From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hou Pengyang Subject: Re: [PATCH 2/2] f2fs:use fggc_set_cold_data instead of set_cold_data in fggc Date: Thu, 23 Feb 2017 10:04:18 +0800 Message-ID: <58AE4322.60800@huawei.com> References: <20170222102900.152902-1-houpengyang@huawei.com> <20170222102900.152902-2-houpengyang@huawei.com> <20170223003957.GE4003@jaegeuk.local> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from sog-mx-2.v43.ch3.sourceforge.com ([172.29.43.192] helo=mx.sourceforge.net) by sfs-ml-2.v29.ch3.sourceforge.com with esmtp (Exim 4.76) (envelope-from ) id 1cgilu-0002Po-AL for linux-f2fs-devel@lists.sourceforge.net; Thu, 23 Feb 2017 02:04:42 +0000 Received: from [45.249.212.188] (helo=dggrg02-dlp.huawei.com) by sog-mx-2.v43.ch3.sourceforge.com with esmtps (TLSv1:RC4-SHA:128) (Exim 4.76) id 1cgils-0000gP-Lm for linux-f2fs-devel@lists.sourceforge.net; Thu, 23 Feb 2017 02:04:42 +0000 In-Reply-To: <20170223003957.GE4003@jaegeuk.local> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: linux-f2fs-devel-bounces@lists.sourceforge.net To: Jaegeuk Kim Cc: linux-f2fs-devel@lists.sourceforge.net On 2017/2/23 8:39, Jaegeuk Kim wrote: > Hi Pengyang, > > On 02/22, Hou Pengyang wrote: >> When no dirty colde segment for fggc ssr, we do NOT set_cold_data. > > The get_ssr_segment() finds a victim segment across all the hot/warm/cold data > logs. When does this happen? > Oh, I miss the travel across hot/warm/cold, sorry. how about the following filter in get_victim_segment, which can avoid unnecessary afterwards bitmap scanning. Thanks, diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index a0eed23..00791d2 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -290,9 +290,16 @@ static int get_victim_by_default(struct f2fs_sb_info *sbi, unsigned int secno, last_victim; unsigned int last_segment = MAIN_SEGS(sbi); unsigned int nsearched = 0; + int nr_dirty; mutex_lock(&dirty_i->seglist_lock); + if (alloc_mode == SSR) { + nr_dirty = dirty_i->nr_dirty[type]; + if (!nr_dirty) + goto out; + } + p.alloc_mode = alloc_mode; select_policy(sbi, gc_type, type, &p); > Thanks, > >> >> Signed-off-by: Hou Pengyang >> --- >> fs/f2fs/gc.c | 21 ++++++++++++++++++++- >> 1 file changed, 20 insertions(+), 1 deletion(-) >> >> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c >> index 4ee749f..a0eed23 100644 >> --- a/fs/f2fs/gc.c >> +++ b/fs/f2fs/gc.c >> @@ -666,6 +666,25 @@ static void move_encrypted_block(struct inode *inode, block_t bidx, >> f2fs_put_page(page, 1); >> } >> >> +/* >> + * If NOT enough dirty cold data segments for fggc, fggc would consume >> + * a valuable free segment, which would slow down the fggc procedure, >> + * If there are NO COLD dirty segment, we do not set_cold_page. >> + */ >> +static void fggc_set_cold_data(struct page *page) >> +{ >> + struct f2fs_sb_info *sbi = F2FS_P_SB(page); >> + struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); >> + int nr_cold_dirties; >> + >> + mutex_lock(&dirty_i->seglist_lock); >> + nr_cold_dirties = dirty_i->nr_dirty[DIRTY_COLD_DATA]; >> + mutex_unlock(&dirty_i->seglist_lock); >> + >> + if (nr_cold_dirties) >> + set_cold_data(page); >> + >> +} >> static void move_data_page(struct inode *inode, block_t bidx, int gc_type, >> unsigned int segno, int off) >> { >> @@ -706,7 +725,7 @@ static void move_data_page(struct inode *inode, block_t bidx, int gc_type, >> remove_dirty_inode(inode); >> } >> >> - set_cold_data(page); >> + fggc_set_cold_data(page); >> >> err = do_write_data_page(&fio); >> if (err == -ENOMEM && is_dirty) { >> -- >> 2.10.1 >> >> >> ------------------------------------------------------------------------------ >> Check out the vibrant tech community on one of the world's most >> engaging tech sites, SlashDot.org! http://sdm.link/slashdot >> _______________________________________________ >> Linux-f2fs-devel mailing list >> Linux-f2fs-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel > > . > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot