linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@redhat.com>
To: Eric Sandeen <sandeen@redhat.com>
Cc: "Jörn Engel" <joern@wohnheim.fh-wedel.de>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] make last_inode counter in new_inode 32-bit on kernels that offer x86 compatability
Date: Tue, 07 Nov 2006 10:56:06 -0500	[thread overview]
Message-ID: <1162914966.28425.24.camel@dantu.rdu.redhat.com> (raw)
In-Reply-To: <454FAAF8.8080707@redhat.com>

On Mon, 2006-11-06 at 15:36 -0600, Eric Sandeen wrote:
> *shrug* I dunno, it's probably not worth arguing this point, it needs to
> be fixed properly in any case. :)
> 
> -Eric

Here's a more involved (again, untested) patch. I turned new_inode into
a wrapper for a new function (new_inode_autonum). If the autonum arg is
set, then we should get back a unique i_ino.

This also converts the iunique function to use the per sb s_lastino
counter. I think that's the right thing to do here, but I'm new to the
hlist stuff and so feedback would be appreciated there.

Current behavior for new_inode shouldn't be changed (aside from the
per-sb s_lastino counter). The idea here would be to move filesystems
that don't later overwrite the i_ino field, to use new_inode_autonum
rather than new_inode and set autonum to true.

Not sure how bad performance with new inode creation will be when that
is set to true, but this is the best way I could see to do this without
adding something like an radix tree to track used inodes.

Signed-off-by: Jeff Layton <jlayton@redhat.com>

diff --git a/fs/inode.c b/fs/inode.c
index 26cdb11..83b9bab 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -517,14 +517,14 @@ repeat:
 }
 
 /**
- *	new_inode 	- obtain an inode
+ *	new_inode_autonum 	- obtain an inode
  *	@sb: superblock
+ *      @autonum: if true, make sure that i_ino is unique
  *
  *	Allocates a new inode for given superblock.
  */
-struct inode *new_inode(struct super_block *sb)
+struct inode *new_inode_autonum(struct super_block *sb, int autonum)
 {
-	static unsigned long last_ino;
 	struct inode * inode;
 
 	spin_lock_prefetch(&inode_lock);
@@ -535,13 +535,26 @@ struct inode *new_inode(struct super_blo
 		inodes_stat.nr_inodes++;
 		list_add(&inode->i_list, &inode_in_use);
 		list_add(&inode->i_sb_list, &sb->s_inodes);
-		inode->i_ino = ++last_ino;
+		if (autonum)
+			inode->i_ino = iunique(sb, 0);
+		else
+			inode->i_ino = ++sb->s_lastino;
 		inode->i_state = 0;
 		spin_unlock(&inode_lock);
 	}
 	return inode;
 }
 
+/**
+ *	new_inode 	- obtain an inode
+ *	@sb: superblock
+ *
+ *	Allocates a new inode for given superblock.
+ */
+struct inode *new_inode(struct super_block *sb)
+{
+	new_inode_autonum(sb, 0);
+}
 EXPORT_SYMBOL(new_inode);
 
 void unlock_new_inode(struct inode *inode)
@@ -683,22 +696,21 @@ static unsigned long hash(struct super_b
  */
 ino_t iunique(struct super_block *sb, ino_t max_reserved)
 {
-	static ino_t counter;
 	struct inode *inode;
 	struct hlist_head * head;
 	ino_t res;
 	spin_lock(&inode_lock);
 retry:
-	if (counter > max_reserved) {
-		head = inode_hashtable + hash(sb,counter);
-		res = counter++;
+	if (sb->s_lastino >= max_reserved) {
+		head = inode_hashtable + hash(sb,++sb->s_lastino);
+		res = sb->s_lastino;
 		inode = find_inode_fast(sb, head, res);
 		if (!inode) {
 			spin_unlock(&inode_lock);
 			return res;
 		}
 	} else {
-		counter = max_reserved + 1;
+		 sb->s_lastino = max_reserved;
 	}
 	goto retry;
 	
diff --git a/fs/super.c b/fs/super.c
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 2fe6e3f..1010e64 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -961,6 +961,14 @@ #endif
 	/* Granularity of c/m/atime in ns.
 	   Cannot be worse than a second */
 	u32		   s_time_gran;
+
+	/* per-sb inode counter for new_inode. Make it a 32-bit counter when
+	   we have the possibility of dealing with 32-bit apps */
+#ifdef CONFIG_COMPAT
+	unsigned int		s_lastino;
+#else
+	unsigned long		s_lastino;
+#endif
 };
 
 extern struct timespec current_fs_time(struct super_block *sb);



  parent reply	other threads:[~2006-11-07 15:56 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-11-06 18:12 [PATCH] make last_inode counter in new_inode 32-bit on kernels that offer x86 compatability Jeff Layton
2006-11-06 18:22 ` Matthew Wilcox
2006-11-06 18:31   ` Jeff Layton
2006-11-06 18:47   ` Jeff Layton
2006-11-06 20:23     ` Jörn Engel
2006-11-06 20:31       ` Eric Dumazet
2006-11-06 20:56         ` Jeff Layton
2006-11-06 20:50       ` Eric Sandeen
2006-11-06 21:11         ` Jörn Engel
2006-11-06 21:36           ` Eric Sandeen
2006-11-06 22:01             ` Andreas Dilger
2006-11-07 15:56             ` Jeff Layton [this message]
2006-11-07 16:07               ` Jeff Layton
2006-11-07 16:10               ` Matthew Wilcox
2006-11-07 17:04               ` Jörn Engel
2006-11-07 17:28               ` Jörn Engel
2006-11-07 17:42                 ` Jörn Engel
2006-11-07 17:53                   ` Dave Kleikamp
2006-11-07 18:07                     ` Jörn Engel
2006-11-07 17:56                   ` Jörn Engel
2006-11-07 18:01                     ` Jörn Engel
2006-11-07 18:10                       ` Eric Sandeen
2006-11-07 19:41                     ` Jeff Layton
2006-11-07 20:41                       ` Jörn Engel
2006-11-07 21:13                         ` Jeff Layton
2006-11-07 21:20                           ` Matthew Wilcox
2006-11-07 22:09                             ` Jeff Layton
2006-11-07 18:01                 ` Jeff Layton
2006-11-07 18:14                   ` Jörn Engel
2006-11-07 18:23                     ` Jeff Layton

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=1162914966.28425.24.camel@dantu.rdu.redhat.com \
    --to=jlayton@redhat.com \
    --cc=joern@wohnheim.fh-wedel.de \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sandeen@redhat.com \
    /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).