svn commit: r193041 - head/sys/sys

[Scott Long]

Occasionally it's useful to be able to run older binaries on newer 
kernels.  One particularly useful place is for making releases.  Does
this change to struct mount break that?  If so, could there be any other
way to achieve what you want without the breakage?

Scott


Edward Tomasz Napierala wrote:
> Author: trasz
> Date: Fri May 29 15:00:04 2009
> New Revision: 193041
> URL: http://svn.freebsd.org/changeset/base/193041
> 
> Log:
>   There is only one spare MNT_ flag left, and I want to use it for NFSv4 ACLs.
>   Make room for additional filesystem flags now, to avoid breaking ABI later.
>   
>   Reviewed by:	kib@
> 
> Modified:
>   head/sys/sys/mount.h
> 
> Modified: head/sys/sys/mount.h
> ==============================================================================
> --- head/sys/sys/mount.h	Fri May 29 14:42:02 2009	(r193040)
> +++ head/sys/sys/mount.h	Fri May 29 15:00:04 2009	(r193041)
> @@ -167,6 +167,7 @@ struct mount {
>  	int		mnt_writeopcount;	/* (i) write syscalls pending */
>  	int		mnt_kern_flag;		/* (i) kernel only flags */
>  	u_int		mnt_flag;		/* (i) flags shared with user */
> +	u_int		mnt_xflag;		/* (i) more flags shared with user */
>  	u_int		mnt_noasync;		/* (i) # noasync overrides */
>  	struct vfsoptlist *mnt_opt;		/* current mount options */
>  	struct vfsoptlist *mnt_optnew;		/* new options passed to fs */
> @@ -221,7 +222,7 @@ void          __mnt_vnode_markerfree(str
>  #endif /* _KERNEL */
>  
>  /*
> - * User specifiable flags.
> + * User specifiable flags, stored in mnt_flag.
>   */
>  #define	MNT_RDONLY	0x00000001	/* read only filesystem */
>  #define	MNT_SYNCHRONOUS	0x00000002	/* filesystem written synchronously */