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=-6.8 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 6E1B8ECE561 for ; Tue, 18 Sep 2018 03:36:15 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 0CEE4206B7 for ; Tue, 18 Sep 2018 03:36:15 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0CEE4206B7 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 S1729010AbeIRJGq (ORCPT ); Tue, 18 Sep 2018 05:06:46 -0400 Received: from szxga07-in.huawei.com ([45.249.212.35]:33776 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1727233AbeIRJGq (ORCPT ); Tue, 18 Sep 2018 05:06:46 -0400 Received: from DGGEMS404-HUB.china.huawei.com (unknown [172.30.72.60]) by Forcepoint Email with ESMTP id F139BFE001A8C; Tue, 18 Sep 2018 11:36:09 +0800 (CST) Received: from [127.0.0.1] (10.134.22.195) by DGGEMS404-HUB.china.huawei.com (10.3.19.204) with Microsoft SMTP Server id 14.3.399.0; Tue, 18 Sep 2018 11:36:06 +0800 Subject: Re: [PATCH] f2fs: allow out-place-update for direct IO in LFS mode To: Jaegeuk Kim , Chao Yu CC: , References: <20180916133407.26802-1-chao@kernel.org> <20180918012701.GC79604@jaegeuk-macbookpro.roam.corp.google.com> From: Chao Yu Message-ID: <7ba874d6-9ba4-4247-e31d-488d8ffe7cca@huawei.com> Date: Tue, 18 Sep 2018 11:36:04 +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: <20180918012701.GC79604@jaegeuk-macbookpro.roam.corp.google.com> Content-Type: text/plain; charset="windows-1252" 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 2018/9/18 9:27, Jaegeuk Kim wrote: > On 09/16, Chao Yu wrote: >> From: Chao Yu >> >> Normally, DIO uses in-pllace-update, but in LFS mode, f2fs doesn't >> allow triggering any in-place-update writes, so we fallback direct >> write to buffered write, result in bad performance of large size >> write. > > I remember this is to serialize all the IOs in one buffered IO path to keep > the IO sequence espeicially useful for SMR drivers. Could you confirm whether > there is any IO reordering problem? IMO, buffered IO will encounter reordering issue more easily, for example: User write 2mb data via DIO, but according to policy on LFS mode, it will fallback to buffered IO. Before dio triggers writeback, an async writeback from background is triggered, so all IO distribution can be: page LBA range async IO: 0 - 255 x + 0, x + 255 sync IO: 256 - 511 x + 256, x + 511 Then block IO scheduler can issue sync IO before async IO, result IO reordering. Anyway, in DIO path we will suffer less IO reordering issue. Thanks, > >> >> This patch adds to support triggering out-place-update for direct IO >> to enhance its performance. >> >> Note that it needs to exclude direct read IO during direct write, >> since new data writing to new block address will no be valid until >> write finished. >> >> storage: zram >> >> time xfs_io -f -d /mnt/f2fs/file -c "pwrite 0 1073741824" -c "fsync" >> >> Before: >> real 0m13.061s >> user 0m0.327s >> sys 0m12.486s >> >> After: >> real 0m6.448s >> user 0m0.228s >> sys 0m6.212s >> >> Signed-off-by: Chao Yu >> --- >> fs/f2fs/data.c | 41 +++++++++++++++++++++++++++++++++-------- >> fs/f2fs/f2fs.h | 39 +++++++++++++++++++++++++++++++++++---- >> fs/f2fs/file.c | 3 ++- >> 3 files changed, 70 insertions(+), 13 deletions(-) >> >> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c >> index 048f9cf1d589..9937dc2f384a 100644 >> --- a/fs/f2fs/data.c >> +++ b/fs/f2fs/data.c >> @@ -894,7 +894,7 @@ static int __allocate_data_block(struct dnode_of_data *dn, int seg_type) >> >> dn->data_blkaddr = datablock_addr(dn->inode, >> dn->node_page, dn->ofs_in_node); >> - if (dn->data_blkaddr == NEW_ADDR) >> + if (dn->data_blkaddr != NULL_ADDR) >> goto alloc; >> >> if (unlikely((err = inc_valid_block_count(sbi, dn->inode, &count)))) >> @@ -950,7 +950,7 @@ int f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *from) >> >> if (direct_io) { >> map.m_seg_type = f2fs_rw_hint_to_seg_type(iocb->ki_hint); >> - flag = f2fs_force_buffered_io(inode, WRITE) ? >> + flag = f2fs_force_buffered_io(inode, iocb, from) ? >> F2FS_GET_BLOCK_PRE_AIO : >> F2FS_GET_BLOCK_PRE_DIO; >> goto map_blocks; >> @@ -1069,7 +1069,15 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, >> goto sync_out; >> } >> >> - if (!is_valid_data_blkaddr(sbi, blkaddr)) { >> + if (is_valid_data_blkaddr(sbi, blkaddr)) { >> + /* use out-place-update for driect IO under LFS mode */ >> + if (test_opt(sbi, LFS) && create && >> + flag == F2FS_GET_BLOCK_DEFAULT) { >> + err = __allocate_data_block(&dn, map->m_seg_type); >> + if (!err) >> + set_inode_flag(inode, FI_APPEND_WRITE); >> + } >> + } else { >> if (create) { >> if (unlikely(f2fs_cp_error(sbi))) { >> err = -EIO; >> @@ -2491,36 +2499,53 @@ static ssize_t f2fs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) >> struct address_space *mapping = iocb->ki_filp->f_mapping; >> struct inode *inode = mapping->host; >> struct f2fs_sb_info *sbi = F2FS_I_SB(inode); >> + struct f2fs_inode_info *fi = F2FS_I(inode); >> size_t count = iov_iter_count(iter); >> loff_t offset = iocb->ki_pos; >> int rw = iov_iter_rw(iter); >> int err; >> enum rw_hint hint = iocb->ki_hint; >> int whint_mode = F2FS_OPTION(sbi).whint_mode; >> + bool lock_read; >> >> err = check_direct_IO(inode, iter, offset); >> if (err) >> return err < 0 ? err : 0; >> >> - if (f2fs_force_buffered_io(inode, rw)) >> + if (f2fs_force_buffered_io(inode, iocb, iter)) >> return 0; >> >> + lock_read = allow_outplace_dio(inode, iocb, iter); >> + >> trace_f2fs_direct_IO_enter(inode, offset, count, rw); >> >> if (rw == WRITE && whint_mode == WHINT_MODE_OFF) >> iocb->ki_hint = WRITE_LIFE_NOT_SET; >> >> - if (!down_read_trylock(&F2FS_I(inode)->i_gc_rwsem[rw])) { >> - if (iocb->ki_flags & IOCB_NOWAIT) { >> + if (iocb->ki_flags & IOCB_NOWAIT) { >> + if (!down_read_trylock(&fi->i_gc_rwsem[rw])) { >> + iocb->ki_hint = hint; >> + err = -EAGAIN; >> + goto out; >> + } >> + if (lock_read && !down_read_trylock(&fi->i_gc_rwsem[READ])) { >> + up_read(&fi->i_gc_rwsem[rw]); >> iocb->ki_hint = hint; >> err = -EAGAIN; >> goto out; >> } >> - down_read(&F2FS_I(inode)->i_gc_rwsem[rw]); >> + } else { >> + down_read(&fi->i_gc_rwsem[rw]); >> + if (lock_read) >> + down_read(&fi->i_gc_rwsem[READ]); >> } >> >> err = blockdev_direct_IO(iocb, inode, iter, get_data_block_dio); >> - up_read(&F2FS_I(inode)->i_gc_rwsem[rw]); >> + >> + if (lock_read) >> + up_read(&fi->i_gc_rwsem[READ]); >> + >> + up_read(&fi->i_gc_rwsem[rw]); >> >> if (rw == WRITE) { >> if (whint_mode == WHINT_MODE_OFF) >> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h >> index 6f59ad180be7..a6dc9d581ee4 100644 >> --- a/fs/f2fs/f2fs.h >> +++ b/fs/f2fs/f2fs.h >> @@ -8,6 +8,7 @@ >> #ifndef _LINUX_F2FS_H >> #define _LINUX_F2FS_H >> >> +#include >> #include >> #include >> #include >> @@ -3461,11 +3462,41 @@ static inline bool f2fs_may_encrypt(struct inode *inode) >> #endif >> } >> >> -static inline bool f2fs_force_buffered_io(struct inode *inode, int rw) >> +static inline int block_unaligned_IO(struct inode *inode, >> + struct kiocb *iocb, struct iov_iter *iter) >> { >> - return (f2fs_post_read_required(inode) || >> - (rw == WRITE && test_opt(F2FS_I_SB(inode), LFS)) || >> - F2FS_I_SB(inode)->s_ndevs); >> + unsigned int i_blkbits = READ_ONCE(inode->i_blkbits); >> + unsigned int blocksize_mask = (1 << i_blkbits) - 1; >> + loff_t offset = iocb->ki_pos; >> + unsigned long align = offset | iov_iter_alignment(iter); >> + >> + return align & blocksize_mask; >> +} >> + >> +static inline int allow_outplace_dio(struct inode *inode, >> + struct kiocb *iocb, struct iov_iter *iter) >> +{ >> + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); >> + int rw = iov_iter_rw(iter); >> + >> + return (test_opt(sbi, LFS) && (rw == WRITE) && >> + !block_unaligned_IO(inode, iocb, iter)); >> +} >> + >> +static inline bool f2fs_force_buffered_io(struct inode *inode, >> + struct kiocb *iocb, struct iov_iter *iter) >> +{ >> + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); >> + int rw = iov_iter_rw(iter); >> + >> + if (f2fs_post_read_required(inode)) >> + return true; >> + if (sbi->s_ndevs) >> + return true; >> + if (test_opt(sbi, LFS) && (rw == WRITE) && >> + block_unaligned_IO(inode, iocb, iter)) >> + return true; >> + return false; >> } >> >> #ifdef CONFIG_F2FS_FAULT_INJECTION >> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c >> index 428e7398bd89..70491631e040 100644 >> --- a/fs/f2fs/file.c >> +++ b/fs/f2fs/file.c >> @@ -3018,7 +3018,8 @@ static ssize_t f2fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from) >> if (!f2fs_overwrite_io(inode, iocb->ki_pos, >> iov_iter_count(from)) || >> f2fs_has_inline_data(inode) || >> - f2fs_force_buffered_io(inode, WRITE)) { >> + f2fs_force_buffered_io(inode, >> + iocb, from)) { >> clear_inode_flag(inode, >> FI_NO_PREALLOC); >> inode_unlock(inode); >> -- >> 2.18.0 > > . >