Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 983b22e6 authored by Minchan Kim's avatar Minchan Kim Committed by Vinayak Menon
Browse files

mm, swap: skip swapcache for swapin of synchronous device 

With fast swap storage, the platforms want to use swap more aggressively
and swap-in is crucial to application latency.

The rw_page() based synchronous devices like zram, pmem and btt are such
fast storage.  When I profile swapin performance with zram lz4
decompress test, S/W overhead is more than 70%.  Maybe, it would be
bigger in nvdimm.

This patch aims to reduce swap-in latency by skipping swapcache if the
swap device is synchronous device like rw_page based device.  It
enhances 45% my swapin test(5G sequential swapin, no readahead, from
2.41sec to 1.64sec).

Change-Id: I3f8d0c3b4487331e6e0c02a09bea3077223534b8
Link: http://lkml.kernel.org/r/1505886205-9671-5-git-send-email-minchan@kernel.org


Signed-off-by: default avatarMinchan Kim <minchan@kernel.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Huang Ying <ying.huang@intel.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
Git-commit: 0bcac06f27d7528591c27ac2b093ccd71c5d0168
Git-repo: git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git


Signed-off-by: default avatarVinayak Menon <vinmenon@codeaurora.org>
parent 66cb4625

include/linux/swap.h

+11 −0
Original line number Original line Diff line number Diff line
@@ -466,6 +466,7 @@ extern int page_swapcount(struct page *); 
extern int __swp_swapcount(swp_entry_t entry);

extern int __swp_swapcount(swp_entry_t entry);

extern int swp_swapcount(swp_entry_t entry);

extern int swp_swapcount(swp_entry_t entry);

extern struct swap_info_struct *page_swap_info(struct page *);

extern struct swap_info_struct *page_swap_info(struct page *);

extern struct swap_info_struct *swp_swap_info(swp_entry_t entry);

extern bool reuse_swap_page(struct page *, int *);

extern bool reuse_swap_page(struct page *, int *);

extern int try_to_free_swap(struct page *);

extern int try_to_free_swap(struct page *);

struct backing_dev_info;

struct backing_dev_info;
@@ -474,6 +475,16 @@ extern void exit_swap_address_space(unsigned int type); 




#else /* CONFIG_SWAP */

#else /* CONFIG_SWAP */





static inline int swap_readpage(struct page *page, bool do_poll)

{

	return 0;

}



static inline struct swap_info_struct *swp_swap_info(swp_entry_t entry)

{

	return NULL;

}



#define swap_address_space(entry)		(NULL)

#define swap_address_space(entry)		(NULL)

#define get_nr_swap_pages()			0L

#define get_nr_swap_pages()			0L

#define total_swap_pages			0L

#define total_swap_pages			0L

mm/memory.c

+36 −16
Original line number Original line Diff line number Diff line
@@ -2842,7 +2842,7 @@ EXPORT_SYMBOL(unmap_mapping_range); 
int do_swap_page(struct vm_fault *vmf)

int do_swap_page(struct vm_fault *vmf)

{

{

	struct vm_area_struct *vma = vmf->vma;

	struct vm_area_struct *vma = vmf->vma;

	struct page *page = NULL, *swapcache;

	struct page *page = NULL, *swapcache = NULL;

	struct mem_cgroup *memcg;

	struct mem_cgroup *memcg;

	struct vma_swap_readahead swap_ra;

	struct vma_swap_readahead swap_ra;

	swp_entry_t entry;

	swp_entry_t entry;
@@ -2881,17 +2881,35 @@ int do_swap_page(struct vm_fault *vmf) 
		}

		}

		goto out;

		goto out;

	}

	}





	delayacct_set_flag(DELAYACCT_PF_SWAPIN);

	delayacct_set_flag(DELAYACCT_PF_SWAPIN);

	if (!page)

	if (!page)

		page = lookup_swap_cache(entry, vma_readahead ? vma : NULL,

		page = lookup_swap_cache(entry, vma_readahead ? vma : NULL,

					 vmf->address);

					 vmf->address);

	if (!page) {

	if (!page) {

		struct swap_info_struct *si = swp_swap_info(entry);



		if (!(si->flags & SWP_SYNCHRONOUS_IO)) {

			if (vma_readahead)

			if (vma_readahead)

				page = do_swap_page_readahead(entry,

				page = do_swap_page_readahead(entry,

					GFP_HIGHUSER_MOVABLE, vmf, &swap_ra);

					GFP_HIGHUSER_MOVABLE, vmf, &swap_ra);

			else

			else

				page = swapin_readahead(entry,

				page = swapin_readahead(entry,

					GFP_HIGHUSER_MOVABLE, vma, vmf->address);

					GFP_HIGHUSER_MOVABLE, vma, vmf->address);

			swapcache = page;

		} else {

			/* skip swapcache */

			page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, vmf->address);

			if (page) {

				__SetPageLocked(page);

				__SetPageSwapBacked(page);

				set_page_private(page, entry.val);

				lru_cache_add_anon(page);

				swap_readpage(page, true);

			}

		}



		if (!page) {

		if (!page) {

			/*

			/*

			 * Back out if somebody else faulted in this pte

			 * Back out if somebody else faulted in this pte
@@ -2920,7 +2938,6 @@ int do_swap_page(struct vm_fault *vmf) 
		goto out_release;

		goto out_release;

	}

	}





	swapcache = page;

	locked = lock_page_or_retry(page, vma->vm_mm, vmf->flags);

	locked = lock_page_or_retry(page, vma->vm_mm, vmf->flags);





	delayacct_clear_flag(DELAYACCT_PF_SWAPIN);

	delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
@@ -2935,7 +2952,8 @@ int do_swap_page(struct vm_fault *vmf) 
	 * test below, are not enough to exclude that.  Even if it is still

	 * test below, are not enough to exclude that.  Even if it is still

	 * swapcache, we need to check that the page's swap has not changed.

	 * swapcache, we need to check that the page's swap has not changed.

	 */

	 */

	if (unlikely(!PageSwapCache(page) || page_private(page) != entry.val))

	if (unlikely((!PageSwapCache(page) ||

			page_private(page) != entry.val)) && swapcache)

		goto out_page;

		goto out_page;





	page = ksm_might_need_to_copy(page, vma, vmf->address);

	page = ksm_might_need_to_copy(page, vma, vmf->address);
@@ -2988,14 +3006,16 @@ int do_swap_page(struct vm_fault *vmf) 
		pte = pte_mksoft_dirty(pte);

		pte = pte_mksoft_dirty(pte);

	set_pte_at(vma->vm_mm, vmf->address, vmf->pte, pte);

	set_pte_at(vma->vm_mm, vmf->address, vmf->pte, pte);

	vmf->orig_pte = pte;

	vmf->orig_pte = pte;

	if (page == swapcache) {



		do_page_add_anon_rmap(page, vma, vmf->address, exclusive);

	/* ksm created a completely new copy */

		mem_cgroup_commit_charge(page, memcg, true, false);

	if (unlikely(page != swapcache && swapcache)) {

		activate_page(page);

	} else { /* ksm created a completely new copy */

		page_add_new_anon_rmap(page, vma, vmf->address, false);

		page_add_new_anon_rmap(page, vma, vmf->address, false);

		mem_cgroup_commit_charge(page, memcg, false, false);

		mem_cgroup_commit_charge(page, memcg, false, false);

		lru_cache_add_active_or_unevictable(page, vma);

		lru_cache_add_active_or_unevictable(page, vma);

	} else {

		do_page_add_anon_rmap(page, vma, vmf->address, exclusive);

		mem_cgroup_commit_charge(page, memcg, true, false);

		activate_page(page);

	}

	}





	swap_free(entry);

	swap_free(entry);
@@ -3003,7 +3023,7 @@ int do_swap_page(struct vm_fault *vmf) 
	    (vma->vm_flags & VM_LOCKED) || PageMlocked(page))

	    (vma->vm_flags & VM_LOCKED) || PageMlocked(page))

		try_to_free_swap(page);

		try_to_free_swap(page);

	unlock_page(page);

	unlock_page(page);

	if (page != swapcache) {

	if (page != swapcache && swapcache) {

		/*

		/*

		 * Hold the lock to avoid the swap entry to be reused

		 * Hold the lock to avoid the swap entry to be reused

		 * until we take the PT lock for the pte_same() check

		 * until we take the PT lock for the pte_same() check
@@ -3036,7 +3056,7 @@ int do_swap_page(struct vm_fault *vmf) 
	unlock_page(page);

	unlock_page(page);

out_release:

out_release:

	put_page(page);

	put_page(page);

	if (page != swapcache) {

	if (page != swapcache && swapcache) {

		unlock_page(swapcache);

		unlock_page(swapcache);

		put_page(swapcache);

		put_page(swapcache);

	}

	}

mm/page_io.c

+3 −3
Original line number Original line Diff line number Diff line
@@ -348,7 +348,7 @@ int __swap_writepage(struct page *page, struct writeback_control *wbc, 
	return ret;

	return ret;

}

}





int swap_readpage(struct page *page, bool do_poll)

int swap_readpage(struct page *page, bool synchronous)

{

{

	struct bio *bio;

	struct bio *bio;

	int ret = 0;

	int ret = 0;
@@ -356,7 +356,7 @@ int swap_readpage(struct page *page, bool do_poll) 
	blk_qc_t qc;

	blk_qc_t qc;

	struct gendisk *disk;

	struct gendisk *disk;





	VM_BUG_ON_PAGE(!PageSwapCache(page), page);

	VM_BUG_ON_PAGE(!PageSwapCache(page) && !synchronous, page);

	VM_BUG_ON_PAGE(!PageLocked(page), page);

	VM_BUG_ON_PAGE(!PageLocked(page), page);

	VM_BUG_ON_PAGE(PageUptodate(page), page);

	VM_BUG_ON_PAGE(PageUptodate(page), page);

	if (frontswap_load(page) == 0) {

	if (frontswap_load(page) == 0) {
@@ -404,7 +404,7 @@ int swap_readpage(struct page *page, bool do_poll) 
	count_vm_event(PSWPIN);

	count_vm_event(PSWPIN);

	bio_get(bio);

	bio_get(bio);

	qc = submit_bio(bio);

	qc = submit_bio(bio);

	while (do_poll) {

	while (synchronous) {

		set_current_state(TASK_UNINTERRUPTIBLE);

		set_current_state(TASK_UNINTERRUPTIBLE);

		if (!READ_ONCE(bio->bi_private))

		if (!READ_ONCE(bio->bi_private))

			break;

			break;

mm/swapfile.c

+7 −4
Original line number Original line Diff line number Diff line
@@ -3455,10 +3455,15 @@ int swapcache_prepare(swp_entry_t entry) 
	return __swap_duplicate(entry, SWAP_HAS_CACHE);

	return __swap_duplicate(entry, SWAP_HAS_CACHE);

}

}





struct swap_info_struct *swp_swap_info(swp_entry_t entry)

{

	return swap_info[swp_type(entry)];

}



struct swap_info_struct *page_swap_info(struct page *page)

struct swap_info_struct *page_swap_info(struct page *page)

{

{

	swp_entry_t swap = { .val = page_private(page) };

	swp_entry_t entry = { .val = page_private(page) };

	return swap_info[swp_type(swap)];

	return swp_swap_info(entry);

}

}





/*

/*
@@ -3466,7 +3471,6 @@ struct swap_info_struct *page_swap_info(struct page *page) 
 */

 */

struct address_space *__page_file_mapping(struct page *page)

struct address_space *__page_file_mapping(struct page *page)

{

{

	VM_BUG_ON_PAGE(!PageSwapCache(page), page);

	return page_swap_info(page)->swap_file->f_mapping;

	return page_swap_info(page)->swap_file->f_mapping;

}

}

EXPORT_SYMBOL_GPL(__page_file_mapping);

EXPORT_SYMBOL_GPL(__page_file_mapping);
@@ -3474,7 +3478,6 @@ EXPORT_SYMBOL_GPL(__page_file_mapping); 
pgoff_t __page_file_index(struct page *page)

pgoff_t __page_file_index(struct page *page)

{

{

	swp_entry_t swap = { .val = page_private(page) };

	swp_entry_t swap = { .val = page_private(page) };

	VM_BUG_ON_PAGE(!PageSwapCache(page), page);

	return swp_offset(swap);

	return swp_offset(swap);

}

}

EXPORT_SYMBOL_GPL(__page_file_index);

EXPORT_SYMBOL_GPL(__page_file_index);