Why do some ufs i-node fields have 2 copies?

[Rick C. Petty]

On Sun, Apr 26, 2009 at 04:56:03PM -0400, Rick Macklem wrote:
> 
> I was just wondering if anyone conversant with ufs/ffs could tell me why
> the following fields of the i-node have off-disk and on-disk copies?
> (One thought I had was that these fields are sometimes set to values
>  that shouldn't get saved on-disk, but it was just a hunch.)

> 	/*
> 	 * Copies from the on-disk dinode itself.
> 	 */
> 	u_int16_t i_mode;	/* IFMT, permissions; see below. */
> 	int16_t	  i_nlink;	/* File link count. */
> 	u_int64_t i_size;	/* File byte count. */
> 	u_int32_t i_flags;	/* Status flags (chflags). */
> 	int64_t	  i_gen;	/* Generation number. */
> 	u_int32_t i_uid;	/* File owner. */
> 	u_int32_t i_gid;	/* File group. */
> 	/*
> 	 * The real copy of the on-disk inode.

You missed a few lines:

       union {
                struct ufs1_dinode *din1;       /* UFS1 on-disk dinode. */
                struct ufs2_dinode *din2;       /* UFS2 on-disk dinode. */
        } dinode_u;

The reason is that the first set (the "copies") are what are referenced by
the ffs/ufs code.  The real copies are what came from the disk (hence the
"on-disk").  Because UFS1 & UFS2 have such different dinode structures, the
copies are moved to/from the dinode structures when the inode is read to or
written from the disk.  At all other times, the "copies" are used.

-- Rick C. Petty