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 Received: from lists.sourceforge.net (lists.sourceforge.net [216.105.38.7]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id D014BF436BF for ; Fri, 17 Apr 2026 16:55:04 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.sourceforge.net; s=beta; h=Content-Transfer-Encoding:Content-Type:Cc: Reply-To:From:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:Subject:In-Reply-To:MIME-Version:References: Message-ID:To:Date:Sender:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=hR43EEOzhzcGMm3/5D+VZk+z5uBU7iRdFPSqJmgnIwU=; b=fryx0ge9Bi8NvUHD5+Mx+XC7rD ED1EJcOLDHkZj3bqXh9m0Zrhn8rs1v6d77R03lk9HP5fKvezt22CntXYyyVOIpiKBiksHiU0PAWih iFDrYbWPjb02+wNe2xx8SmCaZWG0SaG1O/+onideRIJRAIbQZ7f+Mav1mejB3AmzHzs4=; Received: from [127.0.0.1] (helo=sfs-ml-1.v29.lw.sourceforge.com) by sfs-ml-1.v29.lw.sourceforge.com with esmtp (Exim 4.95) (envelope-from ) id 1wDmT8-0002cw-Nh; Fri, 17 Apr 2026 16:55:02 +0000 Received: from [172.30.29.66] (helo=mx.sourceforge.net) by sfs-ml-1.v29.lw.sourceforge.com with esmtps (TLS1.2) tls TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (Exim 4.95) (envelope-from ) id 1wDmSs-0002bv-Lc for linux-f2fs-devel@lists.sourceforge.net; Fri, 17 Apr 2026 16:54:46 +0000 DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=sourceforge.net; s=x; h=In-Reply-To:Content-Type:MIME-Version:References: Message-ID:Subject:Cc:To:From:Date:Sender:Reply-To:Content-Transfer-Encoding: Content-ID:Content-Description:Resent-Date:Resent-From:Resent-Sender: Resent-To:Resent-Cc:Resent-Message-ID:List-Id:List-Help:List-Unsubscribe: List-Subscribe:List-Post:List-Owner:List-Archive; bh=w+Q+qsu5eZzgG+zqqs5cnPSd/WOLqs1mT2p+JGf3FVc=; b=TSDTEr/jXUtP+S4VXsFYAb8jj+ bzF9kewWCOxmQcYC17LHf2kduRiszP1vGjg9Or+y9iYtNgb8a8odfcNKYePmyrTgeHAUmUKyT5shl U2uwKkU0CHMY0RePQZc0pLP+QKRYWDbCLk3WuERKGDFl8ZO4Q80NUPUNHOPq+YTcmzqQ=; DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=sf.net; s=x ; h=In-Reply-To:Content-Type:MIME-Version:References:Message-ID:Subject:Cc:To :From:Date:Sender:Reply-To:Content-Transfer-Encoding:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:List-Id:List-Help:List-Unsubscribe:List-Subscribe: List-Post:List-Owner:List-Archive; bh=w+Q+qsu5eZzgG+zqqs5cnPSd/WOLqs1mT2p+JGf3FVc=; b=XUME7LuqPaSV5S61do2t5wN6nr 4uKKLJy02BhzOuigY16gVrGirW9vMV2BHu7z8p1RaKONMLTz79KFKUMa5cbloD/jR/yhyx8fTuKux +TC9u1y1YhaKQ/XRo5/IshFv4tplt1DxFBPRdc/HvvfiCpzLiVNz1qsloE2qzs9VQ/ss=; Received: from sea.source.kernel.org ([172.234.252.31]) by sfi-mx-2.v28.lw.sourceforge.com with esmtps (TLS1.2:ECDHE-RSA-AES256-GCM-SHA384:256) (Exim 4.95) id 1wDmSr-0008KZ-Nh for linux-f2fs-devel@lists.sourceforge.net; Fri, 17 Apr 2026 16:54:46 +0000 Received: from smtp.kernel.org (transwarp.subspace.kernel.org [100.75.92.58]) by sea.source.kernel.org (Postfix) with ESMTP id 5777344245; Fri, 17 Apr 2026 16:54:40 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 11296C19425; Fri, 17 Apr 2026 16:54:40 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1776444880; bh=QWx9eTo7FLv+Py3JhSnantsahTAE2aK/MBCnTx5n/RY=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=Y11KFfJzqw8istHrktvdLN2Pcrqz53wrY7oJEDu7FZQwE5imjzh8yXS7YtqO0UwUm 2deK4ZTcHym3grT00s8YSAtmmw2w79J0Dgq59dDFPd15zCtp/4A43RV58OS6kRfWBx O9KN7td+eI+RlACljRHtFJ1n1usVbIdNhodXgO7mFJqHbYSiEHX/we594figX5SX1Q q52naWjT/8aBV+seFKgN7wR4wgQBg8rf2sUvZoePlpBQSbw3QTaP4IMqnxcTLauCHr sppGLicbZvUCpMuWsA8lwilmSbkVoObq2eSuiFFC3kUXD1ldPKy+WIpgs8ri5mBFbB 595BRWXJzRQSA== Date: Fri, 17 Apr 2026 16:54:38 +0000 To: Chao Yu Message-ID: References: <20260409134538.3692605-1-jaegeuk@kernel.org> <7b6334a4-4aa1-4489-8fe1-499f0d32e38f@kernel.org> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <7b6334a4-4aa1-4489-8fe1-499f0d32e38f@kernel.org> X-Headers-End: 1wDmSr-0008KZ-Nh Subject: Re: [f2fs-dev] [PATCH v2] f2fs: another way to set large folio by remembering inode number X-BeenThere: linux-f2fs-devel@lists.sourceforge.net X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , From: Jaegeuk Kim via Linux-f2fs-devel Reply-To: Jaegeuk Kim Cc: Akilesh Kailash , linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: linux-f2fs-devel-bounces@lists.sourceforge.net On 04/17, Chao Yu wrote: > On 4/10/2026 9:16 AM, Jaegeuk Kim via Linux-f2fs-devel wrote: > > enum { > > F2FS_XATTR_FADV_LARGEFOLIO, > > }; > > > > unsigned int value = BIT(F2FS_XATTR_FADV_LARGEFOLIO); > > > > 1. setxattr(file, "user.fadvise", &value, sizeof(unsigned int), 0) > > -> register the inode number for large folio > > 2. chmod(0400, file) > > -> make Read-Only > > 3. fsync() && close() && open(READ) > > -> f2fs_iget() with large folio > > 4. open(WRITE), mkwrite on mmap, chmod(WRITE) > > -> return error > > 5. close() and open() > > -> goto #3 > > 6. unlink > > -> deregister the inode number > > > > Suggested-by: Akilesh Kailash > > Signed-off-by: Jaegeuk Kim > > --- > > Log from v1: > > - add a condition in f2fs_drop_inode > > - add Doc > > > > Documentation/filesystems/f2fs.rst | 41 ++++++++++++++++++++++++++---- > > fs/f2fs/checkpoint.c | 2 +- > > fs/f2fs/data.c | 2 +- > > fs/f2fs/f2fs.h | 1 + > > fs/f2fs/file.c | 11 ++++++-- > > fs/f2fs/inode.c | 19 +++++++++++--- > > fs/f2fs/super.c | 7 +++++ > > fs/f2fs/xattr.c | 35 ++++++++++++++++++++++++- > > fs/f2fs/xattr.h | 6 +++++ > > 9 files changed, 111 insertions(+), 13 deletions(-) > > > > diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst > > index 7e4031631286..de899d0d3088 100644 > > --- a/Documentation/filesystems/f2fs.rst > > +++ b/Documentation/filesystems/f2fs.rst > > @@ -1044,11 +1044,14 @@ page allocation for significant performance gains. To minimize code complexity, > > this support is currently excluded from the write path, which requires handling > > complex optimizations such as compression and block allocation modes. > > -This optional feature is triggered only when a file's immutable bit is set. > > -Consequently, F2FS will return EOPNOTSUPP if a user attempts to open a cached > > -file with write permissions, even immediately after clearing the bit. Write > > -access is only restored once the cached inode is dropped. The usage flow is > > -demonstrated below: > > +This optional feature is triggered by two mechanisms: the file's immutable bit > > +or a specific xattr flag. In both cases, F2FS ensures data integrity by > > +restricting the file to a read-only state while large folios are active. > > + > > +1. Immutable Bit Approach: > > +Triggered when the FS_IMMUTABLE_FL is set. This is a strict enforcement > > +where the file cannot be modified at all until the bit is cleared and > > +the cached inode is dropped. > > .. code-block:: > > @@ -1078,3 +1081,31 @@ demonstrated below: > > Written 4096 bytes with pattern = zero, total_time = 29 us, max_latency = 28 us > > # rm /data/testfile_read_seq > > + > > +2. XATTR fadvise Approach: > > +A more flexible registration via extended attributes. > > + > > +.. code-block:: > > + > > + enum { > > + F2FS_XATTR_FADV_LARGEFOLIO, > > + }; > > + unsigned int value = BIT(F2FS_XATTR_FADV_LARGEFOLIO); > > + > > + /* Registers the inode number for large folio support in the subsystem.*/ > > + # setxattr(file, "user.fadvise", &value, sizeof(unsigned int), 0) > > + > > + /* The file must be made Read-Only to transition into the large folio path. */ > > + # fchmod(0400, fd) > > + > > + /* clean up dirty inode state. */ > > + # fsync(fd) > > + > > + /* Drop the inode cache. > > + # close(fd) > > If drop_cache or memory reclaim was called here, then LARGE_FOLIO_INO will > be dropped in f2fs_evict_inode(), do we need to keep LARGE_FOLIO_INO if > inode->i_nlink is valid in f2fs_evict_inode()? We remove it only when we delete the file? > > Thanks, > > > + > > + /* f2fs_iget() instantiates the inode with large folio support.*/ > > + # open() > > + > > + /* Returns -EOPNOTSUPP or error to protect the large folio cache.*/ > > + # open(WRITE), mkwrite on mmap, or chmod(WRITE) > > diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c > > index 01e1ba77263e..fdd62ddc3ed6 100644 > > --- a/fs/f2fs/checkpoint.c > > +++ b/fs/f2fs/checkpoint.c > > @@ -778,7 +778,7 @@ void f2fs_remove_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type) > > __remove_ino_entry(sbi, ino, type); > > } > > -/* mode should be APPEND_INO, UPDATE_INO or TRANS_DIR_INO */ > > +/* mode should be APPEND_INO, UPDATE_INO, LARGE_FOLIO_IO, or TRANS_DIR_INO */ > > bool f2fs_exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode) > > { > > struct inode_management *im = &sbi->im[mode]; > > diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c > > index 965d4e6443c6..5e46230398d7 100644 > > --- a/fs/f2fs/data.c > > +++ b/fs/f2fs/data.c > > @@ -2494,7 +2494,7 @@ static int f2fs_read_data_large_folio(struct inode *inode, > > int ret = 0; > > bool folio_in_bio; > > - if (!IS_IMMUTABLE(inode) || f2fs_compressed_file(inode)) { > > + if (f2fs_compressed_file(inode)) { > > if (folio) > > folio_unlock(folio); > > return -EOPNOTSUPP; > > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h > > index e40b6b2784ee..02bc6eb96a59 100644 > > --- a/fs/f2fs/f2fs.h > > +++ b/fs/f2fs/f2fs.h > > @@ -381,6 +381,7 @@ enum { > > /* for the list of ino */ > > enum { > > ORPHAN_INO, /* for orphan ino list */ > > + LARGE_FOLIO_INO, /* for large folio case */ > > APPEND_INO, /* for append ino list */ > > UPDATE_INO, /* for update ino list */ > > TRANS_DIR_INO, /* for transactions dir ino list */ > > diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c > > index c0220cd7b332..64ba900410fc 100644 > > --- a/fs/f2fs/file.c > > +++ b/fs/f2fs/file.c > > @@ -2068,9 +2068,16 @@ static long f2fs_fallocate(struct file *file, int mode, > > static int f2fs_release_file(struct inode *inode, struct file *filp) > > { > > - if (atomic_dec_and_test(&F2FS_I(inode)->open_count)) > > + if (atomic_dec_and_test(&F2FS_I(inode)->open_count)) { > > f2fs_remove_donate_inode(inode); > > - > > + /* > > + * In order to get large folio as soon as possible, let's drop > > + * inode cache asap. See also f2fs_drop_inode. > > + */ > > + if (f2fs_exist_written_data(F2FS_I_SB(inode), > > + inode->i_ino, LARGE_FOLIO_INO)) > > + d_drop(filp->f_path.dentry); > > + } > > /* > > * f2fs_release_file is called at every close calls. So we should > > * not drop any inmemory pages by close called by other process. > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c > > index 89240be8cc59..e100bc5a378c 100644 > > --- a/fs/f2fs/inode.c > > +++ b/fs/f2fs/inode.c > > @@ -565,6 +565,20 @@ static bool is_meta_ino(struct f2fs_sb_info *sbi, unsigned int ino) > > ino == F2FS_COMPRESS_INO(sbi); > > } > > +static void f2fs_mapping_set_large_folio(struct inode *inode) > > +{ > > + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); > > + > > + if (f2fs_compressed_file(inode)) > > + return; > > + if (f2fs_quota_file(sbi, inode->i_ino)) > > + return; > > + if (IS_IMMUTABLE(inode) || > > + (f2fs_exist_written_data(sbi, inode->i_ino, LARGE_FOLIO_INO) && > > + !(inode->i_mode & S_IWUGO))) > > + mapping_set_folio_min_order(inode->i_mapping, 0); > > +} > > + > > struct inode *f2fs_iget(struct super_block *sb, unsigned long ino) > > { > > struct f2fs_sb_info *sbi = F2FS_SB(sb); > > @@ -620,9 +634,7 @@ struct inode *f2fs_iget(struct super_block *sb, unsigned long ino) > > inode->i_op = &f2fs_file_inode_operations; > > inode->i_fop = &f2fs_file_operations; > > inode->i_mapping->a_ops = &f2fs_dblock_aops; > > - if (IS_IMMUTABLE(inode) && !f2fs_compressed_file(inode) && > > - !f2fs_quota_file(sbi, inode->i_ino)) > > - mapping_set_folio_min_order(inode->i_mapping, 0); > > + f2fs_mapping_set_large_folio(inode); > > } else if (S_ISDIR(inode->i_mode)) { > > inode->i_op = &f2fs_dir_inode_operations; > > inode->i_fop = &f2fs_dir_operations; > > @@ -895,6 +907,7 @@ void f2fs_evict_inode(struct inode *inode) > > f2fs_remove_ino_entry(sbi, inode->i_ino, APPEND_INO); > > f2fs_remove_ino_entry(sbi, inode->i_ino, UPDATE_INO); > > f2fs_remove_ino_entry(sbi, inode->i_ino, FLUSH_INO); > > + f2fs_remove_ino_entry(sbi, inode->i_ino, LARGE_FOLIO_INO); > > if (!is_sbi_flag_set(sbi, SBI_IS_FREEZING)) { > > sb_start_intwrite(inode->i_sb); > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > > index ccf806b676f5..11d1e0c99ac1 100644 > > --- a/fs/f2fs/super.c > > +++ b/fs/f2fs/super.c > > @@ -1844,6 +1844,13 @@ static int f2fs_drop_inode(struct inode *inode) > > return 1; > > } > > } > > + /* > > + * In order to get large folio as soon as possible, let's drop > > + * inode cache asap. See also f2fs_release_file. > > + */ > > + if (f2fs_exist_written_data(sbi, inode->i_ino, LARGE_FOLIO_INO) && > > + !is_inode_flag_set(inode, FI_DIRTY_INODE)) > > + return 1; > > /* > > * This is to avoid a deadlock condition like below. > > diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c > > index 941dc62a6d6f..0c0e44c2dcdd 100644 > > --- a/fs/f2fs/xattr.c > > +++ b/fs/f2fs/xattr.c > > @@ -44,6 +44,16 @@ static void xattr_free(struct f2fs_sb_info *sbi, void *xattr_addr, > > kfree(xattr_addr); > > } > > +static int f2fs_xattr_fadvise_get(struct inode *inode, void *buffer) > > +{ > > + if (!buffer) > > + goto out; > > + if (mapping_large_folio_support(inode->i_mapping)) > > + *((unsigned int *)buffer) |= BIT(F2FS_XATTR_FADV_LARGEFOLIO); > > +out: > > + return sizeof(unsigned int); > > +} > > + > > static int f2fs_xattr_generic_get(const struct xattr_handler *handler, > > struct dentry *unused, struct inode *inode, > > const char *name, void *buffer, size_t size) > > @@ -61,10 +71,29 @@ static int f2fs_xattr_generic_get(const struct xattr_handler *handler, > > default: > > return -EINVAL; > > } > > + if (handler->flags == F2FS_XATTR_INDEX_USER && > > + !strcmp(name, "fadvise")) > > + return f2fs_xattr_fadvise_get(inode, buffer); > > + > > return f2fs_getxattr(inode, handler->flags, name, > > buffer, size, NULL); > > } > > +static int f2fs_xattr_fadvise_set(struct inode *inode, const void *value) > > +{ > > + unsigned int new_fadvise; > > + > > + new_fadvise = *(unsigned int *)value; > > + > > + if (new_fadvise & BIT(F2FS_XATTR_FADV_LARGEFOLIO)) > > + f2fs_add_ino_entry(F2FS_I_SB(inode), > > + inode->i_ino, LARGE_FOLIO_INO); > > + else > > + f2fs_remove_ino_entry(F2FS_I_SB(inode), > > + inode->i_ino, LARGE_FOLIO_INO); > > + return 0; > > +} > > + > > static int f2fs_xattr_generic_set(const struct xattr_handler *handler, > > struct mnt_idmap *idmap, > > struct dentry *unused, struct inode *inode, > > @@ -84,6 +113,10 @@ static int f2fs_xattr_generic_set(const struct xattr_handler *handler, > > default: > > return -EINVAL; > > } > > + if (handler->flags == F2FS_XATTR_INDEX_USER && > > + !strcmp(name, "fadvise")) > > + return f2fs_xattr_fadvise_set(inode, value); > > + > > return f2fs_setxattr(inode, handler->flags, name, > > value, size, NULL, flags); > > } > > @@ -842,4 +875,4 @@ int __init f2fs_init_xattr_cache(void) > > void f2fs_destroy_xattr_cache(void) > > { > > kmem_cache_destroy(inline_xattr_slab); > > -} > > \ No newline at end of file > > +} > > diff --git a/fs/f2fs/xattr.h b/fs/f2fs/xattr.h > > index bce3d93e4755..455f460d014e 100644 > > --- a/fs/f2fs/xattr.h > > +++ b/fs/f2fs/xattr.h > > @@ -24,6 +24,7 @@ > > #define F2FS_XATTR_REFCOUNT_MAX 1024 > > /* Name indexes */ > > +#define F2FS_USER_FADVISE_NAME "user.fadvise" > > #define F2FS_SYSTEM_ADVISE_NAME "system.advise" > > #define F2FS_XATTR_INDEX_USER 1 > > #define F2FS_XATTR_INDEX_POSIX_ACL_ACCESS 2 > > @@ -39,6 +40,11 @@ > > #define F2FS_XATTR_NAME_ENCRYPTION_CONTEXT "c" > > #define F2FS_XATTR_NAME_VERITY "v" > > +/* used for F2FS_USER_FADVISE_NAME */ > > +enum { > > + F2FS_XATTR_FADV_LARGEFOLIO, > > +}; > > + > > struct f2fs_xattr_header { > > __le32 h_magic; /* magic number for identification */ > > __le32 h_refcount; /* reference count */ _______________________________________________ Linux-f2fs-devel mailing list Linux-f2fs-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel