On Wed, 25 May 2005 02:22:18 +0200
Juan RP <juan@xtrarom.org> wrote:

> Shouldn't vfs.lfs.stats.segs_reclaimed be 0 in this case?
> 
> Index: lfs_syscalls.c
> ===================================================================
> RCS file: /cvsroot/src/sys/ufs/lfs/lfs_syscalls.c,v
> retrieving revision 1.106
> diff -u -r1.106 lfs_syscalls.c
> --- lfs_syscalls.c      20 May 2005 19:48:25 -0000      1.106
> +++ lfs_syscalls.c      25 May 2005 00:15:22 -0000
> @@ -877,6 +877,7 @@
>  int
>  lfs_do_segclean(struct lfs *fs, unsigned long segnum)
>  {
> +      extern int lfs_dostats;
>         struct buf *bp;
>         CLEANERINFO *cip;
>         SEGUSE *sup;
> @@ -933,7 +934,8 @@
>         (void) LFS_BWRITE_LOG(bp);
>         wakeup(&fs->lfs_avail);
>  
> -       ++lfs_stats.segs_reclaimed;
> +      if (lfs_dostats)
> +               ++lfs_stats.segs_reclaimed;
>  
>         return (0);
>  }

Hmm, looks like we'll have to include <sys/sysctl.h> to get the value
of lfs_dostats.