Re: [PATCH] hotplug: fix oops when adding cpu.

["tj@kernel.org" <tj@kernel.org>]

Hello,

On Wed, Jun 17, 2015 at 06:04:06PM +0800, songxiumiao@inspur.com wrote:
> Subject: [PATCH] hotplug: fix oops when adding cpu
> From: SongXiumiao <songxiumiao@inspur.com>
> 
> If memory is not in node0 and a cpu is logically hotadded, the kernel oopses. 
> 
> By analysing the bug function call trace,we find that create_worker function 
> will alloc the memory from node0.Because node0 is offline,the allocation is 
> failed.Then we add a condition to ensure the node is online and 
> system can alloc memory from a node that is online. 
> 
> Here's the backtrace:
> [root@localhost ~]# echo 1 > /sys/devices/system/cpu/cpu90/online
...
> [  225.755126] BUG: unable to handle kernel paging request at 0000000000001b08
> [  225.762931] IP: [<ffffffff81182597>] __alloc_pages_nodemask+0xb7/0x940
> [  225.770247] PGD 449bb0067 PUD 46110e067 PMD 0
> [  225.775248] Oops: 0000 [#1] SMP
> [  225.778875] Modules linked in: xt_CHECKSUM ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 ipt_REJECT nf_reject_ipv4 nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntracd
> [  225.868198] CPU: 43 PID: 5400 Comm: bash Not tainted 4.0.0-rc4-bug-fixed-remove #16
> [  225.876754] Hardware name: Insyde Brickland/Type2 - Board Product Name1, BIOS Brickland.05.04.15.0024 02/28/2015
> [  225.888122] task: ffff88045a3d8da0 ti: ffff880446120000 task.ti: ffff880446120000
> [  225.896484] RIP: 0010:[<ffffffff81182597>]  [<ffffffff81182597>] __alloc_pages_nodemask+0xb7/0x940
...
> [  226.002904] Call Trace:
> [  226.019173]  [<ffffffff811d3957>] new_slab+0xa7/0x460
> [  226.024826]  [<ffffffff81678c75>] __slab_alloc+0x310/0x470
> [  226.043812]  [<ffffffff811d46c1>] kmem_cache_alloc_node_trace+0x91/0x250
> [  226.051299]  [<ffffffff8108dd31>] alloc_worker+0x21/0x50
> [  226.057236]  [<ffffffff8108ff23>] create_worker+0x53/0x1e0
> [  226.063357]  [<ffffffff81092092>] alloc_unbound_pwq+0x2a2/0x510
> [  226.069974]  [<ffffffff810924b4>] wq_update_unbound_numa+0x1b4/0x220
> [  226.077076]  [<ffffffff81092828>] workqueue_cpu_up_callback+0x308/0x3d0
> [  226.084468]  [<ffffffff8109784e>] notifier_call_chain+0x4e/0x80
> [  226.091084]  [<ffffffff8109796e>] __raw_notifier_call_chain+0xe/0x10
> [  226.098189]  [<ffffffff810774f3>] cpu_notify+0x23/0x50
...
> diff --git a/kernel/workqueue.c b/kernel/workqueue.c
> index 586ad91..22d194c 100644
> --- a/kernel/workqueue.c
> +++ b/kernel/workqueue.c
> @@ -3253,7 +3253,8 @@ static struct worker_pool *get_unbound_pool(const struct workqueue_attrs *attrs)
> if (wq_numa_enabled) {
> for_each_node(node) {
> if (cpumask_subset(pool->attrs->cpumask,
> -    wq_numa_possible_cpumask[node])) {
> +     wq_numa_possible_cpumask[node])
> +        && node_online(node)) {
> pool->node = node;
> break;

Umm... can you please talk with the fujitsu guys?  They're working on
the same problem and this isn't the full solution.  What happens if
the node goes down after pool is initialized?

Thanks.

-- 
tejun