/** * Copyright (C) 2018-present MongoDB, Inc. * * This program is free software: you can redistribute it and/or modify * it under the terms of the Server Side Public License, version 1, * as published by MongoDB, Inc. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * Server Side Public License for more details. * * You should have received a copy of the Server Side Public License * along with this program. If not, see * . * * As a special exception, the copyright holders give permission to link the * code of portions of this program with the OpenSSL library under certain * conditions as described in each individual source file and distribute * linked combinations including the program with the OpenSSL library. You * must comply with the Server Side Public License in all respects for * all of the code used other than as permitted herein. If you modify file(s) * with this exception, you may extend this exception to your version of the * file(s), but you are not obligated to do so. If you do not wish to do so, * delete this exception statement from your version. If you delete this * exception statement from all source files in the program, then also delete * it in the license file. */ #include "mongo/platform/basic.h" #include "mongo/db/query/plan_yield_policy.h" #include "mongo/db/concurrency/write_conflict_exception.h" #include "mongo/db/curop.h" #include "mongo/db/curop_failpoint_helpers.h" #include "mongo/db/operation_context.h" #include "mongo/db/query/query_knobs_gen.h" #include "mongo/db/service_context.h" #include "mongo/util/fail_point_service.h" #include "mongo/util/scopeguard.h" #include "mongo/util/time_support.h" namespace mongo { namespace { MONGO_FAIL_POINT_DEFINE(setInterruptOnlyPlansCheckForInterruptHang); } // namespace PlanYieldPolicy::PlanYieldPolicy(PlanExecutor* exec, PlanExecutor::YieldPolicy policy) : _policy(exec->getOpCtx()->lockState()->isGlobalLockedRecursively() ? PlanExecutor::NO_YIELD : policy), _forceYield(false), _elapsedTracker(exec->getOpCtx()->getServiceContext()->getFastClockSource(), internalQueryExecYieldIterations.load(), Milliseconds(internalQueryExecYieldPeriodMS.load())), _planYielding(exec) {} PlanYieldPolicy::PlanYieldPolicy(PlanExecutor::YieldPolicy policy, ClockSource* cs) : _policy(policy), _forceYield(false), _elapsedTracker(cs, internalQueryExecYieldIterations.load(), Milliseconds(internalQueryExecYieldPeriodMS.load())), _planYielding(nullptr) {} bool PlanYieldPolicy::shouldYieldOrInterrupt() { if (_policy == PlanExecutor::INTERRUPT_ONLY) { return _elapsedTracker.intervalHasElapsed(); } if (!canAutoYield()) return false; invariant(!_planYielding->getOpCtx()->lockState()->inAWriteUnitOfWork()); if (_forceYield) return true; return _elapsedTracker.intervalHasElapsed(); } void PlanYieldPolicy::resetTimer() { _elapsedTracker.resetLastTime(); } Status PlanYieldPolicy::yieldOrInterrupt(stdx::function whileYieldingFn) { invariant(_planYielding); if (_policy == PlanExecutor::INTERRUPT_ONLY) { ON_BLOCK_EXIT([this]() { resetTimer(); }); OperationContext* opCtx = _planYielding->getOpCtx(); invariant(opCtx); // If the 'setInterruptOnlyPlansCheckForInterruptHang' fail point is enabled, set the 'msg' // field of this operation's CurOp to signal that we've hit this point. if (MONGO_FAIL_POINT(setInterruptOnlyPlansCheckForInterruptHang)) { CurOpFailpointHelpers::waitWhileFailPointEnabled( &setInterruptOnlyPlansCheckForInterruptHang, opCtx, "setInterruptOnlyPlansCheckForInterruptHang"); } return opCtx->checkForInterruptNoAssert(); } invariant(canAutoYield()); // After we finish yielding (or in any early return), call resetTimer() to prevent yielding // again right away. We delay the resetTimer() call so that the clock doesn't start ticking // until after we return from the yield. ON_BLOCK_EXIT([this]() { resetTimer(); }); _forceYield = false; OperationContext* opCtx = _planYielding->getOpCtx(); invariant(opCtx); invariant(!opCtx->lockState()->inAWriteUnitOfWork()); // Can't use writeConflictRetry since we need to call saveState before reseting the transaction. for (int attempt = 1; true; attempt++) { try { try { _planYielding->saveState(); } catch (const WriteConflictException&) { invariant(!"WriteConflictException not allowed in saveState"); } if (_policy == PlanExecutor::WRITE_CONFLICT_RETRY_ONLY) { // Just reset the snapshot. Leave all LockManager locks alone. opCtx->recoveryUnit()->abandonSnapshot(); } else { // Release and reacquire locks. _yieldAllLocks(opCtx, whileYieldingFn, _planYielding->nss()); } _planYielding->restoreStateWithoutRetrying(); return Status::OK(); } catch (const WriteConflictException&) { CurOp::get(opCtx)->debug().additiveMetrics.incrementWriteConflicts(1); WriteConflictException::logAndBackoff( attempt, "plan execution restoreState", _planYielding->nss().ns()); // retry } catch (...) { // Errors other than write conflicts don't get retried, and should instead result in the // PlanExecutor dying. We propagate all such errors as status codes. return exceptionToStatus(); } } } namespace { MONGO_FAIL_POINT_DEFINE(setYieldAllLocksHang); MONGO_FAIL_POINT_DEFINE(setYieldAllLocksWait); } // namespace void PlanYieldPolicy::_yieldAllLocks(OperationContext* opCtx, stdx::function whileYieldingFn, const NamespaceString& planExecNS) { // Things have to happen here in a specific order: // * Release lock mgr locks // * Check for interrupt (kill flag is set) // * Call the whileYieldingFn // * Reacquire lock mgr locks Locker* locker = opCtx->lockState(); Locker::LockSnapshot snapshot; auto unlocked = locker->saveLockStateAndUnlock(&snapshot); // Attempt to check for interrupt while locks are not held, in order to discourage the // assumption that locks will always be held when a Plan Executor returns an error. if (_policy == PlanExecutor::YIELD_AUTO) { opCtx->checkForInterrupt(); // throws } if (!unlocked) { // Nothing was unlocked, just return, yielding is pointless. return; } // Top-level locks are freed, release any potential low-level (storage engine-specific // locks). If we are yielding, we are at a safe place to do so. opCtx->recoveryUnit()->abandonSnapshot(); // Track the number of yields in CurOp. CurOp::get(opCtx)->yielded(); MONGO_FAIL_POINT_BLOCK(setYieldAllLocksHang, config) { StringData ns{config.getData().getStringField("namespace")}; if (ns.empty() || ns == planExecNS.ns()) { MONGO_FAIL_POINT_PAUSE_WHILE_SET(setYieldAllLocksHang); } } MONGO_FAIL_POINT_BLOCK(setYieldAllLocksWait, customWait) { const BSONObj& data = customWait.getData(); BSONElement customWaitNS = data["namespace"]; if (!customWaitNS || planExecNS.ns() == customWaitNS.str()) { sleepFor(Milliseconds(data["waitForMillis"].numberInt())); } } if (whileYieldingFn) { whileYieldingFn(); } locker->restoreLockState(opCtx, snapshot); } } // namespace mongo