From nobody Thu Sep 19 01:39:35 2024 Received: from szxga07-in.huawei.com (szxga07-in.huawei.com [45.249.212.35]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 323B919D8BB; Tue, 27 Aug 2024 09:36:00 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=45.249.212.35 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724751364; cv=none; b=b83qgsqu7Elty9J+ffNcHGF2iwWuNDqHpexVaseGk47NVdsF0GkmljAi0w4NNoWT8XCuXj/X5nAGSFuO7CY4CYqWnbap3OUdnsF8zPDFDPt9VP0Ipm/XCBODX1Iy4VjH9t4W9CwS3YdYRJGxtdnHDtJXyL1EjfNtLn7VLkXfkKc= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724751364; c=relaxed/simple; bh=8/yhPNP2fc16XKIyoBg5uaMAZ24LIogx7cVJhDKd0NA=; h=From:To:CC:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=bl4mz8aHwpSIuG+tZdwTBL2MH2t6l92U2vHmf7IP26PZb+IJ78eFMSv395+i8HRtbT5hD3ooQqtvyoTQaEONyvafoPFF9g13WJ0XYdeVMPncDmVHPbrAZBH8cU5GXpAwYWOZsbczPh6QAb/YPjg1M9eZtlJpa91U7KTYApM/9fU= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=huawei.com; spf=pass smtp.mailfrom=huawei.com; arc=none smtp.client-ip=45.249.212.35 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=huawei.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=huawei.com Received: from mail.maildlp.com (unknown [172.19.88.214]) by szxga07-in.huawei.com (SkyGuard) with ESMTP id 4WtMq81ZjWz1S8tJ; Tue, 27 Aug 2024 17:35:48 +0800 (CST) Received: from kwepemd100024.china.huawei.com (unknown [7.221.188.41]) by mail.maildlp.com (Postfix) with ESMTPS id BCA7D1A016C; Tue, 27 Aug 2024 17:35:58 +0800 (CST) Received: from huawei.com (10.175.124.27) by kwepemd100024.china.huawei.com (7.221.188.41) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.1544.11; Tue, 27 Aug 2024 17:35:58 +0800 From: yangyun To: Miklos Szeredi CC: , , , Subject: [PATCH v3 1/2] fuse: move fuse_forget_link allocation inside fuse_queue_forget() Date: Tue, 27 Aug 2024 17:35:02 +0800 Message-ID: <20240827093503.3397562-2-yangyun50@huawei.com> X-Mailer: git-send-email 2.33.0 In-Reply-To: <20240827093503.3397562-1-yangyun50@huawei.com> References: <20240827093503.3397562-1-yangyun50@huawei.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-ClientProxiedBy: dggems706-chm.china.huawei.com (10.3.19.183) To kwepemd100024.china.huawei.com (7.221.188.41) Content-Type: text/plain; charset="utf-8" The `struct fuse_forget_link` is allocated outside `fuse_queue_forget()` before this patch. This requires the allocation in advance. In some cases (e.g., directory operations without errors), this struct is not needed but allocated, which contributes to memory usage and performance degradation. Besides, this messes up the code to some extent. So move the `fuse_forget_link` allocation inside `fuse_queue_forget()` with __GFP_NOFAIL in these cases. Since the inode eviction path can be triggered from memory reclaim and may block the system, it is better to preallocate it on inode creation. `fuse_force_forget()` is used by `readdirplus` before this patch for the reason that we do not know how many 'fuse_forget_link' structures will be allocated in advance when error happens. After this patch, this function is not needed any more and can be removed (The `noreply` fuse_args can also be removed). By this way, all FUSE_FORGET requests are sent by using `fuse_queue_forget()` function as e.g. virtiofs handles them differently from regular requests. Signed-off-by: yangyun --- fs/fuse/dev.c | 13 ++++++++--- fs/fuse/dir.c | 59 +++++++++++------------------------------------ fs/fuse/fuse_i.h | 10 ++++---- fs/fuse/inode.c | 9 ++------ fs/fuse/readdir.c | 30 ++++++------------------ 5 files changed, 38 insertions(+), 83 deletions(-) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 9eb191b5c4de..31ca97b1b32c 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -234,10 +234,18 @@ __releases(fiq->lock) } =20 void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forg= et, - u64 nodeid, u64 nlookup) + u64 nodeid, u64 nlookup) { struct fuse_iqueue *fiq =3D &fc->iq; =20 + /* + * The nullptr means that fuse_queue_forget() is used in error cases. + * Avoid preallocating this structure because it is unlikely used. + * Use __GFP_NOFAIL to make memory allocation always succeed. + */ + if (forget =3D=3D NULL) + forget =3D fuse_alloc_forget(GFP_KERNEL_ACCOUNT | __GFP_NOFAIL); + forget->forget_one.nodeid =3D nodeid; forget->forget_one.nlookup =3D nlookup; =20 @@ -508,8 +516,7 @@ ssize_t fuse_simple_request(struct fuse_mount *fm, stru= ct fuse_args *args) fuse_adjust_compat(fc, args); fuse_args_to_req(req, args); =20 - if (!args->noreply) - __set_bit(FR_ISREPLY, &req->flags); + __set_bit(FR_ISREPLY, &req->flags); __fuse_request_send(req); ret =3D req->out.h.error; if (!ret && args->out_argvar) { diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 2b0d4781f394..583362492ce0 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -207,7 +207,6 @@ static int fuse_dentry_revalidate(struct dentry *entry,= unsigned int flags) (flags & (LOOKUP_EXCL | LOOKUP_REVAL | LOOKUP_RENAME_TARGET))) { struct fuse_entry_out outarg; FUSE_ARGS(args); - struct fuse_forget_link *forget; u64 attr_version; =20 /* For negative dentries, always do a fresh lookup */ @@ -220,11 +219,6 @@ static int fuse_dentry_revalidate(struct dentry *entry= , unsigned int flags) =20 fm =3D get_fuse_mount(inode); =20 - forget =3D fuse_alloc_forget(); - ret =3D -ENOMEM; - if (!forget) - goto out; - attr_version =3D fuse_get_attr_version(fm->fc); =20 parent =3D dget_parent(entry); @@ -239,15 +233,13 @@ static int fuse_dentry_revalidate(struct dentry *entr= y, unsigned int flags) fi =3D get_fuse_inode(inode); if (outarg.nodeid !=3D get_node_id(inode) || (bool) IS_AUTOMOUNT(inode) !=3D (bool) (outarg.attr.flags & FUSE_AT= TR_SUBMOUNT)) { - fuse_queue_forget(fm->fc, forget, - outarg.nodeid, 1); + fuse_queue_forget(fm->fc, NULL, outarg.nodeid, 1); goto invalid; } spin_lock(&fi->lock); fi->nlookup++; spin_unlock(&fi->lock); } - kfree(forget); if (ret =3D=3D -ENOMEM || ret =3D=3D -EINTR) goto out; if (ret || fuse_invalid_attr(&outarg.attr) || @@ -365,7 +357,6 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid= , const struct qstr *name { struct fuse_mount *fm =3D get_fuse_mount_super(sb); FUSE_ARGS(args); - struct fuse_forget_link *forget; u64 attr_version; int err; =20 @@ -374,23 +365,17 @@ int fuse_lookup_name(struct super_block *sb, u64 node= id, const struct qstr *name if (name->len > FUSE_NAME_MAX) goto out; =20 - - forget =3D fuse_alloc_forget(); - err =3D -ENOMEM; - if (!forget) - goto out; - attr_version =3D fuse_get_attr_version(fm->fc); =20 fuse_lookup_init(fm->fc, &args, nodeid, name, outarg); err =3D fuse_simple_request(fm, &args); /* Zero nodeid is same as -ENOENT, but with valid timeout */ if (err || !outarg->nodeid) - goto out_put_forget; + goto out; =20 err =3D -EIO; if (fuse_invalid_attr(&outarg->attr)) - goto out_put_forget; + goto out; if (outarg->nodeid =3D=3D FUSE_ROOT_ID && outarg->generation !=3D 0) { pr_warn_once("root generation should be zero\n"); outarg->generation =3D 0; @@ -401,13 +386,11 @@ int fuse_lookup_name(struct super_block *sb, u64 node= id, const struct qstr *name attr_version); err =3D -ENOMEM; if (!*inode) { - fuse_queue_forget(fm->fc, forget, outarg->nodeid, 1); + fuse_queue_forget(fm->fc, NULL, outarg->nodeid, 1); goto out; } err =3D 0; =20 - out_put_forget: - kfree(forget); out: return err; } @@ -617,7 +600,6 @@ static int fuse_create_open(struct inode *dir, struct d= entry *entry, struct inode *inode; struct fuse_mount *fm =3D get_fuse_mount(dir); FUSE_ARGS(args); - struct fuse_forget_link *forget; struct fuse_create_in inarg; struct fuse_open_out *outopenp; struct fuse_entry_out outentry; @@ -628,15 +610,10 @@ static int fuse_create_open(struct inode *dir, struct= dentry *entry, /* Userspace expects S_IFREG in create mode */ BUG_ON((mode & S_IFMT) !=3D S_IFREG); =20 - forget =3D fuse_alloc_forget(); - err =3D -ENOMEM; - if (!forget) - goto out_err; - err =3D -ENOMEM; ff =3D fuse_file_alloc(fm, true); if (!ff) - goto out_put_forget_req; + goto out_err; =20 if (!fm->fc->dont_mask) mode &=3D ~current_umask(); @@ -670,7 +647,7 @@ static int fuse_create_open(struct inode *dir, struct d= entry *entry, =20 err =3D get_create_ext(&args, dir, entry, mode); if (err) - goto out_put_forget_req; + goto out_err; =20 err =3D fuse_simple_request(fm, &args); free_ext_value(&args); @@ -690,11 +667,10 @@ static int fuse_create_open(struct inode *dir, struct= dentry *entry, if (!inode) { flags &=3D ~(O_CREAT | O_EXCL | O_TRUNC); fuse_sync_release(NULL, ff, flags); - fuse_queue_forget(fm->fc, forget, outentry.nodeid, 1); + fuse_queue_forget(fm->fc, NULL, outentry.nodeid, 1); err =3D -ENOMEM; goto out_err; } - kfree(forget); d_instantiate(entry, inode); fuse_change_entry_timeout(entry, &outentry); fuse_dir_changed(dir); @@ -716,8 +692,6 @@ static int fuse_create_open(struct inode *dir, struct d= entry *entry, =20 out_free_ff: fuse_file_free(ff); -out_put_forget_req: - kfree(forget); out_err: return err; } @@ -782,15 +756,10 @@ static int create_new_entry(struct fuse_mount *fm, st= ruct fuse_args *args, struct inode *inode; struct dentry *d; int err; - struct fuse_forget_link *forget; =20 if (fuse_is_bad(dir)) return -EIO; =20 - forget =3D fuse_alloc_forget(); - if (!forget) - return -ENOMEM; - memset(&outarg, 0, sizeof(outarg)); args->nodeid =3D get_node_id(dir); args->out_numargs =3D 1; @@ -800,28 +769,27 @@ static int create_new_entry(struct fuse_mount *fm, st= ruct fuse_args *args, if (args->opcode !=3D FUSE_LINK) { err =3D get_create_ext(args, dir, entry, mode); if (err) - goto out_put_forget_req; + goto out_err; } =20 err =3D fuse_simple_request(fm, args); free_ext_value(args); if (err) - goto out_put_forget_req; + goto out_err; =20 err =3D -EIO; if (invalid_nodeid(outarg.nodeid) || fuse_invalid_attr(&outarg.attr)) - goto out_put_forget_req; + goto out_err; =20 if ((outarg.attr.mode ^ mode) & S_IFMT) - goto out_put_forget_req; + goto out_err; =20 inode =3D fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation, &outarg.attr, ATTR_TIMEOUT(&outarg), 0); if (!inode) { - fuse_queue_forget(fm->fc, forget, outarg.nodeid, 1); + fuse_queue_forget(fm->fc, NULL, outarg.nodeid, 1); return -ENOMEM; } - kfree(forget); =20 d_drop(entry); d =3D d_splice_alias(inode, entry); @@ -837,10 +805,9 @@ static int create_new_entry(struct fuse_mount *fm, str= uct fuse_args *args, fuse_dir_changed(dir); return 0; =20 - out_put_forget_req: + out_err: if (err =3D=3D -EEXIST) fuse_invalidate_entry(entry); - kfree(forget); return err; } =20 diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index f23919610313..0b9a02b3155f 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -298,7 +298,6 @@ struct fuse_args { uint8_t out_numargs; uint8_t ext_idx; bool force:1; - bool noreply:1; bool nocreds:1; bool in_pages:1; bool out_pages:1; @@ -1048,10 +1047,13 @@ int fuse_lookup_name(struct super_block *sb, u64 no= deid, const struct qstr *name /** * Send FORGET command */ -void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forg= et, - u64 nodeid, u64 nlookup); +static inline struct fuse_forget_link *fuse_alloc_forget(gfp_t flags) +{ + return kzalloc(sizeof(struct fuse_forget_link), flags); +} =20 -struct fuse_forget_link *fuse_alloc_forget(void); +void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link* forg= et, + u64 nodeid, u64 nlookup); =20 struct fuse_forget_link *fuse_dequeue_forget(struct fuse_iqueue *fiq, unsigned int max, diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 99e44ea7d875..ecc867e21acd 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -63,11 +63,6 @@ MODULE_PARM_DESC(max_user_congthresh, static struct file_system_type fuseblk_fs_type; #endif =20 -struct fuse_forget_link *fuse_alloc_forget(void) -{ - return kzalloc(sizeof(struct fuse_forget_link), GFP_KERNEL_ACCOUNT); -} - static struct fuse_submount_lookup *fuse_alloc_submount_lookup(void) { struct fuse_submount_lookup *sl; @@ -75,7 +70,7 @@ static struct fuse_submount_lookup *fuse_alloc_submount_l= ookup(void) sl =3D kzalloc(sizeof(struct fuse_submount_lookup), GFP_KERNEL_ACCOUNT); if (!sl) return NULL; - sl->forget =3D fuse_alloc_forget(); + sl->forget =3D fuse_alloc_forget(GFP_KERNEL_ACCOUNT); if (!sl->forget) goto out_free; =20 @@ -104,7 +99,7 @@ static struct inode *fuse_alloc_inode(struct super_block= *sb) fi->submount_lookup =3D NULL; mutex_init(&fi->mutex); spin_lock_init(&fi->lock); - fi->forget =3D fuse_alloc_forget(); + fi->forget =3D fuse_alloc_forget(GFP_KERNEL_ACCOUNT); if (!fi->forget) goto out_free; =20 diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c index 0377b6dc24c8..70d161f420c6 100644 --- a/fs/fuse/readdir.c +++ b/fs/fuse/readdir.c @@ -262,27 +262,6 @@ static int fuse_direntplus_link(struct file *file, return 0; } =20 -static void fuse_force_forget(struct file *file, u64 nodeid) -{ - struct inode *inode =3D file_inode(file); - struct fuse_mount *fm =3D get_fuse_mount(inode); - struct fuse_forget_in inarg; - FUSE_ARGS(args); - - memset(&inarg, 0, sizeof(inarg)); - inarg.nlookup =3D 1; - args.opcode =3D FUSE_FORGET; - args.nodeid =3D nodeid; - args.in_numargs =3D 1; - args.in_args[0].size =3D sizeof(inarg); - args.in_args[0].value =3D &inarg; - args.force =3D true; - args.noreply =3D true; - - fuse_simple_request(fm, &args); - /* ignore errors */ -} - static int parse_dirplusfile(char *buf, size_t nbytes, struct file *file, struct dir_context *ctx, u64 attr_version) { @@ -320,8 +299,13 @@ static int parse_dirplusfile(char *buf, size_t nbytes,= struct file *file, nbytes -=3D reclen; =20 ret =3D fuse_direntplus_link(file, direntplus, attr_version); - if (ret) - fuse_force_forget(file, direntplus->entry_out.nodeid); + if (ret) { + struct inode *inode =3D file_inode(file); + struct fuse_mount *fm =3D get_fuse_mount(inode); + + fuse_queue_forget(fm->fc, NULL, + direntplus->entry_out.nodeid, 1); + } } =20 return 0; --=20 2.33.0 From nobody Thu Sep 19 01:39:35 2024 Received: from szxga01-in.huawei.com (szxga01-in.huawei.com [45.249.212.187]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 7993B19DF68; Tue, 27 Aug 2024 09:36:02 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=45.249.212.187 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724751365; cv=none; b=ej+HnagEp8+yjKqMTBtXtUmE7gTq7sZThlWEloXxOekcuT52ObQOyQCi+2zSPy/5T9N3JlX5GTh9DQcGG7Ymn9PPul7B8fi7xb2meRTfdyUlcGRsFRQ27hPBuD0Zp4W/hMU48/Xl8Mgl2HyR+lsXlXxfPGKfCWs3Qb1Bk7nwl3k= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724751365; c=relaxed/simple; bh=UKUT/e0oqRfWxV+jsjzmtFBvnIGKU9JLgLs3/DXdaAw=; h=From:To:CC:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=V9+IgzVkgiaZ1K/5a+Yk6kPzTLc/A9QUf/5yIK6txTFZ94Z2irMvA3UvLn4buA0AN4ZbL3Cq1Q61JzuhP4Xkrg1UavGNLzprxNIryi4xWrQN805/s7Bf+8ifS5K67dAqPXTUWanORmGQ5Cls9kf8KDyxOqbVZrFEYxt3NS7CMVM= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=huawei.com; spf=pass smtp.mailfrom=huawei.com; arc=none smtp.client-ip=45.249.212.187 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=huawei.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=huawei.com Received: from mail.maildlp.com (unknown [172.19.88.105]) by szxga01-in.huawei.com (SkyGuard) with ESMTP id 4WtMpS65HGzyQYV; Tue, 27 Aug 2024 17:35:12 +0800 (CST) Received: from kwepemd100024.china.huawei.com (unknown [7.221.188.41]) by mail.maildlp.com (Postfix) with ESMTPS id E3F13140137; Tue, 27 Aug 2024 17:35:59 +0800 (CST) Received: from huawei.com (10.175.124.27) by kwepemd100024.china.huawei.com (7.221.188.41) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.1544.11; Tue, 27 Aug 2024 17:35:59 +0800 From: yangyun To: Miklos Szeredi CC: , , , Subject: [PATCH v3 2/2] fuse: add support for no forget requests Date: Tue, 27 Aug 2024 17:35:03 +0800 Message-ID: <20240827093503.3397562-3-yangyun50@huawei.com> X-Mailer: git-send-email 2.33.0 In-Reply-To: <20240827093503.3397562-1-yangyun50@huawei.com> References: <20240827093503.3397562-1-yangyun50@huawei.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-ClientProxiedBy: dggems706-chm.china.huawei.com (10.3.19.183) To kwepemd100024.china.huawei.com (7.221.188.41) Content-Type: text/plain; charset="utf-8" FUSE_FORGET requests are not used if the fuse file system does not implement the forget operation in userspace (e.g., fuse file system does not cache any inodes). However, the kernel is invisible to the userspace implementation and always sends FUSE_FORGET requests, which can lead to performance degradation because of useless contex switch and memory copy in some cases (e.g., many inodes are evicted from icache which was described in commit 07e77dca8a1f ("fuse: separate queue for FORGET requests")). Just like 'no_interrupt' in 'struct fuse_conn', we add 'no_forget'. But since FUSE_FORGET request does not have a reply from userspce, we can not use ENOSYS to reflect the 'no_forget' assignment. So add the FUSE_NO_FORGET_SUPPORT init flag. Besides, if no_forget is enabled, 'nlookup' in 'struct fuse_inode' does not used and its value change can be disabled which are protected by spin_lock to reduce lock contention. Signed-off-by: yangyun --- fs/fuse/dev.c | 3 +++ fs/fuse/dir.c | 4 +--- fs/fuse/fuse_i.h | 23 ++++++++++++++++++++ fs/fuse/inode.c | 46 +++++++++++++++++++++++---------------- fs/fuse/readdir.c | 8 ++----- include/uapi/linux/fuse.h | 3 +++ 6 files changed, 59 insertions(+), 28 deletions(-) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 31ca97b1b32c..7832f70baf65 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -238,6 +238,9 @@ void fuse_queue_forget(struct fuse_conn *fc, struct fus= e_forget_link *forget, { struct fuse_iqueue *fiq =3D &fc->iq; =20 + if (fc->no_forget) + return; + /* * The nullptr means that fuse_queue_forget() is used in error cases. * Avoid preallocating this structure because it is unlikely used. diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 583362492ce0..afc56c8ed7d6 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -236,9 +236,7 @@ static int fuse_dentry_revalidate(struct dentry *entry,= unsigned int flags) fuse_queue_forget(fm->fc, NULL, outarg.nodeid, 1); goto invalid; } - spin_lock(&fi->lock); - fi->nlookup++; - spin_unlock(&fi->lock); + fuse_inc_nlookup(fm->fc, fi); } if (ret =3D=3D -ENOMEM || ret =3D=3D -EINTR) goto out; diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 0b9a02b3155f..857357be5c8a 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -859,6 +859,9 @@ struct fuse_conn { /** Passthrough support for read/write IO */ unsigned int passthrough:1; =20 + /** Is forget not implemented by fs? */ + unsigned int no_forget:1; + /** Maximum stack depth for passthrough backing files */ int max_stack_depth; =20 @@ -1028,6 +1031,26 @@ static inline void fuse_sync_bucket_dec(struct fuse_= sync_bucket *bucket) rcu_read_unlock(); } =20 +static inline void fuse_inc_nlookup(struct fuse_conn *fc, struct fuse_inod= e *fi) +{ + if (fc->no_forget) + return; + + spin_lock(&fi->lock); + fi->nlookup++; + spin_lock(&fi->lock); +} + +static inline void fuse_dec_nlookup(struct fuse_conn *fc, struct fuse_inod= e *fi) +{ + if (fc->no_forget) + return; + + spin_lock(&fi->lock); + fi->nlookup--; + spin_lock(&fi->lock); +} + /** Device operations */ extern const struct file_operations fuse_dev_operations; =20 diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index ecc867e21acd..d2771ae515ea 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -84,6 +84,7 @@ static struct fuse_submount_lookup *fuse_alloc_submount_l= ookup(void) static struct inode *fuse_alloc_inode(struct super_block *sb) { struct fuse_inode *fi; + struct fuse_conn *fc =3D get_fuse_conn_super(sb); =20 fi =3D alloc_inode_sb(sb, fuse_inode_cachep, GFP_KERNEL); if (!fi) @@ -97,11 +98,14 @@ static struct inode *fuse_alloc_inode(struct super_bloc= k *sb) fi->orig_ino =3D 0; fi->state =3D 0; fi->submount_lookup =3D NULL; + fi->forget =3D NULL; mutex_init(&fi->mutex); spin_lock_init(&fi->lock); - fi->forget =3D fuse_alloc_forget(GFP_KERNEL_ACCOUNT); - if (!fi->forget) - goto out_free; + if (!fc->no_forget) { + fi->forget =3D fuse_alloc_forget(GFP_KERNEL_ACCOUNT); + if (!fi->forget) + goto out_free; + } =20 if (IS_ENABLED(CONFIG_FUSE_DAX) && !fuse_dax_inode_alloc(sb, fi)) goto out_free_forget; @@ -445,13 +449,15 @@ struct inode *fuse_iget(struct super_block *sb, u64 n= odeid, fuse_init_inode(inode, attr, fc); fi =3D get_fuse_inode(inode); fi->nodeid =3D nodeid; - fi->submount_lookup =3D fuse_alloc_submount_lookup(); - if (!fi->submount_lookup) { - iput(inode); - return NULL; + if (!fc->no_forget) { + fi->submount_lookup =3D fuse_alloc_submount_lookup(); + if (!fi->submount_lookup) { + iput(inode); + return NULL; + } + /* Sets nlookup =3D 1 on fi->submount_lookup->nlookup */ + fuse_init_submount_lookup(fi->submount_lookup, nodeid); } - /* Sets nlookup =3D 1 on fi->submount_lookup->nlookup */ - fuse_init_submount_lookup(fi->submount_lookup, nodeid); inode->i_flags |=3D S_AUTOMOUNT; goto done; } @@ -478,9 +484,7 @@ struct inode *fuse_iget(struct super_block *sb, u64 nod= eid, } } fi =3D get_fuse_inode(inode); - spin_lock(&fi->lock); - fi->nlookup++; - spin_unlock(&fi->lock); + fuse_inc_nlookup(fc, fi); done: fuse_change_attributes(inode, attr, NULL, attr_valid, attr_version); =20 @@ -1326,6 +1330,8 @@ static void process_init_reply(struct fuse_mount *fm,= struct fuse_args *args, } if (flags & FUSE_NO_EXPORT_SUPPORT) fm->sb->s_export_op =3D &fuse_export_fid_operations; + if (flags & FUSE_NO_FORGET_SUPPORT) + fc->no_forget =3D 1; } else { ra_pages =3D fc->max_read / PAGE_SIZE; fc->no_lock =3D 1; @@ -1373,7 +1379,7 @@ void fuse_send_init(struct fuse_mount *fm) FUSE_HANDLE_KILLPRIV_V2 | FUSE_SETXATTR_EXT | FUSE_INIT_EXT | FUSE_SECURITY_CTX | FUSE_CREATE_SUPP_GROUP | FUSE_HAS_EXPIRE_ONLY | FUSE_DIRECT_IO_ALLOW_MMAP | - FUSE_NO_EXPORT_SUPPORT | FUSE_HAS_RESEND; + FUSE_NO_EXPORT_SUPPORT | FUSE_HAS_RESEND | FUSE_NO_FORGET_SUPPORT; #ifdef CONFIG_FUSE_DAX if (fm->fc->dax) flags |=3D FUSE_MAP_ALIGNMENT; @@ -1588,7 +1594,7 @@ static int fuse_fill_super_submount(struct super_bloc= k *sb, * that, though, so undo it here. */ fi =3D get_fuse_inode(root); - fi->nlookup--; + fuse_dec_nlookup(fm->fc, fi); =20 sb->s_d_op =3D &fuse_dentry_operations; sb->s_root =3D d_make_root(root); @@ -1601,11 +1607,13 @@ static int fuse_fill_super_submount(struct super_bl= ock *sb, * prevent the last forget for this nodeid from getting * triggered until all users have finished with it. */ - sl =3D parent_fi->submount_lookup; - WARN_ON(!sl); - if (sl) { - refcount_inc(&sl->count); - fi->submount_lookup =3D sl; + if (!fm->fc->no_forget) { + sl =3D parent_fi->submount_lookup; + WARN_ON(!sl); + if (sl) { + refcount_inc(&sl->count); + fi->submount_lookup =3D sl; + } } =20 return 0; diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c index 70d161f420c6..4c8164d46409 100644 --- a/fs/fuse/readdir.c +++ b/fs/fuse/readdir.c @@ -218,9 +218,7 @@ static int fuse_direntplus_link(struct file *file, } =20 fi =3D get_fuse_inode(inode); - spin_lock(&fi->lock); - fi->nlookup++; - spin_unlock(&fi->lock); + fuse_inc_nlookup(fc, fi); =20 forget_all_cached_acls(inode); fuse_change_attributes(inode, &o->attr, NULL, @@ -247,9 +245,7 @@ static int fuse_direntplus_link(struct file *file, if (!IS_ERR(inode)) { struct fuse_inode *fi =3D get_fuse_inode(inode); =20 - spin_lock(&fi->lock); - fi->nlookup--; - spin_unlock(&fi->lock); + fuse_dec_nlookup(fc, fi); } return PTR_ERR(dentry); } diff --git a/include/uapi/linux/fuse.h b/include/uapi/linux/fuse.h index d08b99d60f6f..bf660880bc7a 100644 --- a/include/uapi/linux/fuse.h +++ b/include/uapi/linux/fuse.h @@ -217,6 +217,7 @@ * - add backing_id to fuse_open_out, add FOPEN_PASSTHROUGH open flag * - add FUSE_NO_EXPORT_SUPPORT init flag * - add FUSE_NOTIFY_RESEND, add FUSE_HAS_RESEND init flag + * - add FUSE_NO_FORGET_SUPPORT init flag */ =20 #ifndef _LINUX_FUSE_H @@ -421,6 +422,7 @@ struct fuse_file_lock { * FUSE_NO_EXPORT_SUPPORT: explicitly disable export support * FUSE_HAS_RESEND: kernel supports resending pending requests, and the hi= gh bit * of the request ID indicates resend requests + * FUSE_NO_FORGET_SUPPORT: disable forget requests */ #define FUSE_ASYNC_READ (1 << 0) #define FUSE_POSIX_LOCKS (1 << 1) @@ -463,6 +465,7 @@ struct fuse_file_lock { #define FUSE_PASSTHROUGH (1ULL << 37) #define FUSE_NO_EXPORT_SUPPORT (1ULL << 38) #define FUSE_HAS_RESEND (1ULL << 39) +#define FUSE_NO_FORGET_SUPPORT (1ULL << 40) =20 /* Obsolete alias for FUSE_DIRECT_IO_ALLOW_MMAP */ #define FUSE_DIRECT_IO_RELAX FUSE_DIRECT_IO_ALLOW_MMAP --=20 2.33.0