mm/swap: clean up and optimize swap cache index
Subsystems (17)
v4 cover mm/swap: clean up and optimize swap cache index
From: Kairui Song <[email protected]>
This is based on latest mm-unstable. Patch 1/12 is not needed if f2fs converted .readahead to use folio, I included it for easier test and review.
Currently we use one swap_address_space for every 64M chunk to reduce lock contention, this is like having a set of smaller files inside a swap device. But when doing swap cache look up or insert, we are still using the offset of the whole large swap device. This is OK for correctness, as the offset (key) is unique.
But Xarray is specially optimized for small indexes, it creates the redix tree levels lazily to be just enough to fit the largest key stored in one Xarray. So we are wasting tree nodes unnecessarily.
For 64M chunk it should only take at most 3 level to contain everything. But if we are using the offset from the whole swap device, the offset (key) value will be way beyond 64M, and so will the tree level.
Optimize this by reduce the swap cache search space into 64M scope.
Test with time memhog 128G inside a 8G memcg using 128G swap (ramdisk with SWP_SYNCHRONOUS_IO dropped, tested 3 times, results are stable. The test result is similar but the improvement is smaller if SWP_SYNCHRONOUS_IO is enabled, as swap out path can never skip swap cache):
Before:
6.07user 250.74system 4:17.26elapsed 99%CPU (0avgtext+0avgdata 8373376maxresident)k 0inputs+0outputs (55major+33555018minor)pagefaults 0swaps
After (+1.8% faster):
6.08user 246.09system 4:12.58elapsed 99%CPU (0avgtext+0avgdata 8373248maxresident)k 0inputs+0outputs (54major+33555027minor)pagefaults 0swaps
Similar result with MySQL and sysbench using swap:
Before:
94055.61 qps
After (+0.8% faster):
94834.91 qps
There is alse a very slight drop of radix tree node slab usage:
Before: 303952K After: 302224K
For this series:
There are multiple places that expect mixed type of pages (page cache or swap cache), eg. migration, huge memory split; There are four helpers for that:
- page_index
- page_file_offset
- folio_index
- folio_file_pos
To keep the code clean and compatible, this series first cleaned up usage of them.
page_file_offset and folio_file_pos are historical helpes that can be simply dropped after clean up. And page_index can be all converted to folio_index or folio->index.
Then introduce two new helpers swap_cache_index and swap_dev_pos for swap. Replace swp_offset with swap_cache_index when used to retrieve folio from swap cache, and use swap_dev_pos when needed to retrieve the device position of a swap entry. This way, swap_cache_index can return the optimized value with no compatibility issue.
The result is better performance and reduced LOC.
Idealy, in the future, we may want to reduce SWAP_ADDRESS_SPACE_SHIFT from 14 to 12: Default Xarray chunk offset is 6, so we have 3 level trees instead of 2 level trees just for 2 extra bits. But swap cache is based on address_space struct, with 4 times more metadata sparsely distributed in memory it waste more cacheline, the performance gain from this series is almost canceled according to my test. So first, just have a cleaner seperation of offsets and smaller search space.
Patch 1/12 - 11/12: Clean up usage of above helpers. Patch 12/12: Apply the optmization.
V3: https://lore.kernel.org/all/[email protected]/
Update from V3:
- Help remove a redundant loop in nilfs2 [Matthew Wilcox]
- Update commit message, use the term swap device instead of swap file to avoid confusion [Huang, Ying]
- Add more details in commit message about folio_file_pos usage in NFS.
- Fix a shadow leak in clear_shadow_from_swap_cache.
V2: https://lore.kernel.org/linux-mm/[email protected]/
Update from V2:
- Clean up usage of page_file_offset and folio_file_pos [Matthew Wilcox]
https://lore.kernel.org/linux-mm/[email protected]/
- Use folio in nilfs_bmap_data_get_key [Ryusuke Konishi]
V1: https://lore.kernel.org/all/[email protected]/
Update from V1:
- Convert more users to use folio directly when possible [Matthew Wilcox]
- Rename swap_file_pos to swap_dev_pos [Huang, Ying]
- Update comments and commit message.
- Adjust headers and add dummy function to fix build error.
This series is part of effort to reduce swap cache overhead, and ultimately remove SWP_SYNCHRONOUS_IO and unify swap cache usage as proposed before:
https://lore.kernel.org/lkml/[email protected]/
Kairui Song (12):
f2fs: drop usage of page_index
nilfs2: drop usage of page_index
ceph: drop usage of page_index
NFS: remove nfs_page_lengthg and usage of page_index
cifs: drop usage of page_file_offset
afs: drop usage of folio_file_pos
netfs: drop usage of folio_file_pos
nfs: drop usage of folio_file_pos
mm/swap: get the swap device offset directly
mm: remove page_file_offset and folio_file_pos
mm: drop page_index and convert folio_index to use folio mm/swap: reduce swap cache search space
fs/afs/dir.c | 6 +++--- fs/afs/dir_edit.c | 4 ++-- fs/ceph/dir.c | 2 +- fs/ceph/inode.c | 2 +- fs/f2fs/data.c | 2 +- fs/netfs/buffered_read.c | 4 ++-- fs/netfs/buffered_write.c | 2 +- fs/nfs/file.c | 2 +- fs/nfs/internal.h | 19 ------------------- fs/nfs/nfstrace.h | 4 ++-- fs/nfs/write.c | 6 +++--- fs/nilfs2/bmap.c | 10 ++-------- fs/smb/client/file.c | 2 +- include/linux/mm.h | 13 ------------- include/linux/pagemap.h | 25 ++++--------------------- mm/huge_memory.c | 2 +- mm/memcontrol.c | 2 +- mm/mincore.c | 2 +- mm/page_io.c | 6 +++--- mm/shmem.c | 2 +- mm/swap.h | 24 ++++++++++++++++++++++++ mm/swap_state.c | 19 ++++++++++--------- mm/swapfile.c | 11 +++++------ 23 files changed, 70 insertions(+), 101 deletions(-)
Signature · 1 line
2.44.0
PATCH v4 1/12 f2fs: drop usage of page_index
From: Kairui Song <[email protected]>
page_index is needed for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use page->index instead.
It can't be a swap cache page here, so just drop it.
[ This commit will not be needed once f2fs converted
f2fs_mpage_readpages() to use folio]
--- fs/f2fs/data.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/f2fs/data.c +1 -1
@@ -2057,7 +2057,7 @@ static int f2fs_read_single_page(struct inode *inode, struct page *page,sector_t block_nr;int ret =0 ; - block_in_file = (sector_t )page_index (page); + block_in_file = (sector_t )page->index; last_block = block_in_file + nr_pages; last_block_in_file =bytes_to_blks (inode,f2fs_readpage_limit (inode) + blocksize -1 );
Signature · 1 line
2.44.0
PATCH v4 2/12 nilfs2: drop usage of page_index
In reply to Kairui SongFrom: Kairui Song <[email protected]>
page_index is only for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use page->index instead.
It can't be a swap cache page here (being part of buffer head), so just drop it. And while we are at it, optimize the code by retrieving the offset of the buffer head within the folio directly using bh_offset, and get rid of the loop and usage of page helpers.
--- fs/nilfs2/bmap.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-)
fs/nilfs2/bmap.c +2 -8
@@ -450,15 +450,9 @@ int nilfs_bmap_test_and_clear_dirty(struct nilfs_bmap *bmap)__u64 nilfs_bmap_data_get_key (const struct nilfs_bmap *bmap,const struct buffer_head *bh) { -struct buffer_head *pbh; -__u64 key; +loff_t pos =folio_pos (bh->b_folio) +bh_offset (bh); - key =page_index (bh->b_page) << (PAGE_SHIFT - - bmap->b_inode->i_blkbits); -for (pbh =page_buffers (bh->b_page); pbh != bh; pbh = pbh->b_this_page) - key++; - -return key; +return pos >> bmap->b_inode->i_blkbits; }__u64 nilfs_bmap_find_target_seq (const struct nilfs_bmap *bmap,__u64 key)
Signature · 1 line
2.44.0
Re: [PATCH v4 02/12] nilfs2: drop usage of page_index
Quoted 42 lines · Kairui Song
From: Kairui Song <[email protected]>
page_indexis only for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just usepage->indexinstead.It can't be a swap cache page here (being part of buffer head), so just drop it. And while we are at it, optimize the code by retrieving the offset of the buffer head within the folio directly using
bh_offset, and get rid of the loop and usage of page helpers.Suggested-by Matthew Wilcox <[email protected]>Signed-off-by Kairui Song <[email protected]>Cc Ryusuke Konishi <[email protected]>--- fs/nilfs2/bmap.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-)fs/nilfs2/bmap.c +2
-9@@ -450,15 +450,9 @@ int nilfs_bmap_test_and_clear_dirty(struct nilfs_bmap *bmap)__u64 nilfs_bmap_data_get_key (const struct nilfs_bmap *bmap,const struct buffer_head *bh) { -struct buffer_head *pbh; -__u64 key; +loff_t pos =folio_pos (bh->b_folio) +bh_offset (bh); - key =page_index (bh->b_page) << (PAGE_SHIFT - - bmap->b_inode->i_blkbits); -for (pbh =page_buffers (bh->b_page); pbh != bh; pbh = pbh->b_this_page) - key++; - -return key; +return pos >> bmap->b_inode->i_blkbits; }__u64 nilfs_bmap_find_target_seq (const struct nilfs_bmap *bmap,__u64 key) --
Quoted 1 line
2.44.0
Looks good. Feel free to add:
Just to be sure, I also tested this change in different environments, including 4k (page size) and smaller block sizes. And of course it's working as expected and so far hasn't broken anything.
Thanks,
Ryusuke Konishi
PATCH v4 3/12 ceph: drop usage of page_index
In reply to Kairui SongFrom: Kairui Song <[email protected]>
page_index is needed for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use page->index instead.
It can't be a swap cache page here, so just drop it.
--- fs/ceph/dir.c | 2 +- fs/ceph/inode.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
fs/ceph/dir.c +1 -1
@@ -141,7 +141,7 @@ __dcache_find_get_entry(struct dentry *parent, u64 idx,if (ptr_pos >=i_size_read (dir))return NULL ; -if (!cache_ctl->page || ptr_pgoff !=page_index (cache_ctl->page)) { +if (!cache_ctl->page || ptr_pgoff != cache_ctl->page->index) {ceph_readdir_cache_release (cache_ctl); cache_ctl->page =find_lock_page (&dir->i_data, ptr_pgoff);if (!cache_ctl->page) {
fs/ceph/inode.c +1 -1
@@ -1861,7 +1861,7 @@ static int fill_readdir_cache(struct inode *dir, struct dentry *dn,unsigned idx = ctl->index % nsize;pgoff_t pgoff = ctl->index / nsize; -if (!ctl->page || pgoff !=page_index (ctl->page)) { +if (!ctl->page || pgoff != ctl->page->index) {ceph_readdir_cache_release (ctl);if (idx ==0 ) ctl->page =grab_cache_page (&dir->i_data, pgoff);
Signature · 1 line
2.44.0
PATCH v4 4/12 NFS: remove nfs_page_lengthg and usage of page_index
In reply to Kairui SongFrom: Kairui Song <[email protected]>
This function is no longer used after
commit 4fa7a717b432 ("NFS: Fix up nfs_vm_page_mkwrite() for folios"), all users have been converted to use folio instead, just delete it to remove usage of page_index.
--- fs/nfs/internal.h | 19 ------------------- 1 file changed, 19 deletions(-)
fs/nfs/internal.h +0 -19
@@ -790,25 +790,6 @@ static inline void nfs_folio_mark_unstable(struct folio *folio, } } -/* -* Determine the number of bytes of data the page contains -*/ -static inline -unsigned int nfs_page_length (struct page *page) -{ -loff_t i_size =i_size_read (page_file_mapping (page)->host); - -if (i_size >0 ) { -pgoff_t index =page_index (page); -pgoff_t end_index = (i_size -1 ) >>PAGE_SHIFT ; -if (index < end_index) -return PAGE_SIZE ; -if (index == end_index) -return ((i_size -1 ) & ~PAGE_MASK ) +1 ; - } -return 0 ; -} -/* * Determine the number of bytes of data the page contains */
Signature · 1 line
2.44.0
PATCH v4 5/12 cifs: drop usage of page_file_offset
In reply to Kairui SongFrom: Kairui Song <[email protected]>
page_file_offset is only needed for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use page_offset instead.
It can't be a swap cache page here, so just drop it and convert it to use folio.
--- fs/smb/client/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/file.c +1 -1
@@ -4828,7 +4828,7 @@ static int cifs_readpage_worker(struct file *file, struct page *page,static int cifs_read_folio (struct file *file,struct folio *folio) {struct page *page = &folio->page; -loff_t offset =page_file_offset (page); +loff_t offset =folio_pos (folio);int rc = -EACCES ;unsigned int xid;
Signature · 1 line
2.44.0
PATCH v4 6/12 afs: drop usage of folio_file_pos
In reply to Kairui SongFrom: Kairui Song <[email protected]>
folio_file_pos is only needed for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use folio_pos instead.
It can't be a swap cache page here. Swap mapping may only call into fs through swap_rw and that is not supported for afs. So just drop it and use folio_pos instead.
--- fs/afs/dir.c | 6 +++--- fs/afs/dir_edit.c | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-)
fs/afs/dir.c +3 -3
@@ -533,14 +533,14 @@ static int afs_dir_iterate(struct inode *dir, struct dir_context *ctx,break ; } - offset =round_down (ctx->pos,sizeof (*dblock)) -folio_file_pos (folio); + offset =round_down (ctx->pos,sizeof (*dblock)) -folio_pos (folio); size =min_t (loff_t,folio_size (folio), - req->actual_len -folio_file_pos (folio)); + req->actual_len -folio_pos (folio)); do { dblock = kmap_local_folio(folio, offset); ret = afs_dir_iterate_block(dvnode, ctx, dblock, - folio_file_pos(folio) + offset); + folio_pos(folio) + offset); kunmap_local(dblock); if (ret != 1) goto out;
fs/afs/dir_edit.c +2 -2
@@ -256,7 +256,7 @@ void afs_edit_dir_add(struct afs_vnode *vnode, folio = folio0; } - block =kmap_local_folio (folio, b *AFS_DIR_BLOCK_SIZE -folio_file_pos (folio)); + block =kmap_local_folio (folio, b *AFS_DIR_BLOCK_SIZE -folio_pos (folio));/* Abandon the edit if we got a callback break. */ if (!test_bit (AFS_VNODE_DIR_VALID , &vnode->flags))
@@ -417,7 +417,7 @@ void afs_edit_dir_remove(struct afs_vnode *vnode, folio = folio0; } - block =kmap_local_folio (folio, b *AFS_DIR_BLOCK_SIZE -folio_file_pos (folio)); + block =kmap_local_folio (folio, b *AFS_DIR_BLOCK_SIZE -folio_pos (folio));/* Abandon the edit if we got a callback break. */ if (!test_bit (AFS_VNODE_DIR_VALID , &vnode->flags))
Signature · 1 line
2.44.0
PATCH v4 7/12 netfs: drop usage of folio_file_pos
In reply to Kairui SongFrom: Kairui Song <[email protected]>
folio_file_pos is only needed for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use folio_pos instead.
It can't be a swap cache page here. Swap mapping may only call into fs through swap_rw and that is not supported for netfs. So just drop it and use folio_pos instead.
--- fs/netfs/buffered_read.c | 4 ++-- fs/netfs/buffered_write.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-)
fs/netfs/buffered_read.c +2 -2
@@ -255,7 +255,7 @@ int netfs_read_folio(struct file *file, struct folio *folio)_enter ("%lx" , folio->index); rreq =netfs_alloc_request (mapping, file, -folio_file_pos (folio),folio_size (folio), +folio_pos (folio),folio_size (folio),NETFS_READPAGE );if (IS_ERR (rreq)) { ret =PTR_ERR (rreq);
@@ -454,7 +454,7 @@ int netfs_write_begin(struct netfs_inode *ctx, } rreq =netfs_alloc_request (mapping, file, -folio_file_pos (folio),folio_size (folio), +folio_pos (folio),folio_size (folio),NETFS_READ_FOR_WRITE );if (IS_ERR (rreq)) { ret =PTR_ERR (rreq);
fs/netfs/buffered_write.c +1 -1
@@ -63,7 +63,7 @@ static enum netfs_how_to_modify netfs_how_to_modify(struct netfs_inode *ctx, bool maybe_trouble) {struct netfs_folio *finfo =netfs_folio_info (folio); -loff_t pos =folio_file_pos (folio); +loff_t pos =folio_pos (folio);_enter ("" );
Signature · 1 line
2.44.0
PATCH v4 8/12 nfs: drop usage of folio_file_pos
In reply to Kairui SongFrom: Kairui Song <[email protected]>
folio_file_pos is only needed for mixed usage of page cache and swap cache, for pure page cache usage, the caller can just use folio_pos instead.
After commit e1209d3a7a67 ("mm: introduce ->swap_rw and use it for reads from SWP_FS_OPS swap-space"), swap cache should never be exposed to nfs.
So remove the usage of folio_file_pos in following NFS functions / helpers:
- nfs_vm_page_mkwrite
It's only used by nfs_file_vm_ops.page_mkwrite
- trace event helper: nfs_folio_event
- trace event helper: nfs_folio_event_done
These two are used through DEFINE_NFS_FOLIO_EVENT and DEFINE_NFS_FOLIO_EVENT_DONE, which defined following events:
- trace_nfs_aop_readpage{_done}: only called by nfs_read_folio
- trace_nfs_writeback_folio: only called by nfs_wb_folio
- trace_nfs_invalidate_folio: only called by nfs_invalidate_folio
- trace_nfs_launder_folio_done: only called by nfs_launder_folio
None of them could possibly be used on swap cache folio, nfs_read_folio only called by:
.write_begin -> nfs_read_folio
.read_folio
nfs_wb_folio only called by nfs mapping:
.release_folio -> nfs_wb_folio
.launder_folio -> nfs_wb_folio
.write_begin -> nfs_read_folio -> nfs_wb_folio
.read_folio -> nfs_wb_folio
.write_end -> nfs_update_folio -> nfs_writepage_setup -> nfs_setup_write_request -> nfs_try_to_update_request -> nfs_wb_folio .page_mkwrite -> nfs_update_folio -> nfs_writepage_setup -> nfs_setup_write_request -> nfs_try_to_update_request -> nfs_wb_folio .write_begin -> nfs_flush_incompatible -> nfs_wb_folio
.page_mkwrite -> nfs_vm_page_mkwrite -> nfs_flush_incompatible -> nfs_wb_folio
nfs_invalidate_folio is only called by .invalidate_folio. nfs_launder_folio is only called by .launder_folio
- nfs_grow_file
- nfs_update_folio
nfs_grow_file is only called by nfs_update_folio, and all possible callers of them are:
.write_end -> nfs_update_folio
.page_mkwrite -> nfs_update_folio
- nfs_wb_folio_cancel
.invalidate_folio -> nfs_wb_folio_cancel
Also, seeing from the swap side, swap_rw is now the only interface calling into fs, the offset info is always in iocb.ki_pos now.
So we can remove all these folio_file_pos call safely.
--- fs/nfs/file.c | 2 +- fs/nfs/nfstrace.h | 4 ++-- fs/nfs/write.c | 6 +++--- 3 files changed, 6 insertions(+), 6 deletions(-)
fs/nfs/file.c +1 -1
@@ -588,7 +588,7 @@ static vm_fault_t nfs_vm_page_mkwrite(struct vm_fault *vmf)dfprintk (PAGECACHE ,"NFS: vm_page_mkwrite(%pD2(%lu), offset %lld)\n" , filp, filp->f_mapping->host->i_ino, - (long long )folio_file_pos (folio)); + (long long )folio_pos (folio));sb_start_pagefault (inode->i_sb);
fs/nfs/nfstrace.h +2 -2
@@ -960,7 +960,7 @@ DECLARE_EVENT_CLASS(nfs_folio_event, __entry->fileid = nfsi->fileid; __entry->fhandle =nfs_fhandle_hash (&nfsi->fh); __entry->version =inode_peek_iversion_raw (inode); - __entry->offset =folio_file_pos (folio); + __entry->offset =folio_pos (folio); __entry->count =nfs_folio_length (folio); ),
@@ -1008,7 +1008,7 @@ DECLARE_EVENT_CLASS(nfs_folio_event_done, __entry->fileid = nfsi->fileid; __entry->fhandle =nfs_fhandle_hash (&nfsi->fh); __entry->version =inode_peek_iversion_raw (inode); - __entry->offset =folio_file_pos (folio); + __entry->offset =folio_pos (folio); __entry->count =nfs_folio_length (folio); __entry->ret = ret; ),
fs/nfs/write.c +3 -3
@@ -281,7 +281,7 @@ static void nfs_grow_file(struct folio *folio, unsigned int offset, end_index = ((i_size -1 ) >>folio_shift (folio)) <<folio_order (folio);if (i_size >0 &&folio_index (folio) < end_index) goto out; - end =folio_file_pos (folio) + (loff_t )offset + (loff_t )count; + end =folio_pos (folio) + (loff_t )offset + (loff_t )count;if (i_size >= end) goto out;trace_nfs_size_grow (inode, end);
@@ -1362,7 +1362,7 @@ int nfs_update_folio(struct file *file, struct folio *folio,nfs_inc_stats (inode,NFSIOS_VFSUPDATEPAGE );dprintk ("NFS: nfs_update_folio(%pD2 %d@%lld)\n" , file, count, - (long long )(folio_file_pos (folio) + offset)); + (long long )(folio_pos (folio) + offset));if (!count) goto out;
@@ -2073,7 +2073,7 @@ int nfs_wb_folio_cancel(struct inode *inode, struct folio *folio)*/ int nfs_wb_folio (struct inode *inode,struct folio *folio) { -loff_t range_start =folio_file_pos (folio); +loff_t range_start =folio_pos (folio);loff_t range_end = range_start + (loff_t )folio_size (folio) -1 ;struct writeback_control wbc = { .sync_mode =WB_SYNC_ALL ,
Signature · 1 line
2.44.0
PATCH v4 9/12 mm/swap: get the swap device offset directly
In reply to Kairui SongFrom: Kairui Song <[email protected]>
folio_file_pos and page_file_offset are for mixed usage of swap cache and page cache, it can't be page cache here, so introduce a new helper to get the swap offset in swap device directly.
Need to include swapops.h in mm/swap.h to ensure swp_offset is always defined before use.
--- mm/page_io.c | 6 +++--- mm/swap.h | 9 +++++++++ 2 files changed, 12 insertions(+), 3 deletions(-)
mm/page_io.c +3 -3
@@ -280,7 +280,7 @@ static void sio_write_complete(struct kiocb *iocb, long ret)* be temporary. */ pr_err_ratelimited ("Write error %ld on dio swapfile (%llu)\n" , - ret,page_file_offset (page)); + ret,swap_dev_pos (page_swap_entry (page)));for (p =0 ; p < sio->pages; p++) { page = sio->bvec[p].bv_page;set_page_dirty (page);
@@ -299,7 +299,7 @@ static void swap_writepage_fs(struct folio *folio, struct writeback_control *wbcstruct swap_iocb *sio =NULL ;struct swap_info_struct *sis =swp_swap_info (folio->swap);struct file *swap_file = sis->swap_file; -loff_t pos =folio_file_pos (folio); +loff_t pos =swap_dev_pos (folio->swap);count_swpout_vm_event (folio);folio_start_writeback (folio);
@@ -430,7 +430,7 @@ static void swap_read_folio_fs(struct folio *folio, struct swap_iocb **plug) {struct swap_info_struct *sis =swp_swap_info (folio->swap);struct swap_iocb *sio =NULL ; -loff_t pos =folio_file_pos (folio); +loff_t pos =swap_dev_pos (folio->swap);if (plug) sio = *plug;
mm/swap.h +9 -0
@@ -5,6 +5,7 @@struct mempolicy ;#ifdef CONFIG_SWAP +#include <linux/swapops.h> /* for swp_offset */ #include <linux/blk_types.h> /* for bio_end_io_t */ /* linux/mm/page_io.c */
@@ -31,6 +32,14 @@ extern struct address_space *swapper_spaces[]; (&swapper_spaces[swp_type (entry)][swp_offset (entry) \ >>SWAP_ADDRESS_SPACE_SHIFT ]) +/* +* Return the swap device position of the swap entry. + */ +static inline loff_t swap_dev_pos (swp_entry_t entry) +{ +return ((loff_t )swp_offset (entry)) <<PAGE_SHIFT ; +} +void show_swap_cache_info (void );bool add_to_swap (struct folio *folio);void *get_shadow_from_swap_cache (swp_entry_t entry);
Signature · 1 line
2.44.0
Re: [PATCH v4 09/12] mm/swap: get the swap device offset directly
Quoted 10 lines · Kairui Song
From: Kairui Song <[email protected]>
folio_file_posandpage_file_offsetare for mixed usage of swap cache and page cache, it can't be page cache here, so introduce a new helper to get the swap offset in swap device directly.Need to include
swapops.hinmm/swap.hto ensureswp_offsetis always defined before use.Signed-off-by Kairui Song <[email protected]>
LGTM! Thanks!
Quoted 63 lines
--- mm/page_io.c | 6 +++--- mm/swap.h | 9 +++++++++ 2 files changed, 12 insertions(+), 3 deletions(-)mm/page_io.c +3
-3@@ -280,7 +280,7 @@ static void sio_write_complete(struct kiocb *iocb, long ret)* be temporary. */ pr_err_ratelimited ("Write error %ld on dio swapfile (%llu)\n" , - ret,page_file_offset (page)); + ret,swap_dev_pos (page_swap_entry (page)));for (p =0 ; p < sio->pages; p++) { page = sio->bvec[p].bv_page;set_page_dirty (page);@@ -299,7 +299,7 @@ static void swap_writepage_fs(struct folio *folio, struct writeback_control *wbcstruct swap_iocb *sio =NULL ;struct swap_info_struct *sis =swp_swap_info (folio->swap);struct file *swap_file = sis->swap_file; -loff_t pos =folio_file_pos (folio); +loff_t pos =swap_dev_pos (folio->swap);count_swpout_vm_event (folio);folio_start_writeback (folio);@@ -430,7 +430,7 @@ static void swap_read_folio_fs(struct folio *folio, struct swap_iocb **plug) {struct swap_info_struct *sis =swp_swap_info (folio->swap);struct swap_iocb *sio =NULL ; -loff_t pos =folio_file_pos (folio); +loff_t pos =swap_dev_pos (folio->swap);if (plug) sio = *plug;mm/swap.h +9
-0@@ -5,6 +5,7 @@struct mempolicy ;#ifdef CONFIG_SWAP +#include <linux/swapops.h> /* for swp_offset */ #include <linux/blk_types.h> /* for bio_end_io_t */ /* linux/mm/page_io.c */ @@ -31,6 +32,14 @@ extern struct address_space *swapper_spaces[]; (&swapper_spaces[swp_type (entry)][swp_offset (entry) \ >>SWAP_ADDRESS_SPACE_SHIFT ]) +/* +* Return the swap device position of the swap entry. + */ +static inline loff_t swap_dev_pos (swp_entry_t entry) +{ +return ((loff_t )swp_offset (entry)) <<PAGE_SHIFT ; +} +void show_swap_cache_info (void );bool add_to_swap (struct folio *folio);void *get_shadow_from_swap_cache (swp_entry_t entry);
PATCH v4 10/12 mm: remove page_file_offset and folio_file_pos
In reply to Kairui SongFrom: Kairui Song <[email protected]>
These two helpers were useful for mixed usage of swap cache and page cache, which help retrieve the corresponding file or swap device offset of a page or folio.
They were introduced in commit f981c5950fa8 ("mm: methods for teaching filesystems about PG_swapcache pages") and used in commit d56b4ddf7781 ("nfs: teach the NFS client how to treat PG_swapcache pages"), suppose to be used with direct_IO for swap over fs.
But after commit e1209d3a7a67 ("mm: introduce ->swap_rw and use it for reads from SWP_FS_OPS swap-space"), swap with direct_IO is no more, and swap cache mapping is never exposed to fs.
Now we have dropped all users of page_file_offset and folio_file_pos, so they can be deleted.
--- include/linux/pagemap.h | 17 ----------------- 1 file changed, 17 deletions(-)
include/linux/pagemap.h +0 -17
@@ -918,11 +918,6 @@ static inline loff_t page_offset(struct page *page)return ((loff_t )page->index) <<PAGE_SHIFT ; } -static inline loff_t page_file_offset (struct page *page) -{ -return ((loff_t )page_index (page)) <<PAGE_SHIFT ; -} -/** * folio_pos - Returns the byte position of this folio in its file. * @folio: The folio.
@@ -932,18 +927,6 @@ static inline loff_t folio_pos(struct folio *folio)return page_offset (&folio->page); } -/** -* folio_file_pos - Returns the byte position of this folio in its file. -* @folio: The folio. -* -* This differs from folio_pos() for folios which belong to a swap file. -* NFS is the only filesystem today which needs to use folio_file_pos(). -*/ -static inline loff_t folio_file_pos (struct folio *folio) -{ -return page_file_offset (&folio->page); -} -/* * Get the offset in PAGE_SIZE (even for hugetlb folios). */
Signature · 1 line
2.44.0
Re: [PATCH v4 10/12] mm: remove page_file_offset and folio_file_pos
Quoted 11 lines · Kairui Song
From: Kairui Song <[email protected]>
These two helpers were useful for mixed usage of swap cache and page cache, which help retrieve the corresponding file or swap device offset of a page or folio.
They were introduced in commit
f981c5950fa8("mm: methods for teaching filesystems aboutPG_swapcachepages") and used in commitd56b4ddf7781("nfs: teach the NFS client how to treatPG_swapcachepages"), suppose to be used withdirect_IOfor swap over fs.
Quoted 8 lines
But after commit
e1209d3a7a67("mm: introduce ->swap_rw and use it for reads fromSWP_FS_OPSswap-space"), swap withdirect_IOis no more, and swap cache mapping is never exposed to fs.Now we have dropped all users of
page_file_offsetandfolio_file_pos, so they can be deleted.Signed-off-by Kairui Song <[email protected]>
LGTM, Thanks!
Quoted 39 lines
--- include/linux/pagemap.h | 17 ----------------- 1 file changed, 17 deletions(-)include/linux/pagemap.h +0
-17@@ -918,11 +918,6 @@ static inline loff_t page_offset(struct page *page)return ((loff_t )page->index) <<PAGE_SHIFT ; } -static inline loff_t page_file_offset (struct page *page) -{ -return ((loff_t )page_index (page)) <<PAGE_SHIFT ; -} -/** * folio_pos - Returns the byte position of this folio in its file. * @folio: The folio. @@ -932,18 +927,6 @@ static inline loff_t folio_pos(struct folio *folio)return page_offset (&folio->page); } -/** -* folio_file_pos - Returns the byte position of this folio in its file. -* @folio: The folio. -* -* This differs from folio_pos() for folios which belong to a swap file. -* NFS is the only filesystem today which needs to use folio_file_pos(). -*/ -static inline loff_t folio_file_pos (struct folio *folio) -{ -return page_file_offset (&folio->page); -} -/* * Get the offset in PAGE_SIZE (even for hugetlb folios). */
PATCH v4 11/12 mm: drop page_index and convert folio_index to use folio
In reply to Kairui SongFrom: Kairui Song <[email protected]>
There are two helpers for retrieving the index within address space for mixed usage of swap cache and page cache:
- page_index
- folio_index (wrapper of page_index)
This commit drops page_index, as we have eliminated all users, and converts folio_index to use folio internally.
--- include/linux/mm.h | 13 ------------- include/linux/pagemap.h | 8 ++++---- mm/swapfile.c | 7 +++---- 3 files changed, 7 insertions(+), 21 deletions(-)
include/linux/mm.h +0 -13
@@ -2290,19 +2290,6 @@ static inline void *folio_address(const struct folio *folio)return page_address (&folio->page); } -extern pgoff_t __page_file_index (struct page *page); - -/* -* Return the pagecache index of the passed page. Regular pagecache pages -* use ->index whereas swapcache pages use swp_offset(->private) -*/ -static inline pgoff_t page_index (struct page *page) -{ -if (unlikely (PageSwapCache (page))) -return __page_file_index (page); -return page->index; -} -/* * Return true only if the page has been allocated with * ALLOC_NO_WATERMARKS and the low watermark was not
include/linux/pagemap.h +4 -4
@@ -778,7 +778,7 @@ static inline struct page *grab_cache_page_nowait(struct address_space *mapping, mapping_gfp_mask(mapping)); } -#define swapcache_index (folio) __page_file_index(&(folio)->page) +extern pgoff_t __folio_swap_cache_index (struct folio *folio);/** * folio_index - File index of a folio.
@@ -793,9 +793,9 @@ static inline struct page *grab_cache_page_nowait(struct address_space *mapping,*/ static inline pgoff_t folio_index (struct folio *folio) { -if (unlikely (folio_test_swapcache (folio))) -return swapcache_index (folio); -return folio->index; +if (unlikely (folio_test_swapcache (folio))) +return __folio_swap_cache_index (folio); +return folio->index; }/**
mm/swapfile.c +3 -4
@@ -3474,12 +3474,11 @@ struct address_space *swapcache_mapping(struct folio *folio) }EXPORT_SYMBOL_GPL (swapcache_mapping); -pgoff_t __page_file_index (struct page *page) +pgoff_t __folio_swap_cache_index (struct folio *folio) { -swp_entry_t swap =page_swap_entry (page); -return swp_offset (swap); +return swp_offset (folio->swap); } -EXPORT_SYMBOL_GPL (__page_file_index); +EXPORT_SYMBOL_GPL (__folio_swap_cache_index);/* * add_swap_count_continuation - called when a swap count is duplicated
Signature · 1 line
2.44.0
Re: [PATCH v4 11/12] mm: drop page_index and convert folio_index to use folio
Quoted 10 lines · Kairui Song
From: Kairui Song <[email protected]>
There are two helpers for retrieving the index within address space for mixed usage of swap cache and page cache:
-
page_index
-folio_index(wrapper ofpage_index)This commit drops
page_index, as we have eliminated all users, and convertsfolio_indexto use folio internally.
The latter does not make sense. folio_index() already is using a folio internally. Maybe a leftover from reshuffling/reworking patches?
Signature · 3 lines
Cheers, David / dhildenb
Re: [PATCH v4 11/12] mm: drop page_index and convert folio_index to use folio
Quoted 7 lines · David Hildenbrand
Quoted 5 lines · Kairui Song
From: Kairui Song <[email protected]>
There are two helpers for retrieving the index within address space for mixed usage of swap cache and page cache:
Quoted 8 lines
Quoted 5 lines
-
page_index
-folio_index(wrapper ofpage_index)This commit drops
page_index, as we have eliminated all users, and convertsfolio_indexto use folio internally.The latter does not make sense.
folio_index()already is using a folio internally. Maybe a leftover from reshuffling/reworking patches?
Hi, David,
folio_index calls swapcache_index, and swapcache_index is defined as:
#define swapcache_index(folio) __page_file_index(&(folio)->page)
Where it casts the folio to page first, then call __page_file_index, __page_file_index is a function and works on pages.
After this commit __page_file_index is converted to __folio_swap_cache_index. This change is a bit of trivial but we get rid of the internal page conversion.
I can simplify the commit message, just say drop page_index to make the code cleaner, if this is confusing.
Re: [PATCH v4 11/12] mm: drop page_index and convert folio_index to use folio
Quoted 27 lines · Kairui Song
Quoted 7 lines · David Hildenbrand
Quoted 5 lines · Kairui Song
From: Kairui Song <[email protected]>
There are two helpers for retrieving the index within address space for mixed usage of swap cache and page cache:
Quoted 8 lines
Quoted 5 lines
-
page_index
-folio_index(wrapper ofpage_index)This commit drops
page_index, as we have eliminated all users, and convertsfolio_indexto use folio internally.The latter does not make sense.
folio_index()already is using a folio internally. Maybe a leftover from reshuffling/reworking patches?Hi, David,
folio_indexcallsswapcache_index, andswapcache_indexis defined as:#define
swapcache_index(folio)__page_file_index(&(folio)->page)Where it casts the folio to page first, then call
__page_file_index,__page_file_indexis a function and works on pages.
Quoted 6 lines
After this commit
__page_file_indexis converted to__folio_swap_cache_index. This change is a bit of trivial but we get rid of the internal page conversion.I can simplify the commit message, just say drop
page_indexto make the code cleaner, if this is confusing.
Ah, you are right folio_index is not a simple wrapper of page_index indeed, that sentence in the commit message doesn't make sense, so it should be deleted, my bad for this leftover.
PATCH v4 12/12 mm/swap: reduce swap cache search space
In reply to Kairui SongFrom: Kairui Song <[email protected]>
Currently we use one swap_address_space for every 64M chunk to reduce lock contention, this is like having a set of smaller swap files inside one swap device. But when doing swap cache look up or insert, we are still using the offset of the whole large swap device. This is OK for correctness, as the offset (key) is unique.
But Xarray is specially optimized for small indexes, it creates the radix tree levels lazily to be just enough to fit the largest key stored in one Xarray. So we are wasting tree nodes unnecessarily.
For 64M chunk it should only take at most 3 levels to contain everything. But if we are using the offset from the whole swap device, the offset (key) value will be way beyond 64M, and so will the tree level.
Optimize this by using a new helper swap_cache_index to get a swap entry's unique offset in its own 64M swap_address_space.
I see a ~1% performance gain in benchmark and actual workload with high memory pressure.
Test with time memhog 128G inside a 8G memcg using 128G swap (ramdisk with SWP_SYNCHRONOUS_IO dropped, tested 3 times, results are stable. The test result is similar but the improvement is smaller if SWP_SYNCHRONOUS_IO is enabled, as swap out path can never skip swap cache):
Before:
6.07user 250.74system 4:17.26elapsed 99%CPU (0avgtext+0avgdata 8373376maxresident)k 0inputs+0outputs (55major+33555018minor)pagefaults 0swaps
After (1.8% faster):
6.08user 246.09system 4:12.58elapsed 99%CPU (0avgtext+0avgdata 8373248maxresident)k 0inputs+0outputs (54major+33555027minor)pagefaults 0swaps
Similar result with MySQL and sysbench using swap:
Before:
94055.61 qps
After (0.8% faster):
94834.91 qps
Radix tree slab usage is also very slightly lower.
--- mm/huge_memory.c | 2 +- mm/memcontrol.c | 2 +- mm/mincore.c | 2 +- mm/shmem.c | 2 +- mm/swap.h | 15 +++++++++++++++ mm/swap_state.c | 19 ++++++++++--------- mm/swapfile.c | 6 +++--- 7 files changed, 32 insertions(+), 16 deletions(-)
mm/huge_memory.c +1 -1
@@ -2918,7 +2918,7 @@ static void __split_huge_page(struct page *page, struct list_head *list,split_page_memcg (head, order, new_order);if (folio_test_anon (folio) &&folio_test_swapcache (folio)) { - offset =swp_offset (folio->swap); + offset =swap_cache_index (folio->swap); swap_cache =swap_address_space (folio->swap);xa_lock (&swap_cache->i_pages); }
mm/memcontrol.c +1 -1
@@ -6165,7 +6165,7 @@ static struct page *mc_handle_swap_pte(struct vm_area_struct *vma,* Because swap_cache_get_folio() updates some statistics counter, * we call find_get_page() with swapper_space directly. */ - page =find_get_page (swap_address_space (ent),swp_offset (ent)); + page =find_get_page (swap_address_space (ent),swap_cache_index (ent)); entry->val = ent.val;return page;
mm/mincore.c +1 -1
@@ -139,7 +139,7 @@ static int mincore_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, }else {#ifdef CONFIG_SWAP *vec =mincore_page (swap_address_space (entry), -swp_offset (entry)); +swap_cache_index (entry));#else WARN_ON (1 ); *vec =1 ;
mm/shmem.c +1 -1
@@ -1756,7 +1756,7 @@ static int shmem_replace_folio(struct folio **foliop, gfp_t gfp, old = *foliop; entry = old->swap; - swap_index =swp_offset (entry); + swap_index =swap_cache_index (entry); swap_mapping =swap_address_space (entry);/*
mm/swap.h +15 -0
@@ -27,6 +27,7 @@ void __swap_writepage(struct folio *folio, struct writeback_control *wbc);/* One swap address space for each 64M swap space */ #define SWAP_ADDRESS_SPACE_SHIFT 14#define SWAP_ADDRESS_SPACE_PAGES (1 << SWAP_ADDRESS_SPACE_SHIFT) +#define SWAP_ADDRESS_SPACE_MASK (BIT(SWAP_ADDRESS_SPACE_SHIFT) - 1)extern struct address_space *swapper_spaces[];#define swap_address_space (entry) \ (&swapper_spaces[swp_type(entry)][swp_offset(entry) \
@@ -40,6 +41,15 @@ static inline loff_t swap_dev_pos(swp_entry_t entry)return ((loff_t )swp_offset (entry)) <<PAGE_SHIFT ; } +/* +* Return the swap cache index of the swap entry. +*/ +static inline pgoff_t swap_cache_index (swp_entry_t entry) +{ +BUILD_BUG_ON ((SWP_OFFSET_MASK |SWAP_ADDRESS_SPACE_MASK ) !=SWP_OFFSET_MASK ); +return swp_offset (entry) &SWAP_ADDRESS_SPACE_MASK ; +} +void show_swap_cache_info (void );bool add_to_swap (struct folio *folio);void *get_shadow_from_swap_cache (swp_entry_t entry);
@@ -86,6 +96,11 @@ static inline struct address_space *swap_address_space(swp_entry_t entry)return NULL ; } +static inline pgoff_t swap_cache_index (swp_entry_t entry) +{ +return 0 ; +} +static inline void show_swap_cache_info (void ) { }
mm/swap_state.c +10 -9
@@ -72,7 +72,7 @@ void show_swap_cache_info(void)void *get_shadow_from_swap_cache (swp_entry_t entry) {struct address_space *address_space =swap_address_space (entry); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);void *shadow; shadow =xa_load (&address_space->i_pages, idx);
@@ -89,7 +89,7 @@ int add_to_swap_cache(struct folio *folio, swp_entry_t entry, gfp_t gfp, void **shadowp) { struct address_space *address_space = swap_address_space(entry); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);XA_STATE_ORDER (xas, &address_space->i_pages, idx,folio_order (folio));unsigned long i, nr =folio_nr_pages (folio);void *old;
@@ -144,7 +144,7 @@ void __delete_from_swap_cache(struct folio *folio,struct address_space *address_space =swap_address_space (entry);int i;long nr =folio_nr_pages (folio); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);XA_STATE (xas, &address_space->i_pages, idx);xas_set_update (&xas, workingset_update_node);
@@ -248,18 +248,19 @@ void delete_from_swap_cache(struct folio *folio)void clear_shadow_from_swap_cache (int type,unsigned long begin,unsigned long end) { -unsigned long curr = begin; +unsigned long curr = begin, offset;void *old;for (;;) { + offset = curr &SWAP_ADDRESS_SPACE_MASK ;swp_entry_t entry =swp_entry (type, curr);struct address_space *address_space =swap_address_space (entry); -XA_STATE (xas, &address_space->i_pages, curr); +XA_STATE (xas, &address_space->i_pages, offset);xas_set_update (&xas, workingset_update_node);xa_lock_irq (&address_space->i_pages); -xas_for_each (&xas, old, end) { +xas_for_each (&xas, old, offset +min (end - curr,SWAP_ADDRESS_SPACE_PAGES )) {if (!xa_is_value (old))continue ;xas_store (&xas,NULL );
@@ -350,7 +351,7 @@ struct folio *swap_cache_get_folio(swp_entry_t entry, {struct folio *folio; - folio =filemap_get_folio (swap_address_space (entry),swp_offset (entry)); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (!IS_ERR (folio)) {bool vma_ra =swap_use_vma_readahead ();bool readahead;
@@ -420,7 +421,7 @@ struct folio *filemap_get_incore_folio(struct address_space *mapping, si =get_swap_device (swp);if (!si)return ERR_PTR (-ENOENT ); - index =swp_offset (swp); + index =swap_cache_index (swp); folio =filemap_get_folio (swap_address_space (swp), index);put_swap_device (si);return folio;
@@ -447,7 +448,7 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,* that would confuse statistics. */ folio =filemap_get_folio (swap_address_space (entry), -swp_offset (entry)); +swap_cache_index (entry));if (!IS_ERR (folio)) goto got_folio;
mm/swapfile.c +3 -3
@@ -142,7 +142,7 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,struct folio *folio;int ret =0 ; - folio =filemap_get_folio (swap_address_space (entry), offset); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (IS_ERR (folio))return 0 ;/*
@@ -2158,7 +2158,7 @@ static int try_to_unuse(unsigned int type) (i = find_next_to_unuse(si, i)) != 0) { entry =swp_entry (type, i); - folio =filemap_get_folio (swap_address_space (entry), i); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (IS_ERR (folio))continue ;
@@ -3476,7 +3476,7 @@ EXPORT_SYMBOL_GPL(swapcache_mapping);pgoff_t __folio_swap_cache_index (struct folio *folio) { -return swp_offset (folio->swap); +return swap_cache_index (folio->swap); }EXPORT_SYMBOL_GPL (__folio_swap_cache_index);
Signature · 1 line
2.44.0
Re: [PATCH v4 12/12] mm/swap: reduce swap cache search space
Quoted 112 lines · Kairui Song
From: Kairui Song <[email protected]>
Currently we use one
swap_address_spacefor every 64M chunk to reduce lock contention, this is like having a set of smaller swap files inside one swap device. But when doing swap cache look up or insert, we are still using the offset of the whole large swap device. This is OK for correctness, as the offset (key) is unique.But Xarray is specially optimized for small indexes, it creates the radix tree levels lazily to be just enough to fit the largest key stored in one Xarray. So we are wasting tree nodes unnecessarily.
For 64M chunk it should only take at most 3 levels to contain everything. But if we are using the offset from the whole swap device, the offset (key) value will be way beyond 64M, and so will the tree level.
Optimize this by using a new helper
swap_cache_indexto get a swap entry's unique offset in its own 64Mswap_address_space.I see a ~1% performance gain in benchmark and actual workload with high memory pressure.
Test with
time memhog 128Ginside a 8G memcg using 128G swap (ramdisk withSWP_SYNCHRONOUS_IOdropped, tested 3 times, results are stable. The test result is similar but the improvement is smaller ifSWP_SYNCHRONOUS_IOis enabled, as swap out path can never skip swap cache):Before:
6.07user 250.74system 4:17.26elapsed 99%CPU (0avgtext+0avgdata 8373376maxresident)k 0inputs+0outputs (55major+33555018minor)pagefaults 0swapsAfter (1.8% faster):
6.08user 246.09system 4:12.58elapsed 99%CPU (0avgtext+0avgdata 8373248maxresident)k 0inputs+0outputs (54major+33555027minor)pagefaults 0swapsSimilar result with MySQL and sysbench using swap:
Before:
94055.61 qpsAfter (0.8% faster):
94834.91 qpsRadix tree slab usage is also very slightly lower.
Signed-off-by Kairui Song <[email protected]>--- mm/huge_memory.c | 2 +- mm/memcontrol.c | 2 +- mm/mincore.c | 2 +- mm/shmem.c | 2 +- mm/swap.h | 15 +++++++++++++++ mm/swap_state.c | 19 ++++++++++--------- mm/swapfile.c | 6 +++--- 7 files changed, 32 insertions(+), 16 deletions(-)mm/huge_memory.c +1
-1@@ -2918,7 +2918,7 @@ static void __split_huge_page(struct page *page, struct list_head *list,split_page_memcg (head, order, new_order);if (folio_test_anon (folio) &&folio_test_swapcache (folio)) { - offset =swp_offset (folio->swap); + offset =swap_cache_index (folio->swap); swap_cache =swap_address_space (folio->swap);xa_lock (&swap_cache->i_pages); }mm/memcontrol.c +1
-1@@ -6165,7 +6165,7 @@ static struct page *mc_handle_swap_pte(struct vm_area_struct *vma,* Because swap_cache_get_folio() updates some statistics counter, * we call find_get_page() with swapper_space directly. */ - page =find_get_page (swap_address_space (ent),swp_offset (ent)); + page =find_get_page (swap_address_space (ent),swap_cache_index (ent)); entry->val = ent.val;return page;mm/mincore.c +1
-1@@ -139,7 +139,7 @@ static int mincore_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, }else {#ifdef CONFIG_SWAP *vec =mincore_page (swap_address_space (entry), -swp_offset (entry)); +swap_cache_index (entry));#else WARN_ON (1 ); *vec =1 ;mm/shmem.c +1
-1@@ -1756,7 +1756,7 @@ static int shmem_replace_folio(struct folio **foliop, gfp_t gfp, old = *foliop; entry = old->swap; - swap_index =swp_offset (entry); + swap_index =swap_cache_index (entry); swap_mapping =swap_address_space (entry);/* mm/swap.h +0
-0
mm/swap.h void __swap_writepage(struct folio *folio, struct writeback_control *wbc); · line 30
Quoted 4 lines
mm/swap.h +1
-0@@ -27,6 +27,7 @@ void __swap_writepage(struct folio *folio, struct writeback_control *wbc);/* One swap address space for each 64M swap space */ #define SWAP_ADDRESS_SPACE_SHIFT 14#define SWAP_ADDRESS_SPACE_PAGES (1 << SWAP_ADDRESS_SPACE_SHIFT) +#define SWAP_ADDRESS_SPACE_MASK (BIT(SWAP_ADDRESS_SPACE_SHIFT) - 1)
#define SWAP_ADDRESS_SPACE_MASK (SWAP_ADDRESS_SPACE_PAGES - 1)
?
We can use BIT() in SWAP_ADDRESS_SPACE_PAGES definition.
Quoted 63 lines
mm/swap.h +14
-0extern struct address_space *swapper_spaces[];#define swap_address_space (entry) \ (&swapper_spaces[swp_type(entry)][swp_offset(entry) \@@ -40,6 +41,15 @@ static inline loff_t swap_dev_pos(swp_entry_t entry)return ((loff_t )swp_offset (entry)) <<PAGE_SHIFT ; } +/* +* Return the swap cache index of the swap entry. +*/ +static inline pgoff_t swap_cache_index (swp_entry_t entry) +{ +BUILD_BUG_ON ((SWP_OFFSET_MASK |SWAP_ADDRESS_SPACE_MASK ) !=SWP_OFFSET_MASK ); +return swp_offset (entry) &SWAP_ADDRESS_SPACE_MASK ; +} +void show_swap_cache_info (void );bool add_to_swap (struct folio *folio);void *get_shadow_from_swap_cache (swp_entry_t entry);@@ -86,6 +96,11 @@ static inline struct address_space *swap_address_space(swp_entry_t entry)return NULL ; } +static inline pgoff_t swap_cache_index (swp_entry_t entry) +{ +return 0 ; +} +static inline void show_swap_cache_info (void ) { }mm/swap_state.c +3
-3@@ -72,7 +72,7 @@ void show_swap_cache_info(void)void *get_shadow_from_swap_cache (swp_entry_t entry) {struct address_space *address_space =swap_address_space (entry); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);void *shadow; shadow =xa_load (&address_space->i_pages, idx);@@ -89,7 +89,7 @@ int add_to_swap_cache(struct folio *folio, swp_entry_t entry, gfp_t gfp, void **shadowp) { struct address_space *address_space = swap_address_space(entry); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);XA_STATE_ORDER (xas, &address_space->i_pages, idx,folio_order (folio));unsigned long i, nr =folio_nr_pages (folio);void *old;@@ -144,7 +144,7 @@ void __delete_from_swap_cache(struct folio *folio,struct address_space *address_space =swap_address_space (entry);int i;long nr =folio_nr_pages (folio); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);XA_STATE (xas, &address_space->i_pages, idx);xas_set_update (&xas, workingset_update_node);
mm/swap_state.c void delete_from_swap_cache(struct folio *folio) · line 251
Quoted 5 lines
mm/swap_state.c +1
-1@@ -248,18 +248,19 @@ void delete_from_swap_cache(struct folio *folio)void clear_shadow_from_swap_cache (int type,unsigned long begin,unsigned long end) { -unsigned long curr = begin; +unsigned long curr = begin, offset;
Better to rename "offset" as "index" to avoid confusion?
Quoted 4 lines
mm/swap_state.c +1
-0void *old;for (;;) { + offset = curr &SWAP_ADDRESS_SPACE_MASK ;
mm/swap_state.c void delete_from_swap_cache(struct folio *folio) · line 263
Quoted 10 lines
mm/swap_state.c +2
-2swp_entry_t entry =swp_entry (type, curr);struct address_space *address_space =swap_address_space (entry); -XA_STATE (xas, &address_space->i_pages, curr); +XA_STATE (xas, &address_space->i_pages, offset);xas_set_update (&xas, workingset_update_node);xa_lock_irq (&address_space->i_pages); -xas_for_each (&xas, old, end) { +xas_for_each (&xas, old, offset +min (end - curr,SWAP_ADDRESS_SPACE_PAGES )) {
Is there a bug in the original code? It doesn't check SWAP_ADDRESS_SPACE_PAGES.
And should it be changed to
xas_for_each(&xas, old, min(offset + end - curr, SWAP_ADDRESS_SPACE_PAGES))
?
Quoted 53 lines
mm/swap_state.c +3
-3if (!xa_is_value (old))continue ;xas_store (&xas,NULL );@@ -350,7 +351,7 @@ struct folio *swap_cache_get_folio(swp_entry_t entry, {struct folio *folio; - folio =filemap_get_folio (swap_address_space (entry),swp_offset (entry)); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (!IS_ERR (folio)) {bool vma_ra =swap_use_vma_readahead ();bool readahead;@@ -420,7 +421,7 @@ struct folio *filemap_get_incore_folio(struct address_space *mapping, si =get_swap_device (swp);if (!si)return ERR_PTR (-ENOENT ); - index =swp_offset (swp); + index =swap_cache_index (swp); folio =filemap_get_folio (swap_address_space (swp), index);put_swap_device (si);return folio;@@ -447,7 +448,7 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,* that would confuse statistics. */ folio =filemap_get_folio (swap_address_space (entry), -swp_offset (entry)); +swap_cache_index (entry));if (!IS_ERR (folio)) goto got_folio;mm/swapfile.c +2
-2@@ -142,7 +142,7 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,struct folio *folio;int ret =0 ; - folio =filemap_get_folio (swap_address_space (entry), offset); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (IS_ERR (folio))return 0 ;/* @@ -2158,7 +2158,7 @@ static int try_to_unuse(unsigned int type) (i = find_next_to_unuse(si, i)) != 0) { entry =swp_entry (type, i); - folio =filemap_get_folio (swap_address_space (entry), i); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (IS_ERR (folio))continue ;
mm/swapfile.c EXPORT_SYMBOL_GPL(swapcache_mapping); · line 3481
Quoted 7 lines
mm/swapfile.c +1
-1@@ -3476,7 +3476,7 @@ EXPORT_SYMBOL_GPL(swapcache_mapping);pgoff_t __folio_swap_cache_index (struct folio *folio) { -return swp_offset (folio->swap); +return swap_cache_index (folio->swap); }EXPORT_SYMBOL_GPL (__folio_swap_cache_index);
--
Best Regards,
Huang, Ying
Re: [PATCH v4 12/12] mm/swap: reduce swap cache search space
Quoted 123 lines · Ying
Quoted 112 lines · Kairui Song
From: Kairui Song <[email protected]>
Currently we use one
swap_address_spacefor every 64M chunk to reduce lock contention, this is like having a set of smaller swap files inside one swap device. But when doing swap cache look up or insert, we are still using the offset of the whole large swap device. This is OK for correctness, as the offset (key) is unique.But Xarray is specially optimized for small indexes, it creates the radix tree levels lazily to be just enough to fit the largest key stored in one Xarray. So we are wasting tree nodes unnecessarily.
For 64M chunk it should only take at most 3 levels to contain everything. it should only take at most 3 levels to contain everything. But if we are using the offset from the whole swap device, the offset (key)
value will be way beyond 64M, and so will the tree level.Optimize this by using a new helper
swap_cache_indexto get a swap entry's unique offset in its own 64Mswap_address_space.I see a ~1% performance gain in benchmark and actual workload with high memory pressure.
Test with
time memhog 128Ginside a 8G memcg using 128G swap (ramdisk withSWP_SYNCHRONOUS_IOdropped, tested 3 times, results are stable. The test result is similar but the improvement is smaller ifSWP_SYNCHRONOUS_IOis enabled, as swap out path can never skip swap cache):Before:
6.07user 250.74system 4:17.26elapsed 99%CPU (0avgtext+0avgdata 8373376maxresident)k 0inputs+0outputs (55major+33555018minor)pagefaults 0swapsAfter (1.8% faster):
6.08user 246.09system 4:12.58elapsed 99%CPU (0avgtext+0avgdata 8373248maxresident)k 0inputs+0outputs (54major+33555027minor)pagefaults 0swapsSimilar result with MySQL and sysbench using swap:
Before:
94055.61 qpsAfter (0.8% faster):
94834.91 qpsRadix tree slab usage is also very slightly lower.
Signed-off-by Kairui Song <[email protected]>--- mm/huge_memory.c | 2 +- mm/memcontrol.c | 2 +- mm/mincore.c | 2 +- mm/shmem.c | 2 +- mm/swap.h | 15 +++++++++++++++ mm/swap_state.c | 19 ++++++++++--------- mm/swapfile.c | 6 +++--- 7 files changed, 32 insertions(+), 16 deletions(-)mm/huge_memory.c +1
-1@@ -2918,7 +2918,7 @@ static void __split_huge_page(struct page *page, struct list_head *list,split_page_memcg (head, order, new_order);if (folio_test_anon (folio) &&folio_test_swapcache (folio)) { - offset =swp_offset (folio->swap); + offset =swap_cache_index (folio->swap); swap_cache =swap_address_space (folio->swap);xa_lock (&swap_cache->i_pages); }mm/memcontrol.c +1
-1@@ -6165,7 +6165,7 @@ static struct page *mc_handle_swap_pte(struct vm_area_struct *vma,* Because swap_cache_get_folio() updates some statistics counter, * we call find_get_page() with swapper_space directly. */ - page =find_get_page (swap_address_space (ent),swp_offset (ent)); + page =find_get_page (swap_address_space (ent),swap_cache_index (ent)); entry->val = ent.val;return page;mm/mincore.c +1
-1@@ -139,7 +139,7 @@ static int mincore_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, }else {#ifdef CONFIG_SWAP *vec =mincore_page (swap_address_space (entry), -swp_offset (entry)); +swap_cache_index (entry));#else WARN_ON (1 ); *vec =1 ;mm/shmem.c +1
-1@@ -1756,7 +1756,7 @@ static int shmem_replace_folio(struct folio **foliop, gfp_t gfp, old = *foliop; entry = old->swap; - swap_index =swp_offset (entry); + swap_index =swap_cache_index (entry); swap_mapping =swap_address_space (entry);/* mm/swap.h +0
-0mm/swap.h void __swap_writepage(struct folio *folio, struct writeback_control *wbc); · line 30
Quoted 4 lines
mm/swap.h +1
-0@@ -27,6 +27,7 @@ void __swap_writepage(struct folio *folio, struct writeback_control *wbc);/* One swap address space for each 64M swap space */ #define SWAP_ADDRESS_SPACE_SHIFT 14#define SWAP_ADDRESS_SPACE_PAGES (1 << SWAP_ADDRESS_SPACE_SHIFT) +#define SWAP_ADDRESS_SPACE_MASK (BIT(SWAP_ADDRESS_SPACE_SHIFT) - 1)#define
SWAP_ADDRESS_SPACE_MASK(SWAP_ADDRESS_SPACE_PAGES- 1)
?
Quoted 2 lines
We can use
BIT()inSWAP_ADDRESS_SPACE_PAGESdefinition.
I'll just use (SWAP_ADDRESS_SPACE_PAGES - 1) then, I was trying to make the changes minimal, but prefered the BIT macro, a trivial change.
Quoted 61 lines
Quoted 61 lines
mm/swap.h +14
-0extern struct address_space *swapper_spaces[];#define swap_address_space (entry) \ (&swapper_spaces[swp_type(entry)][swp_offset(entry) \@@ -40,6 +41,15 @@ static inline loff_t swap_dev_pos(swp_entry_t entry)return ((loff_t )swp_offset (entry)) <<PAGE_SHIFT ; } +/* +* Return the swap cache index of the swap entry. +*/ +static inline pgoff_t swap_cache_index (swp_entry_t entry) +{ +BUILD_BUG_ON ((SWP_OFFSET_MASK |SWAP_ADDRESS_SPACE_MASK ) !=SWP_OFFSET_MASK ); +return swp_offset (entry) &SWAP_ADDRESS_SPACE_MASK ; +} +void show_swap_cache_info (void );bool add_to_swap (struct folio *folio);void *get_shadow_from_swap_cache (swp_entry_t entry);@@ -86,6 +96,11 @@ static inline struct address_space *swap_address_space(swp_entry_t entry)return NULL ; } +static inline pgoff_t swap_cache_index (swp_entry_t entry) +{ +return 0 ; +} +static inline void show_swap_cache_info (void ) { }mm/swap_state.c +3
-3@@ -72,7 +72,7 @@ void show_swap_cache_info(void)void *get_shadow_from_swap_cache (swp_entry_t entry) {struct address_space *address_space =swap_address_space (entry); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);void *shadow; shadow =xa_load (&address_space->i_pages, idx);@@ -89,7 +89,7 @@ int add_to_swap_cache(struct folio *folio, swp_entry_t entry, gfp_t gfp, void **shadowp) { struct address_space *address_space = swap_address_space(entry); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);XA_STATE_ORDER (xas, &address_space->i_pages, idx,folio_order (folio));unsigned long i, nr =folio_nr_pages (folio);void *old;@@ -144,7 +144,7 @@ void __delete_from_swap_cache(struct folio *folio,struct address_space *address_space =swap_address_space (entry);int i;long nr =folio_nr_pages (folio); -pgoff_t idx =swp_offset (entry); +pgoff_t idx =swap_cache_index (entry);XA_STATE (xas, &address_space->i_pages, idx);
Quoted 9 lines
mm/swap_state.c void delete_from_swap_cache(struct folio *folio) · line 251
Quoted 7 lines
mm/swap_state.c +1
-1xas_set_update (&xas, workingset_update_node);@@ -248,18 +248,19 @@ void delete_from_swap_cache(struct folio *folio)void clear_shadow_from_swap_cache (int type,unsigned long begin,unsigned long end) { -unsigned long curr = begin; +unsigned long curr = begin, offset;Better to rename "offset" as "index" to avoid confusion?
Good idea.
Quoted 9 lines
Quoted 9 lines
mm/swap_state.c +2
-1void *old;for (;;) { + offset = curr &SWAP_ADDRESS_SPACE_MASK ;swp_entry_t entry =swp_entry (type, curr);struct address_space *address_space =swap_address_space (entry); -XA_STATE (xas, &address_space->i_pages, curr); +XA_STATE (xas, &address_space->i_pages, offset);
Quoted 7 lines
mm/swap_state.c void delete_from_swap_cache(struct folio *folio) · line 263
Quoted 5 lines
mm/swap_state.c +1
-1xas_set_update (&xas, workingset_update_node);xa_lock_irq (&address_space->i_pages); -xas_for_each (&xas, old, end) { +xas_for_each (&xas, old, offset +min (end - curr,SWAP_ADDRESS_SPACE_PAGES )) {Is there a bug in the original code? It doesn't check
SWAP_ADDRESS_SPACE_PAGES.
That's OK, if the (end - curr) goes above SWAP_ADDRESS_SPACE_PAGES, it means all content in current address_space needs to be purged. xas_for_each will stop after it iterated all content in the current address space. This is a bit hackish though.
Quoted 4 lines
And should it be changed to
xas_for_each(&xas, old, min(offset + end - curr, SWAP_ADDRESS_SPACE_PAGES))
It should be equivalent, as described above, but yeah, this looks cleaner. I'll use your suggested code.
Quoted 54 lines
?
Quoted 52 lines
mm/swap_state.c +3
-3if (!xa_is_value (old))continue ;xas_store (&xas,NULL );@@ -350,7 +351,7 @@ struct folio *swap_cache_get_folio(swp_entry_t entry, {struct folio *folio; - folio =filemap_get_folio (swap_address_space (entry),swp_offset (entry)); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (!IS_ERR (folio)) {bool vma_ra =swap_use_vma_readahead ();bool readahead;@@ -420,7 +421,7 @@ struct folio *filemap_get_incore_folio(struct address_space *mapping, si =get_swap_device (swp);if (!si)return ERR_PTR (-ENOENT ); - index =swp_offset (swp); + index =swap_cache_index (swp); folio =filemap_get_folio (swap_address_space (swp), index);put_swap_device (si);return folio;@@ -447,7 +448,7 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,* that would confuse statistics. */ folio =filemap_get_folio (swap_address_space (entry), -swp_offset (entry)); +swap_cache_index (entry));if (!IS_ERR (folio)) goto got_folio;mm/swapfile.c +2
-2@@ -142,7 +142,7 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,struct folio *folio;int ret =0 ; - folio =filemap_get_folio (swap_address_space (entry), offset); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (IS_ERR (folio))return 0 ;/* @@ -2158,7 +2158,7 @@ static int try_to_unuse(unsigned int type) (i = find_next_to_unuse(si, i)) != 0) { entry =swp_entry (type, i); - folio =filemap_get_folio (swap_address_space (entry), i); + folio =filemap_get_folio (swap_address_space (entry),swap_cache_index (entry));if (IS_ERR (folio))continue ;
Quoted 8 lines
mm/swapfile.c EXPORT_SYMBOL_GPL(swapcache_mapping); · line 3481
Quoted 8 lines
mm/swapfile.c +1
-1@@ -3476,7 +3476,7 @@ EXPORT_SYMBOL_GPL(swapcache_mapping);pgoff_t __folio_swap_cache_index (struct folio *folio) { -return swp_offset (folio->swap); +return swap_cache_index (folio->swap); }EXPORT_SYMBOL_GPL (__folio_swap_cache_index);
Thanks for the suggestions!
Quoted 4 lines
--
Best Regards,
Huang, Ying