Revert "hugetlb: make free_huge_page irq safe"

	int ret = 0;

	struct hstate *h = &default_hstate;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	nid = vma->vm_flags >> CHECKNODE_BITS;

	}

err:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return ret;

}

static int hugetlb_acct_memory(struct hstate *h, long delta,

			       struct dhugetlb_pool *hpool);

static inline void unlock_or_release_subpool(struct hugepage_subpool *spool,

						unsigned long irq_flags)

static inline void unlock_or_release_subpool(struct hugepage_subpool *spool)

{

	bool free = (spool->count == 0) && (spool->used_hpages == 0);

	spin_unlock_irqrestore(&spool->lock, irq_flags);

	spin_unlock(&spool->lock);

	/* If no pages are used, and no other handles to the subpool

	 * remain, give up any reservations mased on minimum size and

void hugepage_put_subpool(struct hugepage_subpool *spool)

{

	unsigned long flags;

	spin_lock_irqsave(&spool->lock, flags);

	spin_lock(&spool->lock);

	BUG_ON(!spool->count);

	spool->count--;

	unlock_or_release_subpool(spool, flags);

	unlock_or_release_subpool(spool);

}

/*

	if (dhugetlb_enabled && hpool)

		return ret;

	spin_lock_irq(&spool->lock);

	spin_lock(&spool->lock);

	if (spool->max_hpages != -1) {		/* maximum size accounting */

		if ((spool->used_hpages + delta) <= spool->max_hpages)

	}

unlock_ret:

	spin_unlock_irq(&spool->lock);

	spin_unlock(&spool->lock);

	return ret;

}

				       long delta, struct dhugetlb_pool *hpool)

{

	long ret = delta;

	unsigned long flags;

	if (!spool)

		return delta;

	if (dhugetlb_enabled && hpool)

		return ret;

	spin_lock_irqsave(&spool->lock, flags);

	spin_lock(&spool->lock);

	if (spool->max_hpages != -1)		/* maximum size accounting */

		spool->used_hpages -= delta;

	 * If hugetlbfs_put_super couldn't free spool due to an outstanding

	 * quota reference, free it now.

	 */

	unlock_or_release_subpool(spool, flags);

	unlock_or_release_subpool(spool);

	return ret;

}

}

#endif

void free_huge_page(struct page *page)

static void __free_huge_page(struct page *page)

{

	/*

	 * Can't pass hstate in here because it is called from the

	struct hugepage_subpool *spool =

		(struct hugepage_subpool *)page_private(page);

	bool restore_reserve;

	unsigned long flags;

	sp_kmemcg_uncharge_hpage(page);

	set_page_private(page, 0);

	ClearPagePrivate(page);

	if (dhugetlb_enabled && PagePool(page)) {

		spin_lock_irqsave(&hugetlb_lock, flags);

		spin_lock(&hugetlb_lock);

		clear_page_huge_active(page);

		list_del(&page->lru);

		hugetlb_cgroup_uncharge_page(hstate_index(h),

					     pages_per_huge_page(h), page);

		spin_unlock_irqrestore(&hugetlb_lock, flags);

		spin_unlock(&hugetlb_lock);

		free_huge_page_to_dhugetlb_pool(page, restore_reserve);

		return;

	}

			restore_reserve = true;

	}

	spin_lock_irqsave(&hugetlb_lock, flags);

	spin_lock(&hugetlb_lock);

	clear_page_huge_active(page);

	hugetlb_cgroup_uncharge_page(hstate_index(h),

				     pages_per_huge_page(h), page);

		arch_clear_hugepage_flags(page);

		enqueue_huge_page(h, page);

	}

	spin_unlock_irqrestore(&hugetlb_lock, flags);

	spin_unlock(&hugetlb_lock);

}

/*

 * As free_huge_page() can be called from a non-task context, we have

 * to defer the actual freeing in a workqueue to prevent potential

 * hugetlb_lock deadlock.

 *

 * free_hpage_workfn() locklessly retrieves the linked list of pages to

 * be freed and frees them one-by-one. As the page->mapping pointer is

 * going to be cleared in __free_huge_page() anyway, it is reused as the

 * llist_node structure of a lockless linked list of huge pages to be freed.

 */

static LLIST_HEAD(hpage_freelist);

static void free_hpage_workfn(struct work_struct *work)

{

	struct llist_node *node;

	struct page *page;

	node = llist_del_all(&hpage_freelist);

	while (node) {

		page = container_of((struct address_space **)node,

				     struct page, mapping);

		node = node->next;

		__free_huge_page(page);

	}

}

static DECLARE_WORK(free_hpage_work, free_hpage_workfn);

void free_huge_page(struct page *page)

{

	/*

	 * Defer freeing if in non-task context to avoid hugetlb_lock deadlock.

	 */

	if (!in_task()) {

		/*

		 * Only call schedule_work() if hpage_freelist is previously

		 * empty. Otherwise, schedule_work() had been called but the

		 * workfn hasn't retrieved the list yet.

		 */

		if (llist_add((struct llist_node *)&page->mapping,

			      &hpage_freelist))

			schedule_work(&free_hpage_work);

		return;

	}

	__free_huge_page(page);

}

static void prep_new_huge_page(struct hstate *h, struct page *page, int nid)

{

	INIT_LIST_HEAD(&page->lru);

	set_compound_page_dtor(page, HUGETLB_PAGE_DTOR);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	set_hugetlb_cgroup(page, NULL);

	h->nr_huge_pages++;

	h->nr_huge_pages_node[nid]++;

	ClearPageHugeFreed(page);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

}

static void prep_compound_gigantic_page(struct page *page, unsigned int order)

	if (page_belong_to_dynamic_hugetlb(page))

		return -EBUSY;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (!PageHuge(page)) {

		rc = 0;

		goto out;

		 * when it is dissolved.

		 */

		if (unlikely(!PageHugeFreed(head))) {

			spin_unlock_irq(&hugetlb_lock);

			spin_unlock(&hugetlb_lock);

			cond_resched();

			/*

		rc = 0;

	}

out:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return rc;

}

	if (hstate_is_gigantic(h))

		return NULL;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (h->surplus_huge_pages >= h->nr_overcommit_huge_pages)

		goto out_unlock;

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	page = alloc_fresh_huge_page(h, gfp_mask, nid, nmask, NULL);

	if (!page)

		return NULL;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	/*

	 * We could have raced with the pool size change.

	 * Double check that and simply deallocate the new page

	 */

	if (h->surplus_huge_pages >= h->nr_overcommit_huge_pages) {

		SetPageHugeTemporary(page);

		spin_unlock_irq(&hugetlb_lock);

		spin_unlock(&hugetlb_lock);

		put_page(page);

		return NULL;

	} else {

	}

out_unlock:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return page;

}

	if (nid != NUMA_NO_NODE)

		gfp_mask |= __GFP_THISNODE;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (h->free_huge_pages - h->resv_huge_pages > 0)

		page = dequeue_huge_page_nodemask(h, gfp_mask, nid, NULL, NULL);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	if (!page) {

		if (enable_charge_mighp)

{

	gfp_t gfp_mask = htlb_alloc_mask(h);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (h->free_huge_pages - h->resv_huge_pages > 0) {

		struct page *page;

		page = dequeue_huge_page_nodemask(h, gfp_mask, preferred_nid,

						  nmask, NULL);

		if (page) {

			spin_unlock_irq(&hugetlb_lock);

			spin_unlock(&hugetlb_lock);

			return page;

		}

	}

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return alloc_migrate_huge_page(h, gfp_mask, preferred_nid, nmask);

}

	ret = -ENOMEM;

retry:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	for (i = 0; i < needed; i++) {

		page = alloc_surplus_huge_page(h, htlb_alloc_mask(h),

				NUMA_NO_NODE, NULL);

	 * After retaking hugetlb_lock, we need to recalculate 'needed'

	 * because either resv_huge_pages or free_huge_pages may have changed.

	 */

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	needed = (h->resv_huge_pages + delta) -

			(h->free_huge_pages + allocated);

	if (needed > 0) {

		enqueue_huge_page(h, page);

	}

free:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	/* Free unnecessary surplus pages to the buddy allocator */

	list_for_each_entry_safe(page, tmp, &surplus_list, lru)

		put_page(page);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	return ret;

}

			 * Use hugetlb_lock to manage the account of

			 * hugetlb cgroup.

			 */

			spin_lock_irq(&hugetlb_lock);

			spin_lock(&hugetlb_lock);

			list_add(&page->lru, &h->hugepage_activelist);

			hugetlb_cgroup_commit_charge(idx,

				pages_per_huge_page(hstate_vma(vma)),

				h_cg, page);

			spin_unlock_irq(&hugetlb_lock);

			spin_unlock(&hugetlb_lock);

			goto out;

		}

		goto out_uncharge_cgroup;

	}

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	/*

	 * glb_chg is passed to indicate whether or not a page must be taken

	 * from the global free pool (global change).  gbl_chg == 0 indicates

	 */

	page = dequeue_huge_page_vma(h, vma, addr, avoid_reserve, gbl_chg);

	if (!page) {

		spin_unlock_irq(&hugetlb_lock);

		spin_unlock(&hugetlb_lock);

		page = alloc_buddy_huge_page_with_mpol(h, vma, addr);

		if (!page)

			goto out_uncharge_cgroup;

		spin_lock_irq(&hugetlb_lock);

		spin_lock(&hugetlb_lock);

		if (!avoid_reserve && vma_has_reserves(vma, gbl_chg)) {

			SetPagePrivate(page);

			h->resv_huge_pages--;

		/* Fall through */

	}

	hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h), h_cg, page);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

out:

	set_page_private(page, (unsigned long)spool);

		return h->max_huge_pages;

	}

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	/*

	 * Check for a node specific request.

		 * page, free_huge_page will handle it by freeing the page

		 * and reducing the surplus.

		 */

		spin_unlock_irq(&hugetlb_lock);

		spin_unlock(&hugetlb_lock);

		/* yield cpu to avoid soft lockup */

		cond_resched();

		ret = alloc_pool_huge_page(h, nodes_allowed,

						node_alloc_noretry);

		spin_lock_irq(&hugetlb_lock);

		spin_lock(&hugetlb_lock);

		if (!ret)

			goto out;

	}

out:

	ret = persistent_huge_pages(h);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	NODEMASK_FREE(node_alloc_noretry);

	if (err)

		return err;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	h->nr_overcommit_huge_pages = input;

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return count;

}

		return -ENOMEM;

	spin_lock(&hpool->lock);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (h->free_huge_pages_node[nid] < size) {

		ret = -ENOMEM;

		goto out_unlock;

	}

	ret = 0;

out_unlock:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	spin_unlock(&hpool->lock);

	return ret;

}

	if (!h)

		return;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	list_for_each_entry_safe(page, page_next,

				 &hpool->dhugetlb_1G_freelists, lru) {

		nr_pages = 1 << huge_page_order(h);

	hpool->free_reserved_1G = 0;

	hpool->total_reserved_1G = 0;

	INIT_LIST_HEAD(&hpool->dhugetlb_1G_freelists);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

}

bool free_dhugetlb_pool(struct dhugetlb_pool *hpool)

		goto out;

	if (write) {

		spin_lock_irq(&hugetlb_lock);

		spin_lock(&hugetlb_lock);

		h->nr_overcommit_huge_pages = tmp;

		spin_unlock_irq(&hugetlb_lock);

		spin_unlock(&hugetlb_lock);

	}

out:

	return ret;

	if (dhugetlb_enabled && hpool)

		return dhugetlb_acct_memory(h, delta, hpool);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	/*

	 * When cpuset is configured, it breaks the strict hugetlb page

	 * reservation as the accounting is done on a global variable. Such

		return_unused_surplus_pages(h, (unsigned long) -delta);

out:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return ret;

}

{

	bool ret = true;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (!PageHeadHuge(page) || !page_huge_active(page) ||

	    !get_page_unless_zero(page)) {

		ret = false;

	clear_page_huge_active(page);

	list_move_tail(&page->lru, list);

unlock:

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return ret;

}

void putback_active_hugepage(struct page *page)

{

	VM_BUG_ON_PAGE(!PageHead(page), page);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	set_page_huge_active(page);

	list_move_tail(&page->lru, &(page_hstate(page))->hugepage_activelist);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	put_page(page);

}

		SetPageHugeTemporary(oldpage);

		ClearPageHugeTemporary(newpage);

		spin_lock_irq(&hugetlb_lock);

		spin_lock(&hugetlb_lock);

		if (h->surplus_huge_pages_node[old_nid]) {

			h->surplus_huge_pages_node[old_nid]--;

			h->surplus_huge_pages_node[new_nid]++;

		}

		spin_unlock_irq(&hugetlb_lock);

		spin_unlock(&hugetlb_lock);

	}

}

{

	struct page *page = NULL;

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	if (h->free_huge_pages - h->resv_huge_pages > 0)

		page = dequeue_huge_page_nodemask(h, gfp_mask, nid, NULL, NULL);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return page;

}

	do {

		for_each_hstate(h) {

			spin_lock_irq(&hugetlb_lock);

			spin_lock(&hugetlb_lock);

			list_for_each_entry(page, &h->hugepage_activelist, lru)

				hugetlb_cgroup_move_parent(idx, h_cg, page);

			spin_unlock_irq(&hugetlb_lock);

			spin_unlock(&hugetlb_lock);

			idx++;

		}

		cond_resched();

		return;

	VM_BUG_ON_PAGE(!PageHuge(oldhpage), oldhpage);

	spin_lock_irq(&hugetlb_lock);

	spin_lock(&hugetlb_lock);

	h_cg = hugetlb_cgroup_from_page(oldhpage);

	set_hugetlb_cgroup(oldhpage, NULL);

	/* move the h_cg details to new cgroup */

	set_hugetlb_cgroup(newhpage, h_cg);

	list_move(&newhpage->lru, &h->hugepage_activelist);

	spin_unlock_irq(&hugetlb_lock);

	spin_unlock(&hugetlb_lock);

	return;

}