From: "Jose R. Santos" <jrs@us.ibm.com>
Subject: Re: [PATCH 4/4][e2fsprogs] New bitmap and inode table allocation
 for FLEX_BG
Date: Mon, 5 Nov 2007 09:09:46 -0600
Message-ID: <20071105090946.7668a20e@gara>
References: <20070814043245.32206.34785.stgit@gara>
	<20070814043314.32206.28698.stgit@gara>
	<20071104005208.GA6455@thunk.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Cc: linux-ext4@vger.kernel.org
To: Theodore Tso <tytso@mit.edu>
In-Reply-To: <20071104005208.GA6455@thunk.org>
Sender: linux-ext4-owner@vger.kernel.org

On Sat, 3 Nov 2007 20:52:08 -0400
Theodore Tso <tytso@mit.edu> wrote:

> On Mon, Aug 13, 2007 at 11:33:14PM -0500, Jose R. Santos wrote:
> > +	if (EXT2_HAS_INCOMPAT_FEATURE (fs->super, 
> > +				       EXT4_FEATURE_INCOMPAT_FLEX_BG)) 
> > +		ext2fs_allocate_flex_groups(fs);
> > +	
> > +	else {
> > +		for (i = 0; i < fs->group_desc_count; i++) {
> > +			retval = ext2fs_allocate_group_table(fs, i, fs->block_map);
> > +			if (retval)
> > +				return retval;
> > +		}
> 
> The code to handle flex groups needs to be moved into
> ext2fs_allocate_group_table(), and not put in
> ext2fs_allocate_tables(), since resize2fs calls
> ext2fs_allocate_group_table(), and we want resize2fs to do the right
> thing for filesystems that have both the FLEX_BG and META_BG flags
> set.  Speaking of which, we need to fix the kernel on-line resizing
> code to allocate new blocks for filesystem metadata blocks using the
> new placement algorithm for FLEX_BG && META_BG filesystems.
> 
> I'll fix this up for the 'pu' branch of e2fsprogs, as well as making
> ext2fs_allocate_flex_groups static, but I'm not going to graduate this
> to the 'next' branch just yet, since it needs a bit more testing,
> specifically with resize2fs.

Yes, this code is not ready for 'next' branch.  This was mostly
intended as RFC and testing.  Resizing is not handle at all in this
code and fsck needs more testing.

> 						- Ted


-JRS