Re: btrfs_inode_item's otime?

[Martin Steigerwald <Martin@lichtvoll.de>]

Am Mittwoch, 7. Januar 2015, 06:42:50 schrieb Christoph Hellwig:
> On Wed, Jan 07, 2015 at 02:57:35PM +0100, Lennart Poettering wrote:
> > Exposig this as xattr sounds great to me too.
> 
> NAK - exposing random stat data as xattr only creates problems.
> 
> Given that we don't seem to be able to get a new stat format anytime
> soon we should add a generic ioctl to expose it, reading it from struct
> kstat which all filesystem that support this attribute should fill out.
> And there's quite a lot of them.

What is the issue with the xstat, new stat format stuff? Basically I have the 
feeling that this is in discussion for at least 5 years or so and I always 
wondered where it is stuck. I am not aware of any discussions of it recently, 
but I may have overseen those.

Ext4, I think, already supports birth time, but cannot expose it either. And 
even the stat command already seems to have some support for it, except that 
it doesn´t work yet (this is on BTRFS, but on Ext4 its the same).

merkaba:~> LANG=C stat /bin/ls
  File: '/bin/ls'
  Size: 118280          Blocks: 232        IO Block: 4096   regular file
Device: 14h/20d Inode: 3574391     Links: 1
Access: (0755/-rwxr-xr-x)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2014-10-31 11:29:37.000000000 +0100
Modify: 2014-10-30 03:43:06.000000000 +0100
Change: 2014-10-31 11:29:41.440773090 +0100
 Birth: -

Any pointers?

-- 
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA  B82F 991B EAAC A599 84C7