sql: make BtCursor's memory aligned

Vdbe at runtime allocates VdbeCursor structure using
allocateCursor() function. Inside there is a pointer at BtCursor
structure. To make the allocation faster and improve cache
locality, both cursors are allocated in one memory block + some
extra memory for uint32_t array, where BtCursor followed
VdbeCursor and the array without any padding:

   VdbeCursor + uint32_t * N + BtCursor

The problem is that BtCursor needs 8 byte alignment. When it
followed VdbeCursor (aligned by 8) + some uint32_t values, its
actual alignment could become 4 bytes. That led to a crash when
alignment sanitizer is enabled in clang.

The patch makes BtCursor offset aligned by 8 bytes.

Part of #4609

src/box/sql/vdbe.c

@@ -234,10 +234,9 @@ allocateCursor(
 	 */
 	Mem *pMem = iCur>0 ? &p->aMem[p->nMem-iCur] : p->aMem;
 
-	int nByte;
 	VdbeCursor *pCx = 0;
-	nByte =
-		ROUND8(sizeof(VdbeCursor)) + sizeof(u32)*nField +
+	int bt_offset = ROUND8(sizeof(VdbeCursor) + sizeof(uint32_t) * nField);
+	int nByte = bt_offset +
 		(eCurType==CURTYPE_TARANTOOL ? ROUND8(sizeof(BtCursor)) : 0);
 
 	assert(iCur>=0 && iCur<p->nCursor);
@@ -251,8 +250,7 @@ allocateCursor(
 		pCx->eCurType = eCurType;
 		pCx->nField = nField;
 		if (eCurType==CURTYPE_TARANTOOL) {
-			pCx->uc.pCursor = (BtCursor*)
-				&pMem->z[ROUND8(sizeof(VdbeCursor))+sizeof(u32)*nField];
+			pCx->uc.pCursor = (BtCursor*)&pMem->z[bt_offset];
 			sqlCursorZero(pCx->uc.pCursor);
 		}
 	}