postgres 源码解析30 内存管理--2

本小节将着重讲解PostgreSQL内存上下文的工作机制，包括初始化创建、内存上下文的分配与重分配
知识回顾：postgres 源码解析29 内存管理–1

源码解析

1 内存上下文的初始化
每个backend 进程在使用内存上下文之前必须进行初始化，该操作入口为MemoryContextInit，主要工作是创建TopMemoryContext和 ErrorContex，后续操作所需的内存空间将从TopMemoryContext下继续分配。

1. TopMemoryContext：该节点分配后将一直存在，直至系统退出后释放。该节点下面分配的子节点占用的空间均来自本身TopMemoryContext占用空间。
2. ErrorContex：该节点是TopMemoryContext第一个节点，是错误恢复处理的永久性内存上下文。

void
MemoryContextInit(void)
{
	AssertState(TopMemoryContext == NULL);

	/*
	 * First, initialize TopMemoryContext, which is the parent of all others.
	 */
	 // 初始化TopMemoryContext，所有子节点的父节点
	TopMemoryContext = AllocSetContextCreate((MemoryContext) NULL,
											 "TopMemoryContext",
											 ALLOCSET_DEFAULT_SIZES);

	/*
	 * Not having any other place to point CurrentMemoryContext, make it point
	 * to TopMemoryContext.  Caller should change this soon!
	 */
	CurrentMemoryContext = TopMemoryContext;

	/*
	 * Initialize ErrorContext as an AllocSetContext with slow growth rate ---
	 * we don't really expect much to be allocated in it. More to the point,
	 * require it to contain at least 8K at all times. This is the only case
	 * where retained memory in a context is *essential* --- we want to be
	 * sure ErrorContext still has some memory even if we've run out
	 * elsewhere! Also, allow allocations in ErrorContext within a critical
	 * section. Otherwise a PANIC will cause an assertion failure in the error
	 * reporting code, before printing out the real cause of the failure.
	 *
	 * This should be the last step in this function, as elog.c assumes memory
	 * management works once ErrorContext is non-null.
	 */
	ErrorContext = AllocSetContextCreate(TopMemoryContext,
										 "ErrorContext",
										 8 * 1024,
										 8 * 1024,
										 8 * 1024);
	MemoryContextAllowInCriticalSection(ErrorContext, true);
}

/*

2 内存上下文的创建
AllocSetContextCreateInternal总体流程：
1）首先判断在空闲内存块数组中是否存在所需内存块，存在则可使用，并将其从空闲数组中移除，指向父节点；
2）如未找到，则初始化一个指定大小的内存块，设置 AllocSet相关结构体信息，最后调用 MemoryContextCreate函数在父节点下创建内存上下文；

/*
 * AllocSetContextCreateInternal
 *		Create a new AllocSet context.
 *
 * parent: parent context, or NULL if top-level context		// 父节点上下文
 * name: name of context (must be statically allocated)		// debug 用
 * minContextSize: minimum context size						// 最小分配块大小
 * initBlockSize: initial allocation block size            	// 初始分配块大小
 * maxBlockSize: maximum allocation block size				// 最大分配块大小
 *
 * Most callers should abstract the context size parameters using a macro
 * such as ALLOCSET_DEFAULT_SIZES.
 *
 * Note: don't call this directly; go through the wrapper macro
 * AllocSetContextCreate.
 */
MemoryContext
AllocSetContextCreateInternal(MemoryContext parent,
							  const char *name,
							  Size minContextSize,
							  Size initBlockSize,
							  Size maxBlockSize)
{
	int			freeListIndex;
	Size		firstBlockSize;
	AllocSet	set;
	AllocBlock	block;

	/* Assert we padded AllocChunkData properly */
	// 校验
	StaticAssertStmt(ALLOC_CHUNKHDRSZ == MAXALIGN(ALLOC_CHUNKHDRSZ),
					 "sizeof(AllocChunkData) is not maxaligned");
	StaticAssertStmt(offsetof(AllocChunkData, aset) + sizeof(MemoryContext) ==
					 ALLOC_CHUNKHDRSZ,
					 "padding calculation in AllocChunkData is wrong");

	/*
	 * First, validate allocation parameters.  Once these were regular runtime
	 * test and elog's, but in practice Asserts seem sufficient because nobody
	 * varies their parameters at runtime.  We somewhat arbitrarily enforce a
	 * minimum 1K block size.
	 */
	Assert(initBlockSize == MAXALIGN(initBlockSize) &&
		   initBlockSize >= 1024);
	Assert(maxBlockSize == MAXALIGN(maxBlockSize) &&
		   maxBlockSize >= initBlockSize &&
		   AllocHugeSizeIsValid(maxBlockSize)); /* must be safe to double */
	Assert(minContextSize == 0 ||
		   (minContextSize == MAXALIGN(minContextSize) &&
			minContextSize >= 1024 &&
			minContextSize <= maxBlockSize));

	/*
	 * Check whether the parameters match either available freelist.  We do
	 * not need to demand a match of maxBlockSize.
	 */
	if (minContextSize == ALLOCSET_DEFAULT_MINSIZE &&
		initBlockSize == ALLOCSET_DEFAULT_INITSIZE)
		freeListIndex = 0;
	else if (minContextSize == ALLOCSET_SMALL_MINSIZE &&
			 initBlockSize == ALLOCSET_SMALL_INITSIZE)
		freeListIndex = 1;
	else
		freeListIndex = -1;

	/*
	 * If a suitable freelist entry exists, just recycle that context.
	 */
	 /// 在空闲链获取相应的上下文
	if (freeListIndex >= 0)
	{
		AllocSetFreeList *freelist = &context_freelists[freeListIndex];

		if (freelist->first_free != NULL)
		{
			/* Remove entry from freelist */
			set = freelist->first_free;
			freelist->first_free = (AllocSet) set->header.nextchild;
			freelist->num_free--;

			/* Update its maxBlockSize; everything else should be OK */
			set->maxBlockSize = maxBlockSize;

			/* Reinitialize its header, installing correct name and parent */
			MemoryContextCreate((MemoryContext) set,
								T_AllocSetContext,
								&AllocSetMethods,
								parent,
								name);

			((MemoryContext) set)->mem_allocated =
				set->keeper->endptr - ((char *) set);

			return (MemoryContext) set;
		}
	}

	/* Determine size of initial block */
	firstBlockSize = MAXALIGN(sizeof(AllocSetContext)) +
		ALLOC_BLOCKHDRSZ + ALLOC_CHUNKHDRSZ;
	if (minContextSize != 0)
		firstBlockSize = Max(firstBlockSize, minContextSize);
	else
		firstBlockSize = Max(firstBlockSize, initBlockSize);

	/*
	 * Allocate the initial block.  Unlike other aset.c blocks, it starts with
	 * the context header and its block header follows that.
	 */
	set = (AllocSet) malloc(firstBlockSize);
	if (set == NULL)
	{
		if (TopMemoryContext)
			MemoryContextStats(TopMemoryContext);
		ereport(ERROR,
				(errcode(ERRCODE_OUT_OF_MEMORY),
				 errmsg("out of memory"),
				 errdetail("Failed while creating memory context \"%s\".",
						   name)));
	}

	/*
	 * Avoid writing code that can fail between here and MemoryContextCreate;
	 * we'd leak the header/initial block if we ereport in this stretch.
	 */
     // 设置相关字段信息
	/* Fill in the initial block's block header */
	block = (AllocBlock) (((char *) set) + MAXALIGN(sizeof(AllocSetContext)));
	block->aset = set;
	block->freeptr = ((char *) block) + ALLOC_BLOCKHDRSZ;
	block->endptr = ((char *) set) + firstBlockSize;
	block->prev = NULL;
	block->next = NULL;

	/* Mark unallocated space NOACCESS; leave the block header alone. */
	VALGRIND_MAKE_MEM_NOACCESS(block->freeptr, block->endptr - block->freeptr);

	/* Remember block as part of block list */
	set->blocks = block;
	/* Mark block as not to be released at reset time */
	set->keeper = block;

	/* Finish filling in aset-specific parts of the context header */
	MemSetAligned(set->freelist, 0, sizeof(set->freelist));

	set->initBlockSize = initBlockSize;
	set->maxBlockSize = maxBlockSize;
	set->nextBlockSize = initBlockSize;
	set->freeListIndex = freeListIndex;

	/*
	 * Compute the allocation chunk size limit for this context.  It can't be
	 * more than ALLOC_CHUNK_LIMIT because of the fixed number of freelists.
	 * If maxBlockSize is small then requests exceeding the maxBlockSize, or
	 * even a significant fraction of it, should be treated as large chunks
	 * too.  For the typical case of maxBlockSize a power of 2, the chunk size
	 * limit will be at most 1/8th maxBlockSize, so that given a stream of
	 * requests that are all the maximum chunk size we will waste at most
	 * 1/8th of the allocated space.
	 *
	 * We have to have allocChunkLimit a power of two, because the requested
	 * and actually-allocated sizes of any chunk must be on the same side of
	 * the limit, else we get confused about whether the chunk is "big".
	 *
	 * Also, allocChunkLimit must not exceed ALLOCSET_SEPARATE_THRESHOLD.
	 */
	StaticAssertStmt(ALLOC_CHUNK_LIMIT == ALLOCSET_SEPARATE_THRESHOLD,
					 "ALLOC_CHUNK_LIMIT != ALLOCSET_SEPARATE_THRESHOLD");

	set->allocChunkLimit = ALLOC_CHUNK_LIMIT;
	while ((Size) (set->allocChunkLimit + ALLOC_CHUNKHDRSZ) >
		   (Size) ((maxBlockSize - ALLOC_BLOCKHDRSZ) / ALLOC_CHUNK_FRACTION))
		set->allocChunkLimit >>= 1;

	/* Finally, do the type-independent part of context creation */
	MemoryContextCreate((MemoryContext) set,
						T_AllocSetContext,
						&AllocSetMethods,
						parent,
						name);

	((MemoryContext) set)->mem_allocated = firstBlockSize;

	return (MemoryContext) set;
}

3 MemoryContextCreate
该函数是内存上下文创建的真正执行函数，其将已申请的内存块链接至父节点的firstchild位置

/*
 * MemoryContextCreate
 *		Context-type-independent part of context creation.
 *
 * This is only intended to be called by context-type-specific
 * context creation routines, not by the unwashed masses.
 *
 * The memory context creation procedure goes like this:
 *	1.  Context-type-specific routine makes some initial space allocation,
 *		including enough space for the context header.  If it fails,
 *		it can ereport() with no damage done.
 *	2.	Context-type-specific routine sets up all type-specific fields of
 *		the header (those beyond MemoryContextData proper), as well as any
 *		other management fields it needs to have a fully valid context.
 *		Usually, failure in this step is impossible, but if it's possible
 *		the initial space allocation should be freed before ereport'ing.
 *	3.	Context-type-specific routine calls MemoryContextCreate() to fill in
 *		the generic header fields and link the context into the context tree.
 *	4.  We return to the context-type-specific routine, which finishes
 *		up type-specific initialization.  This routine can now do things
 *		that might fail (like allocate more memory), so long as it's
 *		sure the node is left in a state that delete will handle.
 *
 * node: the as-yet-uninitialized common part of the context header node.
 * tag: NodeTag code identifying the memory context type.
 * methods: context-type-specific methods (usually statically allocated).
 * parent: parent context, or NULL if this will be a top-level context.
 * name: name of context (must be statically allocated).
 *
 * Context routines generally assume that MemoryContextCreate can't fail,
 * so this can contain Assert but not elog/ereport.
 */
void
MemoryContextCreate(MemoryContext node,
					NodeTag tag,
					const MemoryContextMethods *methods,
					MemoryContext parent,
					const char *name)
{
	/* Creating new memory contexts is not allowed in a critical section */
	Assert(CritSectionCount == 0);

	/* Initialize all standard fields of memory context header */
	node->type = tag;
	node->isReset = true;
	node->methods = methods;
	node->parent = parent;
	node->firstchild = NULL;
	node->mem_allocated = 0;
	node->prevchild = NULL;
	node->name = name;
	node->ident = NULL;
	node->reset_cbs = NULL;

	/* OK to link node into context tree */
	if (parent)
	{
		node->nextchild = parent->firstchild;
		if (parent->firstchild != NULL)
			parent->firstchild->prevchild = node;
		parent->firstchild = node;
		/* inherit allowInCritSection flag from parent */
		node->allowInCritSection = parent->allowInCritSection;
	}
	else
	{
		node->nextchild = NULL;
		node->allowInCritSection = false;
	}

	VALGRIND_CREATE_MEMPOOL(node, 0, false);
}

4 内存上下文中内存的分配
该操作定义在MemoryContextAlloc函数中，其本质是调用上下文函数中alloc方法来实现指定内存的分配。

/*
 * MemoryContextAlloc
 *		Allocate space within the specified context.
 *
 * This could be turned into a macro, but we'd have to import
 * nodes/memnodes.h into postgres.h which seems a bad idea.
 */
void *
MemoryContextAlloc(MemoryContext context, Size size)
{
	void	   *ret;

	AssertArg(MemoryContextIsValid(context));
	AssertNotInCriticalSection(context);

	if (!AllocSizeIsValid(size))
		elog(ERROR, "invalid memory alloc request size %zu", size);

	context->isReset = false;

	ret = context->methods->alloc(context, size);
	if (unlikely(ret == NULL))
	{
		MemoryContextStats(TopMemoryContext);   // OOM

		/*
		 * Here, and elsewhere in this module, we show the target context's
		 * "name" but not its "ident" (if any) in user-visible error messages.
		 * The "ident" string might contain security-sensitive data, such as
		 * values in SQL commands.
		 */
		ereport(ERROR,
				(errcode(ERRCODE_OUT_OF_MEMORY),
				 errmsg("out of memory"),
				 errdetail("Failed on request of size %zu in memory context \"%s\".",
						   size, context->name)));
	}

	VALGRIND_MEMPOOL_ALLOC(context, ret, size);

	return ret;
}

5 内存上下文的重分配
该操作定义在AllocSetReset函数中，其主要功能是充分配指定内存上下文的内存

/*
 * AllocSetReset
 *		Frees all memory which is allocated in the given set.
 *
 * Actually, this routine has some discretion about what to do.
 * It should mark all allocated chunks freed, but it need not necessarily
 * give back all the resources the set owns.  Our actual implementation is
 * that we give back all but the "keeper" block (which we must keep, since
 * it shares a malloc chunk with the context header).  In this way, we don't
 * thrash malloc() when a context is repeatedly reset after small allocations,
 * which is typical behavior for per-tuple contexts.
 */
static void
AllocSetReset(MemoryContext context)
{
	AllocSet	set = (AllocSet) context;
	AllocBlock	block;
	Size		keepersize PG_USED_FOR_ASSERTS_ONLY
	= set->keeper->endptr - ((char *) set);

	AssertArg(AllocSetIsValid(set));

#ifdef MEMORY_CONTEXT_CHECKING
	/* Check for corruption and leaks before freeing */
	AllocSetCheck(context);
#endif

	/* Clear chunk freelists */
	MemSetAligned(set->freelist, 0, sizeof(set->freelist));

	block = set->blocks;

	/* New blocks list will be just the keeper block */
	set->blocks = set->keeper;

	while (block != NULL)
	{
		AllocBlock	next = block->next;

		if (block == set->keeper)
		{
			/* Reset the block, but don't return it to malloc */
			char	   *datastart = ((char *) block) + ALLOC_BLOCKHDRSZ;

#ifdef CLOBBER_FREED_MEMORY
			wipe_mem(datastart, block->freeptr - datastart);
#else
			/* wipe_mem() would have done this */
			VALGRIND_MAKE_MEM_NOACCESS(datastart, block->freeptr - datastart);
#endif
			block->freeptr = datastart;
			block->prev = NULL;
			block->next = NULL;
		}
		else
		{
			/* Normal case, release the block */
			context->mem_allocated -= block->endptr - ((char *) block);

#ifdef CLOBBER_FREED_MEMORY
			wipe_mem(block, block->freeptr - ((char *) block));
#endif
			free(block);
		}
		block = next;
	}

	Assert(context->mem_allocated == keepersize);

	/* Reset block size allocation sequence, too */
	set->nextBlockSize = set->initBlockSize;
}

6 内存上下文的释放
1 AllocSetFree
释放指定内存上下中的内存片操作定义在AllocSetFree函数中，其主要功能是释放指定上下文已分配的内存片，并将其添加至相应大小的Freelist链表中；但针对大于 allocChunkLimit的内存块会直接从上下文中释放。

/*
 * AllocSetFree
 *		Frees allocated memory; memory is removed from the set.
 */
static void
AllocSetFree(MemoryContext context, void *pointer)
{
	AllocSet	set = (AllocSet) context;
	AllocChunk	chunk = AllocPointerGetChunk(pointer);

	/* Allow access to private part of chunk header. */
	VALGRIND_MAKE_MEM_DEFINED(chunk, ALLOCCHUNK_PRIVATE_LEN);
	
	if (chunk->size > set->allocChunkLimit)
	{
		/*
		 * Big chunks are certain to have been allocated as single-chunk
		 * blocks.  Just unlink that block and return it to malloc().
		 */
		AllocBlock	block = (AllocBlock) (((char *) chunk) - ALLOC_BLOCKHDRSZ);

		/*
		 * Try to verify that we have a sane block pointer: it should
		 * reference the correct aset, and freeptr and endptr should point
		 * just past the chunk.
		 */
		if (block->aset != set ||
			block->freeptr != block->endptr ||
			block->freeptr != ((char *) block) +
			(chunk->size + ALLOC_BLOCKHDRSZ + ALLOC_CHUNKHDRSZ))
			elog(ERROR, "could not find block containing chunk %p", chunk);

		/* OK, remove block from aset's list and free it */
		if (block->prev)
			block->prev->next = block->next;
		else
			set->blocks = block->next;
		if (block->next)
			block->next->prev = block->prev;

		context->mem_allocated -= block->endptr - ((char *) block);

		free(block);
	}
	else
	{
		/* Normal case, put the chunk into appropriate freelist */
		int			fidx = AllocSetFreeIndex(chunk->size);

		chunk->aset = (void *) set->freelist[fidx];

		set->freelist[fidx] = chunk;
	}
}

2 AllocSetDelete
释放指定内存上下文所有的内存块，该函数定义在 AllocSetDelete中

/*
 * AllocSetDelete
 *		Frees all memory which is allocated in the given set,
 *		in preparation for deletion of the set.
 *
 * Unlike AllocSetReset, this *must* free all resources of the set.
 */
static void
AllocSetDelete(MemoryContext context)
{
	AllocSet	set = (AllocSet) context;
	AllocBlock	block = set->blocks;
	Size		keepersize PG_USED_FOR_ASSERTS_ONLY
	= set->keeper->endptr - ((char *) set);

	AssertArg(AllocSetIsValid(set));

	/*
	 * If the context is a candidate for a freelist, put it into that freelist
	 * instead of destroying it.
	 */
	 // 首先找到空闲链表，进行内存释放
	if (set->freeListIndex >= 0)
	{
		AllocSetFreeList *freelist = &context_freelists[set->freeListIndex];

		/*
		 * Reset the context, if it needs it, so that we aren't hanging on to
		 * more than the initial malloc chunk.
		 */
		if (!context->isReset)
			MemoryContextResetOnly(context);

		/*
		 * If the freelist is full, just discard what's already in it.  See
		 * comments with context_freelists[].
		 */
		if (freelist->num_free >= MAX_FREE_CONTEXTS)
		{
			while (freelist->first_free != NULL)
			{
				AllocSetContext *oldset = freelist->first_free;

				freelist->first_free = (AllocSetContext *) oldset->header.nextchild;
				freelist->num_free--;

				/* All that remains is to free the header/initial block */
				free(oldset);
			}
			Assert(freelist->num_free == 0);
		}

		/* Now add the just-deleted context to the freelist. */
		set->header.nextchild = (MemoryContext) freelist->first_free;
		freelist->first_free = set;
		freelist->num_free++;

		return;
	}

	/* Free all blocks, except the keeper which is part of context header */
	//后续释放所有快
	while (block != NULL)
	{
		AllocBlock	next = block->next;

		if (block != set->keeper)
			context->mem_allocated -= block->endptr - ((char *) block);

		if (block != set->keeper)
			free(block);

		block = next;
	}

	Assert(context->mem_allocated == keepersize);

	/* Finally, free the context header, including the keeper block */
	free(set);
}