Hi Hanjun,

a update here, tested on 4.9,

>
>  - Applied Ard's two patches only
>  - Applied Robert's patch only
>
> Both of them can work fine on D05 with NUMA enabled, which means
> boot ok and LTP MM stress test is passed.
>

It's not related to this patch set.
LTP "cpuset01" test  crashes with latest 4.9,  4.10-rc1 and 4.10-rc2
kernels on Thunderx 2S .  Do you see any such behaviour on D05.
Any idea what might be causing this issue.


  227.627546] cpuset01: page allocation stalls for 10096ms, order:0,
mode:0x24200ca(GFP_HIGHUSER_MOVABLE)
[  227.627586] CPU: 53 PID: 11017 Comm: cpuset01 Not tainted 4.9.04kNUMA+ #2
[  227.627591] Hardware name: www.cavium.com ThunderX Unknown/ThunderX
Unknown, BIOS 0.3 Aug 24 2016
[  227.627599] Call trace:
[  227.627623] [<ffff000008089f10>] dump_backtrace+0x0/0x238
[  227.627640] [<ffff00000808a16c>] show_stack+0x24/0x30
[  227.627656] [<ffff00000846fb50>] dump_stack+0x94/0xb4
[  227.627679] [<ffff0000081eb4f8>] warn_alloc+0x138/0x150
[  227.627686] [<ffff0000081ec0a4>] __alloc_pages_nodemask+0xb04/0xcf0
[  227.627697] [<ffff000008245988>] alloc_pages_vma+0xc8/0x270
[  227.627715] [<ffff00000821f604>] handle_mm_fault+0xc8c/0xfd8
[  227.627732] [<ffff00000809a488>] do_page_fault+0x2c0/0x368
[  227.627744] [<ffff0000080812ec>] do_mem_abort+0x6c/0xe0
[  227.627752] Exception stack(0xffff801f55823e00 to 0xffff801f55823f30)
[  227.627763] 3e00: 0000000000000000 0000ffff92682000
ffffffffffffffff 0000ffff9252b3e8
[  227.627774] 3e20: 0000000020000000 0000000000000000
000000000000a000 0000000000000003
[  227.627785] 3e40: 0000000000000022 ffffffffffffffff
0000000000000123 00000000000000de
[  227.627793] 3e60: ffff000008972000 0000000000000015
ffff801f55823e90 0000000000040900
[  227.627800] 3e80: 0000000000000000 ffff0000080836f0
0000000000000000 0000ffff92682000
[  227.627809] 3ea0: ffffffffffffffff 0000ffff92575d8c
0000000000000000 0000000000040900
[  227.627819] 3ec0: 0000ffff92682000 00000000000000f7
0000000000004fc0 0000000000000022
[  227.627828] 3ee0: 0000000000000000 0000000000000000
0000ffff925f5508 f7f7f7f7f7f7f7f7
[  227.627838] 3f00: 0000ffff92686ff0 0000000000002ab8
0101010101010101 0000000000000020
[  227.627847] 3f20: 0000000000000000 0000000000000000
[  227.627858] [<ffff000008083324>] el0_da+0x18/0x1c
[  227.627865] Mem-Info:
[  227.627899] active_anon:38613 inactive_anon:8174 isolated_anon:0
                active_file:25148 inactive_file:64173 isolated_file:0
                unevictable:742 dirty:0 writeback:0 unstable:0
                slab_reclaimable:29066 slab_unreclaimable:67304
                mapped:22876 shmem:2597 pagetables:1240 bounce:0
                free:65582521 free_pcp:1834 free_cma:0



>
> I'm not familiar with memory management, it's up to you guys to make
> a decision :)
>
>
> Thanks
> Hanjun
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>