xfs
[Top] [All Lists]

Re: XFS setting custom extent size: real-time section only?

To: Heilige Gheist <hgheist@xxxxxxxxx>
Subject: Re: XFS setting custom extent size: real-time section only?
From: Nathan Scott <nathans@xxxxxxx>
Date: Mon, 24 Jul 2006 10:00:30 +1000
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20060723110919.79373.qmail@xxxxxxxxxxxxxxxxxxxxxxxxxxx>; from hgheist@xxxxxxxxx on Sun, Jul 23, 2006 at 04:09:18AM -0700
References: <20060723110919.79373.qmail@xxxxxxxxxxxxxxxxxxxxxxxxxxx>
Sender: xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.2.5i
On Sun, Jul 23, 2006 at 04:09:18AM -0700, Heilige Gheist wrote:
> I'm trying to get custom extent size per file working, but
> XFS_IOC_FSSETXATTR ioctl returns with EINVAL,
> both from my code or from xfs_io.
> The environment is SLES9 SP2, 2.6.5-7.252 kernel, xfsprogs-2.6.25-0.6
> After examining the kernel code in xfs_vnodeops.c we found out that
> custom extent size is supported on files in real-time section only.
> This is not mentioned anywhere else in the documentation, so I'm
> wondering whether this limitation is removed in further kernel
> versions.
> I've downloaded the recent xfsprogs-2.8.4, it uses a new flag
> XFS_XFLAG_EXTSIZE when setting the custom extent size.
> I'm not sure what kernel version do I need to work with 2.8.4, but it's
> certainly a sign that kernel support has changed.

Indeed it has.  I'm not sure which kernel version this was merged into,
but the rough timeframe can be seen on http://oss.sgi.com/projects/xfs/
on the "recent changes" table.  It wont be there in any SLES9, but IIRC
SLES10 does now include that code.

cheers.

-- 
Nathan


<Prev in Thread] Current Thread [Next in Thread>