* [f2fs-dev] [PATCH] f2fs: optimize the way of traversing free_nid_bitmap [not found] <CGME20171107111512epcas2p2bb846cc19c3d8fa6f2858874183c446c@epcas2p2.samsung.com> @ 2017-11-07 11:14 ` Fan Li 2017-11-08 8:52 ` Chao Yu 0 siblings, 1 reply; 2+ messages in thread From: Fan Li @ 2017-11-07 11:14 UTC (permalink / raw) To: 'Chao Yu', 'Chao Yu', 'Jaegeuk Kim' Cc: linux-kernel, linux-f2fs-devel We call scan_free_nid_bits only when there isn't many free nids left, it means that marked bits in free_nid_bitmap are supposed to be few, use find_next_bit_le is more efficient in such case. According to my tests, use find_next_bit_le instead of test_bit_le will cut down the traversal time to one third of its original. Signed-off-by: Fan li <fanofcode.li@samsung.com> --- fs/f2fs/node.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index fef5c68..d234c6e 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1955,6 +1955,7 @@ static void scan_free_nid_bits(struct f2fs_sb_info *sbi) struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); struct f2fs_journal *journal = curseg->journal; unsigned int i, idx; + nid_t nid; down_read(&nm_i->nat_tree_lock); @@ -1964,10 +1965,10 @@ static void scan_free_nid_bits(struct f2fs_sb_info *sbi) if (!nm_i->free_nid_count[i]) continue; for (idx = 0; idx < NAT_ENTRY_PER_BLOCK; idx++) { - nid_t nid; - - if (!test_bit_le(idx, nm_i->free_nid_bitmap[i])) - continue; + idx = find_next_bit_le(nm_i->free_nid_bitmap[i], + NAT_ENTRY_PER_BLOCK, idx); + if (idx >= NAT_ENTRY_PER_BLOCK) + break; nid = i * NAT_ENTRY_PER_BLOCK + idx; add_free_nid(sbi, nid, true); @@ -1980,7 +1981,6 @@ static void scan_free_nid_bits(struct f2fs_sb_info *sbi) down_read(&curseg->journal_rwsem); for (i = 0; i < nats_in_cursum(journal); i++) { block_t addr; - nid_t nid; addr = le32_to_cpu(nat_in_journal(journal, i).block_addr); nid = le32_to_cpu(nid_in_journal(journal, i)); -- 2.7.4 ^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [f2fs-dev] [PATCH] f2fs: optimize the way of traversing free_nid_bitmap 2017-11-07 11:14 ` [f2fs-dev] [PATCH] f2fs: optimize the way of traversing free_nid_bitmap Fan Li @ 2017-11-08 8:52 ` Chao Yu 0 siblings, 0 replies; 2+ messages in thread From: Chao Yu @ 2017-11-08 8:52 UTC (permalink / raw) To: Fan Li, 'Chao Yu', 'Jaegeuk Kim' Cc: linux-kernel, linux-f2fs-devel On 2017/11/7 19:14, Fan Li wrote: > We call scan_free_nid_bits only when there isn't many > free nids left, it means that marked bits in free_nid_bitmap > are supposed to be few, use find_next_bit_le is more > efficient in such case. > According to my tests, use find_next_bit_le instead of > test_bit_le will cut down the traversal time to one > third of its original. > > Signed-off-by: Fan li <fanofcode.li@samsung.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Thanks, > --- > fs/f2fs/node.c | 10 +++++----- > 1 file changed, 5 insertions(+), 5 deletions(-) > > diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c > index fef5c68..d234c6e 100644 > --- a/fs/f2fs/node.c > +++ b/fs/f2fs/node.c > @@ -1955,6 +1955,7 @@ static void scan_free_nid_bits(struct f2fs_sb_info *sbi) > struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA); > struct f2fs_journal *journal = curseg->journal; > unsigned int i, idx; > + nid_t nid; > > down_read(&nm_i->nat_tree_lock); > > @@ -1964,10 +1965,10 @@ static void scan_free_nid_bits(struct f2fs_sb_info *sbi) > if (!nm_i->free_nid_count[i]) > continue; > for (idx = 0; idx < NAT_ENTRY_PER_BLOCK; idx++) { > - nid_t nid; > - > - if (!test_bit_le(idx, nm_i->free_nid_bitmap[i])) > - continue; > + idx = find_next_bit_le(nm_i->free_nid_bitmap[i], > + NAT_ENTRY_PER_BLOCK, idx); > + if (idx >= NAT_ENTRY_PER_BLOCK) > + break; > > nid = i * NAT_ENTRY_PER_BLOCK + idx; > add_free_nid(sbi, nid, true); > @@ -1980,7 +1981,6 @@ static void scan_free_nid_bits(struct f2fs_sb_info *sbi) > down_read(&curseg->journal_rwsem); > for (i = 0; i < nats_in_cursum(journal); i++) { > block_t addr; > - nid_t nid; > > addr = le32_to_cpu(nat_in_journal(journal, i).block_addr); > nid = le32_to_cpu(nid_in_journal(journal, i)); > ^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2017-11-08 8:52 UTC | newest] Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- [not found] <CGME20171107111512epcas2p2bb846cc19c3d8fa6f2858874183c446c@epcas2p2.samsung.com> 2017-11-07 11:14 ` [f2fs-dev] [PATCH] f2fs: optimize the way of traversing free_nid_bitmap Fan Li 2017-11-08 8:52 ` Chao Yu
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).