linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
* [patch] bug of pgdat_list connection in init_bootmem()
@ 2005-09-28 13:50 Yasunori Goto
  2005-09-28 16:42 ` Dave Hansen
  0 siblings, 1 reply; 3+ messages in thread
From: Yasunori Goto @ 2005-09-28 13:50 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linux-mm

I think 2.6.14-rc2 has a bug in init_bootmem().
There are just 2 pgdat in pgdat_list even if node num is 3 or more,
because pgdat_last is not updated.

Bye.

P.S.
  I would like to remove this pgdat_list, to simplify hot-add/remove
  a node. and posted patch before.
   http://marc.theaimsgroup.com/?l=linux-mm&m=111596924629564&w=2
   http://marc.theaimsgroup.com/?l=linux-mm&m=111596953711780&w=2

  I would like to repost after getting performance impact by this.
  But it is very hard that I can get time to use big NUMA machine now.
  So, I don't know when I will be able to repost it.

  Anyway, this should be modified before remove pgdat_list.


Signed-off-by Yasunori Goto <y-goto@jp.fujitsu.com>

Index: bootmem_new/mm/bootmem.c
===================================================================
--- bootmem_new.orig/mm/bootmem.c	2005-09-23 17:42:06.000000000 +0900
+++ bootmem_new/mm/bootmem.c	2005-09-23 17:44:59.000000000 +0900
@@ -66,9 +66,10 @@ static unsigned long __init init_bootmem
 	pgdat->pgdat_next = NULL;
 	/* Add new nodes last so that bootmem always starts
 	   searching in the first nodes, not the last ones */
-	if (pgdat_last)
+	if (pgdat_last){
 		pgdat_last->pgdat_next = pgdat;
-	else {
+		pgdat_last = pgdat;
+	} else {
 		pgdat_list = pgdat; 	
 		pgdat_last = pgdat;
 	}

-- 
Yasunori Goto 

--
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>

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [patch] bug of pgdat_list connection in init_bootmem()
  2005-09-28 13:50 [patch] bug of pgdat_list connection in init_bootmem() Yasunori Goto
@ 2005-09-28 16:42 ` Dave Hansen
  2005-09-29  1:06   ` Yasunori Goto
  0 siblings, 1 reply; 3+ messages in thread
From: Dave Hansen @ 2005-09-28 16:42 UTC (permalink / raw)
  To: Yasunori Goto; +Cc: Andrew Morton, linux-mm

On Wed, 2005-09-28 at 22:50 +0900, Yasunori Goto wrote:
>   I would like to remove this pgdat_list, to simplify hot-add/remove
>   a node. and posted patch before.
>    http://marc.theaimsgroup.com/?l=linux-mm&m=111596924629564&w=2
>    http://marc.theaimsgroup.com/?l=linux-mm&m=111596953711780&w=2
> 
>   I would like to repost after getting performance impact by this.
>   But it is very hard that I can get time to use big NUMA machine now.
>   So, I don't know when I will be able to repost it.
> 
>   Anyway, this should be modified before remove pgdat_list.

Could you resync those to a current kernel and resend them?  I'll take
them into -mhp for a bit.

I'd be very skeptical that it would hurt performance.  If nothing else,
it just makes the pgdat smaller, and the likelyhood of having the next
bit in a bitmask and the NODE_DATA() entry in your cache is slightly
higher than some random pgdat->list.

-- Dave

--
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>

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [patch] bug of pgdat_list connection in init_bootmem()
  2005-09-28 16:42 ` Dave Hansen
@ 2005-09-29  1:06   ` Yasunori Goto
  0 siblings, 0 replies; 3+ messages in thread
From: Yasunori Goto @ 2005-09-29  1:06 UTC (permalink / raw)
  To: Dave Hansen; +Cc: Andrew Morton, linux-mm

> On Wed, 2005-09-28 at 22:50 +0900, Yasunori Goto wrote:
> >   I would like to remove this pgdat_list, to simplify hot-add/remove
> >   a node. and posted patch before.
> >    http://marc.theaimsgroup.com/?l=linux-mm&m=111596924629564&w=2
> >    http://marc.theaimsgroup.com/?l=linux-mm&m=111596953711780&w=2
> > 
> >   I would like to repost after getting performance impact by this.
> >   But it is very hard that I can get time to use big NUMA machine now.
> >   So, I don't know when I will be able to repost it.
> > 
> >   Anyway, this should be modified before remove pgdat_list.
> 
> Could you resync those to a current kernel and resend them?  I'll take
> them into -mhp for a bit.
> 
> I'd be very skeptical that it would hurt performance.  If nothing else,
> it just makes the pgdat smaller, and the likelyhood of having the next
> bit in a bitmask and the NODE_DATA() entry in your cache is slightly
> higher than some random pgdat->list.

Ok! I'll do it. :-)

Thanks.

-- 
Yasunori Goto 

--
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>

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2005-09-29  1:06 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-09-28 13:50 [patch] bug of pgdat_list connection in init_bootmem() Yasunori Goto
2005-09-28 16:42 ` Dave Hansen
2005-09-29  1:06   ` Yasunori Goto

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox