Re: migration cache, updated

[Ray Bryant <raybry@sgi.com>]

Marcello and Hirokazu,

I've finally gotten around to working on my user controlled page migration 
project.  (What I'm trying to implement is a way for a suitably authorized
user program to request that some or all of the pages or a particular address
space be migrated from one NUMA node to another.)

The first such thing to try to migrate is anonymous, private pages so that
is what I am working on.  To keep things simple, the user program is trying
to migrate part of its own address space.

What I have found out is that this works correctly with the page-migration
patch (extracted from the memory hotplug patch, e. g.:

http://sr71.net/patches/2.6.10/2.6.10-mm1-mhp-test7/page_migration/

) but when I apply the migration cache patch on top what happens is that
the migration appears to occur, but then as soon as I return from the
system call, I get:

VM: killing process test
Killed

(This message comes from ia64_do_page_fault() and appears to because
handle_mm_fault() returned FAULT_OOM....)

I haven't looked into this further, but was wondering if perhaps one of
you would understand why the migrate cache patch would fail in this way?

Thanks,
-- 
Best Regards,
Ray
-----------------------------------------------
                   Ray Bryant
512-453-9679 (work)         512-507-7807 (cell)
raybry@sgi.com             raybry@austin.rr.com
The box said: "Requires Windows 98 or better",
            so I installed Linux.
-----------------------------------------------
--
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:"aart@kvack.org"> aart@kvack.org </a>