From: Nishanth Aravamudan <nacc@us.ibm.com>
To: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Cc: Christoph Lameter <clameter@sgi.com>,
anton@samba.org, akpm@linux-foundation.org, linux-mm@kvack.org
Subject: Re: [PATCH] populated_map: fix !NUMA case, remove comment
Date: Wed, 13 Jun 2007 10:58:02 -0700 [thread overview]
Message-ID: <20070613175802.GP3798@us.ibm.com> (raw)
In-Reply-To: <1181748606.6148.19.camel@localhost>
On 13.06.2007 [11:30:06 -0400], Lee Schermerhorn wrote:
> On Tue, 2007-06-12 at 13:01 -0700, Nishanth Aravamudan wrote:
> > On 12.06.2007 [12:58:16 -0700], Christoph Lameter wrote:
> > > On Tue, 12 Jun 2007, Christoph Lameter wrote:
> > >
> > > > Uhhh... Right there is another special case. The recently
> > > > introduces zonelist swizzle makes the DMA zone come last and if a
> > > > node had only a DMA zone then it may become swizzled to the end of
> > > > the zonelist.
> > >
> > > Maybe we can ignore that case for now:
> > >
> I wish we wouldn't. We need the "DMA zone comes last" for both HP and
> Fujitsu platforms. That's why Kame and I worked on that patch
> together.
Right. I interpreted the "for now" as for this first stack of patches.
We'll need a fix for your platform on top, but it seems to be a minority
case? Not saying it shouldn't be fixed, by any means, just trying to get
a handle on it.
> > > Fix GFP_THISNODE behavior for memoryless nodes
> > >
> > > GFP_THISNODE checks that the zone selected is within the pgdat (node) of the
> > > first zone of a nodelist. That only works if the node has memory. A
> > > memoryless node will have its first node on another pgdat (node).
> > >
> > > GFP_THISNODE currently will return simply memory on the first pgdat.
> > > Thus it is returning memory on other nodes. GFP_THISNODE should fail
> > > if there is no local memory on a node.
> > >
> > > So we add a check to verify that the node specified has memory in
> > > alloc_pages_node(). If the node has no memory then return NULL.
> > >
> > > The case of alloc_pages(GFP_THISNODE) is not changed. alloc_pages() (with no memory
> > > policies in effect)
> > >
> > > Signed-off-by: Christoph Lameter <clameter@sgi.com>
> > > Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
> > >
> > > Index: linux-2.6.22-rc4-mm2/include/linux/gfp.h
> > > ===================================================================
> > > --- linux-2.6.22-rc4-mm2.orig/include/linux/gfp.h 2007-06-12 12:33:37.000000000 -0700
> > > +++ linux-2.6.22-rc4-mm2/include/linux/gfp.h 2007-06-12 12:38:37.000000000 -0700
> > > @@ -175,6 +175,13 @@ static inline struct page *alloc_pages_n
> > > if (nid < 0)
> > > nid = numa_node_id();
> > >
> > > + /*
> > > + * Check for the special case that GFP_THISNODE is used on a
> > > + * memoryless node
> > > + */
> > > + if ((gfp_mask & __GFP_THISNODE) && !node_memory(nid))
> > > + return NULL;
> > > +
> >
> > Yep, this seems to be the right thing to do, and was in my rolled-up
> > patch.
>
> I think that the "node has memory" mask is fine for scanning nodes
> that might have memory in the zone of interest--including in the
> hugetlb alloc_fresh_huge_page() loop. However, I think that to
> support all platforms in a generic way, alloc_pages_node() and
> alloc_page_interleave() [both take a node id arg] should be more
> strict when the gfp mask includes 'THISNODE and not assume that a
> populated node always has on-node memory in the zone of interest.
Hrm, perhaps.
> E.g., something like:
>
> pgdat_t *pgdat;
> struct zonelist *zonelist;
>
> ...
>
> /*
> * after validating nid, ...
> * Note that we need to fetch these values anyway for the
> * [likely?] call to __alloc_pages().
> */
> pgdat = NODE_DATA(nid);
> zonelist = pgdat->node_zonelists + gfp_zone(gfp_mask);
>
> if ((gfp_mask & __GFP_THISNODE) &&
> zonelist->zones[0]->zone_pgdat != pgdat)
> return NULL;
>
> return __alloc_pages(gfp_mask, order, zonelist);
>
>
> I see you've submitted a new patch set. I grab it [when Nish reposts]
> and test it as is and modified to look something like the above, if
> needed.
I think your code above makes sense -- I'd still leave in the earlier
check, though.
So it probably should be:
pgdat = NODE_DATA(nid);
zonelist = pgdat->node_zonelists + gfp_zone(gfp_mask);
if (unlikely((gfp_mask & __GFP_THISNODE) &&
(!node_memory(nid) ||
zonelist->zones[0]->zone_pgdat != pgdat)))
return NULL;
That way, if the node has no memory whatsoever, we don't bother checking
the pgdat of the relevant zone?
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>
next prev parent reply other threads:[~2007-06-13 17:58 UTC|newest]
Thread overview: 140+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-06-11 20:27 [PATCH] Add populated_map to account for memoryless nodes Nishanth Aravamudan, Lee Schermerhorn
2007-06-11 21:25 ` Christoph Lameter
2007-06-11 22:10 ` [PATCH v2] " Nishanth Aravamudan
2007-06-11 22:42 ` Christoph Lameter
2007-06-11 22:52 ` [PATCH v3] " Nishanth Aravamudan
2007-06-11 23:00 ` Christoph Lameter
2007-06-11 23:41 ` [PATCH v4] " Nishanth Aravamudan
2007-06-11 23:45 ` Christoph Lameter
2007-06-12 0:07 ` [PATCH] populated_map: fix !NUMA case, remove comment Nishanth Aravamudan
2007-06-12 0:41 ` Christoph Lameter
2007-06-12 1:43 ` Nishanth Aravamudan
2007-06-12 1:45 ` Christoph Lameter
2007-06-12 1:52 ` Nishanth Aravamudan
2007-06-12 2:39 ` Nishanth Aravamudan
2007-06-12 2:02 ` Nishanth Aravamudan
2007-06-12 2:20 ` Christoph Lameter
2007-06-12 2:32 ` Nishanth Aravamudan
2007-06-12 2:54 ` Christoph Lameter
2007-06-12 3:20 ` Nishanth Aravamudan
2007-06-12 3:21 ` Christoph Lameter
2007-06-12 3:31 ` Nishanth Aravamudan
2007-06-12 15:06 ` Lee Schermerhorn
2007-06-12 17:28 ` Nishanth Aravamudan
2007-06-12 18:43 ` Christoph Lameter
2007-06-12 18:48 ` Lee Schermerhorn
2007-06-12 18:51 ` Christoph Lameter
2007-06-12 19:44 ` Lee Schermerhorn
2007-06-12 19:48 ` Christoph Lameter
2007-06-12 19:58 ` Christoph Lameter
2007-06-12 20:01 ` Nishanth Aravamudan
2007-06-13 15:30 ` Lee Schermerhorn
2007-06-13 17:58 ` Nishanth Aravamudan [this message]
2007-06-13 18:21 ` Lee Schermerhorn
2007-06-13 19:01 ` Nishanth Aravamudan
2007-06-13 22:51 ` Christoph Lameter
2007-06-14 15:50 ` Lee Schermerhorn
2007-06-14 15:57 ` Christoph Lameter
2007-06-14 16:54 ` Lee Schermerhorn
2007-06-14 16:09 ` Nishanth Aravamudan
2007-06-14 16:15 ` Christoph Lameter
2007-06-14 17:07 ` Lee Schermerhorn
2007-06-14 17:16 ` Christoph Lameter
2007-06-14 18:04 ` Lee Schermerhorn
2007-06-14 22:35 ` Nishanth Aravamudan
2007-06-13 22:50 ` Christoph Lameter
2007-06-13 23:09 ` Nishanth Aravamudan
2007-06-13 23:12 ` Christoph Lameter
2007-06-13 23:18 ` Nishanth Aravamudan
2007-06-13 23:26 ` Christoph Lameter
2007-06-13 23:56 ` Nishanth Aravamudan
2007-06-14 14:23 ` Lee Schermerhorn
2007-06-13 22:49 ` Christoph Lameter
2007-06-12 19:55 ` Nishanth Aravamudan
2007-06-12 18:41 ` Christoph Lameter
2007-06-12 19:07 ` Lee Schermerhorn
2007-06-12 19:13 ` Christoph Lameter
2007-06-11 23:08 ` [PATCH][RFC] Fix INTERLEAVE with memoryless nodes Nishanth Aravamudan
2007-06-11 23:10 ` [PATCH v6][RFC] Fix hugetlb pool allocation with empty nodes Nishanth Aravamudan
2007-06-11 23:11 ` [PATCH][RFC] hugetlb: numafy several functions Nishanth Aravamudan
2007-06-11 23:13 ` [PATCH][RFC] hugetlb: add per-node nr_hugepages sysfs attribute Nishanth Aravamudan
2007-06-11 23:40 ` Christoph Lameter
2007-06-11 23:42 ` Christoph Lameter
2007-06-12 0:19 ` Nishanth Aravamudan
2007-06-12 0:43 ` Christoph Lameter
2007-06-12 2:19 ` Nishanth Aravamudan
2007-06-12 2:22 ` Christoph Lameter
2007-06-12 2:34 ` Nishanth Aravamudan
2007-06-11 23:38 ` [PATCH][RFC] hugetlb: numafy several functions Christoph Lameter
2007-06-11 23:17 ` [PATCH v6][RFC] Fix hugetlb pool allocation with empty nodes Christoph Lameter
2007-06-12 0:15 ` Nishanth Aravamudan
2007-06-12 0:47 ` Christoph Lameter
2007-06-12 2:12 ` Nishanth Aravamudan
2007-06-12 2:21 ` Christoph Lameter
2007-06-12 2:25 ` Christoph Lameter
2007-06-12 2:34 ` Nishanth Aravamudan
2007-06-12 2:55 ` Christoph Lameter
2007-06-12 3:17 ` Nishanth Aravamudan
2007-06-12 3:19 ` Christoph Lameter
2007-06-12 3:30 ` Nishanth Aravamudan
2007-06-12 3:48 ` Christoph Lameter
2007-06-12 5:07 ` Nishanth Aravamudan
2007-06-12 18:47 ` Christoph Lameter
2007-06-12 17:43 ` Nishanth Aravamudan
2007-06-12 18:49 ` Christoph Lameter
2007-06-12 2:33 ` Nishanth Aravamudan
2007-06-12 3:44 ` William Lee Irwin III
2007-06-12 3:50 ` Christoph Lameter
2007-06-12 3:53 ` William Lee Irwin III
2007-06-12 3:53 ` Christoph Lameter
2007-06-12 4:14 ` William Lee Irwin III
2007-06-12 5:09 ` Nishanth Aravamudan
2007-06-12 5:15 ` William Lee Irwin III
2007-06-12 17:36 ` Nishanth Aravamudan
2007-06-12 18:50 ` Christoph Lameter
2007-06-12 17:45 ` Nishanth Aravamudan
2007-06-12 19:13 ` William Lee Irwin III
2007-06-13 0:04 ` [PATCH v7][RFC] " Nishanth Aravamudan
2007-06-13 15:26 ` [PATCH v3][RFC] hugetlb: numafy several functions Nishanth Aravamudan
2007-06-13 15:28 ` [PATCH v3][RFC] hugetlb: add per-node nr_hugepages sysfs attribute Nishanth Aravamudan
2007-06-13 18:23 ` Lee Schermerhorn
2007-06-13 19:19 ` [PATCH v4][RFC] " Nishanth Aravamudan
2007-06-13 20:05 ` Lee Schermerhorn
2007-06-13 20:29 ` Nishanth Aravamudan
2007-06-13 21:02 ` Lee Schermerhorn
2007-07-23 19:23 ` Christoph Lameter
2007-07-23 20:14 ` Lee Schermerhorn
2007-06-13 21:04 ` [PATCH v7][RFC] Fix hugetlb pool allocation with empty nodes Lee Schermerhorn
2007-06-13 21:50 ` [PATCH v7][UPDATE][RFC] " Nishanth Aravamudan
2007-06-12 14:28 ` [PATCH v6][RFC] " Lee Schermerhorn
2007-06-11 23:15 ` [PATCH][RFC] Fix INTERLEAVE with memoryless nodes Christoph Lameter
2007-06-12 0:14 ` [PATCH v2][RFC] " Nishanth Aravamudan
2007-06-12 0:42 ` Christoph Lameter
2007-06-12 0:57 ` Andrew Morton
2007-06-12 1:12 ` Christoph Lameter
2007-06-12 1:41 ` Nishanth Aravamudan
2007-06-12 1:52 ` Andrew Morton
2007-06-12 2:03 ` Nishanth Aravamudan
2007-06-12 14:19 ` [PATCH v2] Add populated_map to account for " Lee Schermerhorn
2007-06-12 17:32 ` Nishanth Aravamudan
2007-06-12 18:45 ` Christoph Lameter
2007-06-12 19:17 ` Lee Schermerhorn
2007-06-12 19:22 ` Christoph Lameter
2007-06-12 19:49 ` Nishanth Aravamudan
2007-06-12 19:51 ` Christoph Lameter
2007-06-12 20:00 ` Nishanth Aravamudan
2007-06-12 20:03 ` Christoph Lameter
2007-06-12 20:10 ` Christoph Lameter
2007-06-12 19:52 ` Christoph Lameter
2007-06-12 19:58 ` Christoph Lameter
2007-06-12 20:00 ` Nishanth Aravamudan
2007-06-12 20:06 ` Christoph Lameter
2007-06-12 14:10 ` [PATCH] " Lee Schermerhorn
2007-06-12 17:35 ` Nishanth Aravamudan
2007-06-12 18:39 ` Christoph Lameter
2007-06-12 18:54 ` Lee Schermerhorn
2007-06-12 19:00 ` Christoph Lameter
2007-06-12 2:27 ` KAMEZAWA Hiroyuki
2007-06-12 2:46 ` Nishanth Aravamudan
2007-06-12 2:53 ` Christoph Lameter
2007-06-12 3:04 ` KAMEZAWA Hiroyuki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20070613175802.GP3798@us.ibm.com \
--to=nacc@us.ibm.com \
--cc=Lee.Schermerhorn@hp.com \
--cc=akpm@linux-foundation.org \
--cc=anton@samba.org \
--cc=clameter@sgi.com \
--cc=linux-mm@kvack.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox