!252 并行查询增加NUMA绑核

Merge pull request !252 from TotaJ/feature/parallel_performance
This commit is contained in:
opengauss-bot 2020-09-27 11:24:20 +08:00 committed by Gitee
commit 7662afaa4c
25 changed files with 3532 additions and 3254 deletions

View File

@ -1,61 +1,61 @@
#
# Copyright (c) 2020 Huawei Technologies Co.,Ltd.
#
# openGauss is licensed under Mulan PSL v2.
# You can use this software according to the terms and conditions of the Mulan PSL v2.
# You may obtain a copy of Mulan PSL v2 at:
#
# http://license.coscl.org.cn/MulanPSL2
#
# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND,
# EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT,
# MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE.
# See the Mulan PSL v2 for more details.
# ---------------------------------------------------------------------------------------
#
# Makefile
# Makefile for the mysql_fdw
#
# IDENTIFICATION
# contrib/mysql_fdw/Makefile
#
# ---------------------------------------------------------------------------------------
all:mysql_fdw_target
install:install-data
top_builddir ?= ../../
MYSQL_FDW_DIR=$(top_builddir)/third_party/dependency/mysql_fdw
MYSQL_FDW_PACKAGE=mysql_fdw-REL-2_5_3
MYSQL_FDW_PATCH=huawei_mysql_fdw-2.5.3_patch
MYSQL_FDW_MEGRED_SOURCES_DIR=$(MYSQL_FDW_DIR)/code
.PHONY: mysql_fdw_target
mysql_fdw_target:
@$(call create_mysql_fdw_sources)
@make -C $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE)
.PHONY: install-data
install-data: mysql_fdw_target
@make -C $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE) install
uninstall distclean clean:
@rm -rf $(MYSQL_FDW_MEGRED_SOURCES_DIR)
MYSQL_FDW_RELEVANT_SOURCES = connection.c deparse.c mysql_fdw.c mysql_query.c option.c
define create_mysql_fdw_sources
rm -rf $(MYSQL_FDW_MEGRED_SOURCES_DIR); \
mkdir $(MYSQL_FDW_MEGRED_SOURCES_DIR); \
tar xfzv $(MYSQL_FDW_DIR)/$(MYSQL_FDW_PACKAGE).tar.gz -C $(MYSQL_FDW_MEGRED_SOURCES_DIR) &> /dev/null; \
for ((i=1;i<=99;i++)); \
do \
file_name="$(MYSQL_FDW_DIR)/$$i-mysql_fdw-2.5.3_patch.patch"; \
if [ ! -f "$$file_name" ]; then \
exit 0; \
fi; \
patch -p0 -d $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE) < $$file_name &> /dev/null; \
done
rename ".c" ".cpp" $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE)/*.c; \
patch -p0 -d $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE) < $(MYSQL_FDW_DIR)/$(MYSQL_FDW_PATCH).patch &> /dev/null;
endef
#
# Copyright (c) 2020 Huawei Technologies Co.,Ltd.
#
# openGauss is licensed under Mulan PSL v2.
# You can use this software according to the terms and conditions of the Mulan PSL v2.
# You may obtain a copy of Mulan PSL v2 at:
#
# http://license.coscl.org.cn/MulanPSL2
#
# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND,
# EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT,
# MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE.
# See the Mulan PSL v2 for more details.
# ---------------------------------------------------------------------------------------
#
# Makefile
# Makefile for the mysql_fdw
#
# IDENTIFICATION
# contrib/mysql_fdw/Makefile
#
# ---------------------------------------------------------------------------------------
all:mysql_fdw_target
install:install-data
top_builddir ?= ../../
MYSQL_FDW_DIR=$(top_builddir)/third_party/dependency/mysql_fdw
MYSQL_FDW_PACKAGE=mysql_fdw-REL-2_5_3
MYSQL_FDW_PATCH=openGauss_mysql_fdw-2.5.3_patch
MYSQL_FDW_MEGRED_SOURCES_DIR=$(MYSQL_FDW_DIR)/code
.PHONY: mysql_fdw_target
mysql_fdw_target:
@$(call create_mysql_fdw_sources)
@make -C $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE)
.PHONY: install-data
install-data: mysql_fdw_target
@make -C $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE) install
uninstall distclean clean:
@rm -rf $(MYSQL_FDW_MEGRED_SOURCES_DIR)
MYSQL_FDW_RELEVANT_SOURCES = connection.c deparse.c mysql_fdw.c mysql_query.c option.c
define create_mysql_fdw_sources
rm -rf $(MYSQL_FDW_MEGRED_SOURCES_DIR); \
mkdir $(MYSQL_FDW_MEGRED_SOURCES_DIR); \
tar xfzv $(MYSQL_FDW_DIR)/$(MYSQL_FDW_PACKAGE).tar.gz -C $(MYSQL_FDW_MEGRED_SOURCES_DIR) &> /dev/null; \
for ((i=1;i<=99;i++)); \
do \
file_name="$(MYSQL_FDW_DIR)/$$i-mysql_fdw-2.5.3_patch.patch"; \
if [ ! -f "$$file_name" ]; then \
exit 0; \
fi; \
patch -p0 -d $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE) < $$file_name &> /dev/null; \
done
rename ".c" ".cpp" $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE)/*.c; \
patch -p0 -d $(MYSQL_FDW_MEGRED_SOURCES_DIR)/$(MYSQL_FDW_PACKAGE) < $(MYSQL_FDW_DIR)/$(MYSQL_FDW_PATCH).patch &> /dev/null;
endef

View File

@ -1,59 +1,59 @@
#
# Copyright (c) 2020 Huawei Technologies Co.,Ltd.
#
# openGauss is licensed under Mulan PSL v2.
# You can use this software according to the terms and conditions of the Mulan PSL v2.
# You may obtain a copy of Mulan PSL v2 at:
#
# http://license.coscl.org.cn/MulanPSL2
#
# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND,
# EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT,
# MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE.
# See the Mulan PSL v2 for more details.
# ---------------------------------------------------------------------------------------
#
# Makefile
# Makefile for the oracle_fdw
#
# IDENTIFICATION
# contrib/oracle_fdw/Makefile
#
# ---------------------------------------------------------------------------------------
all:oracle_fdw_target
install:install-data
top_builddir ?= ../../
ORACLE_FDW_DIR=$(top_builddir)/third_party/dependency/oracle_fdw
ORACLE_FDW_PACKAGE=oracle_fdw-ORACLE_FDW_2_2_0
ORACLE_FDW_PATCH=huawei_oracle_fdw-2.2.0_patch
ORACLE_FDW_MEGRED_SOURCES_DIR=$(ORACLE_FDW_DIR)/code
.PHONY: oracle_fdw_target
oracle_fdw_target:
@$(call create_oracle_fdw_sources)
@make -C $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) NO_PGXS=1
.PHONY: install-data
install-data: oracle_fdw_target
@make -C $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) NO_PGXS=1 install
uninstall distclean clean:
@rm -rf $(ORACLE_FDW_MEGRED_SOURCES_DIR)
define create_oracle_fdw_sources
rm -rf $(ORACLE_FDW_MEGRED_SOURCES_DIR); \
mkdir $(ORACLE_FDW_MEGRED_SOURCES_DIR); \
tar xfzv $(ORACLE_FDW_DIR)/$(ORACLE_FDW_PACKAGE).tar.gz -C $(ORACLE_FDW_MEGRED_SOURCES_DIR) &> /dev/null; \
for ((i=1;i<=99;i++)); \
do \
file_name="$(ORACLE_FDW_DIR)/$$i-oracle_fdw-2.2.0_patch.patch"; \
if [ ! -f "$$file_name" ]; then \
exit 0; \
fi; \
patch -p0 -d $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) < $$file_name &> /dev/null; \
done
rename ".c" ".cpp" $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE)/*.c; \
patch -p0 -d $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) < $(ORACLE_FDW_DIR)/$(ORACLE_FDW_PATCH).patch &> /dev/null;
endef
#
# Copyright (c) 2020 Huawei Technologies Co.,Ltd.
#
# openGauss is licensed under Mulan PSL v2.
# You can use this software according to the terms and conditions of the Mulan PSL v2.
# You may obtain a copy of Mulan PSL v2 at:
#
# http://license.coscl.org.cn/MulanPSL2
#
# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND,
# EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT,
# MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE.
# See the Mulan PSL v2 for more details.
# ---------------------------------------------------------------------------------------
#
# Makefile
# Makefile for the oracle_fdw
#
# IDENTIFICATION
# contrib/oracle_fdw/Makefile
#
# ---------------------------------------------------------------------------------------
all:oracle_fdw_target
install:install-data
top_builddir ?= ../../
ORACLE_FDW_DIR=$(top_builddir)/third_party/dependency/oracle_fdw
ORACLE_FDW_PACKAGE=oracle_fdw-ORACLE_FDW_2_2_0
ORACLE_FDW_PATCH=openGauss_oracle_fdw-2.2.0_patch
ORACLE_FDW_MEGRED_SOURCES_DIR=$(ORACLE_FDW_DIR)/code
.PHONY: oracle_fdw_target
oracle_fdw_target:
@$(call create_oracle_fdw_sources)
@make -C $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) NO_PGXS=1
.PHONY: install-data
install-data: oracle_fdw_target
@make -C $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) NO_PGXS=1 install
uninstall distclean clean:
@rm -rf $(ORACLE_FDW_MEGRED_SOURCES_DIR)
define create_oracle_fdw_sources
rm -rf $(ORACLE_FDW_MEGRED_SOURCES_DIR); \
mkdir $(ORACLE_FDW_MEGRED_SOURCES_DIR); \
tar xfzv $(ORACLE_FDW_DIR)/$(ORACLE_FDW_PACKAGE).tar.gz -C $(ORACLE_FDW_MEGRED_SOURCES_DIR) &> /dev/null; \
for ((i=1;i<=99;i++)); \
do \
file_name="$(ORACLE_FDW_DIR)/$$i-oracle_fdw-2.2.0_patch.patch"; \
if [ ! -f "$$file_name" ]; then \
exit 0; \
fi; \
patch -p0 -d $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) < $$file_name &> /dev/null; \
done
rename ".c" ".cpp" $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE)/*.c; \
patch -p0 -d $(ORACLE_FDW_MEGRED_SOURCES_DIR)/$(ORACLE_FDW_PACKAGE) < $(ORACLE_FDW_DIR)/$(ORACLE_FDW_PATCH).patch &> /dev/null;
endef

View File

@ -528,7 +528,6 @@ max_inner_tool_connections|int|1,8388607|NULL|NULL|
max_keep_log_seg|int|0,2147483647|NULL|NULL|
max_background_workers|int|0,262143|NULL|NULL|
min_parallel_table_scan_size|int|0,715827882|kB|NULL|
max_parallel_workers|int|0,1024|NULL|NULL|
max_parallel_workers_per_gather|int|0,1024|NULL|NULL|
parallel_tuple_cost|real|0,1.79769e+308|NULL|NULL|
parallel_setup_cost|real|0,1.79769e+308|NULL|NULL|

View File

@ -9229,22 +9229,6 @@ static void init_configure_names_int()
NULL,
NULL
},
{
{
"max_parallel_workers",
PGC_USERSET,
RESOURCES_ASYNCHRONOUS,
gettext_noop("Sets the maximum number of parallel workers that can be active at one time."),
NULL
},
&g_instance.attr.attr_common.max_parallel_workers,
8,
0,
MAX_PARALLEL_WORKER_LIMIT,
NULL,
NULL,
NULL
},
{
{
"max_parallel_workers_per_gather",
@ -9253,7 +9237,7 @@ static void init_configure_names_int()
gettext_noop("Sets the maximum number of parallel processes per executor node."),
NULL
},
&g_instance.attr.attr_common.max_parallel_workers_per_gather,
&u_sess->attr.attr_sql.max_parallel_workers_per_gather,
2,
0,
MAX_PARALLEL_WORKER_LIMIT,

View File

@ -898,17 +898,19 @@ bool HeapTupleSatisfiesMVCC(HeapTuple htup, Snapshot snapshot, Buffer buffer)
TransactionIdStatus hintstatus;
Page page = BufferGetPage(buffer);
ereport(DEBUG1,
(errmsg("HeapTupleSatisfiesMVCC self(%u,%u) ctid(%u,%u) cur_xid " XID_FMT " xmin " XID_FMT
" xmax " XID_FMT " csn " CSN_FMT,
ItemPointerGetBlockNumber(&htup->t_self),
ItemPointerGetOffsetNumber(&htup->t_self),
ItemPointerGetBlockNumber(&tuple->t_ctid),
ItemPointerGetOffsetNumber(&tuple->t_ctid),
GetCurrentTransactionIdIfAny(),
HeapTupleHeaderGetXmin(page, tuple),
HeapTupleHeaderGetXmax(page, tuple),
snapshot->snapshotcsn)));
if (SHOW_DEBUG_MESSAGE()) {
ereport(DEBUG1,
(errmsg("HeapTupleSatisfiesMVCC self(%u,%u) ctid(%u,%u) cur_xid " XID_FMT " xmin " XID_FMT
" xmax " XID_FMT " csn " CSN_FMT,
ItemPointerGetBlockNumber(&htup->t_self),
ItemPointerGetOffsetNumber(&htup->t_self),
ItemPointerGetBlockNumber(&tuple->t_ctid),
ItemPointerGetOffsetNumber(&tuple->t_ctid),
GetCurrentTransactionIdIfAny(),
HeapTupleHeaderGetXmin(page, tuple),
HeapTupleHeaderGetXmax(page, tuple),
snapshot->snapshotcsn)));
}
/*
* Just valid for read-only transaction when u_sess->attr.attr_common.XactReadOnly is true.

View File

@ -899,7 +899,7 @@ static void set_plain_rel_pathlist(PlannerInfo* root, RelOptInfo* rel, RangeTblE
* sophisticated, but we need something here for now.
*/
while (rel->pages > parallel_threshold * 3 &&
parallel_degree < g_instance.attr.attr_common.max_parallel_workers_per_gather) {
parallel_degree < u_sess->attr.attr_sql.max_parallel_workers_per_gather) {
parallel_degree++;
parallel_threshold *= 3;
if (parallel_threshold >= PG_INT32_MAX / 3)

View File

@ -462,7 +462,7 @@ PlannedStmt* standard_planner(Query* parse, int cursorOptions, ParamListInfo bou
*/
glob->parallelModeOK = (cursorOptions & CURSOR_OPT_PARALLEL_OK) != 0 && IsUnderPostmaster &&
parse->commandType == CMD_SELECT && !parse->hasModifyingCTE && parse->utilityStmt == NULL &&
g_instance.attr.attr_common.max_parallel_workers_per_gather > 0 && !IsParallelWorker() &&
u_sess->attr.attr_sql.max_parallel_workers_per_gather > 0 && !IsParallelWorker() &&
!IsolationIsSerializable() && !has_parallel_hazard((Node *)parse, true);
/*

View File

@ -683,7 +683,7 @@ void StartBackgroundWorker(void* bgWorkerSlotShmAddr)
BackgroundWorker *worker = t_thrd.bgworker_cxt.my_bgworker_entry;
bgworker_main_type entrypt;
t_thrd.proc_cxt.MyProgName = "BackgroundWorker";
knl_thread_set_name("BgWorker");
/*
* Create memory context and buffer used for RowDescription messages. As
* SendRowDescriptionMessage(), via exec_describe_statement_message(), is

File diff suppressed because it is too large Load Diff

View File

@ -1,434 +1,436 @@
/* -------------------------------------------------------------------------
*
* nodeGather.c
* Support routines for scanning a plan via multiple workers.
*
* Portions Copyright (c) 1996-2015, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* A Gather executor launches parallel workers to run multiple copies of a
* plan. It can also run the plan itself, if the workers are not available
* or have not started up yet. It then merges all of the results it produces
* and the results from the workers into a single output stream. Therefore,
* it will normally be used with a plan where running multiple copies of the
* same plan does not produce duplicate output, such as parallel-aware
* SeqScan.
*
* Alternatively, a Gather node can be configured to use just one worker
* and the single-copy flag can be set. In this case, the Gather node will
* run the plan in one worker and will not execute the plan itself. In
* this case, it simply returns whatever tuples were returned by the worker.
* If a worker cannot be obtained, then it will run the plan itself and
* return the results. Therefore, a plan used with a single-copy Gather
* node need not be parallel-aware.
*
* IDENTIFICATION
* src/backend/executor/nodeGather.c
*
* -------------------------------------------------------------------------
*/
#include "postgres.h"
#include "access/relscan.h"
#include "access/xact.h"
#include "executor/execdebug.h"
#include "executor/execParallel.h"
#include "executor/nodeGather.h"
#include "executor/nodeSubplan.h"
#include "executor/tqueue.h"
#include "miscadmin.h"
#include "utils/memutils.h"
#include "utils/rel.h"
static TupleTableSlot *gather_getnext(GatherState *gatherstate);
static HeapTuple gather_readnext(GatherState *gatherstate);
static void ExecShutdownGatherWorkers(GatherState *node);
/* ----------------------------------------------------------------
* ExecInitGather
* ----------------------------------------------------------------
*/
GatherState *ExecInitGather(Gather *node, EState *estate, int eflags)
{
bool hasoid = false;
/* Gather node doesn't have innerPlan node. */
Assert(innerPlan(node) == NULL);
/*
* create state structure
*/
GatherState *gatherstate = makeNode(GatherState);
gatherstate->ps.plan = (Plan *)node;
gatherstate->ps.state = estate;
gatherstate->need_to_scan_locally = !node->single_copy &&
u_sess->attr.attr_sql.parallel_leader_participation;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &gatherstate->ps);
/*
* initialize child expressions
*/
gatherstate->ps.targetlist = (List *)ExecInitExpr((Expr *)node->plan.targetlist, (PlanState *)gatherstate);
gatherstate->ps.qual = (List *)ExecInitExpr((Expr *)node->plan.qual, (PlanState *)gatherstate);
/*
* tuple table initialization
*/
gatherstate->funnel_slot = ExecInitExtraTupleSlot(estate);
ExecInitResultTupleSlot(estate, &gatherstate->ps);
/*
* now initialize outer plan
*/
Plan *outerNode = outerPlan(node);
outerPlanState(gatherstate) = ExecInitNode(outerNode, estate, eflags);
gatherstate->ps.ps_TupFromTlist = false;
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&gatherstate->ps);
ExecAssignProjectionInfo(&gatherstate->ps, NULL);
/*
* Initialize funnel slot to same tuple descriptor as outer plan.
*/
if (!ExecContextForcesOids(&gatherstate->ps, &hasoid))
hasoid = false;
TupleDesc tupDesc = ExecTypeFromTL(outerNode->targetlist, hasoid);
ExecSetSlotDescriptor(gatherstate->funnel_slot, tupDesc);
return gatherstate;
}
/* ----------------------------------------------------------------
* ExecGather(node)
*
* Scans the relation via multiple workers and returns
* the next qualifying tuple.
* ----------------------------------------------------------------
*/
TupleTableSlot *ExecGather(GatherState *node)
{
TupleTableSlot *fslot = node->funnel_slot;
int i;
TupleTableSlot *slot = NULL;
TupleTableSlot *resultSlot = NULL;
ExprDoneCond isDone;
CHECK_FOR_INTERRUPTS();
/*
* Initialize the parallel context and workers on first execution. We do
* this on first execution rather than during node initialization, as it
* needs to allocate large dynamic segement, so it is better to do if it
* is really needed.
*/
if (!node->initialized) {
EState *estate = node->ps.state;
Gather *gather = (Gather *)node->ps.plan;
t_thrd.subrole = BACKGROUND_LEADER;
/*
* Sometimes we might have to run without parallelism; but if
* parallel mode is active then we can try to fire up some workers.
*/
if (gather->num_workers > 0 && IsInParallelMode()) {
bool got_any_worker = false;
/* Initialize the workers required to execute Gather node. */
if (!node->pei)
node->pei = ExecInitParallelPlan(node->ps.lefttree, estate, gather->num_workers);
/*
* Register backend workers. We might not get as many as we
* requested, or indeed any at all.
*/
ParallelContext *pcxt = node->pei->pcxt;
LaunchParallelWorkers(pcxt);
/* Set up tuple queue readers to read the results. */
if (pcxt->nworkers > 0) {
node->nreaders = 0;
node->reader = (TupleQueueReader **)palloc(pcxt->nworkers * sizeof(TupleQueueReader *));
for (i = 0; i < pcxt->nworkers; ++i) {
if (pcxt->worker[i].bgwhandle == NULL)
continue;
shm_mq_set_handle(node->pei->tqueue[i], pcxt->worker[i].bgwhandle);
node->reader[node->nreaders++] =
CreateTupleQueueReader(node->pei->tqueue[i], fslot->tts_tupleDescriptor);
got_any_worker = true;
}
}
/* No workers? Then never mind. */
if (!got_any_worker)
ExecShutdownGatherWorkers(node);
}
/* Run plan locally if no workers or not single-copy. */
node->need_to_scan_locally = (node->reader == NULL) ||
(!gather->single_copy && u_sess->attr.attr_sql.parallel_leader_participation);
node->initialized = true;
}
/*
* Check to see if we're still projecting out tuples from a previous scan
* tuple (because there is a function-returning-set in the projection
* expressions). If so, try to project another one.
*/
if (node->ps.ps_TupFromTlist) {
resultSlot = ExecProject(node->ps.ps_ProjInfo, &isDone);
if (isDone == ExprMultipleResult)
return resultSlot;
/* Done with that source tuple... */
node->ps.ps_TupFromTlist = false;
}
/*
* Reset per-tuple memory context to free any expression evaluation
* storage allocated in the previous tuple cycle. Note we can't do this
* until we're done projecting. This will also clear any previous tuple
* returned by a TupleQueueReader; to make sure we don't leave a dangling
* pointer around, clear the working slot first.
*/
(void)ExecClearTuple(node->funnel_slot);
ExprContext *econtext = node->ps.ps_ExprContext;
ResetExprContext(econtext);
/* Get and return the next tuple, projecting if necessary. */
for (;;) {
/*
* Get next tuple, either from one of our workers, or by running the
* plan ourselves.
*/
slot = gather_getnext(node);
if (TupIsNull(slot))
return NULL;
/*
* form the result tuple using ExecProject(), and return it --- unless
* the projection produces an empty set, in which case we must loop
* back around for another tuple
*/
econtext->ecxt_outertuple = slot;
resultSlot = ExecProject(node->ps.ps_ProjInfo, &isDone);
if (isDone != ExprEndResult) {
node->ps.ps_TupFromTlist = (isDone == ExprMultipleResult);
return resultSlot;
}
}
return slot;
}
/* ----------------------------------------------------------------
* ExecEndGather
*
* frees any storage allocated through C routines.
* ----------------------------------------------------------------
*/
void ExecEndGather(GatherState *node)
{
ExecShutdownGather(node);
ExecFreeExprContext(&node->ps);
(void)ExecClearTuple(node->ps.ps_ResultTupleSlot);
ExecEndNode(outerPlanState(node));
}
/*
* Read the next tuple. We might fetch a tuple from one of the tuple queues
* using gather_readnext, or if no tuple queue contains a tuple and the
* single_copy flag is not set, we might generate one locally instead.
*/
static TupleTableSlot *gather_getnext(GatherState *gatherstate)
{
PlanState *outerPlan = outerPlanState(gatherstate);
TupleTableSlot *fslot = gatherstate->funnel_slot;
while (gatherstate->reader != NULL || gatherstate->need_to_scan_locally) {
CHECK_FOR_INTERRUPTS();
if (gatherstate->reader != NULL) {
HeapTuple tup = gather_readnext(gatherstate);
if (HeapTupleIsValid(tup)) {
(void)ExecStoreTuple(tup, /* tuple to store */
fslot, /* slot in which to store the tuple */
InvalidBuffer, /* buffer associated with this tuple */
true); /* pfree this pointer if not from heap */
return fslot;
}
}
if (gatherstate->need_to_scan_locally) {
TupleTableSlot *outerTupleSlot = ExecProcNode(outerPlan);
if (!TupIsNull(outerTupleSlot))
return outerTupleSlot;
gatherstate->need_to_scan_locally = false;
}
}
return ExecClearTuple(fslot);
}
/*
* Attempt to read a tuple from one of our parallel workers.
*/
static HeapTuple gather_readnext(GatherState *gatherstate)
{
int nvisited = 0;
for (;;) {
bool readerdone = false;
/* Check for async events, particularly messages from workers. */
CHECK_FOR_INTERRUPTS();
/* Attempt to read a tuple, but don't block if none is available. */
TupleQueueReader *reader = gatherstate->reader[gatherstate->nextreader];
HeapTuple tup = TupleQueueReaderNext(reader, true, &readerdone);
/*
* If this reader is done, remove it. If all readers are done,
* clean up remaining worker state.
*/
if (readerdone) {
Assert(!tup);
DestroyTupleQueueReader(reader);
--gatherstate->nreaders;
if (gatherstate->nreaders == 0) {
ExecShutdownGatherWorkers(gatherstate);
return NULL;
}
Size remainSize = sizeof(TupleQueueReader *) * (gatherstate->nreaders - gatherstate->nextreader);
if (remainSize != 0) {
int rc = memmove_s(&gatherstate->reader[gatherstate->nextreader], remainSize,
&gatherstate->reader[gatherstate->nextreader + 1], remainSize);
securec_check(rc, "", "");
}
if (gatherstate->nextreader >= gatherstate->nreaders) {
gatherstate->nextreader = 0;
}
continue;
}
/* If we got a tuple, return it. */
if (tup)
return tup;
/*
* Advance nextreader pointer in round-robin fashion. Note that we
* only reach this code if we weren't able to get a tuple from the
* current worker. We used to advance the nextreader pointer after
* every tuple, but it turns out to be much more efficient to keep
* reading from the same queue until that would require blocking.
*/
gatherstate->nextreader++;
if (gatherstate->nextreader >= gatherstate->nreaders)
gatherstate->nextreader = 0;
/* Have we visited every (surviving) TupleQueueReader? */
nvisited++;
if (nvisited >= gatherstate->nreaders) {
/*
* If (still) running plan locally, return NULL so caller can
* generate another tuple from the local copy of the plan.
*/
if (gatherstate->need_to_scan_locally)
return NULL;
/* Nothing to do except wait for developments. */
(void)WaitLatch(&t_thrd.proc->procLatch, WL_LATCH_SET, 0);
CHECK_FOR_INTERRUPTS();
ResetLatch(&t_thrd.proc->procLatch);
nvisited = 0;
}
}
}
/* ----------------------------------------------------------------
* ExecShutdownGatherWorkers
*
* Destroy the parallel workers. Collect all the stats after
* workers are stopped, else some work done by workers won't be
* accounted.
* ----------------------------------------------------------------
*/
static void ExecShutdownGatherWorkers(GatherState *node)
{
/* Shut down tuple queue readers before shutting down workers. */
if (node->reader != NULL) {
for (int i = 0; i < node->nreaders; ++i)
DestroyTupleQueueReader(node->reader[i]);
pfree(node->reader);
node->reader = NULL;
}
/* Now shut down the workers. */
if (node->pei != NULL)
ExecParallelFinish(node->pei);
}
/* ----------------------------------------------------------------
* ExecShutdownGather
*
* Destroy the setup for parallel workers including parallel context.
* Collect all the stats after workers are stopped, else some work
* done by workers won't be accounted.
* ----------------------------------------------------------------
*/
void ExecShutdownGather(GatherState *node)
{
ExecShutdownGatherWorkers(node);
/* Now destroy the parallel context. */
if (node->pei != NULL) {
ExecParallelCleanup(node->pei);
node->pei = NULL;
}
}
/* ----------------------------------------------------------------
* Join Support
* ----------------------------------------------------------------
*/
/* ----------------------------------------------------------------
* ExecReScanGather
*
* Re-initialize the workers and rescans a relation via them.
* ----------------------------------------------------------------
*/
void ExecReScanGather(GatherState *node)
{
/*
* Re-initialize the parallel workers to perform rescan of relation.
* We want to gracefully shutdown all the workers so that they
* should be able to propagate any error or other information to master
* backend before dying. Parallel context will be reused for rescan.
*/
ExecShutdownGatherWorkers(node);
node->initialized = false;
if (node->pei)
ExecParallelReinitialize(node->pei);
ExecReScan(node->ps.lefttree);
}
/* -------------------------------------------------------------------------
*
* nodeGather.c
* Support routines for scanning a plan via multiple workers.
*
* Portions Copyright (c) 1996-2015, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* A Gather executor launches parallel workers to run multiple copies of a
* plan. It can also run the plan itself, if the workers are not available
* or have not started up yet. It then merges all of the results it produces
* and the results from the workers into a single output stream. Therefore,
* it will normally be used with a plan where running multiple copies of the
* same plan does not produce duplicate output, such as parallel-aware
* SeqScan.
*
* Alternatively, a Gather node can be configured to use just one worker
* and the single-copy flag can be set. In this case, the Gather node will
* run the plan in one worker and will not execute the plan itself. In
* this case, it simply returns whatever tuples were returned by the worker.
* If a worker cannot be obtained, then it will run the plan itself and
* return the results. Therefore, a plan used with a single-copy Gather
* node need not be parallel-aware.
*
* IDENTIFICATION
* src/backend/executor/nodeGather.c
*
* -------------------------------------------------------------------------
*/
#include "postgres.h"
#include "access/relscan.h"
#include "access/xact.h"
#include "executor/execdebug.h"
#include "executor/execParallel.h"
#include "executor/nodeGather.h"
#include "executor/nodeSubplan.h"
#include "executor/tqueue.h"
#include "miscadmin.h"
#include "utils/memutils.h"
#include "utils/rel.h"
static TupleTableSlot *gather_getnext(GatherState *gatherstate);
static HeapTuple gather_readnext(GatherState *gatherstate);
static void ExecShutdownGatherWorkers(GatherState *node);
/* ----------------------------------------------------------------
* ExecInitGather
* ----------------------------------------------------------------
*/
GatherState *ExecInitGather(Gather *node, EState *estate, int eflags)
{
bool hasoid = false;
/* Gather node doesn't have innerPlan node. */
Assert(innerPlan(node) == NULL);
/*
* create state structure
*/
GatherState *gatherstate = makeNode(GatherState);
gatherstate->ps.plan = (Plan *)node;
gatherstate->ps.state = estate;
gatherstate->need_to_scan_locally = !node->single_copy &&
u_sess->attr.attr_sql.parallel_leader_participation;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &gatherstate->ps);
/*
* initialize child expressions
*/
gatherstate->ps.targetlist = (List *)ExecInitExpr((Expr *)node->plan.targetlist, (PlanState *)gatherstate);
gatherstate->ps.qual = (List *)ExecInitExpr((Expr *)node->plan.qual, (PlanState *)gatherstate);
/*
* tuple table initialization
*/
gatherstate->funnel_slot = ExecInitExtraTupleSlot(estate);
ExecInitResultTupleSlot(estate, &gatherstate->ps);
/*
* now initialize outer plan
*/
Plan *outerNode = outerPlan(node);
outerPlanState(gatherstate) = ExecInitNode(outerNode, estate, eflags);
gatherstate->ps.ps_TupFromTlist = false;
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&gatherstate->ps);
ExecAssignProjectionInfo(&gatherstate->ps, NULL);
/*
* Initialize funnel slot to same tuple descriptor as outer plan.
*/
if (!ExecContextForcesOids(&gatherstate->ps, &hasoid))
hasoid = false;
TupleDesc tupDesc = ExecTypeFromTL(outerNode->targetlist, hasoid);
ExecSetSlotDescriptor(gatherstate->funnel_slot, tupDesc);
return gatherstate;
}
/* ----------------------------------------------------------------
* ExecGather(node)
*
* Scans the relation via multiple workers and returns
* the next qualifying tuple.
* ----------------------------------------------------------------
*/
TupleTableSlot *ExecGather(GatherState *node)
{
TupleTableSlot *fslot = node->funnel_slot;
int i;
TupleTableSlot *slot = NULL;
TupleTableSlot *resultSlot = NULL;
ExprDoneCond isDone;
CHECK_FOR_INTERRUPTS();
/*
* Initialize the parallel context and workers on first execution. We do
* this on first execution rather than during node initialization, as it
* needs to allocate large dynamic segement, so it is better to do if it
* is really needed.
*/
if (!node->initialized) {
EState *estate = node->ps.state;
Gather *gather = (Gather *)node->ps.plan;
/*
* Sometimes we might have to run without parallelism; but if
* parallel mode is active then we can try to fire up some workers.
*/
if (gather->num_workers > 0 && IsInParallelMode()) {
bool got_any_worker = false;
/* Initialize the workers required to execute Gather node. */
if (!node->pei)
node->pei = ExecInitParallelPlan(node->ps.lefttree, estate, gather->num_workers);
/*
* Register backend workers. We might not get as many as we
* requested, or indeed any at all.
*/
ParallelContext *pcxt = node->pei->pcxt;
LaunchParallelWorkers(pcxt);
/* Set up tuple queue readers to read the results. */
if (pcxt->nworkers > 0) {
node->nreaders = 0;
node->reader = (TupleQueueReader **)palloc(pcxt->nworkers * sizeof(TupleQueueReader *));
for (i = 0; i < pcxt->nworkers; ++i) {
if (pcxt->worker[i].bgwhandle == NULL)
continue;
shm_mq_set_handle(node->pei->tqueue[i], pcxt->worker[i].bgwhandle);
node->reader[node->nreaders++] =
CreateTupleQueueReader(node->pei->tqueue[i], fslot->tts_tupleDescriptor);
got_any_worker = true;
}
}
/* No workers? Then never mind. */
if (!got_any_worker) {
ExecShutdownGatherWorkers(node);
} else {
t_thrd.subrole = BACKGROUND_LEADER;
}
}
/* Run plan locally if no workers or not single-copy. */
node->need_to_scan_locally = (node->reader == NULL) ||
(!gather->single_copy && u_sess->attr.attr_sql.parallel_leader_participation);
node->initialized = true;
}
/*
* Check to see if we're still projecting out tuples from a previous scan
* tuple (because there is a function-returning-set in the projection
* expressions). If so, try to project another one.
*/
if (node->ps.ps_TupFromTlist) {
resultSlot = ExecProject(node->ps.ps_ProjInfo, &isDone);
if (isDone == ExprMultipleResult)
return resultSlot;
/* Done with that source tuple... */
node->ps.ps_TupFromTlist = false;
}
/*
* Reset per-tuple memory context to free any expression evaluation
* storage allocated in the previous tuple cycle. Note we can't do this
* until we're done projecting. This will also clear any previous tuple
* returned by a TupleQueueReader; to make sure we don't leave a dangling
* pointer around, clear the working slot first.
*/
(void)ExecClearTuple(node->funnel_slot);
ExprContext *econtext = node->ps.ps_ExprContext;
ResetExprContext(econtext);
/* Get and return the next tuple, projecting if necessary. */
for (;;) {
/*
* Get next tuple, either from one of our workers, or by running the
* plan ourselves.
*/
slot = gather_getnext(node);
if (TupIsNull(slot))
return NULL;
/*
* form the result tuple using ExecProject(), and return it --- unless
* the projection produces an empty set, in which case we must loop
* back around for another tuple
*/
econtext->ecxt_outertuple = slot;
resultSlot = ExecProject(node->ps.ps_ProjInfo, &isDone);
if (isDone != ExprEndResult) {
node->ps.ps_TupFromTlist = (isDone == ExprMultipleResult);
return resultSlot;
}
}
return slot;
}
/* ----------------------------------------------------------------
* ExecEndGather
*
* frees any storage allocated through C routines.
* ----------------------------------------------------------------
*/
void ExecEndGather(GatherState *node)
{
ExecShutdownGather(node);
ExecFreeExprContext(&node->ps);
(void)ExecClearTuple(node->ps.ps_ResultTupleSlot);
ExecEndNode(outerPlanState(node));
}
/*
* Read the next tuple. We might fetch a tuple from one of the tuple queues
* using gather_readnext, or if no tuple queue contains a tuple and the
* single_copy flag is not set, we might generate one locally instead.
*/
static TupleTableSlot *gather_getnext(GatherState *gatherstate)
{
PlanState *outerPlan = outerPlanState(gatherstate);
TupleTableSlot *fslot = gatherstate->funnel_slot;
while (gatherstate->reader != NULL || gatherstate->need_to_scan_locally) {
CHECK_FOR_INTERRUPTS();
if (gatherstate->reader != NULL) {
HeapTuple tup = gather_readnext(gatherstate);
if (HeapTupleIsValid(tup)) {
(void)ExecStoreTuple(tup, /* tuple to store */
fslot, /* slot in which to store the tuple */
InvalidBuffer, /* buffer associated with this tuple */
true); /* pfree this pointer if not from heap */
return fslot;
}
}
if (gatherstate->need_to_scan_locally) {
TupleTableSlot *outerTupleSlot = ExecProcNode(outerPlan);
if (!TupIsNull(outerTupleSlot))
return outerTupleSlot;
gatherstate->need_to_scan_locally = false;
}
}
return ExecClearTuple(fslot);
}
/*
* Attempt to read a tuple from one of our parallel workers.
*/
static HeapTuple gather_readnext(GatherState *gatherstate)
{
int nvisited = 0;
for (;;) {
bool readerdone = false;
/* Check for async events, particularly messages from workers. */
CHECK_FOR_INTERRUPTS();
/* Attempt to read a tuple, but don't block if none is available. */
TupleQueueReader *reader = gatherstate->reader[gatherstate->nextreader];
HeapTuple tup = TupleQueueReaderNext(reader, true, &readerdone);
/*
* If this reader is done, remove it. If all readers are done,
* clean up remaining worker state.
*/
if (readerdone) {
Assert(!tup);
DestroyTupleQueueReader(reader);
--gatherstate->nreaders;
if (gatherstate->nreaders == 0) {
ExecShutdownGatherWorkers(gatherstate);
return NULL;
}
Size remainSize = sizeof(TupleQueueReader *) * (gatherstate->nreaders - gatherstate->nextreader);
if (remainSize != 0) {
int rc = memmove_s(&gatherstate->reader[gatherstate->nextreader], remainSize,
&gatherstate->reader[gatherstate->nextreader + 1], remainSize);
securec_check(rc, "", "");
}
if (gatherstate->nextreader >= gatherstate->nreaders) {
gatherstate->nextreader = 0;
}
continue;
}
/* If we got a tuple, return it. */
if (tup)
return tup;
/*
* Advance nextreader pointer in round-robin fashion. Note that we
* only reach this code if we weren't able to get a tuple from the
* current worker. We used to advance the nextreader pointer after
* every tuple, but it turns out to be much more efficient to keep
* reading from the same queue until that would require blocking.
*/
gatherstate->nextreader++;
if (gatherstate->nextreader >= gatherstate->nreaders)
gatherstate->nextreader = 0;
/* Have we visited every (surviving) TupleQueueReader? */
nvisited++;
if (nvisited >= gatherstate->nreaders) {
/*
* If (still) running plan locally, return NULL so caller can
* generate another tuple from the local copy of the plan.
*/
if (gatherstate->need_to_scan_locally)
return NULL;
/* Nothing to do except wait for developments. */
(void)WaitLatch(&t_thrd.proc->procLatch, WL_LATCH_SET, 0);
CHECK_FOR_INTERRUPTS();
ResetLatch(&t_thrd.proc->procLatch);
nvisited = 0;
}
}
}
/* ----------------------------------------------------------------
* ExecShutdownGatherWorkers
*
* Destroy the parallel workers. Collect all the stats after
* workers are stopped, else some work done by workers won't be
* accounted.
* ----------------------------------------------------------------
*/
static void ExecShutdownGatherWorkers(GatherState *node)
{
/* Shut down tuple queue readers before shutting down workers. */
if (node->reader != NULL) {
for (int i = 0; i < node->nreaders; ++i)
DestroyTupleQueueReader(node->reader[i]);
pfree(node->reader);
node->reader = NULL;
}
/* Now shut down the workers. */
if (node->pei != NULL)
ExecParallelFinish(node->pei);
}
/* ----------------------------------------------------------------
* ExecShutdownGather
*
* Destroy the setup for parallel workers including parallel context.
* Collect all the stats after workers are stopped, else some work
* done by workers won't be accounted.
* ----------------------------------------------------------------
*/
void ExecShutdownGather(GatherState *node)
{
ExecShutdownGatherWorkers(node);
/* Now destroy the parallel context. */
if (node->pei != NULL) {
ExecParallelCleanup(node->pei);
node->pei = NULL;
}
}
/* ----------------------------------------------------------------
* Join Support
* ----------------------------------------------------------------
*/
/* ----------------------------------------------------------------
* ExecReScanGather
*
* Re-initialize the workers and rescans a relation via them.
* ----------------------------------------------------------------
*/
void ExecReScanGather(GatherState *node)
{
/*
* Re-initialize the parallel workers to perform rescan of relation.
* We want to gracefully shutdown all the workers so that they
* should be able to propagate any error or other information to master
* backend before dying. Parallel context will be reused for rescan.
*/
ExecShutdownGatherWorkers(node);
node->initialized = false;
if (node->pei)
ExecParallelReinitialize(node->pei);
ExecReScan(node->ps.lefttree);
}

File diff suppressed because it is too large Load Diff

View File

@ -398,18 +398,16 @@ IndexTuple index_truncate_tuple(TupleDesc tupleDescriptor, IndexTuple olditup, i
TupleDesc itupdesc = CreateTupleDescCopyConstr(tupleDescriptor);
Datum values[INDEX_MAX_KEYS];
bool isnull[INDEX_MAX_KEYS];
IndexTuple newitup;
int indnatts = tupleDescriptor->natts;
Assert(indnatts <= INDEX_MAX_KEYS);
Assert(tupleDescriptor->natts <= INDEX_MAX_KEYS);
Assert(new_indnatts > 0);
Assert(new_indnatts < indnatts);
Assert(new_indnatts < tupleDescriptor->natts);
index_deform_tuple(olditup, tupleDescriptor, values, isnull);
/* form new tuple that will contain only key attributes */
itupdesc->natts = new_indnatts;
newitup = index_form_tuple(itupdesc, values, isnull);
IndexTuple newitup = index_form_tuple(itupdesc, values, isnull);
newitup->t_tid = olditup->t_tid;
FreeTupleDesc(itupdesc);

View File

@ -366,9 +366,11 @@ void heapgetpage(HeapScanDesc scan, BlockNumber page)
scan->rs_vistuples[ntup++] = line_off;
}
ereport(DEBUG1,
(errmsg(
"heapgetpage xid %lu ctid(%u,%d) valid %d", GetCurrentTransactionIdIfAny(), page, line_off, valid)));
if (SHOW_DEBUG_MESSAGE()) {
ereport(DEBUG1,
(errmsg(
"heapgetpage xid %lu ctid(%u,%d) valid %d", GetCurrentTransactionIdIfAny(), page, line_off, valid)));
}
}
}

File diff suppressed because it is too large Load Diff

View File

@ -1,63 +1,79 @@
/* -------------------------------------------------------------------------
*
* dsm.c
* manage dynamic shared memory segments
*
* This file provides a set of services to make programming with dynamic
* shared memory segments more convenient. Unlike the low-level
* facilities provided by dsm_impl.h and dsm_impl.c, mappings and segments
* created using this module will be cleaned up automatically. Mappings
* will be removed when the resource owner under which they were created
* is cleaned up, unless dsm_pin_mapping() is used, in which case they
* have session lifespan. Segments will be removed when there are no
* remaining mappings, or at postmaster shutdown in any case. After a
* hard postmaster crash, remaining segments will be removed, if they
* still exist, at the next postmaster startup.
*
* Portions Copyright (c) 2020 Huawei Technologies Co.,Ltd
* Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
*
* IDENTIFICATION
* src/gausskernel/storage/ipc/dsm.c
*
* -------------------------------------------------------------------------
*/
#include "postgres.h"
#include "storage/dsm.h"
#include "knl/knl_session.h"
#include "utils/memutils.h"
#include "postmaster/bgworker_internals.h"
void dsm_detach(void **seg)
{
Assert(*seg != NULL);
knl_u_parallel_context *ctx = (knl_u_parallel_context *)*seg;
MemoryContextDelete(ctx->memCtx);
ctx->memCtx = NULL;
ctx->pwCtx = NULL;
ctx->used = false;
}
void *dsm_create(void)
{
for (int i = 0; i < DSM_MAX_ITEM_PER_QUERY; i++) {
if (u_sess->parallel_ctx[i].used == false) {
u_sess->parallel_ctx[i].memCtx = AllocSetContextCreate(u_sess->top_mem_cxt, "parallel query",
ALLOCSET_DEFAULT_MINSIZE, ALLOCSET_DEFAULT_INITSIZE, ALLOCSET_DEFAULT_MAXSIZE, SHARED_CONTEXT);
MemoryContext oldContext = MemoryContextSwitchTo(u_sess->parallel_ctx[i].memCtx);
u_sess->parallel_ctx[i].pwCtx = (ParallelInfoContext *)palloc0(sizeof(ParallelInfoContext));
(void)MemoryContextSwitchTo(oldContext);
u_sess->parallel_ctx[i].used = true;
return &(u_sess->parallel_ctx[i]);
}
}
ereport(ERROR, (errcode(ERRCODE_INSUFFICIENT_RESOURCES), errmsg("too many dynamic shared memory segments")));
return NULL;
}
/* -------------------------------------------------------------------------
*
* dsm.c
* manage dynamic shared memory segments
*
* This file provides a set of services to make programming with dynamic
* shared memory segments more convenient. Unlike the low-level
* facilities provided by dsm_impl.h and dsm_impl.c, mappings and segments
* created using this module will be cleaned up automatically. Mappings
* will be removed when the resource owner under which they were created
* is cleaned up, unless dsm_pin_mapping() is used, in which case they
* have session lifespan. Segments will be removed when there are no
* remaining mappings, or at postmaster shutdown in any case. After a
* hard postmaster crash, remaining segments will be removed, if they
* still exist, at the next postmaster startup.
*
* Portions Copyright (c) 2020 Huawei Technologies Co.,Ltd
* Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
*
* IDENTIFICATION
* src/gausskernel/storage/ipc/dsm.c
*
* -------------------------------------------------------------------------
*/
#include "postgres.h"
#include "storage/dsm.h"
#include "knl/knl_session.h"
#include "utils/memutils.h"
#include "postmaster/bgworker_internals.h"
#ifdef __USE_NUMA
static void RestoreCpuAffinity(cpu_set_t *cpuset)
{
/* Resotre CPU affinity after parallel query is done. */
if (cpuset != NULL) {
int rc = pthread_setaffinity_np(t_thrd.proc->pid, sizeof(cpu_set_t), cpuset);
if (rc != 0) {
ereport(WARNING, (errmsg("pthread_setaffinity_np failed:%d", rc)));
}
}
}
#endif
void dsm_detach(void **seg)
{
Assert(*seg != NULL);
knl_u_parallel_context *ctx = (knl_u_parallel_context *)*seg;
#ifdef __USE_NUMA
RestoreCpuAffinity(ctx->pwCtx->cpuset);
#endif
MemoryContextDelete(ctx->memCtx);
ctx->memCtx = NULL;
ctx->pwCtx = NULL;
ctx->used = false;
}
void *dsm_create(void)
{
for (int i = 0; i < DSM_MAX_ITEM_PER_QUERY; i++) {
if (u_sess->parallel_ctx[i].used == false) {
u_sess->parallel_ctx[i].memCtx = AllocSetContextCreate(u_sess->top_mem_cxt, "parallel query",
ALLOCSET_DEFAULT_MINSIZE, ALLOCSET_DEFAULT_INITSIZE, ALLOCSET_DEFAULT_MAXSIZE, SHARED_CONTEXT);
MemoryContext oldContext = MemoryContextSwitchTo(u_sess->parallel_ctx[i].memCtx);
u_sess->parallel_ctx[i].pwCtx = (ParallelInfoContext *)palloc0(sizeof(ParallelInfoContext));
(void)MemoryContextSwitchTo(oldContext);
u_sess->parallel_ctx[i].used = true;
return &(u_sess->parallel_ctx[i]);
}
}
ereport(ERROR, (errcode(ERRCODE_INSUFFICIENT_RESOURCES), errmsg("too many dynamic shared memory segments")));
return NULL;
}

View File

@ -608,8 +608,8 @@ shm_mq_result shm_mq_receive(shm_mq_handle *mqh, Size *nbytesp, void **datap, bo
lengthbytes = sizeof(Size) - mqh->mqh_partial_bytes;
else
lengthbytes = rb;
errno_t rc = memcpy_s(&mqh->mqh_buffer[mqh->mqh_partial_bytes], lengthbytes,
rawdata, lengthbytes);
errno_t rc = memcpy_s(&mqh->mqh_buffer[mqh->mqh_partial_bytes],
mqh->mqh_buflen - mqh->mqh_partial_bytes, rawdata, lengthbytes);
securec_check(rc, "\0", "\0");
mqh->mqh_partial_bytes += lengthbytes;
mqh->mqh_consume_pending += MAXALIGN(lengthbytes);
@ -671,7 +671,8 @@ shm_mq_result shm_mq_receive(shm_mq_handle *mqh, Size *nbytesp, void **datap, bo
/* Copy as much as we can. */
Assert(mqh->mqh_partial_bytes + rb <= nbytes);
if (rb != 0) {
errno_t rc = memcpy_s(&mqh->mqh_buffer[mqh->mqh_partial_bytes], rb, rawdata, rb);
errno_t rc = memcpy_s(&mqh->mqh_buffer[mqh->mqh_partial_bytes],
mqh->mqh_buflen - mqh->mqh_partial_bytes, rawdata, rb);
securec_check(rc, "\0", "\0");
mqh->mqh_partial_bytes += rb;
@ -897,8 +898,8 @@ static shm_mq_result shm_mq_send_bytes(shm_mq_handle *mqh, Size nbytes, const vo
* subsequent write to mq_ring, we need a full barrier here.)
*/
pg_memory_barrier();
errno_t rc = memcpy_s(&mq->mq_ring[mq->mq_ring_offset + offset], sendnow,
(char*)data + sent, sendnow);
errno_t rc = memcpy_s(&mq->mq_ring[mq->mq_ring_offset + offset],
ringsize - offset, (char*)data + sent, sendnow);
securec_check(rc, "\0", "\0");
sent += sendnow;

View File

@ -76,10 +76,6 @@ typedef struct knl_instance_attr_common {
bool enable_alarm;
char* Alarm_component;
char* MOTConfigFileName;
int max_worker_processes;
int max_parallel_workers;
int max_parallel_workers_per_gather;
} knl_instance_attr_common;
#endif /* SRC_INCLUDE_KNL_KNL_INSTANCE_ATTR_COMMON_H_ */

View File

@ -205,6 +205,7 @@ typedef struct knl_session_attr_sql {
int opfusion_debug_mode;
int single_shard_stmt;
int force_parallel_mode;
int max_parallel_workers_per_gather;
} knl_session_attr_sql;
#endif /* SRC_INCLUDE_KNL_KNL_SESSION_ATTR_SQL */

View File

@ -2089,7 +2089,10 @@ typedef struct ParallelInfoContext {
char *tupleQueue;
struct SharedExecutorInstrumentation *instrumentation;
char *namespace_search_path;
#ifdef __USE_NUMA
int numaNode;
cpu_set_t *cpuset;
#endif
/* Mutex protects remaining fields. */
slock_t mutex;
/* Maximum XactLastRecEnd of any worker. */

View File

@ -0,0 +1,167 @@
create table parallel_t1(a int);
insert into parallel_t1 values(generate_series(1,100000));
--normal plan for seq scan
explain (costs off) select count(*) from parallel_t1;
QUERY PLAN
-------------------------------
Aggregate
-> Seq Scan on parallel_t1
(2 rows)
explain (costs off) select count(*) from parallel_t1 where a = 5000;
QUERY PLAN
-------------------------------
Aggregate
-> Seq Scan on parallel_t1
Filter: (a = 5000)
(3 rows)
explain (costs off) select count(*) from parallel_t1 where a > 5000;
QUERY PLAN
-------------------------------
Aggregate
-> Seq Scan on parallel_t1
Filter: (a > 5000)
(3 rows)
explain (costs off) select count(*) from parallel_t1 where a < 5000;
QUERY PLAN
-------------------------------
Aggregate
-> Seq Scan on parallel_t1
Filter: (a < 5000)
(3 rows)
explain (costs off) select count(*) from parallel_t1 where a <> 5000;
QUERY PLAN
-------------------------------
Aggregate
-> Seq Scan on parallel_t1
Filter: (a <> 5000)
(3 rows)
select count(*) from parallel_t1;
count
--------
100000
(1 row)
select count(*) from parallel_t1 where a = 5000;
count
-------
1
(1 row)
select count(*) from parallel_t1 where a > 5000;
count
-------
95000
(1 row)
select count(*) from parallel_t1 where a < 5000;
count
-------
4999
(1 row)
select count(*) from parallel_t1 where a <> 5000;
count
-------
99999
(1 row)
--set parallel parameter
set force_parallel_mode=on;
set parallel_setup_cost=0;
set parallel_tuple_cost=0.000005;
set max_parallel_workers_per_gather=2;
set min_parallel_table_scan_size=0;
set parallel_leader_participation=on;
--parallel plan for seq scan
explain (costs off) select count(*) from parallel_t1;
QUERY PLAN
----------------------------------------------
Aggregate
-> Gather
Number of Workers: 2
-> Parallel Seq Scan on parallel_t1
(4 rows)
explain (costs off) select count(*) from parallel_t1 where a = 5000;
QUERY PLAN
----------------------------------------------
Aggregate
-> Gather
Number of Workers: 2
-> Parallel Seq Scan on parallel_t1
Filter: (a = 5000)
(5 rows)
explain (costs off) select count(*) from parallel_t1 where a > 5000;
QUERY PLAN
----------------------------------------------
Aggregate
-> Gather
Number of Workers: 2
-> Parallel Seq Scan on parallel_t1
Filter: (a > 5000)
(5 rows)
explain (costs off) select count(*) from parallel_t1 where a < 5000;
QUERY PLAN
----------------------------------------------
Aggregate
-> Gather
Number of Workers: 2
-> Parallel Seq Scan on parallel_t1
Filter: (a < 5000)
(5 rows)
explain (costs off) select count(*) from parallel_t1 where a <> 5000;
QUERY PLAN
----------------------------------------------
Aggregate
-> Gather
Number of Workers: 2
-> Parallel Seq Scan on parallel_t1
Filter: (a <> 5000)
(5 rows)
select count(*) from parallel_t1;
count
--------
100000
(1 row)
select count(*) from parallel_t1 where a = 5000;
count
-------
1
(1 row)
select count(*) from parallel_t1 where a > 5000;
count
-------
95000
(1 row)
select count(*) from parallel_t1 where a < 5000;
count
-------
4999
(1 row)
select count(*) from parallel_t1 where a <> 5000;
count
-------
99999
(1 row)
--clean up
drop table parallel_t1;
reset force_parallel_mode;
reset parallel_setup_cost;
reset parallel_tuple_cost;
reset max_parallel_workers_per_gather;
reset min_parallel_table_scan_size;
reset parallel_leader_participation;

View File

@ -595,5 +595,8 @@ test: create_procedure create_function pg_compatibility postgres_fdw
# autonomous transaction Test
test: autonomous_transaction
# parallel query
test: parallel_query
# gs_basebackup
test: gs_basebackup

View File

@ -25,3 +25,6 @@ test: upsert_grammer_test_01 upsert_unlog_test upsert_tmp_test
test: upsert_grammer_test_02 upsert_restriction upsert_composite
test: upsert_trigger_test upsert_explain
test: upsert_clean
# test parallel query
test: parallel_query

View File

@ -0,0 +1,42 @@
create table parallel_t1(a int);
insert into parallel_t1 values(generate_series(1,100000));
--normal plan for seq scan
explain (costs off) select count(*) from parallel_t1;
explain (costs off) select count(*) from parallel_t1 where a = 5000;
explain (costs off) select count(*) from parallel_t1 where a > 5000;
explain (costs off) select count(*) from parallel_t1 where a < 5000;
explain (costs off) select count(*) from parallel_t1 where a <> 5000;
select count(*) from parallel_t1;
select count(*) from parallel_t1 where a = 5000;
select count(*) from parallel_t1 where a > 5000;
select count(*) from parallel_t1 where a < 5000;
select count(*) from parallel_t1 where a <> 5000;
--set parallel parameter
set force_parallel_mode=on;
set parallel_setup_cost=0;
set parallel_tuple_cost=0.000005;
set max_parallel_workers_per_gather=2;
set min_parallel_table_scan_size=0;
set parallel_leader_participation=on;
--parallel plan for seq scan
explain (costs off) select count(*) from parallel_t1;
explain (costs off) select count(*) from parallel_t1 where a = 5000;
explain (costs off) select count(*) from parallel_t1 where a > 5000;
explain (costs off) select count(*) from parallel_t1 where a < 5000;
explain (costs off) select count(*) from parallel_t1 where a <> 5000;
select count(*) from parallel_t1;
select count(*) from parallel_t1 where a = 5000;
select count(*) from parallel_t1 where a > 5000;
select count(*) from parallel_t1 where a < 5000;
select count(*) from parallel_t1 where a <> 5000;
--clean up
drop table parallel_t1;
reset force_parallel_mode;
reset parallel_setup_cost;
reset parallel_tuple_cost;
reset max_parallel_workers_per_gather;
reset min_parallel_table_scan_size;
reset parallel_leader_participation;