I'd go with DEV_BSIZE here too.  in other places the code uses btodb()
in computing changes to fs_pendingblocks, and that uses DEV_BSIZE.
so that expression will work as-is with a redefined dbtofsb().

-Chuck


On Tue, Jun 04, 2002 at 02:57:49PM -0700, Trevin Beattie wrote:
> Question: in sys/ufs/ffs/ffs_vfsops.c:991, is fs_pendingblocks the number
> of logical (DEV_BSIZE) blocks being freed or the number of physical sectors
> being freed?  (This is where the result of "dbtofsb(fs,
> fs->fs_pendingblocks)" is added to cs_nffree and cs_nbfree to get the total
> number of free fragments in a statfs call.)
> 
> -----------------------
> Trevin Beattie          "Do not meddle in the affairs of wizards,
> trevin@xmission.com     for you are crunchy and good with ketchup."
>       {:->                                     --unknown