filemap: dont shrink_page_cache in add_to_page_cache

hulk inclusion
category: feature
bugzilla: https://gitee.com/openeuler/kernel/issues/I4SK3S
CVE: NA

--------------------------------

The reason of disable shrink_page_cache in add_to_page_cache are:

1. Synchronous memory reclamation will affect performance.
2. add_to_page_cache will not increase the number of LRU size in
   HugeTLB situation, so shrink_page_cache will not be triggered.

Now, add_to_page_cache in mm/filemap.c and include/linux/pagemap.h
are same, don't delete add_to_page_cache in mm/filemap.c, just keep
interface for KABI.
Signed-off-by: NChen Wandun <chenwandun@huawei.com>
Reviewed-by: NKefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

mm/filemap.c

@@ -991,8 +991,6 @@ int add_to_page_cache(struct page *page,
 {
 	int error;
 
-	if (vm_cache_limit_mbytes && page_cache_over_limit())
-		shrink_page_cache(gfp_mask);
 	__SetPageLocked(page);
 	error = add_to_page_cache_locked(page, mapping, offset, gfp_mask);
 	if (unlikely(error))