Re: Kernel panic with 3.13.2 and SLAB

[Helge Deller <deller@gmx.de>]

On 02/11/2014 11:52 AM, Guy Martin wrote:
> Hi,
> 
> I've recently upgraded from 3.12.4 to 3.13.2 and the kernel did not boot with the exact same config.

There has often been problems reported with 

WARN_ON(page_to_nid(virt_to_page(objp)) != nodeid);
would be interesting to know what the values of nodeid and page_to_nid() are..

Anyway, I think this needs debugging by someone who knows the details of SLAB.

And, please make sure that this patch is included 
http://git.kernel.org/cgit/linux/kernel/git/deller/parisc-linux.git/commit/?h=for-next&id=57737c49dd72c96cfbcd4f66559f3ffc399aeb4f

Helge


> Here is an excerpt of the panic :
> 
> [    9.080000] ------------[ cut here ]------------
> [    9.208000] WARNING: at mm/slab.c:2614
> [    9.208000] Modules linked in:
> [    9.348000] CPU: 3 PID: 1 Comm: swapper/0 Not tainted 3.13.2 #1
> [    9.348000] task: 000000007fc75178 ti: 000000007fc78000 task.ti: 000000007fc78000
> [    9.488000]
> [    9.488000]      YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
> [    9.624000] PSW: 00001000000001000000000000001110 Not tainted
> [    9.624000] r00-03  000000000804000e 0000000000000000 00000000402862b4 000000007fc78650
> [    9.764000] r04-07  000000004091a300 000000007fc02080 00000001437e40e0 0000000040a64030
> [    9.904000] r08-11  000000007fd4da00 00000001437e4100 000000000000000f 000f000000000000
> [   10.044000] r12-15  000000007fc02088 ffffffffc0000000 0000000000100100 000000004096a300
> [   10.044000] r16-19  0000000040969b00 00000000000001ff 0000000040969b00 0000000000000001
> [   10.184000] r20-23  0000000000000000 000000007fc020a8 0000000000000001 bff405fd00000000
> [   10.320000] r24-27  000000007fc78740 00000000006fc81c 00000000007fc020 000000004091a300
> [   10.460000] r28-31  000000023f804040 00000000409c2740 000000007fc78770 000000007fc020a8
> [   10.604000] sr00-03  0000000000000000 0000000000000000 0000000000000000 0000000000000000
> [   10.604000] sr04-07  0000000000000000 0000000000000000 0000000000000000 0000000000000000
> [   10.740000]
> [   10.740000] IASQ: 0000000000000000 0000000000000000 IAOQ: 00000000402864f8 00000000402864fc
> [   10.880000]  IIR: 03ffe01f    ISR: 0000000010240000  IOR: 0000040df93e40f8
> [   11.016000]  CPU:        3   CR30: 000000007fc78000 CR31: ffffffffdffffaef
> [   11.156000]  ORIG_R28: 000000004096a300
> [   11.156000]  IAOQ[0]: cache_alloc_refill+0x4c8/0xbf8
> [   11.296000]  IAOQ[1]: cache_alloc_refill+0x4cc/0xbf8
> [   11.296000]  RP(r2): cache_alloc_refill+0x284/0xbf8
> [   11.436000] Backtrace:
> [   11.436000]  [<0000000040285fa8>] kmem_cache_alloc+0x1c8/0x250
> [   11.436000]  [<0000000040257100>] kmem_cache_create_memcg+0xd0/0x3a8
> [   11.576000]  [<000000004025742c>] kmem_cache_create+0x54/0x68
> [   11.712000]  [<000000004011daf4>] 0x4011daf4
> [   11.712000]  [<000000004015d0b4>] do_one_initcall+0x1ac/0x230
> [   11.756000] random: nonblocking pool is initialized
> [   11.852000]  [<0000000040101a5c>] 0x40101a5c
> [   11.992000]  [<0000000040152d38>] kernel_init+0x30/0x2a0
> [   11.992000]  [<000000004014b020>] end_fault_vector+0x20/0xc0
> [   12.132000]  [<0000000040152d08>] kernel_init+0x0/0x2a0
> [   12.132000]
> [   12.132000] ---[ end trace 436196ee1e834c21 ]---
> 
> 
> 
> The full kernel logs can be found here : http://bpaste.net/show/177607/
> 
> 
> And kconfig here : http://bpaste.net/show/177929/
> 
> 
> I switched from SLAB to SLUB and the problem went away. However, this used to work in 3.12.4.