Re: r-o bind in nfsd

["NeilBrown" <neilb@suse.de>]

On Wed, March 26, 2008 9:49 am, Al Viro wrote:
> On Wed, Mar 26, 2008 at 09:32:08AM +1100, NeilBrown wrote:
>
>> >   (1) The kernel don't know what operation (open/create/truncate etc.)
>> > will be
>> >       done at the moment of link_path_walk().
>>
>> Though the 'indent' data structure could be used to carry this
>> information.
>
> If it's 'intent',

Yes, sorry.

>                   that mess will be gone next cycle.

Cool.  Any chance of a preview?  Is it in -next or -mm ??

>> >   (3) The rename() and link() operations handle two pathnames.
>> >       But, it is not possible to know both pathnames at the moment of
>> >       link_path_walk().
>>
>> Not an insolvable problem.
>> One could imagine an implementation where a TYPE_RENAME_FROM security
>> check produced a cookie that was consumed by a TYPE_RENAME_TO security
>> check.  The cookie could then be used by the security module to
>> make any connection between the two names that might be appropriate.
>
> alt.tasteless.software is that -> way...
>

While I have no desire to defend that particular design, saying "tasteless"
without suggesting an alternate approach does appear somewhat unhelpful.

NeilBrown