Love wrote:

> Gregory McGarry <g.mcgarry@ieee.org> writes:
> 
> > > I added an updated patch, I changed so you can't dump your own memory
> > > mapping, but then you can't attach to yourself so that shouldn't be a
> > > problem.
> > 
> > I like the functionality provided by pmap(1) on Solaris.  Providing
> > the backing store for each mapping looks useful.  So I'd check the
> > Solaris interface too.  But perhaps this is beyond the scope you
> > intended.
> 
> I can see why you think this is useful. This exists in the linux procfs
> support, if you just the filename. I don't really care about if it should
> be included or not, so its up to others to say so. It should be somewhat
> easy to add.

I think it is useful.

> I don't really like cache_revlookup(), mostly since it only might work. On
> my laptop the failure rate is 0.3%.

getcwd_scandir() should do want you want, shouldn't it?  Seems like
reasonable functionality to factor out, to a function called, say,
vn_scandir().

	-- Gregory McGarry <g.mcgarry@ieee.org>