Subject: Re: kern/21296: more file corruption since UFS2
To: enami tsugutomo <enami@sm.sony.co.jp>
From: Frank van der Linden <fvdl@netbsd.org>
List: netbsd-bugs
Date: 04/24/2003 13:34:59
On Thu, Apr 24, 2003 at 07:57:11PM +0900, enami tsugutomo wrote:
> Gregory McGarry <g.mcgarry@ieee.org> writes:
> 
> > I think it is because ffs1_din is not being cleared during allocation.
> > Before ufs2 went in, memset() on the inode would have also cleared the
> > dinode structure.
> 
> For ffs, it is a job of newfs isn't it?

Correct. newfs(8) initializes all inode blocks to 0; FFS just loads
the inode from disk (ffs_vget doesn't know if it's gettting an
allocated inode or not).

If you're using an UFS2 filesystem, the inode blocks are initialized
on the fly. But the bug report doesn't seem to be about a UFS2 filesystem.

- Frank