Re: x86 patches was Re: -mm merge plans for 2.6.24

!MAILaRCHIVE_VOTE_RePLACE
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Andrew Morton <akpm@...>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@...>, Andi Kleen <andi@...>, <linux-kernel@...>, <mpm@...>, Huang, Ying <ying.huang@...>, Thomas Gleixner <tglx@...>, Christoph Lameter <clameter@...>, <apw@...>
Date: Tuesday, October 2, 2007 - 1:25 pm

On Tue, 2007-10-02 at 00:43 -0700, Andrew Morton wrote:

For the HP platforms, we can configure each cell with from 0% to 100%
"cell local memory".  When we configure with <100% CLM, the "missing
percentages" are interleaved by hardware on a cache-line granularity to
improve bandwidth at the expense of latency for numa-challenged
applications [and OSes, but not our problem ;-)].  When we boot Linux on
such a config, all of the real nodes have no memory--it all resides in a
single interleaved pseudo-node.  

When we boot Linux on a 100% CLM configuration [== NUMA], we still have
the interleaved pseudo-node.  It contains a few hundred MB stolen from
the real nodes to contain the DMA zone.  [Interleaved memory resides at
phys addr 0].  The memoryless-nodes patches, along with the zoneorder
patches, support this config as well.

Also, when we boot a NUMA config with the "mem=" command line,
specifying less memory than actually exists, Linux takes the excluded
memory "off the top" rather than distributing it across the nodes.  This
can result in memoryless nodes, as well.

Lee

-
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
-mm merge plans for 2.6.24, Andrew Morton, (Mon Oct 1, 5:22 pm)
new aops merge [was Re: -mm merge plans for 2.6.24], Hugh Dickins, (Tue Oct 2, 12:21 pm)
x86 patches was Re: -mm merge plans for 2.6.24, Andi Kleen, (Tue Oct 2, 2:18 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andrew Morton, (Tue Oct 2, 2:32 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Ingo Molnar, (Tue Oct 2, 3:37 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andi Kleen, (Tue Oct 2, 3:46 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Thomas Gleixner, (Tue Oct 2, 3:58 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andi Kleen, (Tue Oct 2, 3:01 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andy Whitcroft, (Tue Oct 2, 5:26 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andrew Morton, (Tue Oct 2, 3:18 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, KAMEZAWA Hiroyuki, (Tue Oct 2, 3:36 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Christoph Lameter, (Tue Oct 2, 2:16 pm)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Nish Aravamudan, (Tue Oct 2, 12:40 pm)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andrew Morton, (Tue Oct 2, 3:43 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, KAMEZAWA Hiroyuki, (Tue Oct 2, 4:16 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Christoph Lameter, (Tue Oct 2, 2:18 pm)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Yasunori Goto, (Tue Oct 2, 6:48 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Lee Schermerhorn, (Tue Oct 2, 1:25 pm)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Lee Schermerhorn, (Tue Oct 2, 1:17 pm)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Matt Mackall, (Tue Oct 2, 3:55 am)
Re: x86 patches was Re: -mm merge plans for 2.6.24, Andi Kleen, (Tue Oct 2, 3:59 am)
Re: -mm merge plans for 2.6.24, Pekka Enberg, (Tue Oct 2, 12:12 pm)
v4l-stk11xx* [Was: -mm merge plans for 2.6.24], Jiri Slaby, (Tue Oct 2, 3:59 am)
writeback fixes, Fengguang Wu, (Tue Oct 2, 4:39 am)
Re: -mm merge plans for 2.6.24, Borislav Petkov, (Sat Oct 13, 4:44 am)
Re: -mm merge plans for 2.6.24, Andrew Morton, (Sat Oct 13, 4:52 am)
Re: -mm merge plans for 2.6.24, Borislav Petkov, (Sat Oct 13, 7:45 am)
r/o bind mounts, was Re: -mm merge plans for 2.6.24, Christoph Hellwig, (Tue Oct 9, 5:19 am)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Linus Torvalds, (Wed Oct 3, 11:21 am)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Linus Torvalds, (Tue Oct 9, 10:52 am)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Linus Torvalds, (Tue Oct 9, 10:22 pm)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Hugh Dickins, (Wed Oct 10, 12:06 am)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Linus Torvalds, (Wed Oct 10, 1:20 am)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Linus Torvalds, (Wed Oct 10, 11:04 am)
Re: remove zero_page (was Re: -mm merge plans for 2.6.24), Linus Torvalds, (Tue Oct 9, 11:06 pm)
wibbling over the cpuset shed domain connnection, Paul Jackson, (Mon Oct 1, 5:34 pm)
Re: wibbling over the cpuset shed domain connnection, Nick Piggin, (Tue Oct 2, 8:36 am)
Re: wibbling over the cpuset shed domain connnection, Paul Jackson, (Wed Oct 3, 1:21 am)
Re: wibbling over the cpuset shed domain connnection, Nick Piggin, (Tue Oct 2, 9:12 am)
Re: wibbling over the cpuset shed domain connnection, Paul Jackson, (Wed Oct 3, 3:00 am)
Re: wibbling over the cpuset shed domain connnection, Andrew Morton, (Wed Oct 3, 6:57 am)
per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Tue Oct 2, 4:17 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Martin Knoblauch, (Wed Oct 3, 7:00 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Fri Oct 26, 10:48 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Trond Myklebust, (Fri Oct 26, 12:37 pm)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Fri Dec 14, 10:50 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Miklos Szeredi, (Fri Dec 14, 11:14 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Fri Dec 14, 11:54 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Miklos Szeredi, (Fri Oct 26, 11:06 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Fri Oct 26, 11:22 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Fri Oct 26, 11:33 am)
[PATCH] mm: sysfs: expose the BDI object in sysfs, Peter Zijlstra, (Fri Nov 2, 10:59 am)
Re: [PATCH] mm: sysfs: expose the BDI object in sysfs, Kay Sievers, (Fri Nov 2, 11:13 am)
Re: per BDI dirty limit (was Re: -mm merge plans for 2.6.24), Peter Zijlstra, (Sat Oct 27, 12:07 pm)