| Message ID | 9047C53C18267742AB12E43B65C7F9F70BCD2DE9@dggemi505-mbs.china.huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 2018/1/25 15:05, Gaoxiang (OS) wrote: > Previously, we attempt to flush the whole cp pack in a single bio, > however, when suddenly powering off at this time, we could get into > an extreme scenario that cp page1 and cp page2 are updated and > latest, but payload or current summaries are still partially outdated. > (see reliable write in the UFS specification) > > This patch submits the whole cp pack except cp page2 at first, > and then writes the cp page2 with an extra independent > bio with pre-io barrier. > > Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> > --- > Change log from v1: > - Apply the review comments from Chao > - time data from "finish block_ops" to " finish checkpoint" (tested on ARM64 with TOSHIBA 128GB UFS): > Before patch: 0.002273 0.001973 0.002789 0.005159 0.002050 > After patch: 0.002502 0.001624 0.002487 0.003049 0.002696 Looks not very stable. > fs/f2fs/checkpoint.c | 41 ++++++++++++++++++++++++++++++++++++----- > 1 file changed, 36 insertions(+), 5 deletions(-) > > diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c > index 14d2fed..70deb09 100644 > --- a/fs/f2fs/checkpoint.c > +++ b/fs/f2fs/checkpoint.c > @@ -300,6 +300,36 @@ static int f2fs_write_meta_pages(struct address_space *mapping, > return 0; > } > > +static void commit_checkpoint_with_barrier(struct f2fs_sb_info *sbi, Minor, {commit,sync}_checkpoint will be OK to me, since we will decide to keep barrier in more deep places according to NOBARRIER mount option, instead of this function. > + void *src, block_t blk_addr) > +{ > + struct writeback_control wbc = { > + .for_reclaim = 0, > + }; > + struct page *page = grab_meta_page(sbi, blk_addr); > + int err; > + > + memcpy(page_address(page), src, PAGE_SIZE); > + set_page_dirty(page); > + > + f2fs_wait_on_page_writeback(page, META, true); > + BUG_ON(PageWriteback(page)); > + if (unlikely(!clear_page_dirty_for_io(page))) > + BUG(); f2fs_bug_on(sbi, PageWriteback(page)); f2fs_bug_on(sbi, unlikely(!clear_page_dirty_for_io(page))); > + > + /* writeout cp page2 */ Please keep defined terms as the same in all places: s/cp page2/cp pack 2 page > + err = __f2fs_write_meta_page(page, &wbc, FS_CP_META_IO); > + if (err) { > + f2fs_put_page(page, 1); > + f2fs_stop_checkpoint(sbi, false); > + return; How about f2fs_bug_on(sbi, err), since we should not fail in __f2fs_write_meta_page, right? > + } > + f2fs_put_page(page, 0); > + > + /* submit checkpoint with barrier */ > + f2fs_submit_merged_write(sbi, META_FLUSH); > +} > + > long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type, > long nr_to_write, enum iostat_type io_type) > { > @@ -1297,9 +1327,6 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) > start_blk += NR_CURSEG_NODE_TYPE; > } > > - /* writeout checkpoint block */ > - update_meta_page(sbi, ckpt, start_blk); > - > /* wait for previous submitted node/meta pages writeback */ > wait_on_all_pages_writeback(sbi); > > @@ -1313,12 +1340,16 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) > sbi->last_valid_block_count = sbi->total_valid_block_count; > percpu_counter_set(&sbi->alloc_valid_block_count, 0); > > - /* Here, we only have one bio having CP pack */ > - sync_meta_pages(sbi, META_FLUSH, LONG_MAX, FS_CP_META_IO); > + /* Here, we have one bio having CP pack except cp page2 */ > + sync_meta_pages(sbi, META, LONG_MAX, FS_CP_META_IO); > > /* wait for previous submitted meta pages writeback */ > wait_on_all_pages_writeback(sbi); /* wait for previous submitted meta pages writeback */ if (!test_opt(sbi, NOBARRIER) wait_on_all_pages_writeback(sbi); Thanks, > > + /* barrier and flush checkpoint cp page2 */ > + commit_checkpoint_with_barrier(sbi, ckpt, start_blk); > + wait_on_all_pages_writeback(sbi); > + > release_ino_entry(sbi, false); > > if (unlikely(f2fs_cp_error(sbi))) >
Hi Chao, On 2018/1/25 16:18, Chao Yu wrote: > On 2018/1/25 15:05, Gaoxiang (OS) wrote: >> Previously, we attempt to flush the whole cp pack in a single bio, >> however, when suddenly powering off at this time, we could get into >> an extreme scenario that cp page1 and cp page2 are updated and >> latest, but payload or current summaries are still partially outdated. >> (see reliable write in the UFS specification) >> >> This patch submits the whole cp pack except cp page2 at first, >> and then writes the cp page2 with an extra independent >> bio with pre-io barrier. >> >> Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> >> --- >> Change log from v1: >> - Apply the review comments from Chao >> - time data from "finish block_ops" to " finish checkpoint" (tested on ARM64 with TOSHIBA 128GB UFS): >> Before patch: 0.002273 0.001973 0.002789 0.005159 0.002050 >> After patch: 0.002502 0.001624 0.002487 0.003049 0.002696 > > Looks not very stable. > The raw time is as follows, actually I think the time depends on the CPU , CPU schedule and some other conditions... We drop REQ_PREFLUSH and REQ_FUA in the first bio and the second bio is only 4k.. It would not have obvious performance difference, I think. before: kworker/u16:3-220 [003] ...1 378.093580: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:3-220 [003] ...1 378.095468: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:3-220 [003] ...1 378.097741: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (1) kworker/u16:1-97 [003] ...1 438.190720: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-97 [003] ...1 438.192693: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-97 [003] ...1 438.195540: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (2) kworker/u16:4-238 [001] ...1 498.301875: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:4-238 [001] ...1 498.304664: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:4-238 [001] ...1 498.308997: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (3) kworker/u16:4-238 [002] ...1 558.414151: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:4-238 [002] ...1 558.417236: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:4-238 [002] ...1 558.422395: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (4) kworker/u16:4-238 [001] ...1 618.525854: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:4-238 [001] ...1 618.527882: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:4-238 [000] ...1 618.529932: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (5) kworker/u16:4-238 [002] ...1 678.639271: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:4-238 [002] ...1 678.644787: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:4-238 [003] ...1 678.651480: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (6) kworker/u16:1-97 [002] ...1 738.765589: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-97 [002] ...1 738.767638: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-97 [000] ...1 738.770752: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (7) kworker/u16:4-238 [003] ...1 798.895281: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:4-238 [003] ...1 798.897615: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:4-238 [001] ...1 798.902006: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (8) kworker/u16:1-97 [003] ...1 859.021172: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-97 [003] ...1 859.022934: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-97 [003] ...1 859.024095: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (9) kworker/u16:3-220 [002] ...1 919.133620: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:3-220 [002] ...1 919.135558: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:3-220 [002] ...1 919.137182: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (10) after: kworker/u16:3-222 [000] ...1 137.820152: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:3-222 [000] ...1 137.822471: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:3-222 [000] ...1 137.824973: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (1) kworker/u16:1-96 [002] ...1 197.917060: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-96 [002] ...1 197.919031: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-96 [002] ...1 197.920655: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (2) kworker/u16:1-96 [002] ...1 258.013014: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-96 [002] ...1 258.014914: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-96 [002] ...1 258.017401: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (3) kworker/u16:0-6 [002] ...1 318.109930: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:0-6 [002] ...1 318.111908: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:0-6 [002] ...1 318.114957: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (4) kworker/u16:1-96 [003] ...1 378.222490: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-96 [003] ...1 378.224729: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-96 [002] ...1 378.227425: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (5) kworker/u16:1-96 [001] ...1 438.334204: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-96 [001] ...1 438.337002: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-96 [001] ...1 438.341479: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (6) kworker/u16:1-96 [002] ...1 498.462207: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:1-96 [002] ...1 498.465213: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:1-96 [003] ...1 498.469858: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (7) kworker/u16:0-6 [003] ...1 558.589039: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:0-6 [003] ...1 558.591138: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:0-6 [002] ...1 558.594847: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (8) kworker/u16:3-222 [003] ...1 618.702535: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:3-222 [003] ...1 618.708222: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:3-222 [003] ...1 618.724619: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (9) kworker/u16:3-222 [002] ...1 678.830550: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = start block_ops kworker/u16:3-222 [002] ...1 678.832785: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish block_ops kworker/u16:3-222 [002] ...1 678.835356: f2fs_write_checkpoint: dev = (259,44), checkpoint for Sync, state = finish checkpoint (10) >> fs/f2fs/checkpoint.c | 41 ++++++++++++++++++++++++++++++++++++----- >> 1 file changed, 36 insertions(+), 5 deletions(-) >> >> diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c >> index 14d2fed..70deb09 100644 >> --- a/fs/f2fs/checkpoint.c >> +++ b/fs/f2fs/checkpoint.c >> @@ -300,6 +300,36 @@ static int f2fs_write_meta_pages(struct address_space *mapping, >> return 0; >> } >> >> +static void commit_checkpoint_with_barrier(struct f2fs_sb_info *sbi, > > Minor, {commit,sync}_checkpoint will be OK to me, since we will decide to > keep barrier in more deep places according to NOBARRIER mount option, > instead of this function. OK, I agree. > >> + void *src, block_t blk_addr) >> +{ >> + struct writeback_control wbc = { >> + .for_reclaim = 0, >> + }; >> + struct page *page = grab_meta_page(sbi, blk_addr); >> + int err; >> + >> + memcpy(page_address(page), src, PAGE_SIZE); >> + set_page_dirty(page); >> + >> + f2fs_wait_on_page_writeback(page, META, true); >> + BUG_ON(PageWriteback(page)); >> + if (unlikely(!clear_page_dirty_for_io(page))) >> + BUG(); > > f2fs_bug_on(sbi, PageWriteback(page)); Will fix. > f2fs_bug_on(sbi, unlikely(!clear_page_dirty_for_io(page))); > I attempt to use if (unlikely(!clear_page_dirty_for_io(page)) f2fs_bug_on(1); since It is better to contain judgement only for assert-like functions (In case that clear_page_dirty_for_io could be omited by changing BUG_ON definition in the future). >> + >> + /* writeout cp page2 */ > > Please keep defined terms as the same in all places: > > s/cp page2/cp pack 2 page > >> + err = __f2fs_write_meta_page(page, &wbc, FS_CP_META_IO); >> + if (err) { >> + f2fs_put_page(page, 1); >> + f2fs_stop_checkpoint(sbi, false); >> + return; > > How about f2fs_bug_on(sbi, err), since we should not fail in > __f2fs_write_meta_page, right? OK, will fix in that way. > >> + } >> + f2fs_put_page(page, 0); >> + >> + /* submit checkpoint with barrier */ >> + f2fs_submit_merged_write(sbi, META_FLUSH); >> +} >> + >> long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type, >> long nr_to_write, enum iostat_type io_type) >> { >> @@ -1297,9 +1327,6 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) >> start_blk += NR_CURSEG_NODE_TYPE; >> } >> >> - /* writeout checkpoint block */ >> - update_meta_page(sbi, ckpt, start_blk); >> - >> /* wait for previous submitted node/meta pages writeback */ >> wait_on_all_pages_writeback(sbi); >> >> @@ -1313,12 +1340,16 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) >> sbi->last_valid_block_count = sbi->total_valid_block_count; >> percpu_counter_set(&sbi->alloc_valid_block_count, 0); >> >> - /* Here, we only have one bio having CP pack */ >> - sync_meta_pages(sbi, META_FLUSH, LONG_MAX, FS_CP_META_IO); >> + /* Here, we have one bio having CP pack except cp page2 */ >> + sync_meta_pages(sbi, META, LONG_MAX, FS_CP_META_IO); >> >> /* wait for previous submitted meta pages writeback */ >> wait_on_all_pages_writeback(sbi); > > /* wait for previous submitted meta pages writeback */ > if (!test_opt(sbi, NOBARRIER) > wait_on_all_pages_writeback(sbi); > OK. Thanks, > Thanks, > >> >> + /* barrier and flush checkpoint cp page2 */ >> + commit_checkpoint_with_barrier(sbi, ckpt, start_blk); >> + wait_on_all_pages_writeback(sbi); >> + >> release_ino_entry(sbi, false); >> >> if (unlikely(f2fs_cp_error(sbi))) >> >
diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 14d2fed..70deb09 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -300,6 +300,36 @@ static int f2fs_write_meta_pages(struct address_space *mapping, return 0; } +static void commit_checkpoint_with_barrier(struct f2fs_sb_info *sbi, + void *src, block_t blk_addr) +{ + struct writeback_control wbc = { + .for_reclaim = 0, + }; + struct page *page = grab_meta_page(sbi, blk_addr); + int err; + + memcpy(page_address(page), src, PAGE_SIZE); + set_page_dirty(page); + + f2fs_wait_on_page_writeback(page, META, true); + BUG_ON(PageWriteback(page)); + if (unlikely(!clear_page_dirty_for_io(page))) + BUG(); + + /* writeout cp page2 */ + err = __f2fs_write_meta_page(page, &wbc, FS_CP_META_IO); + if (err) { + f2fs_put_page(page, 1); + f2fs_stop_checkpoint(sbi, false); + return; + } + f2fs_put_page(page, 0); + + /* submit checkpoint with barrier */ + f2fs_submit_merged_write(sbi, META_FLUSH); +} + long sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type, long nr_to_write, enum iostat_type io_type) { @@ -1297,9 +1327,6 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) start_blk += NR_CURSEG_NODE_TYPE; } - /* writeout checkpoint block */ - update_meta_page(sbi, ckpt, start_blk); - /* wait for previous submitted node/meta pages writeback */ wait_on_all_pages_writeback(sbi); @@ -1313,12 +1340,16 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) sbi->last_valid_block_count = sbi->total_valid_block_count; percpu_counter_set(&sbi->alloc_valid_block_count, 0); - /* Here, we only have one bio having CP pack */ - sync_meta_pages(sbi, META_FLUSH, LONG_MAX, FS_CP_META_IO); + /* Here, we have one bio having CP pack except cp page2 */ + sync_meta_pages(sbi, META, LONG_MAX, FS_CP_META_IO); /* wait for previous submitted meta pages writeback */ wait_on_all_pages_writeback(sbi); + /* barrier and flush checkpoint cp page2 */ + commit_checkpoint_with_barrier(sbi, ckpt, start_blk); + wait_on_all_pages_writeback(sbi); + release_ino_entry(sbi, false); if (unlikely(f2fs_cp_error(sbi)))
Previously, we attempt to flush the whole cp pack in a single bio, however, when suddenly powering off at this time, we could get into an extreme scenario that cp page1 and cp page2 are updated and latest, but payload or current summaries are still partially outdated. (see reliable write in the UFS specification) This patch submits the whole cp pack except cp page2 at first, and then writes the cp page2 with an extra independent bio with pre-io barrier. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> --- Change log from v1: - Apply the review comments from Chao - time data from "finish block_ops" to " finish checkpoint" (tested on ARM64 with TOSHIBA 128GB UFS): Before patch: 0.002273 0.001973 0.002789 0.005159 0.002050 After patch: 0.002502 0.001624 0.002487 0.003049 0.002696 fs/f2fs/checkpoint.c | 41 ++++++++++++++++++++++++++++++++++++----- 1 file changed, 36 insertions(+), 5 deletions(-)