/* This file is part of drd, a thread error detector. Copyright (C) 2006-2015 Bart Van Assche <bvanassche@acm.org>. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. 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 GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA. The GNU General Public License is contained in the file COPYING. */ #ifndef __DRD_VC_H #define __DRD_VC_H /* * DRD vector clock implementation: * - One counter per thread. * - A vector clock is implemented as multiple pairs of (thread id, counter). * - Pairs are stored in an array sorted by thread id. * * Semantics: * - Each time a thread performs an action that implies an ordering between * intra-thread events, the counter of that thread is incremented. * - Vector clocks are compared by comparing all counters of all threads. * - When a thread synchronization action is performed that guarantees that * new actions of the current thread are executed after the actions of the * other thread, the vector clock of the synchronization object and the * current thread are combined (by taking the component-wise maximum). * - A vector clock is incremented during actions such as * pthread_create(), pthread_mutex_unlock(), sem_post(). (Actions where * an inter-thread ordering "arrow" starts). */ #include "pub_tool_basics.h" /* Addr, SizeT */ #include "drd_basics.h" /* DrdThreadId */ #include "pub_tool_libcassert.h" /* tl_assert() */ #define VC_PREALLOCATED 8 /** Vector clock element. */ typedef struct { DrdThreadId threadid; UInt count; } VCElem; typedef struct { unsigned capacity; /**< number of elements allocated for array vc. */ unsigned size; /**< number of elements used of array vc. */ VCElem* vc; /**< vector clock elements. */ VCElem preallocated[VC_PREALLOCATED]; } VectorClock; void DRD_(vc_init)(VectorClock* const vc, const VCElem* const vcelem, const unsigned size); void DRD_(vc_cleanup)(VectorClock* const vc); void DRD_(vc_copy)(VectorClock* const new, const VectorClock* const rhs); void DRD_(vc_assign)(VectorClock* const lhs, const VectorClock* const rhs); void DRD_(vc_increment)(VectorClock* const vc, DrdThreadId const tid); static __inline__ Bool DRD_(vc_lte)(const VectorClock* const vc1, const VectorClock* const vc2); Bool DRD_(vc_ordered)(const VectorClock* const vc1, const VectorClock* const vc2); void DRD_(vc_min)(VectorClock* const result, const VectorClock* const rhs); void DRD_(vc_combine)(VectorClock* const result, const VectorClock* const rhs); void DRD_(vc_print)(const VectorClock* const vc); HChar* DRD_(vc_aprint)(const VectorClock* const vc); void DRD_(vc_check)(const VectorClock* const vc); void DRD_(vc_test)(void); /** * @return True if all thread id's that are present in vc1 also exist in * vc2, and if additionally all corresponding counters in v2 are higher or * equal. */ static __inline__ Bool DRD_(vc_lte)(const VectorClock* const vc1, const VectorClock* const vc2) { unsigned i; unsigned j = 0; for (i = 0; i < vc1->size; i++) { while (j < vc2->size && vc2->vc[j].threadid < vc1->vc[i].threadid) j++; if (j >= vc2->size || vc2->vc[j].threadid > vc1->vc[i].threadid) return False; #ifdef ENABLE_DRD_CONSISTENCY_CHECKS /* * This assert statement has been commented out because of performance * reasons. */ tl_assert(j < vc2->size && vc2->vc[j].threadid == vc1->vc[i].threadid); #endif if (vc1->vc[i].count > vc2->vc[j].count) return False; } return True; } #endif /* __DRD_VC_H */