From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.0 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 194E9C04EB9 for ; Fri, 30 Nov 2018 03:10:52 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id CE0382081C for ; Fri, 30 Nov 2018 03:10:51 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org CE0382081C Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=huawei.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727284AbeK3OSg (ORCPT ); Fri, 30 Nov 2018 09:18:36 -0500 Received: from szxga06-in.huawei.com ([45.249.212.32]:38416 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1726386AbeK3OSg (ORCPT ); Fri, 30 Nov 2018 09:18:36 -0500 Received: from DGGEMS402-HUB.china.huawei.com (unknown [172.30.72.58]) by Forcepoint Email with ESMTP id 3F7C9AD00A3A5; Fri, 30 Nov 2018 11:10:47 +0800 (CST) Received: from [127.0.0.1] (10.134.22.195) by DGGEMS402-HUB.china.huawei.com (10.3.19.202) with Microsoft SMTP Server id 14.3.408.0; Fri, 30 Nov 2018 11:10:39 +0800 Subject: Re: [f2fs-dev] [PATCH] f2fs: avoid frequent costly fsck triggers To: Sheng Yong , Jaegeuk Kim CC: , References: <20181128073125.39102-1-jaegeuk@kernel.org> <724de929-6ee1-3ad7-cfa7-6e80d4e7a3da@huawei.com> <20181128081035.GA41969@jaegeuk-macbookpro.roam.corp.google.com> <4f43735a-d01f-f194-18ce-4fbbe10ad8d4@huawei.com> <20181128174805.GB41969@jaegeuk-macbookpro.roam.corp.google.com> <234cabcd-a1c1-9fb6-13aa-94fe681731dc@huawei.com> From: Chao Yu Message-ID: Date: Fri, 30 Nov 2018 11:10:37 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <234cabcd-a1c1-9fb6-13aa-94fe681731dc@huawei.com> Content-Type: text/plain; charset="utf-8" Content-Language: en-US Content-Transfer-Encoding: 7bit X-Originating-IP: [10.134.22.195] X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 2018/11/30 10:35, Sheng Yong wrote: > Hi, Jaegeuk and Chao, > > On 2018/11/29 1:48, Jaegeuk Kim wrote: >> On 11/28, Chao Yu wrote: >>> On 2018/11/28 16:10, Jaegeuk Kim wrote: >>>> On 11/28, Chao Yu wrote: >>>>> Hi Jaeguek, >>>>> >>>>> On 2018/11/28 15:31, Jaegeuk Kim wrote: >>>>>> If we want to re-enable nat_bits, we rely on fsck which requires full scan >>>>>> of directory tree. Let's do that by regular fsck or unclean shutdown. >>>>> >>>>> Reviewed-by: Chao Yu >>>>> >>>>> BTW, I have patch made some month ago... >>>>> >>>>> In order to detect nat_bits disabling, could we introduce one more flag for >>>>> fsck? >>>> >>>> Do we have a way to enable nat_bits very quickly in fsck? >>> >>> For image with SBI_NATBIT_NEED_REPAIR flag, can we just check metadata and >>> rebuild nat_bits based on verified nat blocks/journals? >> >> I'm leaning to rely on full scan to enable nat_bits again. We may add a mount >> count or timer to trigger fsck regularly? > > I'm afraid regular full fsck would give us bad experience of booting time. > FYI, 256GB storage, which is filled with small files, costs almost 10 min > to do a full fsck. And it seems larger capacity storages are on the way. Agreed. > So, is it worth doing that only to enable nat_bits (plus checking f2fs > consistent not that necessarily)? In android environment, I think it may be too expensive for adding nat_bits by triggering full scan by fsck during boot time. If we can update all nat bitmap in free time after mount, maybe we can rebuild nat_bits based on full nat bitmap during umount, which can be cheaper than rebuiding in userspace. Thanks, > > Thanks >> >>> >>> Thanks, >>> >>>> >>>>> >>>>> >From 86e8bdb2faeec904944bb6621073f4f7de51cc2d Mon Sep 17 00:00:00 2001 >>>>> From: Chao Yu >>>>> Date: Sun, 9 Sep 2018 05:40:20 +0800 >>>>> Subject: [PATCH] f2fs: set specified flag after invalidating nat_bits >>>>> >>>>> Signed-off-by: Chao Yu >>>>> --- >>>>> fs/f2fs/checkpoint.c | 12 +++++++++++- >>>>> fs/f2fs/f2fs.h | 3 ++- >>>>> fs/f2fs/node.c | 3 +++ >>>>> include/linux/f2fs_fs.h | 1 + >>>>> 4 files changed, 17 insertions(+), 2 deletions(-) >>>>> >>>>> diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c >>>>> index 7e17bb3dfcb1..f7fb14e0f5f9 100644 >>>>> --- a/fs/f2fs/checkpoint.c >>>>> +++ b/fs/f2fs/checkpoint.c >>>>> @@ -1226,13 +1226,16 @@ static void update_ckpt_flags(struct f2fs_sb_info >>>>> *sbi, struct cp_control *cpc) >>>>> unsigned long orphan_num = sbi->im[ORPHAN_INO].ino_num; >>>>> struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi); >>>>> unsigned long flags; >>>>> + bool disable_natbits = false; >>>>> >>>>> spin_lock_irqsave(&sbi->cp_lock, flags); >>>>> >>>>> if ((cpc->reason & CP_UMOUNT) && >>>>> le32_to_cpu(ckpt->cp_pack_total_block_count) > >>>>> - sbi->blocks_per_seg - NM_I(sbi)->nat_bits_blocks) >>>>> + sbi->blocks_per_seg - NM_I(sbi)->nat_bits_blocks) { >>>>> disable_nat_bits(sbi, false); >>>>> + disable_natbits = true; >>>>> + } >>>>> >>>>> if (cpc->reason & CP_TRIMMED) >>>>> __set_ckpt_flags(ckpt, CP_TRIMMED_FLAG); >>>>> @@ -1270,11 +1273,18 @@ static void update_ckpt_flags(struct f2fs_sb_info >>>>> *sbi, struct cp_control *cpc) >>>>> if (is_sbi_flag_set(sbi, SBI_QUOTA_NEED_REPAIR)) >>>>> __set_ckpt_flags(ckpt, CP_QUOTA_NEED_FSCK_FLAG); >>>>> >>>>> + if (is_sbi_flag_set(sbi, SBI_NATBIT_NEED_REPAIR)) >>>>> + __set_ckpt_flags(ckpt, CP_NATBIT_NEED_FSCK_FLAG); >>>>> + >>>>> /* set this flag to activate crc|cp_ver for recovery */ >>>>> __set_ckpt_flags(ckpt, CP_CRC_RECOVERY_FLAG); >>>>> __clear_ckpt_flags(ckpt, CP_NOCRC_RECOVERY_FLAG); >>>>> >>>>> spin_unlock_irqrestore(&sbi->cp_lock, flags); >>>>> + >>>>> + if (disable_natbits) >>>>> + f2fs_msg(sbi->sb, KERN_NOTICE, >>>>> + "No enough space in CP area, disable nat_bits."); >>>>> } >>>>> >>>>> static void commit_checkpoint(struct f2fs_sb_info *sbi, >>>>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h >>>>> index f0cedbe0c536..b55341c269b2 100644 >>>>> --- a/fs/f2fs/f2fs.h >>>>> +++ b/fs/f2fs/f2fs.h >>>>> @@ -1107,6 +1107,7 @@ enum { >>>>> SBI_QUOTA_NEED_FLUSH, /* need to flush quota info in CP */ >>>>> SBI_QUOTA_SKIP_FLUSH, /* skip flushing quota in current CP */ >>>>> SBI_QUOTA_NEED_REPAIR, /* quota file may be corrupted */ >>>>> + SBI_NATBIT_NEED_REPAIR, /* nat full/empty bitmaps need repair */ >>>>> }; >>>>> >>>>> enum { >>>>> @@ -1628,7 +1629,7 @@ static inline void disable_nat_bits(struct >>>>> f2fs_sb_info *sbi, bool lock) >>>>> { >>>>> unsigned long flags; >>>>> >>>>> - set_sbi_flag(sbi, SBI_NEED_FSCK); >>>>> + set_sbi_flag(sbi, SBI_NATBIT_NEED_REPAIR); >>>>> >>>>> if (lock) >>>>> spin_lock_irqsave(&sbi->cp_lock, flags); >>>>> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c >>>>> index e57add1e8966..0c6f8312a087 100644 >>>>> --- a/fs/f2fs/node.c >>>>> +++ b/fs/f2fs/node.c >>>>> @@ -2902,6 +2902,9 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi) >>>>> >>>>> cp_ver |= (cur_cp_crc(ckpt) << 32); >>>>> if (cpu_to_le64(cp_ver) != *(__le64 *)nm_i->nat_bits) { >>>>> + f2fs_msg(sbi->sb, KERN_NOTICE, >>>>> + "Disable nat_bits due to incorrect cp_ver (%llu, %llu)", >>>>> + cp_ver, le64_to_cpu(*(__le64 *)nm_i->nat_bits)); >>>>> disable_nat_bits(sbi, true); >>>>> return 0; >>>>> } >>>>> diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h >>>>> index 7196653833fa..1f3ae1504573 100644 >>>>> --- a/include/linux/f2fs_fs.h >>>>> +++ b/include/linux/f2fs_fs.h >>>>> @@ -117,6 +117,7 @@ struct f2fs_super_block { >>>>> /* >>>>> * For checkpoint >>>>> */ >>>>> +#define CP_NATBIT_NEED_FSCK_FLAG 0X00002000 >>>>> #define CP_DISABLED_FLAG 0x00001000 >>>>> #define CP_QUOTA_NEED_FSCK_FLAG 0x00000800 >>>>> #define CP_LARGE_NAT_BITMAP_FLAG 0x00000400 >>>>> -- >>>>> 2.18.0.rc1 >>>>> >>>>> >>>>> >>>>>> >>>>>> Signed-off-by: Jaegeuk Kim >>>>>> --- >>>>>> fs/f2fs/f2fs.h | 6 +++++- >>>>>> 1 file changed, 5 insertions(+), 1 deletion(-) >>>>>> >>>>>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h >>>>>> index c28a9d1cb278..aa500239baf2 100644 >>>>>> --- a/fs/f2fs/f2fs.h >>>>>> +++ b/fs/f2fs/f2fs.h >>>>>> @@ -1621,7 +1621,11 @@ static inline void disable_nat_bits(struct f2fs_sb_info *sbi, bool lock) >>>>>> { >>>>>> unsigned long flags; >>>>>> >>>>>> - set_sbi_flag(sbi, SBI_NEED_FSCK); >>>>>> + /* >>>>>> + * In order to re-enable nat_bits we need to call fsck.f2fs by >>>>>> + * set_sbi_flag(sbi, SBI_NEED_FSCK). But it may give huge cost, >>>>>> + * so let's rely on regular fsck or unclean shutdown. >>>>>> + */ >>>>>> >>>>>> if (lock) >>>>>> spin_lock_irqsave(&sbi->cp_lock, flags); >>>>>> >>>> >>>> . >>>> >> >> >> _______________________________________________ >> Linux-f2fs-devel mailing list >> Linux-f2fs-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel >> >> . >> > > > . >