Skip to content

T
TDengine

Mr、小李 / TDengine
与 Fork 源项目一致

Fork自 taosdata / TDengine

通知 1
Star 0
Fork 0
T
TDengine

体验新版 GitCode，发现更多精彩内容 >>

提交 10b9968d 编写于 作者: H Haojun Liao
浏览文件
操作

[td-225] fix bugs in resbuf

上级 8e97949a
隐藏空白更改
内联 并排
Showing with 170 addition and 69 deletion
src/query/inc/qResultbuf.h
浏览文件 @ 10b9968d
......@@ -49,20 +49,19 @@ typedef struct SDiskbasedResultBuf {
int32_t numOfRowsPerPage;
int32_t numOfPages;
int64_t totalBufSize;
// int32_t fd;
int64_t diskFileSize; // disk file size
FILE* file;
int32_t allocateId; // allocated page id
// int32_t incStep; // minimum allocated pages
void* pBuf; // mmap buffer pointer
char* path; // file path
int32_t pageSize; // current used page size
int32_t inMemPages; // numOfPages that are allocated in memory
SHashObj* idsTable; // id hash table
SHashObj* groupSet; // id hash table
SHashObj* all;
SList* pPageList;
SList* lruList;
void* handle; // for debug purpose
void* emptyDummyIdList; // dummy id list
bool comp;
bool comp; // compressed before flushed to disk
void* assistBuf; // assistant buffer for compress data
SArray* pFree; // free area in file
int32_t nextPos; // next page flush position
} SDiskbasedResultBuf;
......@@ -95,7 +94,7 @@ tFilePage* getNewDataBuf(SDiskbasedResultBuf* pResultBuf, int32_t groupId, int32
* @param pResultBuf
* @return
*/
int32_t getNumOfRowsPerPage(SDiskbasedResultBuf* pResultBuf);
size_t getNumOfRowsPerPage(const SDiskbasedResultBuf* pResultBuf);
/**
*
......@@ -113,6 +112,11 @@ SIDList getDataBufPagesIdList(SDiskbasedResultBuf* pResultBuf, int32_t groupId);
*/
tFilePage* getResBufPage(SDiskbasedResultBuf* pResultBuf, int32_t id);
/**
* release the referenced buf pages
* @param pResultBuf
* @param page
*/
void releaseResBufPage(SDiskbasedResultBuf* pResultBuf, void* page);
/**
......@@ -120,14 +124,14 @@ void releaseResBufPage(SDiskbasedResultBuf* pResultBuf, void* page);
* @param pResultBuf
* @return
*/
int32_t getResBufSize(SDiskbasedResultBuf* pResultBuf);
size_t getResBufSize(const SDiskbasedResultBuf* pResultBuf);
/**
* get the number of groups in the result buffer
* @param pResultBuf
* @return
*/
int32_t getNumOfResultBufGroupId(SDiskbasedResultBuf* pResultBuf);
size_t getNumOfResultBufGroupId(const SDiskbasedResultBuf* pResultBuf);
/**
* destroy result buffer
......
src/query/src/qResultbuf.c
浏览文件 @ 10b9968d
#include "qResultbuf.h"
#include <stddef.h>
#include <tscompression.h>
#include "hash.h"
#include "qExtbuffer.h"
#include "queryLog.h"
......@@ -24,11 +25,13 @@ int32_t createDiskbasedResultBuffer(SDiskbasedResultBuf** pResultBuf, int32_t nu
pResBuf->totalBufSize = pResBuf->numOfPages * pagesize;
pResBuf->allocateId = -1;
pResBuf->pPageList = tdListNew(POINTER_BYTES);
pResBuf->lruList = tdListNew(POINTER_BYTES);
// init id hash table
pResBuf->idsTable = taosHashInit(10, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), false);
pResBuf->groupSet = taosHashInit(10, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), false);
pResBuf->all = taosHashInit(10, taosGetDefaultHashFunction(TSDB_DATA_TYPE_INT), false);
pResBuf->assistBuf = malloc(pResBuf->pageSize + 2); // EXTRA BYTES
pResBuf->comp = true;
char path[PATH_MAX] = {0};
getTmpfilePath("qbuf", path);
......@@ -43,25 +46,28 @@ int32_t createDiskbasedResultBuffer(SDiskbasedResultBuf** pResultBuf, int32_t nu
return TSDB_CODE_SUCCESS;
}
int32_t getNumOfResultBufGroupId(SDiskbasedResultBuf* pResultBuf) { return taosHashGetSize(pResultBuf->idsTable); }
int32_t getResBufSize(SDiskbasedResultBuf* pResultBuf) { return pResultBuf->totalBufSize; }
#define NUM_OF_PAGES_ON_DISK(_r) ((_r)->numOfPages - (_r)->inMemPages)
#define FILE_SIZE_ON_DISK(_r) (NUM_OF_PAGES_ON_DISK(_r) * (_r)->pageSize)
static int32_t createDiskResidesBuf(SDiskbasedResultBuf* pResultBuf) {
// pResultBuf->fd = open(pResultBuf->path, O_CREAT | O_RDWR, 0666);
pResultBuf->file = fopen(pResultBuf->path, "w");
pResultBuf->file = fopen(pResultBuf->path, "wb+");
if (pResultBuf->file == NULL) {
qError("failed to create tmp file: %s on disk. %s", pResultBuf->path, strerror(errno));
return TAOS_SYSTEM_ERROR(errno);
}
return TSDB_CODE_SUCCESS;
}
static char* doCompressData(void* data, int32_t srcSize, int32_t *dst) { // do nothing
*dst = srcSize;
static char* doCompressData(void* data, int32_t srcSize, int32_t *dst, bool comp, void* assistBuf) { // do nothing
if (!comp) {
*dst = srcSize;
return data;
}
*dst = tsCompressString(data, srcSize, 1, assistBuf, srcSize, ONE_STAGE_COMP, NULL, 0);
memcpy(data, assistBuf, *dst);
return data;
}
......@@ -89,57 +95,64 @@ static int32_t allocatePositionInFile(SDiskbasedResultBuf* pResultBuf, size_t si
}
}
static void doFlushPageToDisk(SDiskbasedResultBuf* pResultBuf, SPageInfo* pg) {
assert(T_REF_VAL_GET(pg) == 0);
static char* doFlushPageToDisk(SDiskbasedResultBuf* pResultBuf, SPageInfo* pg) {
assert(T_REF_VAL_GET(pg) == 0 && pg->pData != NULL);
int32_t size = -1;
char* t = doCompressData(pg->pData + POINTER_BYTES, pResultBuf->pageSize, &size);
char* t = doCompressData(pg->pData + POINTER_BYTES, pResultBuf->pageSize, &size, pResultBuf->comp, pResultBuf->assistBuf);
pg->info.length = size;
// this page is flushed to disk for the first time
if (pg->info.offset == -1) {
int32_t offset = allocatePositionInFile(pResultBuf, size);
pg->info.offset = allocatePositionInFile(pResultBuf, size);
pResultBuf->nextPos += size;
fseek(pResultBuf->file, offset, SEEK_SET);
fwrite(t, size, 1, pResultBuf->file);
fseek(pResultBuf->file, pg->info.offset, SEEK_SET);
int32_t ret = fwrite(t, 1, size, pResultBuf->file);
UNUSED(ret);
} else {
if (pg->info.length < size) { // length becomes greater, current space is not enough, allocate new place.
//1. add current space to free list
taosArrayPush(pResultBuf->pFree, &pg->info);
//2. allocate new position, and update the info
int32_t offset = allocatePositionInFile(pResultBuf, size);
pg->info.offset = allocatePositionInFile(pResultBuf, size);
pResultBuf->nextPos += size;
//3. write to disk.
fseek(pResultBuf->file, offset, SEEK_SET);
fseek(pResultBuf->file, pg->info.offset, SEEK_SET);
fwrite(t, size, 1, pResultBuf->file);
}
}
char* ret = pg->pData;
pg->pData = NULL;
return ret;
}
static int32_t flushPageToDisk(SDiskbasedResultBuf* pResultBuf, SPageInfo* pg) {
static char* flushPageToDisk(SDiskbasedResultBuf* pResultBuf, SPageInfo* pg) {
int32_t ret = TSDB_CODE_SUCCESS;
assert(pResultBuf->numOfPages * pResultBuf->pageSize == pResultBuf->totalBufSize && pResultBuf->numOfPages >= pResultBuf->inMemPages);
if (pResultBuf->pBuf == NULL) {
assert(pResultBuf->file == NULL);
if (pResultBuf->file == NULL) {
if ((ret = createDiskResidesBuf(pResultBuf)) != TSDB_CODE_SUCCESS) {
return ret;
terrno = ret;
return NULL;
}
}
doFlushPageToDisk(pResultBuf, pg);
return TSDB_CODE_SUCCESS;
return doFlushPageToDisk(pResultBuf, pg);
}
#define NO_AVAILABLE_PAGES(_b) ((_b)->numOfPages >= (_b)->inMemPages)
static SIDList addNewGroup(SDiskbasedResultBuf* pResultBuf, int32_t groupId) {
assert(taosHashGet(pResultBuf->idsTable, (const char*) &groupId, sizeof(int32_t)) == NULL);
assert(taosHashGet(pResultBuf->groupSet, (const char*) &groupId, sizeof(int32_t)) == NULL);
SArray* pa = taosArrayInit(1, sizeof(SPageInfo));
int32_t ret = taosHashPut(pResultBuf->idsTable, (const char*)&groupId, sizeof(int32_t), &pa, POINTER_BYTES);
int32_t ret = taosHashPut(pResultBuf->groupSet, (const char*)&groupId, sizeof(int32_t), &pa, POINTER_BYTES);
assert(ret == 0);
return pa;
......@@ -148,7 +161,7 @@ static SIDList addNewGroup(SDiskbasedResultBuf* pResultBuf, int32_t groupId) {
static SPageInfo* registerPage(SDiskbasedResultBuf* pResultBuf, int32_t groupId, int32_t pageId) {
SIDList list = NULL;
char** p = taosHashGet(pResultBuf->idsTable, (const char*)&groupId, sizeof(int32_t));
char** p = taosHashGet(pResultBuf->groupSet, (const char*)&groupId, sizeof(int32_t));
if (p == NULL) { // it is a new group id
list = addNewGroup(pResultBuf, groupId);
} else {
......@@ -162,10 +175,13 @@ static SPageInfo* registerPage(SDiskbasedResultBuf* pResultBuf, int32_t groupId,
}
tFilePage* getNewDataBuf(SDiskbasedResultBuf* pResultBuf, int32_t groupId, int32_t* pageId) {
char* allocPg = NULL;
if (NO_AVAILABLE_PAGES(pResultBuf)) {
// get the last page in linked list
SListIter iter = {0};
tdListInitIter(pResultBuf->pPageList, &iter, TD_LIST_BACKWARD);
tdListInitIter(pResultBuf->lruList, &iter, TD_LIST_BACKWARD);
SListNode* pn = NULL;
while((pn = tdListNext(&iter)) != NULL) {
......@@ -183,11 +199,12 @@ tFilePage* getNewDataBuf(SDiskbasedResultBuf* pResultBuf, int32_t groupId, int32
qWarn("%p in memory buf page not sufficient, expand from %d to %d, page size:%d", pResultBuf, prev,
pResultBuf->inMemPages, pResultBuf->pageSize);
} else {
tdListPopNode(pResultBuf->pPageList, pn);
tdListPopNode(pResultBuf->lruList, pn);
SPageInfo* d = *(SPageInfo**) pn->data;
tfree(pn);
if (flushPageToDisk(pResultBuf, d) != TSDB_CODE_SUCCESS) {
allocPg = flushPageToDisk(pResultBuf, d);
if (allocPg == NULL) {
return NULL;
}
}
......@@ -200,14 +217,19 @@ tFilePage* getNewDataBuf(SDiskbasedResultBuf* pResultBuf, int32_t groupId, int32
SPageInfo* pi = registerPage(pResultBuf, groupId, *pageId);
// add to LRU list
assert(listNEles(pResultBuf->pPageList) < pResultBuf->inMemPages);
tdListPrepend(pResultBuf->pPageList, &pi);
assert(listNEles(pResultBuf->lruList) < pResultBuf->inMemPages);
tdListPrepend(pResultBuf->lruList, &pi);
// add to hash map
taosHashPut(pResultBuf->all, pageId, sizeof(int32_t), &pi, POINTER_BYTES);
// allocate buf
pi->pData = calloc(1, pResultBuf->pageSize + POINTER_BYTES);
if (allocPg == NULL) {
pi->pData = calloc(1, pResultBuf->pageSize + POINTER_BYTES);
} else {
pi->pData = allocPg;
}
pResultBuf->totalBufSize += pResultBuf->pageSize;
T_REF_INC(pi); // add ref count
......@@ -231,41 +253,47 @@ tFilePage* getResBufPage(SDiskbasedResultBuf* pResultBuf, int32_t id) {
SListNode* pnode = NULL; // todo speed up
SListIter iter = {0};
tdListInitIter(pResultBuf->pPageList, &iter, TD_LIST_FORWARD);
tdListInitIter(pResultBuf->lruList, &iter, TD_LIST_FORWARD);
while((pnode = tdListNext(&iter)) != NULL) {
SPageInfo** pInfo = (SPageInfo**) pnode->data;
// remove it and add it into the front of linked-list
if ((*pInfo)->pageId == id) {
tdListPopNode(pResultBuf->pPageList, pnode);
tdListPrependNode(pResultBuf->pPageList, pnode);
tdListPopNode(pResultBuf->lruList, pnode);
tdListPrependNode(pResultBuf->lruList, pnode);
T_REF_INC(*(SPageInfo**)pnode->data);
return ((*(SPageInfo**)pnode->data)->pData + POINTER_BYTES);
}
}
} else { // not in memory
// choose the be flushed page
// get the last page in linked list
assert((*pi)->pData == NULL && (*pi)->info.length >= 0 && (*pi)->info.offset >= 0);
// choose the be flushed page: get the last page in linked list
SListIter iter1 = {0};
tdListInitIter(pResultBuf->pPageList, &iter1, TD_LIST_BACKWARD);
tdListInitIter(pResultBuf->lruList, &iter1, TD_LIST_BACKWARD);
SListNode* pn = NULL;
while((pn = tdListNext(&iter1)) != NULL) {
assert(pn != NULL);
if (T_REF_VAL_GET(*(SPageInfo**)pn->data) == 0) {
if (T_REF_VAL_GET(*(SPageInfo**)(pn->data)) == 0) {
break;
}
}
// all pages are referenced by user, try to allocate new space
if (pn == NULL) {
int32_t prev = pResultBuf->inMemPages;
pResultBuf->inMemPages = pResultBuf->inMemPages * 1.5;
assert(0);
return NULL;
qWarn("%p in memory buf page not sufficient, expand from %d to %d, page size:%d", pResultBuf, prev,
pResultBuf->inMemPages, pResultBuf->pageSize);
(*pi)->pData = calloc(1, pResultBuf->pageSize + POINTER_BYTES);
} else {
tdListPopNode(pResultBuf->pPageList, pn);
tdListPopNode(pResultBuf->lruList, pn);
if (flushPageToDisk(pResultBuf, *(SPageInfo**)pn->data) != TSDB_CODE_SUCCESS) {
return NULL;
}
......@@ -273,15 +301,23 @@ tFilePage* getResBufPage(SDiskbasedResultBuf* pResultBuf, int32_t id) {
char* buf = (*(SPageInfo**)pn->data)->pData;
(*(SPageInfo**)pn->data)->pData = NULL;
// load file in disk
fseek(pResultBuf->file, (*pi)->info.offset, SEEK_SET);
fread(buf, (*pi)->info.length, 1, pResultBuf->file);
(*pi)->pData = buf;
((void**)((*pi)->pData))[0] = (*pi);
tfree(pn);
return (*pi)->pData;
}
// load file in disk
int32_t ret = fseek(pResultBuf->file, (*pi)->info.offset, SEEK_SET);
ret = fread((*pi)->pData + POINTER_BYTES, 1, (*pi)->info.length, pResultBuf->file);
if (ret != (*pi)->info.length) {
terrno = errno;
return NULL;
}
// todo do decomp
return (*pi)->pData + POINTER_BYTES;
}
return NULL;
......@@ -297,12 +333,16 @@ void releaseResBufPage(SDiskbasedResultBuf* pResultBuf, void* page) {
T_REF_DEC(ppi);
}
int32_t getNumOfRowsPerPage(SDiskbasedResultBuf* pResultBuf) { return pResultBuf->numOfRowsPerPage; }
size_t getNumOfRowsPerPage(const SDiskbasedResultBuf* pResultBuf) { return pResultBuf->numOfRowsPerPage; }
size_t getNumOfResultBufGroupId(const SDiskbasedResultBuf* pResultBuf) { return taosHashGetSize(pResultBuf->groupSet); }
size_t getResBufSize(const SDiskbasedResultBuf* pResultBuf) { return pResultBuf->totalBufSize; }
SIDList getDataBufPagesIdList(SDiskbasedResultBuf* pResultBuf, int32_t groupId) {
assert(pResultBuf != NULL);
char** p = taosHashGet(pResultBuf->idsTable, (const char*)&groupId, sizeof(int32_t));
char** p = taosHashGet(pResultBuf->groupSet, (const char*)&groupId, sizeof(int32_t));
if (p == NULL) { // it is a new group id
return pResultBuf->emptyDummyIdList;
} else {
......@@ -320,7 +360,6 @@ void destroyResultBuf(SDiskbasedResultBuf* pResultBuf, void* handle) {
pResultBuf->totalBufSize, pResultBuf->path, FILE_SIZE_ON_DISK(pResultBuf));
fclose(pResultBuf->file);
pResultBuf->pBuf = NULL;
} else {
qDebug("QInfo:%p disk-based output buffer closed, total:%" PRId64 " bytes, no file created", handle,
pResultBuf->totalBufSize);
......@@ -329,19 +368,25 @@ void destroyResultBuf(SDiskbasedResultBuf* pResultBuf, void* handle) {
unlink(pResultBuf->path);
tfree(pResultBuf->path);
SHashMutableIterator* iter = taosHashCreateIter(pResultBuf->idsTable);
SHashMutableIterator* iter = taosHashCreateIter(pResultBuf->groupSet);
while(taosHashIterNext(iter)) {
SArray** p = (SArray**) taosHashIterGet(iter);
size_t n = taosArrayGetSize(*p);
for(int32_t i = 0; i < n; ++i) {
SPageInfo* pi = taosArrayGet(*p, i);
tfree(pi->pData);
}
taosArrayDestroy(*p);
}
taosHashDestroyIter(iter);
tdListFree(pResultBuf->pPageList);
tdListFree(pResultBuf->lruList);
taosArrayDestroy(pResultBuf->emptyDummyIdList);
taosHashCleanup(pResultBuf->idsTable);
taosHashCleanup(pResultBuf->groupSet);
taosHashCleanup(pResultBuf->all);
tfree(pResultBuf->assistBuf);
tfree(pResultBuf);
}
......
src/query/tests/resultBufferTest.cpp
浏览文件 @ 10b9968d
......@@ -29,28 +29,80 @@ void simpleTest() {
tFilePage* pBufPage1 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t = getResBufPage(pResultBuf, pageId);
assert(t == pBufPage1);
ASSERT_TRUE(t == pBufPage1);
tFilePage* pBufPage2 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t1 = getResBufPage(pResultBuf, pageId);
assert(t1 == pBufPage2);
ASSERT_TRUE(t1 == pBufPage2);
tFilePage* pBufPage3 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t2 = getResBufPage(pResultBuf, pageId);
assert(t2 == pBufPage3);
ASSERT_TRUE(t2 == pBufPage3);
tFilePage* pBufPage4 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t3 = getResBufPage(pResultBuf, pageId);
assert(t3 == pBufPage4);
ASSERT_TRUE(t3 == pBufPage4);
tFilePage* pBufPage5 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t4 = getResBufPage(pResultBuf, pageId);
assert(t4 == pBufPage5);
ASSERT_TRUE(t4 == pBufPage5);
destroyResultBuf(pResultBuf, NULL);
}
void writeDownTest() {
SDiskbasedResultBuf* pResultBuf = NULL;
int32_t ret = createDiskbasedResultBuffer(&pResultBuf, 1000, 64, 1024, 4, NULL);
int32_t pageId = 0;
int32_t writePageId = 0;
int32_t groupId = 0;
int32_t nx = 12345;
tFilePage* pBufPage = getNewDataBuf(pResultBuf, groupId, &pageId);
ASSERT_TRUE(pBufPage != NULL);
*(int32_t*)(pBufPage->data) = nx;
writePageId = pageId;
releaseResBufPage(pResultBuf, pBufPage);
tFilePage* pBufPage1 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t1 = getResBufPage(pResultBuf, pageId);
ASSERT_TRUE(t1 == pBufPage1);
ASSERT_TRUE(pageId == 1);
tFilePage* pBufPage2 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t2 = getResBufPage(pResultBuf, pageId);
ASSERT_TRUE(t2 == pBufPage2);
ASSERT_TRUE(pageId == 2);
tFilePage* pBufPage3 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t3 = getResBufPage(pResultBuf, pageId);
ASSERT_TRUE(t3 == pBufPage3);
ASSERT_TRUE(pageId == 3);
tFilePage* pBufPage4 = getNewDataBuf(pResultBuf, groupId, &pageId);
tFilePage* t4 = getResBufPage(pResultBuf, pageId);
ASSERT_TRUE(t4 == pBufPage4);
ASSERT_TRUE(pageId == 4);
releaseResBufPage(pResultBuf, t4);
// flush the written page to disk, and read it out again
tFilePage* pBufPagex = getResBufPage(pResultBuf, writePageId);
ASSERT_EQ(*(int32_t*)pBufPagex->data, nx);
SArray* pa = getDataBufPagesIdList(pResultBuf, groupId);
ASSERT_EQ(taosArrayGetSize(pa), 5);
destroyResultBuf(pResultBuf, NULL);
}
} // namespace
TEST(testCase, resultBufferTest) {
srand(time(NULL));
simpleTest();
writeDownTest();
}
src/util/src/tcompare.c
浏览文件 @ 10b9968d
#include "taosdef.h"
#include "tcompare.h"
#include <tarray.h>
#include "tarray.h"
#include "tutil.h"
int32_t compareInt32Val(const void *pLeft, const void *pRight) {
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或