Re: NetBSD hammer with fuse and hammerread

["Jeremy C. Reed" <reed@xxxxxxxxxxxxx>]

On Thu, 16 Apr 2009, Jeremy C. Reed wrote:

> Problems:
> - doesn't have number of links for directory (maybe causes find problem 
> 	shown below)

I see on a DragonFly system using native hammer and also mounting the test 
image, the number of links for a directory is always 1 (versus 0 I see 
when using the fusehammer). On ufs, the links would show the correct 
amount -- always more than 1 for ufs directories.

Is this documented?

> - doesn't have size of directory (number of blocks used)

On a DragonFly system using native hammer and also mounting the test 
image, the number of blocks used by the directory is always 0. (Where with 
ufs it begins at 512.)

Is this behaviour documented? Okay? Known?

> - timestamps
> - no inode numbers

These do work (of course) on native hammer and also mounting the test 
image.

My ino_t is uint64_t and my time_t is int -- but NetBSD -current is 
__int64_t -- I need to test this on NetBSD-current with 64 bit 
timestamps....

> And a find problem (doesn't descend):
>
> $ find ./mnt -ls
>       0      0 drwxr-xr-x    0 root              wheel                    0 Dec 31  1969 ./mnt
>       0      0 -rw-r--r--    0 root              wheel                    12 Dec 31  1969 ./mnt/test
>       0      0 drwxr-xr-x    0 root              wheel                     0 Dec 31  1969 ./mnt/a

Mounting the image does the find correctly (all entries shown).

I mounted with:

   vnconfig /dev/vn0 hammerdisk.raw
   mount -t hammer /dev/vn0 ./mnt

Using DragonFly 2.2.0-RELEASE