svn commit: r213398 - head/bin/rm

[Ben Kaduk]

On Mon, Oct 4, 2010 at 7:42 AM, Alexander Best <arundel at freebsd.org> wrote:
> On Mon Oct  4 10, Pawel Jakub Dawidek wrote:
>> On Mon, Oct 04, 2010 at 02:35:54PM +0400, Dmitry Morozovsky wrote:
>> > On Mon, 4 Oct 2010, Alexander Best wrote:
>> > AB>      "The -P option assumes that the underlying file system is a fixed-block
>> > AB>      file system.  UFS is a fixed-block file system, LFS is not.  In addition,
>> > AB>      only regular files are overwritten, other types of files are not."
>> >
>> > Maybe s/LFS/ZFS/ then, as LFS is no more relevant for FreeBSD users while ZFS
>> > now is?
>>
>> That's what I thought too.
>
> good point. ZFS should really be added to the list and LFS should go away. are
> there any other relevant filesystems without a fixed-block size that need to be
> mentioned? what about afs? or tmpfs?

I'm not fully up-to-speed on the AFS fileserver backend, but it is
certainly the case that AFS cannot guarantee that rm -P will actually
overwrite the data on-disk.  There are probably several mechanisms by
which this could happen, the easiest to see of which would be if a
filesystem like ZFS was used as the backing store for the fileserver
partitions.

>
> also: is this really something belonging into a BUGS section? personally i
> think the BUGS section in rm(1) should be renamed to CAVEATS.

BUGS is easier to find than CAVEATS, though I guess rm(1) is short
enough that we can expect people to read the whole thing.

-Ben Kaduk