Re: file preallocation without unwritten flag being set

From: Dave Chinner <david@fromorbit.com>
To: p v <pvlogin@yahoo.com>
Cc: Eric Sandeen <sandeen@sandeen.net>, xfs@oss.sgi.com
Subject: Re: file preallocation without unwritten flag being set
Date: Thu, 14 May 2009 08:28:23 +1000	[thread overview]
Message-ID: <20090513222823.GL16929@discord.disaster> (raw)
In-Reply-To: <705795.15734.qm@web65604.mail.ac4.yahoo.com>

On Wed, May 13, 2009 at 02:05:16PM -0700, p v wrote:
> 
> doesn't seem to work - I tried to clear the extflg in the
> versionnum of the superblock (in every copy of it as well) but it
> doesn't work. The flag is still set on all extents.

Sure - that xfs_db command only clears it from the superblock so
that *new* preallocations don't have the unwritten bit set. it
doesn't change existing allocations.

> And once I make the file xfs_repair complains and resets the sb
> flag - my guess is that in the extent allocation path it is
> hardcoded for the version 4.

More likely is that repair is seeing an existing unwritten extent
and setting the flag on the superblock.

> - any extent allocated beyond file size will get the flag .

Allocation beyond EOF does not use unwritten extents unless
it is preallocation.

> Also - 2 questions -
> 
> 1) what is inode64 and where can I find out all of the
> undocumented mkfs/mount options (it's unfortunate that such a good
> fs doesnt' have a correspondingly good documentation)

All the options should be documented.  Try 'man mkfs.xfs', 'man 8
mount' and Documentation/filesystems/xfs.txt

> 2) why is the largest extent size limited to xxx blocks

2^21 blocks. Limited to that because there are 21 bits for
the extent size in the on disk extent record.

> (can't find
> out thenumber - when does the inode get finally flushed? ls -i
> reports 19 as the inode number but even after unmounting inode 19
> in xfs_db still shows a free inode - is it still only in the
> log???)

Might be, or you are seeing stale cached block device data
(xfs_db operates in a different address space to a mounted
filesystem). Try dropping the page cache and then re-read.

> ? I assumed that xfs_bmap gets me the correct number of
> extents but now looking at the inode with xfs_db it's obvious that
> xfs_bmap reports contiguous ranges rather than actual extents in
> the blockmap tree.

Sure it does. You can tell how many extents a specific range is from
their maximum size (e.g. one extent per 8GB for a 4k block size
filesystem).

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs