[k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[MIke Galbraith]

[-- Attachment #1: Type: text/plain, Size: 6695 bytes --]

Greetings,

Excuse me if this is already known, I've been too busy tinkering to read
lkml.

If I compile this kernel as SMP on my P4 box, I receive this BUG on
boot.  Trying to start amaroK tipped me off that something was screwy,
and I thought it was some scheduler changes I'm working on, but I'm
innocent.  The below is Virgin source.

gzipped config attached.  AmaroK isn't very happy after the BUG... and
neither is just about anything else it seems.  Hopefully I'm not just
wasting electrons typing this :)  We'll see.

	-Mike

kernel BUG at include/linux/mm.h:302!
invalid opcode: 0000 [#1]
PREEMPT SMP 
last sysfs file: /devices/pci0000:00/0000:00:1e.0/0000:02:09.0/subsystem_device
Modules linked in: xt_pkttype ipt_LOG xt_limit snd_pcm_oss snd_mixer_oss button battery snd_seq snd_seq_device ac edd ip6t_REJECT xt_tcpudp ipt_REJECT xt_state iptable_mangle iptable_nat ip_nat iptable_filter ip6table_mangle ip_conntrack ip_tables ip6table_filter ip6_tables x_tables tda9887 at76c651 prism54 saa7134 ohci1394 ieee1394 bt878 ir_kbd_i2c snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_pcm snd_timer snd soundcore snd_page_alloc i2c_i801 tuner bttv video_buf firmware_class btcx_risc ir_common tveeprom nls_iso8859_1 nls_cp437 nls_utf8 sd_mod fan thermal processor
CPU:    0
EIP:    0060:[<c01461cb>]    Not tainted VLI
EFLAGS: 00210246   (2.6.16-rc1-mm5 #1) 
EIP is at release_pages+0x165/0x196
eax: 00000000   ebx: c16f3220   ecx: 00000000   edx: 00000008
esi: c18077d0   edi: c18077b4   ebp: 00000000   esp: f6644e74
ds: 007b   es: 007b   ss: 0068
Process knotify (pid: 7482, threadinfo=f6644000 task=f658b030)
Stack: <0>00000008 00000008 00000001 00000000 00000000 c161e920 c161e940 c161e960 
       c161e980 c161e9a0 c057f600 f6644edc 00000001 00000002 f6644ed4 c01446de 
       c161e9a0 c18077d0 00000008 c16f32e0 c18077d0 00000008 00000008 c01521b5 
Call Trace:
 <c01446de> __pagevec_free+0x1f/0x2e   <c01521b5> free_pages_and_swap_cache+0x74/0xb6
 <c014afd4> unmap_vmas+0x5da/0x642   <c014de7d> unmap_region+0xa7/0x131
 <c014e5cd> do_munmap+0x139/0x1cc   <c014e6a0> sys_munmap+0x40/0x59
 <c0102bc3> sysenter_past_esp+0x54/0x75  
Code: e8 65 de fb ff e9 01 ff ff ff 83 ad a4 05 00 00 01 eb 85 85 ed 0f 84 c6 fe ff ff 8d 85 80 05 00 00 e8 05 64 32 00 e9 b6 fe ff ff <0f> 0b 2e 01 e6 fd 49 c0 eb b4 0f 0b 00 01 66 fe 49 c0 e9 14 ff 
 <6>note: knotify[7482] exited with preempt_count 1
BUG: scheduling while atomic: knotify/0x00000001/7482
 <c046aea3> schedule+0xaf1/0xe6e   <c011bd98> __call_console_drivers+0x38/0x44
 <c011c01f> release_console_sem+0x177/0x1e2   <c011ca3c> vprintk+0x266/0x36e
 <c011c058> release_console_sem+0x1b0/0x1e2   <c046bf21> rwsem_down_read_failed+0x92/0x200
 <c011f923> .text.lock.exit+0x7/0x64   <c011e498> do_exit+0x115/0x8d9
 <c0104362> do_simd_coprocessor_error+0x0/0x163   <c010477d> do_invalid_op+0x0/0xab
 <c010481f> do_invalid_op+0xa2/0xab   <c01461cb> release_pages+0x165/0x196
 <c0114715> __wake_up+0x32/0x43   <c03b0168> sock_def_readable+0x3c/0x69
 <c01037bb> error_code+0x4f/0x54   <c01461cb> release_pages+0x165/0x196
 <c01446de> __pagevec_free+0x1f/0x2e   <c01521b5> free_pages_and_swap_cache+0x74/0xb6
 <c014afd4> unmap_vmas+0x5da/0x642   <c014de7d> unmap_region+0xa7/0x131
 <c014e5cd> do_munmap+0x139/0x1cc   <c014e6a0> sys_munmap+0x40/0x59
 <c0102bc3> sysenter_past_esp+0x54/0x75  

SysRq-T

amarokapp     X C0132967     0  7898      1          7929  7767 (L-TLB)
e563ce6c 00000001 00000000 c0132967 7fffffff 00000000 f16b1b9a 000000e1 
       00000202 e563c000 f6531ab0 f669c580 dff0b740 00000a07 f664b164 f664b030 
       e458e030 c18084c0 f174a51c 000000e1 00000000 e563c000 00000000 e458e030 
Call Trace:
 <c0132967> sys_futex+0x50/0x108   <c011d870> exit_mm+0x10c/0x129
 <c011e894> do_exit+0x511/0x8d9   <c011ec85> do_group_exit+0x29/0x86
 <c0127d65> get_signal_to_deliver+0x271/0x4a9   <c0102508> do_notify_resume+0x87/0x673
 <c0418823> unix_ioctl+0x98/0xb8   <c03ac335> sock_ioctl+0xb4/0x1ec
 <c01704b6> sys_select+0xd2/0x1ac   <c0102cae> work_notifysig+0x13/0x19
amarokapp     D 00000001     0  7929      1  7927    7804  7898 (L-TLB)
e4156bf8 00000002 e4156be4 00000001 00001ef9 00000000 f1c78b0c 000000e1 
       00000000 e4156000 00000000 e4156d30 1e2c31bc 0003a1bf e458e6a4 e458e570 
       c0579000 c18084c0 f1cddb08 000000e1 00000000 e4156000 e458e570 f1c78b0c 
Call Trace:
 <c011ec11> do_exit+0x88e/0x8d9   <c01036f0> apic_timer_interrupt+0x1c/0x24
 <c0104362> do_simd_coprocessor_error+0x0/0x163   <c010477d> do_invalid_op+0x0/0xab
 <c010481f> do_invalid_op+0xa2/0xab   <c01461cb> release_pages+0x165/0x196
 <c0143a3d> free_pages_bulk+0x2f/0x2f5   <c01037bb> error_code+0x4f/0x54
 <c014007b> generic_file_buffered_write+0x3e2/0x670   <c01461cb> release_pages+0x165/0x196
 <c01521b5> free_pages_and_swap_cache+0x74/0xb6   <c014afd4> unmap_vmas+0x5da/0x642
 <c014d750> exit_mmap+0x82/0x114   <c01198f3> mmput+0x31/0xef
 <c011e498> do_exit+0x115/0x8d9   <c011ec85> do_group_exit+0x29/0x86
 <c0127d65> get_signal_to_deliver+0x271/0x4a9   <c0102508> do_notify_resume+0x87/0x673
 <c0116cbc> default_wake_function+0x0/0xc   <c0132967> sys_futex+0x50/0x108
 <c0102cae> work_notifysig+0x13/0x19  
amarokapp     X 00000000     0  7927   7929          7928       (L-TLB)
e4fa9e6c 00000000 0000000f 00000000 00000001 e53edaa8 f198dea3 000000e1 
       f38d0480 e4fa9000 e458e570 00000000 f67d7580 00016bd0 e4df6164 e4df6030 
       dffb4ab0 c18104c0 f1a636e9 000000e1 00000001 e4fa9000 00000000 dffb4ab0 
Call Trace:
 <c011e894> do_exit+0x511/0x8d9   <c046a708> schedule+0x356/0xe6e
 <c046a733> schedule+0x381/0xe6e   <c011ec85> do_group_exit+0x29/0x86
 <c0127d65> get_signal_to_deliver+0x271/0x4a9   <c0102508> do_notify_resume+0x87/0x673
 <c016926c> pipe_wait+0x85/0x8c   <c012e9c4> autoremove_wake_function+0x0/0x37
 <c0169ce7> pipe_read+0x0/0x29   <c0169d07> pipe_read+0x20/0x29
 <c015d7dd> vfs_read+0xa2/0x158   <c015e280> sys_read+0x41/0x6a
 <c0102cae> work_notifysig+0x13/0x19  
amarokapp     X 00000000     0  7928   7929                7927 (L-TLB)
e458fe6c 00000000 0000000f 00000000 00000001 e53ed92c f189a1fc 000000e1 
       f627bd80 e458f000 f6531030 dffe3580 dff0b900 00001ccc e458ebe4 e458eab0 
       e4df6030 c18104c0 f18c9ee2 000000e1 00000001 e458f000 00000000 e4df6030 
Call Trace:
 <c011e894> do_exit+0x511/0x8d9   <c046a97d> schedule+0x5cb/0xe6e
 <c011ec85> do_group_exit+0x29/0x86   <c0127d65> get_signal_to_deliver+0x271/0x4a9
 <c0102508> do_notify_resume+0x87/0x673   <c016926c> pipe_wait+0x85/0x8c
 <c012e9c4> autoremove_wake_function+0x0/0x37   <c0169ce7> pipe_read+0x0/0x29
 <c0169d07> pipe_read+0x20/0x29   <c015d7dd> vfs_read+0xa2/0x158
 <c015e280> sys_read+0x41/0x6a   <c0102cae> work_notifysig+0x13/0x19


[-- Attachment #2: config-2.6.16-rc1-mm5.gz --]
[-- Type: application/x-gzip, Size: 12223 bytes --]

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[Nick Piggin]

MIke Galbraith wrote:
> Greetings,
> 
> Excuse me if this is already known, I've been too busy tinkering to read
> lkml.
> 

It should be fixed as of current -git (not sure about the latest
-mm though). It would be good if you could verify that 2.6.16-rc2-git7
works OK for you.

Thanks,
Nick

-- 
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com 

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[Andrew Morton]

MIke Galbraith <efault@gmx.de> wrote:
>
> If I compile this kernel as SMP on my P4 box, I receive this BUG on
>  boot.  Trying to start amaroK tipped me off that something was screwy,
>  and I thought it was some scheduler changes I'm working on, but I'm
>  innocent.  The below is Virgin source.
> 
>  gzipped config attached.  AmaroK isn't very happy after the BUG... and
>  neither is just about anything else it seems.  Hopefully I'm not just
>  wasting electrons typing this :)  We'll see.
> 
>  	-Mike
> 
>  kernel BUG at include/linux/mm.h:302!
>  invalid opcode: 0000 [#1]
>  PREEMPT SMP 
>  last sysfs file: /devices/pci0000:00/0000:00:1e.0/0000:02:09.0/subsystem_device
>  Modules linked in: xt_pkttype ipt_LOG xt_limit snd_pcm_oss snd_mixer_oss button battery snd_seq snd_seq_device ac edd ip6t_REJECT xt_tcpudp ipt_REJECT xt_state iptable_mangle iptable_nat ip_nat iptable_filter ip6table_mangle ip_conntrack ip_tables ip6table_filter ip6_tables x_tables tda9887 at76c651 prism54 saa7134 ohci1394 ieee1394 bt878 ir_kbd_i2c snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_pcm snd_timer snd soundcore snd_page_alloc i2c_i801 tuner bttv video_buf firmware_class btcx_risc ir_common tveeprom nls_iso8859_1 nls_cp437 nls_utf8 sd_mod fan thermal processor
>  CPU:    0
>  EIP:    0060:[<c01461cb>]    Not tainted VLI
>  EFLAGS: 00210246   (2.6.16-rc1-mm5 #1) 
>  EIP is at release_pages+0x165/0x196
>  eax: 00000000   ebx: c16f3220   ecx: 00000000   edx: 00000008
>  esi: c18077d0   edi: c18077b4   ebp: 00000000   esp: f6644e74
>  ds: 007b   es: 007b   ss: 0068
>  Process knotify (pid: 7482, threadinfo=f6644000 task=f658b030)
>  Stack: <0>00000008 00000008 00000001 00000000 00000000 c161e920 c161e940 c161e960 
>         c161e980 c161e9a0 c057f600 f6644edc 00000001 00000002 f6644ed4 c01446de 
>         c161e9a0 c18077d0 00000008 c16f32e0 c18077d0 00000008 00000008 c01521b5 
>  Call Trace:
>   <c01446de> __pagevec_free+0x1f/0x2e   <c01521b5> free_pages_and_swap_cache+0x74/0xb6
>   <c014afd4> unmap_vmas+0x5da/0x642   <c014de7d> unmap_region+0xa7/0x131
>   <c014e5cd> do_munmap+0x139/0x1cc   <c014e6a0> sys_munmap+0x40/0x59

Doesn't happen here (but it never does).

Is it always knotify, always dying in the same manner?

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[Andrew Morton]

Nick Piggin <nickpiggin@yahoo.com.au> wrote:
>
> MIke Galbraith wrote:
> > Greetings,
> > 
> > Excuse me if this is already known, I've been too busy tinkering to read
> > lkml.
> > 
> 
> It should be fixed as of current -git (not sure about the latest
> -mm though). It would be good if you could verify that 2.6.16-rc2-git7
> works OK for you.
> 

This was a -mm kernel - how do we know it's not -mm breakage?

IOW: to what bug are you referring?

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[MIke Galbraith]

On Thu, 2006-02-09 at 19:38 +1100, Nick Piggin wrote:
> MIke Galbraith wrote:
> > Greetings,
> > 
> > Excuse me if this is already known, I've been too busy tinkering to read
> > lkml.
> > 
> 
> It should be fixed as of current -git (not sure about the latest
> -mm though). It would be good if you could verify that 2.6.16-rc2-git7
> works OK for you.

Well shoot.  I'd be happy to if I could find it.  That extension makes
me suspect I need to gitified to get there from here.  Yes?  I'm more
accustomed to digging bits and pieces out of mondo patches... no SCM
thingies here.

	-Mike  (SCM wimp)

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[MIke Galbraith]

On Thu, 2006-02-09 at 00:46 -0800, Andrew Morton wrote:
> MIke Galbraith <efault@gmx.de> wrote:
> >
> >
> >  kernel BUG at include/linux/mm.h:302!
> >  invalid opcode: 0000 [#1]
> >  PREEMPT SMP 
> >  last sysfs file: /devices/pci0000:00/0000:00:1e.0/0000:02:09.0/subsystem_device
> >  Modules linked in: xt_pkttype ipt_LOG xt_limit snd_pcm_oss snd_mixer_oss button battery snd_seq snd_seq_device ac edd ip6t_REJECT xt_tcpudp ipt_REJECT xt_state iptable_mangle iptable_nat ip_nat iptable_filter ip6table_mangle ip_conntrack ip_tables ip6table_filter ip6_tables x_tables tda9887 at76c651 prism54 saa7134 ohci1394 ieee1394 bt878 ir_kbd_i2c snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_pcm snd_timer snd soundcore snd_page_alloc i2c_i801 tuner bttv video_buf firmware_class btcx_risc ir_common tveeprom nls_iso8859_1 nls_cp437 nls_utf8 sd_mod fan thermal processor
> >  CPU:    0
> >  EIP:    0060:[<c01461cb>]    Not tainted VLI
> >  EFLAGS: 00210246   (2.6.16-rc1-mm5 #1) 
> >  EIP is at release_pages+0x165/0x196
> >  eax: 00000000   ebx: c16f3220   ecx: 00000000   edx: 00000008
> >  esi: c18077d0   edi: c18077b4   ebp: 00000000   esp: f6644e74
> >  ds: 007b   es: 007b   ss: 0068
> >  Process knotify (pid: 7482, threadinfo=f6644000 task=f658b030)
> >  Stack: <0>00000008 00000008 00000001 00000000 00000000 c161e920 c161e940 c161e960 
> >         c161e980 c161e9a0 c057f600 f6644edc 00000001 00000002 f6644ed4 c01446de 
> >         c161e9a0 c18077d0 00000008 c16f32e0 c18077d0 00000008 00000008 c01521b5 
> >  Call Trace:
> >   <c01446de> __pagevec_free+0x1f/0x2e   <c01521b5> free_pages_and_swap_cache+0x74/0xb6
> >   <c014afd4> unmap_vmas+0x5da/0x642   <c014de7d> unmap_region+0xa7/0x131
> >   <c014e5cd> do_munmap+0x139/0x1cc   <c014e6a0> sys_munmap+0x40/0x59
> 
> Doesn't happen here (but it never does).
> 
> Is it always knotify, always dying in the same manner?
> 

Yes, tentatively.  I haven't rebooted many times, but every time I have,
It's been knotify.

	-Mike

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[MIke Galbraith]

On Thu, 2006-02-09 at 00:47 -0800, Andrew Morton wrote:
> Nick Piggin <nickpiggin@yahoo.com.au> wrote:
> >
> > MIke Galbraith wrote:
> > > Greetings,
> > > 
> > > Excuse me if this is already known, I've been too busy tinkering to read
> > > lkml.
> > > 
> > 
> > It should be fixed as of current -git (not sure about the latest
> > -mm though). It would be good if you could verify that 2.6.16-rc2-git7
> > works OK for you.
> > 
> 
> This was a -mm kernel - how do we know it's not -mm breakage?

It _appears_ to be mm breakage.  I just built/ran rc1 with the same
config, and it works fine.

RL is calling, so I can't dig right this minute... in a couple hours I
hope to be able to start though.

Before I get to the 'what comes next' compile marathon, any likely
candidates?  (or Nick, do you have the supposed fix handy?)

	-Mike

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[Andrew Morton]

MIke Galbraith <efault@gmx.de> wrote:
>
> On Thu, 2006-02-09 at 00:47 -0800, Andrew Morton wrote:
> > Nick Piggin <nickpiggin@yahoo.com.au> wrote:
> > >
> > > MIke Galbraith wrote:
> > > > Greetings,
> > > > 
> > > > Excuse me if this is already known, I've been too busy tinkering to read
> > > > lkml.
> > > > 
> > > 
> > > It should be fixed as of current -git (not sure about the latest
> > > -mm though). It would be good if you could verify that 2.6.16-rc2-git7
> > > works OK for you.
> > > 
> > 
> > This was a -mm kernel - how do we know it's not -mm breakage?
> 
> It _appears_ to be mm breakage.  I just built/ran rc1 with the same
> config, and it works fine.
> 
> RL is calling, so I can't dig right this minute... in a couple hours I
> hope to be able to start though.
> 
> Before I get to the 'what comes next' compile marathon, any likely
> candidates?

rc2-mm1?

>  (or Nick, do you have the supposed fix handy?)

Yeah, I'm still scratching my head over the mystery fix.

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[Nick Piggin]

Andrew Morton wrote:
> MIke Galbraith <efault@gmx.de> wrote:
> 
>>On Thu, 2006-02-09 at 00:47 -0800, Andrew Morton wrote:

>>>This was a -mm kernel - how do we know it's not -mm breakage?
>>

It's an educated guess. I suppose it could be -mm breakage.

I sent Andrew a patch which tightens up some debug checking, and
that is likely to be causing your BUGs.

>>It _appears_ to be mm breakage.  I just built/ran rc1 with the same
>>config, and it works fine.
>>
>>RL is calling, so I can't dig right this minute... in a couple hours I
>>hope to be able to start though.
>>
>>Before I get to the 'what comes next' compile marathon, any likely
>>candidates?
> 
> 
> rc2-mm1?
> 
> 
>> (or Nick, do you have the supposed fix handy?)
> 
> 
> Yeah, I'm still scratching my head over the mystery fix.
> 
> 

The mm/swap.c hunk from git 8519fb30e438f8088b71a94a7d5a660a814d3872
is the mystery fix (the mm.h hunk is already in there).

I suppose you'd better verify that -mm works fine with the patch as
well, when you get time.

Thanks,

-- 
SUSE Labs, Novell Inc.

Send instant messages to your online friends http://au.messenger.yahoo.com 

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[MIke Galbraith]

On Thu, 2006-02-09 at 23:44 +1100, Nick Piggin wrote:
> Andrew Morton wrote:

> >> (or Nick, do you have the supposed fix handy?)
> > 
> > 
> > Yeah, I'm still scratching my head over the mystery fix.
> > 
> > 
> 
> The mm/swap.c hunk from git 8519fb30e438f8088b71a94a7d5a660a814d3872
> is the mystery fix (the mm.h hunk is already in there).
> 
> I suppose you'd better verify that -mm works fine with the patch as
> well, when you get time.

Verified.  rc2-mm1 worked fine, and plugging the extracted bit below
into rc1-mm5 fixed it's BUG.

	Thanks,

	-Mike

--- linux-2.6.16-rc1-mm5/mm/swap.c	2006-02-09 05:38:11.000000000 +0100
+++ linux-2.6.16-rc2-mm1/mm/swap.c	2006-02-09 13:04:04.000000000 +0100
@@ -34,19 +34,22 @@
 /* How many pages do we try to swap or page in/out together? */
 int page_cluster;
 
-void put_page(struct page *page)
+static void put_compound_page(struct page *page)
 {
-	if (unlikely(PageCompound(page))) {
-		page = (struct page *)page_private(page);
-		if (put_page_testzero(page)) {
-			void (*dtor)(struct page *page);
+	page = (struct page *)page_private(page);
+	if (put_page_testzero(page)) {
+		void (*dtor)(struct page *page);
 
-			dtor = (void (*)(struct page *))page[1].mapping;
-			(*dtor)(page);
-		}
-		return;
+		dtor = (void (*)(struct page *))page[1].mapping;
+		(*dtor)(page);
 	}
-	if (put_page_testzero(page))
+}
+
+void put_page(struct page *page)
+{
+	if (unlikely(PageCompound(page)))
+		put_compound_page(page);
+	else if (put_page_testzero(page))
 		__page_cache_release(page);
 }
 EXPORT_SYMBOL(put_page);
@@ -242,6 +245,15 @@
 	for (i = 0; i < nr; i++) {
 		struct page *page = pages[i];
 
+		if (unlikely(PageCompound(page))) {
+			if (zone) {
+				spin_unlock_irq(&zone->lru_lock);
+				zone = NULL;
+			}
+			put_compound_page(page);
+			continue;
+		}
+
 		if (!put_page_testzero(page))
 			continue;
 
@@ -265,7 +277,7 @@
 			}
 			__pagevec_free(&pages_to_free);
 			pagevec_reinit(&pages_to_free);
-		}
+  		}
 	}
 	if (zone)
 		spin_unlock_irq(&zone->lru_lock);

Re: [k2.6.16-rc1-mm5] kernel BUG at include/linux/mm.h:302!

[Nick Piggin]

MIke Galbraith wrote:
> On Thu, 2006-02-09 at 23:44 +1100, Nick Piggin wrote:
> 
>>Andrew Morton wrote:
> 
> 
>>>>(or Nick, do you have the supposed fix handy?)
>>>
>>>
>>>Yeah, I'm still scratching my head over the mystery fix.
>>>
>>>
>>
>>The mm/swap.c hunk from git 8519fb30e438f8088b71a94a7d5a660a814d3872
>>is the mystery fix (the mm.h hunk is already in there).
>>
>>I suppose you'd better verify that -mm works fine with the patch as
>>well, when you get time.
> 
> 
> Verified.  rc2-mm1 worked fine, and plugging the extracted bit below
> into rc1-mm5 fixed it's BUG.
> 

Thanks. And thanks for reporting.

-- 
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com