensorPath, std::numeric_limits<double>::quiet_NaN()); |
| 959 | } |
| 960 | } |
| 961 | return; |
| 962 | } |
| 963 | |
Sheldon Bailey | 373af75 | 2022-02-21 15:14:00 -0600 | [diff] [blame] | 964 | void Manager::setSensorValueToNonFunctional(uint32_t id) const |
| 965 | { |
| 966 | for (const auto& [sensorPath, occId] : existingSensors) |
| 967 | { |
| 968 | if (occId == id) |
| 969 | { |
| 970 | dbus::OccDBusSensors::getOccDBus().setValue( |
| 971 | sensorPath, std::numeric_limits<double>::quiet_NaN()); |
| 972 | |
| 973 | dbus::OccDBusSensors::getOccDBus().setOperationalStatus(sensorPath, |
| 974 | false); |
| 975 | } |
| 976 | } |
| 977 | return; |
| 978 | } |
| 979 | |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 980 | void Manager::getSensorValues(std::unique_ptr<Status>& occ) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 981 | { |
Chris Cain | e2d0a43 | 2022-03-28 11:08:49 -0500 | [diff] [blame] | 982 | static bool tracedError[8] = {0}; |
| 983 | const fs::path sensorPath = occ->getHwmonPath(); |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 984 | const uint32_t id = occ->getOccInstanceID(); |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 985 | |
Chris Cain | e2d0a43 | 2022-03-28 11:08:49 -0500 | [diff] [blame] | 986 | if (fs::exists(sensorPath)) |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 987 | { |
Chris Cain | e2d0a43 | 2022-03-28 11:08:49 -0500 | [diff] [blame] | 988 | // Read temperature sensors |
| 989 | readTempSensors(sensorPath, id); |
| 990 | |
| 991 | if (occ->isMasterOcc()) |
| 992 | { |
| 993 | // Read power sensors |
| 994 | readPowerSensors(sensorPath, id); |
| 995 | } |
| 996 | tracedError[id] = false; |
| 997 | } |
| 998 | else |
| 999 | { |
| 1000 | if (!tracedError[id]) |
| 1001 | { |
| 1002 | log<level::ERR>( |
| 1003 | fmt::format( |
| 1004 | "Manager::getSensorValues: OCC{} sensor path missing: {}", |
| 1005 | id, sensorPath.c_str()) |
| 1006 | .c_str()); |
| 1007 | tracedError[id] = true; |
| 1008 | } |
Chicago Duan | bb895cb | 2021-06-18 19:37:16 +0800 | [diff] [blame] | 1009 | } |
| 1010 | |
| 1011 | return; |
| 1012 | } |
| 1013 | #endif |
Chris Cain | 1725767 | 2021-10-22 13:41:03 -0500 | [diff] [blame] | 1014 | |
| 1015 | // Read the altitude from DBus |
| 1016 | void Manager::readAltitude() |
| 1017 | { |
| 1018 | static bool traceAltitudeErr = true; |
| 1019 | |
| 1020 | utils::PropertyValue altitudeProperty{}; |
| 1021 | try |
| 1022 | { |
| 1023 | altitudeProperty = utils::getProperty(ALTITUDE_PATH, ALTITUDE_INTERFACE, |
| 1024 | ALTITUDE_PROP); |
| 1025 | auto sensorVal = std::get<double>(altitudeProperty); |
| 1026 | if (sensorVal < 0xFFFF) |
| 1027 | { |
| 1028 | if (sensorVal < 0) |
| 1029 | { |
| 1030 | altitude = 0; |
| 1031 | } |
| 1032 | else |
| 1033 | { |
| 1034 | // Round to nearest meter |
| 1035 | altitude = uint16_t(sensorVal + 0.5); |
| 1036 | } |
| 1037 | log<level::DEBUG>(fmt::format("readAltitude: sensor={} ({}m)", |
| 1038 | sensorVal, altitude) |
| 1039 | .c_str()); |
| 1040 | traceAltitudeErr = true; |
| 1041 | } |
| 1042 | else |
| 1043 | { |
| 1044 | if (traceAltitudeErr) |
| 1045 | { |
| 1046 | traceAltitudeErr = false; |
| 1047 | log<level::DEBUG>( |
| 1048 | fmt::format("Invalid altitude value: {}", sensorVal) |
| 1049 | .c_str()); |
| 1050 | } |
| 1051 | } |
| 1052 | } |
| 1053 | catch (const sdbusplus::exception::exception& e) |
| 1054 | { |
| 1055 | if (traceAltitudeErr) |
| 1056 | { |
| 1057 | traceAltitudeErr = false; |
| 1058 | log<level::INFO>( |
| 1059 | fmt::format("Unable to read Altitude: {}", e.what()).c_str()); |
| 1060 | } |
| 1061 | altitude = 0xFFFF; // not available |
| 1062 | } |
| 1063 | } |
| 1064 | |
| 1065 | // Callback function when ambient temperature changes |
| 1066 | void Manager::ambientCallback(sdbusplus::message::message& msg) |
| 1067 | { |
| 1068 | double currentTemp = 0; |
| 1069 | uint8_t truncatedTemp = 0xFF; |
| 1070 | std::string msgSensor; |
| 1071 | std::map<std::string, std::variant<double>> msgData; |
| 1072 | msg.read(msgSensor, msgData); |
| 1073 | |
| 1074 | auto valPropMap = msgData.find(AMBIENT_PROP); |
| 1075 | if (valPropMap == msgData.end()) |
| 1076 | { |
| 1077 | log<level::DEBUG>("ambientCallback: Unknown ambient property changed"); |
| 1078 | return; |
| 1079 | } |
| 1080 | currentTemp = std::get<double>(valPropMap->second); |
| 1081 | if (std::isnan(currentTemp)) |
| 1082 | { |
| 1083 | truncatedTemp = 0xFF; |
| 1084 | } |
| 1085 | else |
| 1086 | { |
| 1087 | if (currentTemp < 0) |
| 1088 | { |
| 1089 | truncatedTemp = 0; |
| 1090 | } |
| 1091 | else |
| 1092 | { |
| 1093 | // Round to nearest degree C |
| 1094 | truncatedTemp = uint8_t(currentTemp + 0.5); |
| 1095 | } |
| 1096 | } |
| 1097 | |
| 1098 | // If ambient changes, notify OCCs |
| 1099 | if (truncatedTemp != ambient) |
| 1100 | { |
| 1101 | log<level::DEBUG>( |
| 1102 | fmt::format("ambientCallback: Ambient change from {} to {}C", |
| 1103 | ambient, currentTemp) |
| 1104 | .c_str()); |
| 1105 | |
| 1106 | ambient = truncatedTemp; |
| 1107 | if (altitude == 0xFFFF) |
| 1108 | { |
| 1109 | // No altitude yet, try reading again |
| 1110 | readAltitude(); |
| 1111 | } |
| 1112 | |
| 1113 | log<level::DEBUG>( |
| 1114 | fmt::format("ambientCallback: Ambient: {}C, altitude: {}m", ambient, |
| 1115 | altitude) |
| 1116 | .c_str()); |
| 1117 | #ifdef POWER10 |
| 1118 | // Send ambient and altitude to all OCCs |
| 1119 | for (auto& obj : statusObjects) |
| 1120 | { |
| 1121 | if (obj->occActive()) |
| 1122 | { |
| 1123 | obj->sendAmbient(ambient, altitude); |
| 1124 | } |
| 1125 | } |
| 1126 | #endif // POWER10 |
| 1127 | } |
| 1128 | } |
| 1129 | |
| 1130 | // return the current ambient and altitude readings |
| 1131 | void Manager::getAmbientData(bool& ambientValid, uint8_t& ambientTemp, |
| 1132 | uint16_t& altitudeValue) const |
| 1133 | { |
| 1134 | ambientValid = true; |
| 1135 | ambientTemp = ambient; |
| 1136 | altitudeValue = altitude; |
| 1137 | |
| 1138 | if (ambient == 0xFF) |
| 1139 | { |
| 1140 | ambientValid = false; |
| 1141 | } |
| 1142 | } |
| 1143 | |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1144 | #ifdef POWER10 |
| 1145 | void Manager::occsNotAllRunning() |
| 1146 | { |
Chris Cain | 6fa848a | 2022-01-24 14:54:38 -0600 | [diff] [blame] | 1147 | // Function will also gets called when occ-control app gets |
| 1148 | // restarted. (occ active sensors do not change, so the Status |
| 1149 | // object does not call Manager back for all OCCs) |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1150 | |
| 1151 | if (activeCount != statusObjects.size()) |
| 1152 | { |
| 1153 | // Not all OCCs went active |
| 1154 | log<level::WARNING>( |
| 1155 | fmt::format( |
| 1156 | "occsNotAllRunning: Active OCC count ({}) does not match expected count ({})", |
| 1157 | activeCount, statusObjects.size()) |
| 1158 | .c_str()); |
| 1159 | // Procs may be garded, so may not need reset. |
| 1160 | } |
| 1161 | |
| 1162 | validateOccMaster(); |
| 1163 | } |
| 1164 | #endif // POWER10 |
| 1165 | |
| 1166 | // Verify single master OCC and start presence monitor |
| 1167 | void Manager::validateOccMaster() |
| 1168 | { |
| 1169 | int masterInstance = -1; |
| 1170 | for (auto& obj : statusObjects) |
| 1171 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 1172 | #ifdef POWER10 |
| 1173 | if (!obj->occActive()) |
| 1174 | { |
| 1175 | if (utils::isHostRunning()) |
| 1176 | { |
| 1177 | // OCC does not appear to be active yet, check active sensor |
| 1178 | pldmHandle->checkActiveSensor(obj->getOccInstanceID()); |
| 1179 | if (obj->occActive()) |
| 1180 | { |
| 1181 | log<level::INFO>( |
| 1182 | fmt::format( |
| 1183 | "validateOccMaster: OCC{} is ACTIVE after reading sensor", |
| 1184 | obj->getOccInstanceID()) |
| 1185 | .c_str()); |
| 1186 | } |
| 1187 | } |
| 1188 | else |
| 1189 | { |
| 1190 | log<level::WARNING>( |
| 1191 | fmt::format( |
| 1192 | "validateOccMaster: HOST is not running (OCC{})", |
| 1193 | obj->getOccInstanceID()) |
| 1194 | .c_str()); |
| 1195 | return; |
| 1196 | } |
| 1197 | } |
| 1198 | #endif // POWER10 |
| 1199 | |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1200 | if (obj->isMasterOcc()) |
| 1201 | { |
Chris Cain | 5d66a0a | 2022-02-09 08:52:10 -0600 | [diff] [blame] | 1202 | obj->addPresenceWatchMaster(); |
| 1203 | |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1204 | if (masterInstance == -1) |
| 1205 | { |
| 1206 | masterInstance = obj->getOccInstanceID(); |
| 1207 | } |
| 1208 | else |
| 1209 | { |
| 1210 | log<level::ERR>( |
| 1211 | fmt::format( |
| 1212 | "validateOccMaster: Multiple OCC masters! ({} and {})", |
| 1213 | masterInstance, obj->getOccInstanceID()) |
| 1214 | .c_str()); |
| 1215 | // request reset |
| 1216 | obj->deviceError(); |
| 1217 | } |
| 1218 | } |
| 1219 | } |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 1220 | |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1221 | if (masterInstance < 0) |
| 1222 | { |
Chris Cain | bae4d07 | 2022-02-28 09:46:50 -0600 | [diff] [blame] | 1223 | log<level::ERR>( |
| 1224 | fmt::format("validateOccMaster: Master OCC not found! (of {} OCCs)", |
| 1225 | statusObjects.size()) |
| 1226 | .c_str()); |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1227 | // request reset |
| 1228 | statusObjects.front()->deviceError(); |
| 1229 | } |
| 1230 | else |
| 1231 | { |
| 1232 | log<level::INFO>( |
Chris Cain | 36f9cde | 2021-11-22 11:18:21 -0600 | [diff] [blame] | 1233 | fmt::format("validateOccMaster: OCC{} is master of {} OCCs", |
| 1234 | masterInstance, activeCount) |
Chris Cain | a7b74dc | 2021-11-10 17:03:43 -0600 | [diff] [blame] | 1235 | .c_str()); |
| 1236 | } |
| 1237 | } |
| 1238 | |
Chris Cain | 40501a2 | 2022-03-14 17:33:27 -0500 | [diff] [blame] | 1239 | void Manager::updatePcapBounds() const |
| 1240 | { |
| 1241 | if (pcap) |
| 1242 | { |
| 1243 | pcap->updatePcapBounds(); |
| 1244 | } |
| 1245 | } |
| 1246 | |
Vishwanatha Subbanna | dfc7ec7 | 2017-09-07 18:18:01 +0530 | [diff] [blame] | 1247 | } // namespace occ |
| 1248 | } // namespace open_power |