Messages in this thread | | | Date | Fri, 6 May 2011 09:15:35 +0200 | From | Michal Hocko <> | Subject | Re: [PATCH resend] mm: get rid of CONFIG_STACK_GROWSUP || CONFIG_IA64 |
| |
On Thu 05-05-11 18:12:12, Hugh Dickins wrote: > On Wed, May 4, 2011 at 11:30 PM, Michal Hocko <mhocko@suse.cz> wrote: > > > So I think the flag should be used that way. If we ever going to add a > > new architecture like IA64 which uses both ways of expanding we should > > make it easier by minimizing the places which have to be examined. > > If, yes. Let's just agree to disagree. It looks like I'm preferring > to think of the ia64 case as exceptional, and I want to be reminded of > that peculiar case; whereas you are wanting to generalize and make it > not stand out. Both valid.
Probably a call for Andrew? Anyway, whatever is the way we go I think that both declaration and definition should be guarded by the same ifdefs.
> > OK, now, with the cleanup patch, we have expand_stack and > > expand_stack_{downwards,upwards}. I will repost the patch to Andrew with > > up and down cases renamed. Does it work for you? > > Sounds right.
OK, I will repost the updated patch.
> >> But it's always going to be somewhat confusing and asymmetrical > >> because of the ia64 register backing store case. > > > > How come? We would have expand_stack which is pretty much clear that it > > is expanding stack in the architecture specific way. And then we would > > have expand_{upwards,downward} which are clear about way how we expand > > whatever VMA, right? > > Right. I'm preferring to be reminded of the confusion and asymmetry, > you're preferring to smooth over it.
OK
Thanks -- Michal Hocko SUSE Labs SUSE LINUX s.r.o. Lihovarska 1060/12 190 00 Praha 9 Czech Republic
| |