[2.6.26 patch] fuse: fix thinko in max I/O size calucation

[2.6.26 patch] fuse: fix thinko in max I/O size calucation

[Miklos Szeredi]

From: Miklos Szeredi <mszeredi@suse.cz>

Use max not min to enforce a lower limit on the max I/O size.

This bug was introduced by

   commit e5d9a0df07484d6d191756878c974e4307fb24ce
   Author: Miklos Szeredi <mszeredi@suse.cz>
   Date:   Wed Apr 30 00:54:44 2008 -0700
   
       fuse: fix max i/o size calculation

Thanks to Brian Wang for noticing.

Reported-by: Brian Wang <ywang221@hotmail.com>
CC: Szabolcs Szakacsits <szaka@ntfs-3g.org>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
---
 fs/fuse/inode.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Index: linux-2.6/fs/fuse/inode.c
===================================================================
--- linux-2.6.orig/fs/fuse/inode.c	2008-06-17 17:55:38.000000000 +0200
+++ linux-2.6/fs/fuse/inode.c	2008-06-17 17:55:48.000000000 +0200
@@ -591,7 +591,7 @@ static void process_init_reply(struct fu
 		fc->bdi.ra_pages = min(fc->bdi.ra_pages, ra_pages);
 		fc->minor = arg->minor;
 		fc->max_write = arg->minor < 5 ? 4096 : arg->max_write;
-		fc->max_write = min_t(unsigned, 4096, fc->max_write);
+		fc->max_write = max_t(unsigned, 4096, fc->max_write);
 		fc->conn_init = 1;
 	}
 	fuse_put_request(fc, req);
@@ -667,7 +667,7 @@ static int fuse_fill_super(struct super_
 	fc->flags = d.flags;
 	fc->user_id = d.user_id;
 	fc->group_id = d.group_id;
-	fc->max_read = min_t(unsigned, 4096, d.max_read);
+	fc->max_read = max_t(unsigned, 4096, d.max_read);
 
 	/* Used by get_root_inode() */
 	sb->s_fs_info = fc;

Re: [2.6.26 patch] fuse: fix thinko in max I/O size calucation

[Szabolcs Szakacsits]

On Tue, 17 Jun 2008, Miklos Szeredi wrote:

> From: Miklos Szeredi <mszeredi@suse.cz>
> 
> Use max not min to enforce a lower limit on the max I/O size.

Big write is ok now, thanks. NTFS-3G in-memory write performance seems to 
be 6x better on a T9300 and quick tests seem to show that it's at least as 
fast as ext3 in-memory writes (but this needs more testing, etc).

However the max write size seems to be limited to 128kB. Is this 
intentional? If yes then

Acked-by: Szabolcs Szakacsits <szaka@ntfs-3g.org>

	Szaka

> This bug was introduced by
> 
>    commit e5d9a0df07484d6d191756878c974e4307fb24ce
>    Author: Miklos Szeredi <mszeredi@suse.cz>
>    Date:   Wed Apr 30 00:54:44 2008 -0700
>    
>        fuse: fix max i/o size calculation
> 
> Thanks to Brian Wang for noticing.
> 
> Reported-by: Brian Wang <ywang221@hotmail.com>
> CC: Szabolcs Szakacsits <szaka@ntfs-3g.org>
> Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
> ---
>  fs/fuse/inode.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> Index: linux-2.6/fs/fuse/inode.c
> ===================================================================
> --- linux-2.6.orig/fs/fuse/inode.c	2008-06-17 17:55:38.000000000 +0200
> +++ linux-2.6/fs/fuse/inode.c	2008-06-17 17:55:48.000000000 +0200
> @@ -591,7 +591,7 @@ static void process_init_reply(struct fu
>  		fc->bdi.ra_pages = min(fc->bdi.ra_pages, ra_pages);
>  		fc->minor = arg->minor;
>  		fc->max_write = arg->minor < 5 ? 4096 : arg->max_write;
> -		fc->max_write = min_t(unsigned, 4096, fc->max_write);
> +		fc->max_write = max_t(unsigned, 4096, fc->max_write);
>  		fc->conn_init = 1;
>  	}
>  	fuse_put_request(fc, req);
> @@ -667,7 +667,7 @@ static int fuse_fill_super(struct super_
>  	fc->flags = d.flags;
>  	fc->user_id = d.user_id;
>  	fc->group_id = d.group_id;
> -	fc->max_read = min_t(unsigned, 4096, d.max_read);
> +	fc->max_read = max_t(unsigned, 4096, d.max_read);
>  
>  	/* Used by get_root_inode() */
>  	sb->s_fs_info = fc;
> 

--
NTFS-3G:  http://ntfs-3g.org

Re: [2.6.26 patch] fuse: fix thinko in max I/O size calucation

[Miklos Szeredi]

> > From: Miklos Szeredi <mszeredi@suse.cz>
> > 
> > Use max not min to enforce a lower limit on the max I/O size.
> 
> Big write is ok now, thanks. NTFS-3G in-memory write performance seems to 
> be 6x better on a T9300 and quick tests seem to show that it's at least as 
> fast as ext3 in-memory writes (but this needs more testing, etc).
> 
> However the max write size seems to be limited to 128kB. Is this 
> intentional?

Yes, that's due to another limit (max 32 pages per request).

Miklos