Re: [RFC/PATCH] ext3: remove inode constructor

2007-05-05 Thread Andrew Morton 
On Sat, 5 May 2007 11:58:45 +0300 (EEST) Pekka J Enberg [EMAIL PROTECTED] 
wrote:

 On Fri, 4 May 2007, Andrew Morton wrote:
  I got 100% rejects against this because Christoph has already had
  his paws all over the slab constructor code everywhere.
  
  Was going to fix it up but then decided that we ought to make changes
  like this to ext4 as well.  Ideally beforehand, but simultaneously is
  OK as long as it's simple enough.
 
 I'll send you proper patches for them (and will convert other filesystems 
 too).

May as well.

 On Fri, 4 May 2007, Andrew Morton wrote:
  btw, for a benchmark I'd suggest just a silly create-1-files
  tight loop rather than something more complex like postmark.
 
 Do you want me to redo the benchmarks or are you happy enough with the 
 postmark numbers?

I doubt if this is measurable, really.  It'll be something like the
difference between an L1 hit and an L2 hit in amongst all the other stuff
we do on a per-inode basis.

-
To unsubscribe from this list: send the line unsubscribe linux-ext4 in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] ext3: remove inode constructor

2007-05-05 Thread Pekka J Enberg 
On Fri, 4 May 2007, Christoph Lameter wrote:
 Also could this be generalized to also apply to the generic inode 
 allocation in fs/inode.c?

I think we want to stick inode_init_ince() in alloc_inode() but we can't 
do that until all filesystems are converted over.

Pekka
-
To unsubscribe from this list: send the line unsubscribe linux-ext4 in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] ext3: remove inode constructor

2007-05-04 Thread Christoph Lameter 
On Fri, 4 May 2007, Pekka J Enberg wrote:

 Index: 2.6/fs/ext3/super.c
 ===
 --- 2.6.orig/fs/ext3/super.c  2007-05-04 12:57:09.0 +0300
 +++ 2.6/fs/ext3/super.c   2007-05-04 13:01:27.0 +0300
 @@ -444,17 +444,26 @@ static struct kmem_cache *ext3_inode_cac
  static struct inode *ext3_alloc_inode(struct super_block *sb)
  {
   struct ext3_inode_info *ei;
 + struct inode *inode;
  
   ei = kmem_cache_alloc(ext3_inode_cachep, GFP_NOFS);
   if (!ei)
   return NULL;
 +INIT_LIST_HEAD(ei-i_orphan);
 +#ifdef CONFIG_EXT3_FS_XATTR
 +init_rwsem(ei-xattr_sem);
 +#endif
 +mutex_init(ei-truncate_mutex);

^^^ whitespace issues

Also could this be generalized to also apply to the generic inode 
allocation in fs/inode.c?

-
To unsubscribe from this list: send the line unsubscribe linux-ext4 in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] ext3: remove inode constructor

2007-05-04 Thread Andrew Morton 
On Fri, 4 May 2007 13:14:35 +0300 (EEST)
Pekka J Enberg [EMAIL PROTECTED] wrote:

 As explained by Christoph Lameter, ext3_alloc_inode() touches the same
 cache line as init_once() so we gain nothing from using slab
 constructors.  The SLUB allocator will be more effective without it
 (free pointer can be placed inside the free'd object), so move inode
 initialization to ext3_alloc_inode completely.

I got 100% rejects against this because Christoph has already had
his paws all over the slab constructor code everywhere.

Was going to fix it up but then decided that we ought to make changes
like this to ext4 as well.  Ideally beforehand, but simultaneously is
OK as long as it's simple enough.

btw, for a benchmark I'd suggest just a silly create-1-files
tight loop rather than something more complex like postmark.
-
To unsubscribe from this list: send the line unsubscribe linux-ext4 in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html