Re: [PATCH] x86: split e820 reserved entries record to late v4

[Linus Torvalds <torvalds@linux-foundation.org>]

On Fri, 29 Aug 2008, Ingo Molnar wrote:
> > 
> > you will get
> > fec00000 - ffffffff reserved
> >    fed0000 - fed003ff hpet
> >      fed0000 - fed003ff 0000:00:14.0
> 
> ok - because it's fully contained insert_resource() will succeed?

Correct.

> I thought it would only succeed if the new resource was smaller than (a 
> subset of) the existing resource.

No, that's "request_resource()".

Yeah, I know, the resource code is complicated, and I wish it wasn't, but 
the whole issue with nesting resources correctly simply _is_ fairly 
complex.

So "insert_resource()" literally tries to fit a resource into an existing 
tree, at the right level, whatever level that is.

In contrast, "request_resource()" is about exclusivity, and tries to 
request a "leaf" resource - and it refuses to work if there are resources 
it clashes with that cover the same range.

> In the other direction, when a newly inserted resource is a superset of 
> the existing resource, i thought we'd fail.

For request_resource, yes.

> hypothetical scenario, what if we had neither a superset nor a subset 
> scenario, but a partial overlap, between:
> 
> > >  BIOS-e820: 00000000fec00000 - 0000000100000000 (reserved)
> 
> and:
> 
> > >   pci 0000:00:14.0: BAR has HPET at feb0f000-fec01000

Now THAT is somethign that the resource structs simply cannot handle. At 
that point you'd get a failure even from insert_resource(), because it no 
longer nests in the tree. You would have to split the range in order for 
it to nest properly.

			Linus