Re: question about link_mem_sections()

[Robin Holt <holt@sgi.com>]

On Thu, Dec 22, 2011 at 02:18:47AM +0100, Kay Sievers wrote:
> On Mon, Dec 12, 2011 at 14:49, Robin Holt <holt@sgi.com> wrote:
> > On Mon, Dec 12, 2011 at 02:11:28PM +0100, Kay Sievers wrote:
> 
> > If you do decide to rework that logic, we can retest for the next couple
> > months.  We do not normally have a 16TB test machine lying around (IIRC,
> > the memory in something like that costs a couple million dollars), but we
> > do fortunately have one right now until the latter part of next quarter.
> 
> The rework of the sys_devices with the goal to finally removal of all
> sysdev stuff is in Greg's tree now:
>   http://git.kernel.org/?p=linux/kernel/git/gregkh/driver-core.git;a=shortlog;h=refs/heads/driver-core-next
> It will be available in linux-next.
> 
> The hinted search moved from the 'kobject' to the 'bus' as:
>   dev = subsys_find_device_by_id(&memory_subsys, block_id, hintdev);
> 
> Would be nice if you can eventually give it a try, on the big box, to
> make sure we didn't break anything.

If it is in linux-next, we may already have it.

Russ or Jack, IIRC, you had been doing some testing with linux-next
on the 16TB machine.  Is that correct?  If so, do you happen to have a
timed boot log that we could look at?

Thanks,
Robin