From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750892AbdADEAR (ORCPT ); Tue, 3 Jan 2017 23:00:17 -0500 Received: from szxga01-in.huawei.com ([58.251.152.64]:46426 "EHLO szxga01-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750719AbdADEAQ (ORCPT ); Tue, 3 Jan 2017 23:00:16 -0500 Subject: Re: [PATCH] f2fs: fix small discards when se->valid_blocks is zero To: Jaegeuk Kim References: <1483434117-24427-1-git-send-email-yunlong.song@huawei.com> <20170104015510.GB16504@jaegeuk.local> CC: , , , , , , , From: Yunlong Song Message-ID: <586C731F.9020509@huawei.com> Date: Wed, 4 Jan 2017 11:59:27 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:38.0) Gecko/20100101 Thunderbird/38.5.1 MIME-Version: 1.0 In-Reply-To: <20170104015510.GB16504@jaegeuk.local> Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit X-Originating-IP: [10.111.74.205] X-CFilter-Loop: Reflected X-Mirapoint-Virus-RAPID-Raw: score=unknown(0), refid=str=0001.0A020202.586C733B.0152,ss=1,re=0.000,recu=0.000,reip=0.000,cl=1,cld=1,fgs=0, ip=0.0.0.0, so=2013-06-18 04:22:30, dmn=2013-03-21 17:37:32 X-Mirapoint-Loop-Id: c10d84bd39c0a0191e188a30e3837c97 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Kim, Although the blocks of that file will finally be discarded when it is not current segment any more and almost fully invalidate, but the point is that the blocks of that file can only be discarded along with the whole segment now, which violates the meaning of small discard. Look at the case I said in last mail, if the segment which owns the deleted file has no more changing after the file deleting, and its validate blocks are perhaps over 95%, and it may not be easy to be selected as a gc victim. In this case, FTL can not know the "file delete" on time, and the invalidate blocks of that file can not be discarded in FTL layer on time. On 2017/1/4 9:55, Jaegeuk Kim wrote: > Hi Yunlong, > > On 01/03, Yunlong Song wrote: >> In the small discard case, when se->valid_blocks is zero, the add_discard_addrs >> will directly return without __add_discard_entry. This will cause the file >> delete have no small discard. The case is like this: >> >> 1. Allocate free 2M segment >> 2. Write a file (size n blocks < 512) in that 2M segment, se->valid_blocks = n >> 3. Delete that file, se->valid_blocks = 0, add_discard_addrs will return without >> sending any discard of that file, and forever due to cur_map[i] ^ ckpt_map[i] = >> 0 after that checkpoint > During this checkpoint, that'll be discarded as a prefree segment, no? > Note that, if this is a current segment, f2fs won't discard it until it is > fully allocated. > > Thanks, > >> Signed-off-by: Yunlong Song >> --- >> fs/f2fs/segment.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c >> index 0738f48..8610f14 100644 >> --- a/fs/f2fs/segment.c >> +++ b/fs/f2fs/segment.c >> @@ -838,7 +838,7 @@ static void add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc) >> return; >> >> if (!force) { >> - if (!test_opt(sbi, DISCARD) || !se->valid_blocks || >> + if (!test_opt(sbi, DISCARD) || >> SM_I(sbi)->nr_discards >= SM_I(sbi)->max_discards) >> return; >> } >> -- >> 1.8.5.2 > . > -- Thanks, Yunlong Song