Re: [PATCH v6][RFC] Fix hugetlb pool allocation with empty nodes

[Nishanth Aravamudan <nacc@us.ibm.com>]

On 11.06.2007 [20:48:08 -0700], Christoph Lameter wrote:
> On Mon, 11 Jun 2007, Nishanth Aravamudan wrote:
> 
> > > Export a function for the interleave functionality so that we do not
> > > have to replicate the same thing in various locations in the kernel.
> > 
> > But I don't understand this at all.
> > 
> > This is *not* generically available, unless every caller has its own
> > private static variable. I don't know how to do that in C.
> 
> It is already there. Each task has a il_next field in its task struct
> for that purpose.

Ok, I see that. And it represent the next node to use for an interleaved
allocation. Makes sense to me, and I see how it's used in mempolicy.c to
achieve that. But we're running at system boot time, or whenever some
invokes the sysctl /proc/sys/vm/nr_hugepages. Do we really want to muck
with some arbitray bash shell's il_next field to achieve interleaving?
What if it's a C process that is trying to achieve actual interleaving
for other purposes and also allocates some hugepages on the system? It
seems like il_next is very much a process-related field.

When I wrote "caller", I meant calling function, sorry, not calling
process.

I'm not entirely sure how il_next is useful here, sorry.

Thanks,
Nish

-- 
Nishanth Aravamudan <nacc@us.ibm.com>
IBM Linux Technology Center

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>