Hi Peter,
I get the following warning on s390 when using fake NUMA beginning with
your patch
e9d867a67fd0 "sched: Allow per-cpu kernel threads to run on online && !active"
[ 3.162909] WARNING: CPU: 0 PID: 1 at include/linux/cpumask.h:121
select_task_rq+0xe6/0x1a8
[ 3.162911] Modules linked in:
[ 3.162914] CPU: 0 PID: 1 Comm: swapper/0 Not tainted
4.6.0-rc6-00001-ge9d867a67fd0-dirty #28
[ 3.162917] task: 00000001dd270008 ti: 00000001eccb4000 task.ti:
00000001eccb4000
[ 3.162918] Krnl PSW : 0404c00180000000 0000000000176c56
(select_task_rq+0xe6/0x1a8)
[ 3.162923] R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:3 CC:0 PM:0
RI:0 EA:3
Krnl GPRS: 00000000009f3d00 00000001dd270008 0000000000000100
00000000000000f4
[ 3.162927] 0000000000eaf4e2 0000000000000000 00000001eccb7bc0
0400000000000001
[ 3.162929] 00000001ec660950 0000000000000000 00000001ec660520
00000001ec660008
[ 3.162930] 0000000000000100 000000000099c1a0 0000000000176c30
00000001eccb7a70
[ 3.162940] Krnl Code: 0000000000176c4a: a774002f brc 7,176ca8
0000000000176c4e: 92014000 mvi 0(%r4),1
#0000000000176c52: a7f40001 brc
15,176c54
>0000000000176c56: a7f40029 brc
15,176ca8
0000000000176c5a: 95004000 cli 0(%r4),0
0000000000176c5e: a7740006 brc 7,176c6a
0000000000176c62: 92014000 mvi 0(%r4),1
0000000000176c66: a7f40001 brc
15,176c68
[ 3.162958] Call Trace:
[ 3.162961] ([<0000000000176c30>] select_task_rq+0xc0/0x1a8)
[ 3.162963] ([<0000000000177d64>] try_to_wake_up+0x2e4/0x478)
[ 3.162968] ([<000000000015d46c>] create_worker+0x174/0x1c0)
[ 3.162971] ([<0000000000161a98>] alloc_unbound_pwq+0x360/0x438)
[ 3.162973] ([<0000000000162550>] apply_wqattrs_prepare+0x200/0x2a0)
[ 3.162975] ([<000000000016266a>] apply_workqueue_attrs_locked+0x7a/0xb0)
[ 3.162977] ([<0000000000162af0>] apply_workqueue_attrs+0x50/0x78)
[ 3.162979] ([<000000000016441c>] __alloc_workqueue_key+0x304/0x520)
[ 3.162983] ([<0000000000ee3706>] default_bdi_init+0x3e/0x70)
[ 3.162986] ([<0000000000100270>] do_one_initcall+0x140/0x1d8)
[ 3.162990] ([<0000000000ec9da8>] kernel_init_freeable+0x220/0x2d8)
[ 3.162993] ([<0000000000984a7a>] kernel_init+0x2a/0x150)
[ 3.162996] ([<00000000009913fa>] kernel_thread_starter+0x6/0xc)
[ 3.162998] ([<00000000009913f4>] kernel_thread_starter+0x0/0xc)
[ 3.163000] 4 locks held by swapper/0/1:
[ 3.163002] #0: (cpu_hotplug.lock){++++++}, at: [<000000000013ebe0>]
get_online_cpus+0x48/0xb8
[ 3.163010] #1: (wq_pool_mutex){+.+.+.}, at: [<0000000000162ae2>]
apply_workqueue_attrs+0x42/0x78
[ 3.163016] #2: (&pool->lock/1){......}, at: [<000000000015d44a>]
create_worker+0x152/0x1c0
[ 3.163022] #3: (&p->pi_lock){..-...}, at: [<0000000000177ac4>]
try_to_wake_up+0x44/0x478
[ 3.163028] Last Breaking-Event-Address:
[ 3.163030] [<0000000000176c52>] select_task_rq+0xe2/0x1a8
For some unknown reason select_task_rq() gets called with a task that has
nr_cpus_allowed == 0. Hence "cpu = cpumask_any(tsk_cpus_allowed(p));"
within select_task_rq() will set cpu to nr_cpu_ids which in turn causes the
warning later on.
It only happens with more than one node, otherwise it seems to work fine.
Any idea what could be wrong here?
