From 636004199b0e592817c5e3d0eb3f295d16a1e171 Mon Sep 17 00:00:00 2001
From: yihaoDeng <luomoxyz@126.com>
Date: Tue, 27 Sep 2022 21:32:18 +0800
Subject: [PATCH] fix(idx): fix index read crash

---
 source/libs/index/src/indexCache.c   |  1 +
 source/libs/index/src/indexFstFile.c | 42 +++++++++++++++++-----------
 source/libs/index/test/jsonUT.cc     |  4 +--
 3 files changed, 29 insertions(+), 18 deletions(-)

diff --git a/source/libs/index/src/indexCache.c b/source/libs/index/src/indexCache.c
index 7e867db755..39bba4e269 100644
--- a/source/libs/index/src/indexCache.c
+++ b/source/libs/index/src/indexCache.c
@@ -302,6 +302,7 @@ static int32_t cacheSearchCompareFunc_JSON(void* cache, SIndexTerm* term, SIdxTR
         char* p = taosMemoryCalloc(1, strlen(c->colVal) + 1);
         memcpy(p, c->colVal, strlen(c->colVal));
         cond = cmpFn(p + skip, term->colVal, dType);
+        taosMemoryFree(p);
       }
     }
     if (cond == MATCH) {
diff --git a/source/libs/index/src/indexFstFile.c b/source/libs/index/src/indexFstFile.c
index 2a33ddd477..7021fdfae3 100644
--- a/source/libs/index/src/indexFstFile.c
+++ b/source/libs/index/src/indexFstFile.c
@@ -69,6 +69,8 @@ static int idxFileCtxDoReadFrom(IFileCtx* ctx, uint8_t* buf, int len, int32_t of
   int32_t blkOffset = offset % kBlockSize;
   int32_t blkLeft = kBlockSize - blkOffset;
 
+  if (offset >= ctx->file.size) return 0;
+
   do {
     char key[128] = {0};
     idxGenLRUKey(key, ctx->file.buf, blkId);
@@ -80,24 +82,34 @@ static int idxFileCtxDoReadFrom(IFileCtx* ctx, uint8_t* buf, int len, int32_t of
       memcpy(buf + total, blk->buf + blkOffset, nread);
       taosLRUCacheRelease(ctx->lru, h, false);
     } else {
-      int32_t cacheMemSize = sizeof(SDataBlock) + kBlockSize;
+      int32_t left = ctx->file.size - offset;
+      if (left < kBlockSize) {
+        nread = TMIN(left, len);
+        int32_t bytes = taosPReadFile(ctx->file.pFile, buf + total, nread, offset);
+        assert(bytes == nread);
 
-      SDataBlock* blk = taosMemoryCalloc(1, cacheMemSize);
-      blk->blockId = blkId;
-      blk->nread = taosPReadFile(ctx->file.pFile, blk->buf, kBlockSize, blkId * kBlockSize);
-      assert(blk->nread <= kBlockSize);
+        total += bytes;
+        return total;
+      } else {
+        int32_t cacheMemSize = sizeof(SDataBlock) + kBlockSize;
 
-      if (blk->nread < kBlockSize && blk->nread < len) {
-        break;
-      }
+        SDataBlock* blk = taosMemoryCalloc(1, cacheMemSize);
+        blk->blockId = blkId;
+        blk->nread = taosPReadFile(ctx->file.pFile, blk->buf, kBlockSize, blkId * kBlockSize);
+        assert(blk->nread <= kBlockSize);
 
-      nread = TMIN(blkLeft, len);
-      memcpy(buf + total, blk->buf + blkOffset, nread);
+        if (blk->nread < kBlockSize && blk->nread < len) {
+          break;
+        }
 
-      LRUStatus s = taosLRUCacheInsert(ctx->lru, key, strlen(key), blk, cacheMemSize, deleteDataBlockFromLRU, NULL,
-                                       TAOS_LRU_PRIORITY_LOW);
-      if (s != TAOS_LRU_STATUS_OK) {
-        return -1;
+        nread = TMIN(blkLeft, len);
+        memcpy(buf + total, blk->buf + blkOffset, nread);
+
+        LRUStatus s = taosLRUCacheInsert(ctx->lru, key, strlen(key), blk, cacheMemSize, deleteDataBlockFromLRU, NULL,
+                                         TAOS_LRU_PRIORITY_LOW);
+        if (s != TAOS_LRU_STATUS_OK) {
+          return -1;
+        }
       }
     }
     total += nread;
@@ -146,9 +158,7 @@ IFileCtx* idxFileCtxCreate(WriterType type, const char* path, bool readOnly, int
     } else {
       ctx->file.pFile = taosOpenFile(path, TD_FILE_READ);
 
-      int64_t size = 0;
       taosFStatFile(ctx->file.pFile, &ctx->file.size, NULL);
-      ctx->file.size = (int)size;
 #ifdef USE_MMAP
       ctx->file.ptr = (char*)tfMmapReadOnly(ctx->file.pFile, ctx->file.size);
 #endif
diff --git a/source/libs/index/test/jsonUT.cc b/source/libs/index/test/jsonUT.cc
index 1911514d97..8ae3fd4135 100644
--- a/source/libs/index/test/jsonUT.cc
+++ b/source/libs/index/test/jsonUT.cc
@@ -172,9 +172,9 @@ TEST_F(JsonEnv, testWriteMillonData) {
   {
     std::string colName("voltagefdadfa");
     std::string colVal("abxxxxxxxxxxxx");
-    for (int i = 0; i < 10; i++) {
+    for (int i = 0; i < 10000; i++) {
       colVal[i % colVal.size()] = '0' + i % 128;
-      for (size_t i = 0; i < 100; i++) {
+      for (size_t i = 0; i < 10; i++) {
         SIndexTerm* term = indexTermCreateT(1, ADD_VALUE, TSDB_DATA_TYPE_BINARY, colName.c_str(), colName.size(),
                                             colVal.c_str(), colVal.size());