2.6.27-rc8 hot memory remove panic

2.6.27-rc8 hot memory remove panic

[Badari Pulavarty]

Hi,

Ran into this while testing hotplug memory remove on 2.6.27-rc8.
Never saw this earlier.

Any ideas on whats happening.

put_page_bootmem():
        BUG_ON(type >= -1);

Thanks,
Badari


------------[ cut here ]------------
kernel BUG at mm/memory_hotplug.c:78!
Oops: Exception in kernel mode, sig: 5 [#1]
SMP NR_CPUS=32 pSeries
Modules linked in:
NIP: c0000000001015b8 LR: c0000000000f50b8 CTR: c000000000121cbc
REGS: c00000005a5334b0 TRAP: 0700   Not tainted  (2.6.27-rc8)
MSR: 8000000000029032 <EE,ME,IR,DR>  CR: 24008442  XER: 00000010
TASK = c00000006f0655a0[13407] 'drmgr' THREAD: c00000005a530000 CPU: 0
GPR00: 0000000000000001 c00000005a533730 c0000000009f1290 c000000001355930 
GPR04: 0000000001730000 c000000071e00d20 c00000000028c318 0000000000000000 
GPR08: c000000000a62680 6db6db6db6db6db7 c000000000a96020 0000000000000000 
GPR12: 0000000024000482 c000000000a1a400 0000000000000000 0000000000000000 
GPR16: 00000000100d58f8 00000000100d5628 0000000010070000 0000000000000000 
GPR20: 00000000100b0670 00000000ffffffff 0000000000000000 0000000000000000 
GPR24: c0000000008ad000 0000000000000001 0000000000000000 000000000006a000 
GPR28: c000000002a2a000 0000000000000001 c000000000959ed0 c000000001355968 
NIP [c0000000001015b8] .put_page_bootmem+0x24/0x90
LR [c0000000000f50b8] .sparse_remove_one_section+0x1e4/0x21c
Call Trace:
[c00000005a533730] [c00000005a5337c0] 0xc00000005a5337c0 (unreliable)
[c00000005a5337b0] [c0000000000f50b8] .sparse_remove_one_section+0x1e4/0x21c
[c00000005a533850] [c000000000100ef8] .__remove_pages+0xf8/0x160
[c00000005a533900] [c00000000005222c] .pseries_remove_lmb+0x78/0xd0
[c00000005a533980] [c000000000052390] .pseries_memory_notifier+0x10c/0x1f0
[c00000005a533a10] [c0000000000a2874] .notifier_call_chain+0x7c/0xe4
[c00000005a533ab0] [c0000000000a2c84] .__blocking_notifier_call_chain+0x6c/0xa8
[c00000005a533b50] [c000000000049d58] .ofdt_write+0x4f4/0x804
[c00000005a533c60] [c00000000015ae60] .proc_reg_write+0xf4/0x130
[c00000005a533d00] [c000000000106c00] .vfs_write+0xe8/0x1a4
[c00000005a533d90] [c000000000106dc4] .sys_write+0x54/0x98
[c00000005a533e30] [c000000000008748] syscall_exit+0x0/0x40
Instruction dump:
ebe1fff8 7c0803a6 4e800020 7c0802a6 f8010010 f821ff81 8003000c 7c0007b4 
2f80ffff 38000001 409c0008 38000000 <0b000000> 38030008 7c2004ac 7d200028 
---[ end trace 23a803b7faaa91ee ]---


--
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>

Re: 2.6.27-rc8 hot memory remove panic

[Badari Pulavarty]

On Thu, 2008-10-02 at 10:23 -0700, Badari Pulavarty wrote:
> Hi,
> 
> Ran into this while testing hotplug memory remove on 2.6.27-rc8.
> Never saw this earlier.
> 
> Any ideas on whats happening.
> 
> put_page_bootmem():
>         BUG_ON(type >= -1);


It looks like we have undocumented dependency on CONFIG_NUMA=y to get
hotplug memory remove working.

register_page_bootmem_info_node() gets called only if 
CONFIG_NEED_MULTIPLE_NODES=y which gets selected only if CONFIG_NUMA=y.

Thanks,
Badari

> 
> Thanks,
> Badari
> 
> 
> ------------[ cut here ]------------
> kernel BUG at mm/memory_hotplug.c:78!
> Oops: Exception in kernel mode, sig: 5 [#1]
> SMP NR_CPUS=32 pSeries
> Modules linked in:
> NIP: c0000000001015b8 LR: c0000000000f50b8 CTR: c000000000121cbc
> REGS: c00000005a5334b0 TRAP: 0700   Not tainted  (2.6.27-rc8)
> MSR: 8000000000029032 <EE,ME,IR,DR>  CR: 24008442  XER: 00000010
> TASK = c00000006f0655a0[13407] 'drmgr' THREAD: c00000005a530000 CPU: 0
> GPR00: 0000000000000001 c00000005a533730 c0000000009f1290 c000000001355930 
> GPR04: 0000000001730000 c000000071e00d20 c00000000028c318 0000000000000000 
> GPR08: c000000000a62680 6db6db6db6db6db7 c000000000a96020 0000000000000000 
> GPR12: 0000000024000482 c000000000a1a400 0000000000000000 0000000000000000 
> GPR16: 00000000100d58f8 00000000100d5628 0000000010070000 0000000000000000 
> GPR20: 00000000100b0670 00000000ffffffff 0000000000000000 0000000000000000 
> GPR24: c0000000008ad000 0000000000000001 0000000000000000 000000000006a000 
> GPR28: c000000002a2a000 0000000000000001 c000000000959ed0 c000000001355968 
> NIP [c0000000001015b8] .put_page_bootmem+0x24/0x90
> LR [c0000000000f50b8] .sparse_remove_one_section+0x1e4/0x21c
> Call Trace:
> [c00000005a533730] [c00000005a5337c0] 0xc00000005a5337c0 (unreliable)
> [c00000005a5337b0] [c0000000000f50b8] .sparse_remove_one_section+0x1e4/0x21c
> [c00000005a533850] [c000000000100ef8] .__remove_pages+0xf8/0x160
> [c00000005a533900] [c00000000005222c] .pseries_remove_lmb+0x78/0xd0
> [c00000005a533980] [c000000000052390] .pseries_memory_notifier+0x10c/0x1f0
> [c00000005a533a10] [c0000000000a2874] .notifier_call_chain+0x7c/0xe4
> [c00000005a533ab0] [c0000000000a2c84] .__blocking_notifier_call_chain+0x6c/0xa8
> [c00000005a533b50] [c000000000049d58] .ofdt_write+0x4f4/0x804
> [c00000005a533c60] [c00000000015ae60] .proc_reg_write+0xf4/0x130
> [c00000005a533d00] [c000000000106c00] .vfs_write+0xe8/0x1a4
> [c00000005a533d90] [c000000000106dc4] .sys_write+0x54/0x98
> [c00000005a533e30] [c000000000008748] syscall_exit+0x0/0x40
> Instruction dump:
> ebe1fff8 7c0803a6 4e800020 7c0802a6 f8010010 f821ff81 8003000c 7c0007b4 
> 2f80ffff 38000001 409c0008 38000000 <0b000000> 38030008 7c2004ac 7d200028 
> ---[ end trace 23a803b7faaa91ee ]---
> 
> 
> --
> 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>

--
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>

Re: 2.6.27-rc8 hot memory remove panic

[Yasunori Goto]

> On Thu, 2008-10-02 at 10:23 -0700, Badari Pulavarty wrote:
> > Hi,
> > 
> > Ran into this while testing hotplug memory remove on 2.6.27-rc8.
> > Never saw this earlier.
> > 
> > Any ideas on whats happening.
> > 
> > put_page_bootmem():
> >         BUG_ON(type >= -1);
> 
> 
> It looks like we have undocumented dependency on CONFIG_NUMA=y to get
> hotplug memory remove working.
> 
> register_page_bootmem_info_node() gets called only if 
> CONFIG_NEED_MULTIPLE_NODES=y which gets selected only if CONFIG_NUMA=y.

Oops. My bad....

I remember its dependency is removed from Kconfig some month ago.
Then, register_page_bootmem_info_node() should be called when CONFIG_NUMA
is off.
Is this patch enough?

---

register_page_bootmem_info_node() should be called for memory hot-remove
even when CONFIG_NUMA is off.

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

---
 mm/bootmem.c |    1 +
 1 file changed, 1 insertion(+)

Index: current/mm/bootmem.c
===================================================================
--- current.orig/mm/bootmem.c	2008-10-03 09:58:58.000000000 +0900
+++ current/mm/bootmem.c	2008-10-03 11:57:21.000000000 +0900
@@ -222,6 +222,7 @@ unsigned long __init free_all_bootmem_no
  */
 unsigned long __init free_all_bootmem(void)
 {
+	register_page_bootmem_info_node(NODE_DATA(0));
 	return free_all_bootmem_core(NODE_DATA(0)->bdata);
 }
 

-- 
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>