linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Al Viro <viro@zeniv.linux.org.uk>
To: Andreas Dilger <adilger@dilger.ca>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>,
	Linux List Kernel Mailing <linux-kernel@vger.kernel.org>
Subject: Re: [RFC][PATCHSET] sorting out RCU-delayed stuff in ->destroy_inode()
Date: Tue, 30 Apr 2019 05:26:24 +0100	[thread overview]
Message-ID: <20190430042623.GJ23075@ZenIV.linux.org.uk> (raw)
In-Reply-To: <F01D238D-8A6C-4629-ABC5-4A8BAC25951F@dilger.ca>

On Mon, Apr 29, 2019 at 10:18:04PM -0600, Andreas Dilger wrote:
> > 
> > 	void			*i_private; /* fs or device private pointer */
> > +	void (*free_inode)(struct inode *);
> 
> It seems like a waste to increase the size of every struct inode just to access
> a static pointer.  Is this the only place that ->free_inode() is called?  Why
> not move the ->free_inode() pointer into inode->i_fop->free_inode() so that it
> is still directly accessible at this point.

i_op, surely?  In any case, increasing sizeof(struct inode) is not a problem -
if anything, I'd turn ->i_fop into an anon union with that.  As in,

diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting
index 9d80f9e0855e..b8d3ddd8b8db 100644
--- a/Documentation/filesystems/porting
+++ b/Documentation/filesystems/porting
@@ -655,3 +655,11 @@ in your dentry operations instead.
 		* if ->free_inode() is non-NULL, it gets scheduled by call_rcu()
 		* combination of NULL ->destroy_inode and NULL ->free_inode is
 		  treated as NULL/free_inode_nonrcu, to preserve the compatibility.
+
+	Note that the callback (be it via ->free_inode() or explicit call_rcu()
+	in ->destroy_inode()) is *NOT* ordered wrt superblock destruction;
+	as the matter of fact, the superblock and all associated structures
+	might be already gone.  The filesystem driver is guaranteed to be still
+	there, but that's it.  Freeing memory in the callback is fine; doing
+	more than that is possible, but requires a lot of care and is best
+	avoided.
diff --git a/fs/inode.c b/fs/inode.c
index fb45590d284e..627e1766503a 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -211,8 +211,8 @@ EXPORT_SYMBOL(free_inode_nonrcu);
 static void i_callback(struct rcu_head *head)
 {
 	struct inode *inode = container_of(head, struct inode, i_rcu);
-	if (inode->i_sb->s_op->free_inode)
-		inode->i_sb->s_op->free_inode(inode);
+	if (inode->free_inode)
+		inode->free_inode(inode);
 	else
 		free_inode_nonrcu(inode);
 }
@@ -236,6 +236,7 @@ static struct inode *alloc_inode(struct super_block *sb)
 			if (!ops->free_inode)
 				return NULL;
 		}
+		inode->free_inode = ops->free_inode;
 		i_callback(&inode->i_rcu);
 		return NULL;
 	}
@@ -276,6 +277,7 @@ static void destroy_inode(struct inode *inode)
 		if (!ops->free_inode)
 			return;
 	}
+	inode->free_inode = ops->free_inode;
 	call_rcu(&inode->i_rcu, i_callback);
 }
 
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 2e9b9f87caca..92732286b748 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -694,7 +694,10 @@ struct inode {
 #ifdef CONFIG_IMA
 	atomic_t		i_readcount; /* struct files open RO */
 #endif
-	const struct file_operations	*i_fop;	/* former ->i_op->default_file_ops */
+	union {
+		const struct file_operations	*i_fop;	/* former ->i_op->default_file_ops */
+		void (*free_inode)(struct inode *);
+	};
 	struct file_lock_context	*i_flctx;
 	struct address_space	i_data;
 	struct list_head	i_devices;

  reply	other threads:[~2019-04-30  4:26 UTC|newest]

Thread overview: 82+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-16 17:49 [RFC][PATCHSET] sorting out RCU-delayed stuff in ->destroy_inode() Al Viro
2019-04-16 17:52 ` [RFC PATCH 01/62] securityfs: fix use-after-free on symlink traversal Al Viro
2019-04-16 17:52   ` [RFC PATCH 02/62] apparmorfs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 03/62] new inode method: ->free_inode() Al Viro
2019-04-16 17:52   ` [RFC PATCH 04/62] spufs: switch to ->free_inode() Al Viro
2019-04-16 17:52   ` [RFC PATCH 05/62] erofs: " Al Viro
2019-04-18 14:01     ` Gao Xiang
2019-04-16 17:52   ` [RFC PATCH 06/62] 9p: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 07/62] adfs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 08/62] affs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 09/62] befs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 10/62] bfs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 11/62] bdev: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 12/62] cifs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 13/62] debugfs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 14/62] efs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 15/62] ext2: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 16/62] f2fs: " Al Viro
2019-04-20  2:52     ` Chao Yu
2019-04-16 17:52   ` [RFC PATCH 17/62] fat: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 18/62] freevxfs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 19/62] gfs2: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 20/62] hfs: " Al Viro
2019-04-16 17:52   ` [RFC PATCH 21/62] hfsplus: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 22/62] hostfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 23/62] hpfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 24/62] isofs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 25/62] jffs2: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 26/62] minix: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 27/62] nfs{,4}: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 28/62] nilfs2: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 29/62] dlmfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 30/62] ocfs2: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 31/62] openpromfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 32/62] procfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 33/62] qnx4: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 34/62] qnx6: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 35/62] reiserfs: convert " Al Viro
2019-04-16 17:53   ` [RFC PATCH 36/62] romfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 37/62] squashfs: switch " Al Viro
2019-04-16 17:53   ` [RFC PATCH 38/62] ubifs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 39/62] udf: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 40/62] sysv: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 41/62] coda: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 42/62] ufs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 43/62] mqueue: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 44/62] bpf: " Al Viro
2019-04-16 18:07     ` Alexei Starovoitov
2019-04-16 21:34       ` Song Liu
2019-04-16 17:53   ` [RFC PATCH 45/62] rpcpipe: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 46/62] apparmor: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 47/62] securityfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 48/62] ntfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 49/62] dax: make use of ->free_inode() Al Viro
2019-04-18 12:16     ` Jan Kara
2019-04-18 16:58       ` Dan Williams
2019-04-16 17:53   ` [RFC PATCH 50/62] afs: switch to " Al Viro
2019-04-16 17:53   ` [RFC PATCH 51/62] btrfs: use ->free_inode() Al Viro
2019-04-16 17:53   ` [RFC PATCH 52/62] ceph: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 53/62] ecryptfs: make use of ->free_inode() Al Viro
2019-04-16 17:53   ` [RFC PATCH 54/62] ext4: " Al Viro
2019-04-18 12:10     ` Jan Kara
2019-04-16 17:53   ` [RFC PATCH 55/62] fuse: switch to ->free_inode() Al Viro
2019-04-16 17:53   ` [RFC PATCH 56/62] jfs: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 57/62] overlayfs: make use of ->free_inode() Al Viro
2019-04-16 17:53   ` [RFC PATCH 58/62] hugetlb: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 59/62] shmem: " Al Viro
2019-04-16 17:53   ` [RFC PATCH 60/62] orangefs: " Al Viro
2019-04-22 21:14     ` Mike Marshall
2019-04-22 21:56       ` Linus Torvalds
2019-04-22 23:10         ` Al Viro
2019-04-22 23:17           ` Mike Marshall
2019-04-16 17:53   ` [RFC PATCH 61/62] sockfs: switch to ->free_inode() Al Viro
2019-04-16 17:53   ` [RFC PATCH 62/62] coallocate socket->wq with socket itself Al Viro
2019-04-16 18:01 ` [RFC][PATCHSET] sorting out RCU-delayed stuff in ->destroy_inode() Linus Torvalds
2019-04-30  3:09   ` Al Viro
     [not found]     ` <CAHk-=wiMvCR0iENUVorfU-3EMC7G8RNSeHSQrz9tndP1uSg2BQ@mail.gmail.com>
2019-04-30  4:00       ` Al Viro
2019-05-01  1:59         ` Al Viro
2019-04-30  4:18     ` Andreas Dilger
2019-04-30  4:26       ` Al Viro [this message]
2019-04-30  5:26         ` Andreas Dilger
2019-04-17 15:55 ` David Sterba

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190430042623.GJ23075@ZenIV.linux.org.uk \
    --to=viro@zeniv.linux.org.uk \
    --cc=adilger@dilger.ca \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).