Date: Thu, 1 Nov 2012 19:04:20 +0100
From: Tino Reichardt <list-jfs@mcmilk.de>
To: jfs-discussion@lists.sourceforge.net,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>
Subject: Re: [Jfs-discussion] Out of memory on 3.5 kernels
Message-ID: <20121101180420.GA24922@mcmilk.de> (sfid-20121101_191215_298510_21F9FAD7)
References: <505B1FA0.6030100@broadcom.com>
 <20120921194931.GA732@schottelius.org>
 <alpine.DEB.2.00.1209241538310.6356@chino.kir.corp.google.com>
 <20120926060653.GD16575@schottelius.org>
 <20120926085708.GA7839@schottelius.org>
 <20120927055208.GA25252@schottelius.org>
 <20121003212311.GA14018@schottelius.org>
 <22962.1349452084@turing-police.cc.vt.edu>
 <20121005175139.GC831@schottelius.org>
 <20121030103535.GA10526@schottelius.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20121030103535.GA10526@schottelius.org>
Sender: linux-wireless-owner@vger.kernel.org

* Nico Schottelius <nico-kernel20120920@schottelius.org> wrote:
> Good morning,
> 
> update: this problem still exists on 3.6.2-1-ARCH and it got worse:
> 
> I reformatted the external disk to use xfs, but as the my
> root filesystem is still jfs, it still appears:
> 
>  Active / Total Objects (% used)    : 642732 / 692268 (92.8%)
>  Active / Total Slabs (% used)      : 24801 / 24801 (100.0%)
>  Active / Total Caches (% used)     : 79 / 111 (71.2%)
>  Active / Total Size (% used)       : 603522.30K / 622612.05K (96.9%)
>  Minimum / Average / Maximum Object : 0.01K / 0.90K / 15.25K
> 
>   OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME                   
> 475548 467649  98%    1.21K  18722       26    599104K jfs_ip
>  25670  19143  74%    0.05K    302       85      1208K shared_policy_node
>  24612  16861  68%    0.19K   1172       21      4688K dentry
>  24426  19524  79%    0.17K   1062       23      4248K vm_area_struct
>  21636  21180  97%    0.11K    601       36      2404K sysfs_dir_cache
>  12352   9812  79%    0.06K    193       64       772K kmalloc-64
>  11684   9145  78%    0.09K    254       46      1016K anon_vma
>   9855   8734  88%    0.58K    365       27      5840K inode_cache
>   9728   9281  95%    0.01K     19      512        76K kmalloc-8
>   8932   4411  49%    0.55K    319       28      5104K radix_tree_node
>   6336   5760  90%    0.25K    198       32      1584K kmalloc-256
>   5632   5632 100%    0.02K     22      256        88K kmalloc-16
>   4998   2627  52%    0.09K    119       42       476K kmalloc-96
>   4998   3893  77%    0.04K     49      102       196K Acpi-Namespace
>   4736   3887  82%    0.03K     37      128       148K kmalloc-32
>   4144   4144 100%    0.07K     74       56       296K Acpi-ParseExt
>   3740   3740 100%    0.02K     22      170        88K numa_policy
>   3486   3023  86%    0.19K    166       21       664K kmalloc-192
>   3200   2047  63%    0.12K    100       32       400K kmalloc-128
>   2304   2074  90%    0.50K     72       32      1152K kmalloc-512
>   2136   2019  94%    0.64K     89       24      1424K proc_inode_cache
>   2080   2080 100%    0.12K     65       32       260K jfs_mp
>   2024   1890  93%    0.70K     88       23      1408K shmem_inode_cache
>   1632   1556  95%    1.00K     51       32      1632K kmalloc-1024
> 
> 
> I am wondering if anyone is feeling responsible for this bug or if the mid-term
> solution is to move away from jfs?

I also did some tests, when this bug was first reported... but I couln't
re-produce it... currently I have no idea what is wrong there.

I think moving to ext4 or xfs is the best for now...  :(


-- 
regards, TR