mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Christoph Lameter <clameter@sgi.com>
To: Badari Pulavarty <pbadari@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	lkml <linux-kernel@vger.kernel.org>
Subject: Re: 2.6.21-rc5-mm4 (SLUB powerpc)
Date: Wed, 4 Apr 2007 10:13:05 -0700 (PDT)	[thread overview]
Message-ID: <Pine.LNX.4.64.0704041003170.24640@schroedinger.engr.sgi.com> (raw)
In-Reply-To: <1175699559.6185.24.camel@dyn9047017100.beaverton.ibm.com>

On Wed, 4 Apr 2007, Badari Pulavarty wrote:

> Well !! Helps a little, but not enough to boot (hangs little later) :(
> I will try to get stack trace for that.

Great! Thanks for all the debugging help.

 
> Processor 6 found.
> Processor 7 found.
> Brought up 8 CPUs
> mm/memory.c:111: bad pud c0000000f20c0480.

Hmmm... Checking for slabs used in powerpc arch code:

The pgtable cache is configured as


          pgtable_cache[i] = kmem_cache_create(name,
                                                     size, size,
                                                     SLAB_HWCACHE_ALIGN |
                                                     SLAB_MUST_HWCACHE_ALIGN,
                                                     zero_ctor,
                                                     NULL);

Hmmm.... aligned slabs at size and then we MUST_HWCACHE_ALIGN?? Two 
competing alignment requirements and a constructor. Constructor requires
the moving of the free pointer after the slab and thus increases the slab 
size.

Sigh. IF SLAB_HWCACHE_ALIGN is set then SLUB believes this to be the 
ultimate demand that overrides all other alignments and only aligns to the 
cacheline. Try the following fix:



SLUB: Treat SLAB_HWCACHE_ALIGN as a mininum and not as *the* alignment

If the specified alignment is higher than L1_CACHE_BYTES and
SLAB_HWCACHE_ALIGN is set then use the higher alignment.

Signed-off-by: Christoph Lameter <clameter@sgi.com>

Index: linux-2.6.21-rc5-mm4/mm/slub.c
===================================================================
--- linux-2.6.21-rc5-mm4.orig/mm/slub.c	2007-04-04 10:09:20.000000000 -0700
+++ linux-2.6.21-rc5-mm4/mm/slub.c	2007-04-04 10:09:42.000000000 -0700
@@ -1373,10 +1373,7 @@ static int calculate_order(int size)
 static unsigned long calculate_alignment(unsigned long flags,
 		unsigned long align)
 {
-	if (flags & SLAB_HWCACHE_ALIGN)
-		return L1_CACHE_BYTES;
-
-	if (flags & SLAB_MUST_HWCACHE_ALIGN)
+	if (flags & (SLAB_MUST_HWCACHE_ALIGN | SLAB_HWCACHE_ALIGN))
 		return max_t(unsigned long, align, L1_CACHE_BYTES);
 
 	if (align < ARCH_SLAB_MINALIGN)

  parent reply	other threads:[~2007-04-04 17:13 UTC|newest]

Thread overview: 86+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-04-03  5:47 2.6.21-rc5-mm4 Andrew Morton
2007-04-03 12:57 ` 2.6.21-rc5-mm4 Jiri Kosina
2007-04-03 15:11   ` 2.6.21-rc5-mm4 Kok, Auke
2007-04-03 16:30     ` 2.6.21-rc5-mm4 Jiri Kosina
2007-04-03 17:19       ` 2.6.21-rc5-mm4 Kok, Auke
2007-04-03 17:42         ` 2.6.21-rc5-mm4 Jiri Kosina
2007-04-04 16:10           ` 2.6.21-rc5-mm4 Jiri Kosina
2007-04-04 17:22             ` 2.6.21-rc5-mm4 Eric W. Biederman
2007-04-06 13:23               ` 2.6.21-rc5-mm4 Jiri Kosina
2007-04-06 14:32                 ` 2.6.21-rc5-mm4 Eric W. Biederman
2007-04-03 16:11 ` 2.6.21-rc5-mm4 Michal Piotrowski
2007-04-03 22:20   ` 2.6.21-rc5-mm4 Michal Piotrowski
2007-04-03 22:31     ` 2.6.21-rc5-mm4 Con Kolivas
2007-04-03 23:31       ` 2.6.21-rc5-mm4 Michal Piotrowski
2007-04-04 13:22         ` [PATCH] sched: implement staircase deadline cpu scheduler improvements fix Con Kolivas
2007-04-03 19:48 ` 2.6.21-rc5-mm4 Randy Dunlap
2007-04-03 19:51   ` 2.6.21-rc5-mm4 Davide Libenzi
2007-04-03 20:31 ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 20:40   ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-03 22:53     ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 22:59       ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-03 23:40         ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 23:41           ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-04  1:16       ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-04 15:12         ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-04 15:27           ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-04 17:13           ` Christoph Lameter [this message]
2007-04-04 17:33             ` 2.6.21-rc5-mm4 (SLUB powerpc) Badari Pulavarty
2007-04-04 17:35               ` Christoph Lameter
2007-04-04 20:53                 ` Badari Pulavarty
2007-04-03 21:55 ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 23:19   ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 23:27     ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-03 23:38       ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 23:40         ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-03 23:51           ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-03 23:55             ` 2.6.21-rc5-mm4 Christoph Lameter
2007-04-04 14:53               ` 2.6.21-rc5-mm4 Badari Pulavarty
2007-04-04 17:03                 ` 2.6.21-rc5-mm4 (SLUB) Christoph Lameter
2007-04-04 18:02                   ` Badari Pulavarty
2007-04-04 18:04                     ` Christoph Lameter
2007-04-04 18:22                       ` Christoph Lameter
2007-04-04 22:53                         ` Badari Pulavarty
2007-04-04 22:59                           ` Christoph Lameter
2007-04-04 23:42                             ` Badari Pulavarty
2007-04-05  0:31                               ` Christoph Lameter
2007-04-05  2:44                                 ` Badari Pulavarty
2007-04-05  3:15                                   ` Christoph Lameter
2007-04-05  4:29                                   ` Christoph Lameter
2007-04-05 15:16                                     ` Badari Pulavarty
2007-04-05 18:19                                       ` Christoph Lameter
2007-04-13 15:13                       ` [PATCH] cache_k8_northbridges() overflows beyond allocation (Was: 2.6.21-rc5-mm4 (SLUB)) Badari Pulavarty
2007-04-13 15:45                         ` Andi Kleen
2007-04-13 16:42                           ` Chuck Ebbert
2007-04-13 16:52                             ` Andi Kleen
2007-04-13 16:57                           ` Andrew Morton
2007-04-13 18:07                         ` Christoph Lameter
2007-04-04 18:09                     ` 2.6.21-rc5-mm4 (SLUB) Christoph Lameter
2007-04-03 22:40 ` 2.6.21-rc5-mm4 J.A. Magallón
2007-04-03 22:51   ` 2.6.21-rc5-mm4 Andrew Morton
2007-04-03 22:58     ` 2.6.21-rc5-mm4 J.A. Magallón
2007-04-03 23:22       ` 2.6.21-rc5-mm4 Valdis.Kletnieks
2007-04-03 23:43         ` 2.6.21-rc5-mm4 J.A. Magallón
2007-04-04  1:35 ` 2.6.21-rc5-mm4 Valdis.Kletnieks
2007-04-04  2:20   ` 2.6.21-rc5-mm4 Randy Dunlap
2007-04-04  2:46     ` 2.6.21-rc5-mm4 Valdis.Kletnieks
2007-04-04  3:37       ` 2.6.21-rc5-mm4 Randy Dunlap
2007-04-04 17:55         ` 2.6.21-rc5-mm4 Valdis.Kletnieks
2007-04-05 11:14           ` 2.6.21-rc5-mm4 Sam Ravnborg
2007-04-04 15:25 ` [-mm patch] net/: possible cleanups Adrian Bunk
2007-04-04 15:25 ` [-mm patch] unexport pci_proc_attach_device Adrian Bunk
2007-04-04 21:56 ` 2.6.21-rc5-mm4 Antonino A. Daplas
2007-04-04 22:10   ` 2.6.21-rc5-mm4 Andrew Morton
2007-04-04 22:38     ` 2.6.21-rc5-mm4 Con Kolivas
2007-04-04 23:55       ` 2.6.21-rc5-mm4 Antonino A. Daplas
2007-04-05  0:56 ` 2.6.21-rc5-mm4 Antonino A. Daplas
2007-04-05 13:53 ` [-mm patch] pcm_native: lockdep warning when launching jack Frederik Deweerdt
2007-04-05 15:00   ` Takashi Iwai
2007-04-05 16:33 ` RAID1 "out of memory" error, was Re: 2.6.21-rc5-mm4 Reuben Farrelly
2007-04-05 20:21   ` Andrew Morton
2007-04-06  5:34     ` Dan Williams
2007-04-11  3:31   ` Neil Brown
2007-04-05 17:02 ` 2.6.21-rc5-mm4 Valdis.Kletnieks
2007-04-05 20:31   ` 2.6.21-rc5-mm4 Andrew Morton
2007-04-06 15:26     ` 2.6.21-rc5-mm4 Valdis.Kletnieks
2007-04-06 19:33       ` 2.6.21-rc5-mm4 Andrew Morton

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=Pine.LNX.4.64.0704041003170.24640@schroedinger.engr.sgi.com \
    --to=clameter@sgi.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pbadari@gmail.com \
    /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

all inboxes | Powered by JetHome®