Re: current->sched_class->yield_task is NULL, any hint?

From: Lin Ming
Date: Wed Apr 09 2014 - 17:40:32 EST


On Wed, Apr 9, 2014 at 1:08 PM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> On Wed, Apr 09, 2014 at 09:47:42PM +0200, Peter Zijlstra wrote:
>> On Wed, Apr 09, 2014 at 10:43:32AM -0700, Lin Ming wrote:
>> > [12890.586996] [<c007a364>] (sys_sched_yield+0x0/0x90) from [<c03cbf90>] (yield+0x2c/0x30)
>> > [12890.594779] r5:c18f4000 r4:00000000
>> > [12890.598258] [<c03cbf64>] (yield+0x0/0x30) from [<c02eb650>] (netlink_broadcast_filtered+0x3ec/0x400)
>> > [12890.607170] [<c02eb264>] (netlink_broadcast_filtered+0x0/0x400) from [<c02eb68c>] (netlink_broadcast+0x28/0x30)
>> > [12890.616998] [<c02eb664>] (netlink_broadcast+0x0/0x30) from [<c02ed420>] (msg_netlink_broadcast+0xa8/0xb4)
>> > [12890.626337] [<c02ed378>] (msg_netlink_broadcast+0x0/0xb4) from [<c039ad10>] (br_igmp_send_netlink_msg+0xb0/0xbc)
>> > [12890.636256] [<c039ac60>] (br_igmp_send_netlink_msg+0x0/0xbc) from [<c039b4a0>] (br_igmp_mc_fdb_add+0x424/0x6c0)
>> > [12890.646084] [<c039b07c>] (br_igmp_mc_fdb_add+0x0/0x6c0) from [<c039bd34>] (br_igmp_process_v3+0x4a8/0x518)
>> > [12890.655515] [<c039b88c>] (br_igmp_process_v3+0x0/0x518) from [<c039c110>] (br_igmp_snooping+0x36c/0x3e4)
>> > [12890.664762] [<c039bda4>] (br_igmp_snooping+0x0/0x3e4) from [<c039c204>] (br_igmp_mc_forward+0x7c/0x270)
>> > [12890.673918] [<c039c188>] (br_igmp_mc_forward+0x0/0x270) from [<c0393120>] (br_handle_frame_finish+0x160/0x358)
>> > [12890.683654] [<c0392fc0>] (br_handle_frame_finish+0x0/0x358) from [<c0398e7c>] (br_nf_pre_routing_finish+0x348/0x364)
>> > [12890.693940] [<c0398b34>] (br_nf_pre_routing_finish+0x0/0x364) from [<c03994c4>] (br_nf_pre_routing+0x5e0/0x630)
>> > [12890.703767] r7:c2e5c000 r6:00000000 r5:d71899c0 r4:da6ed0d0
>> > [12890.709291] [<c0398ee4>] (br_nf_pre_routing+0x0/0x630) from [<c02eebf8>] (nf_iterate+0x68/0xa4)
>> > [12890.717746] [<c02eeb90>] (nf_iterate+0x0/0xa4) from [<c02eec9c>] (nf_hook_slow+0x68/0x124)
>> > [12890.725803] [<c02eec34>] (nf_hook_slow+0x0/0x124) from [<c03935ac>] (br_handle_frame+0x294/0x2e4)
>> > [12890.734471] [<c0393318>] (br_handle_frame+0x0/0x2e4) from [<c02c9eb0>] (__netif_receive_skb+0x3b0/0x514)
>> > [12890.743718] [<c02c9b00>] (__netif_receive_skb+0x0/0x514) from [<c02cabf0>] (netif_receive_skb+0xa8/0xbc)
>> > [12890.752966] [<c02cab48>] (netif_receive_skb+0x0/0xbc) from [<c02cd730>] (napi_skb_finish+0x34/0x50)
>> > [12890.761756] r5:d71899c0 r4:00000003
>> > [12890.765235] [<c02cd6fc>] (napi_skb_finish+0x0/0x50) from [<c02cd848>] (napi_gro_receive+0xfc/0x100)
>> > [12890.774056] r4:00000000 r3:00000000
>> > [12890.777566] [<c02cd74c>] (napi_gro_receive+0x0/0x100) from [<bf1a1618>] (nss_gmac_receive+0x48/0x4c [qca_nss_gmac])
>> > [12890.787759] [<bf1a15d0>] (nss_gmac_receive+0x0/0x4c [qca_nss_gmac]) from [<bf186e40>] (nss_core_handle_napi+0x440/0xa78 [qca_nss_drv])
>> > [12890.799510] r5:d71899c0 r4:bf18e410
>> > [12890.802989] [<bf186a00>] (nss_core_handle_napi+0x0/0xa78 [qca_nss_drv]) from [<c02cade4>] (net_rx_action+0x84/0x174)
>> > [12890.813244] [<c02cad60>] (net_rx_action+0x0/0x174) from [<c005ad80>] (__do_softirq+0xbc/0x160)
>> > [12890.821637] [<c005acc4>] (__do_softirq+0x0/0x160) from [<c005b250>] (irq_exit+0x58/0xa4)
>> > [12890.829542] [<c005b1f8>] (irq_exit+0x0/0xa4) from [<c000ed60>] (handle_IRQ+0x8c/0xb8)
>>
>> Look at that, its calling yield() from a non-preemptible context as
>> well.
>>
>> That code is full of fail.
>
> *phew* none of that br_igmp_* muck lives in any kernel tree near me so I
> can stop worrying :-)

It's our internal code.

>
> In any case, it wrongly passes a GFP_WAIT along to netlink_broadcast.

I'll fix it.
Thanks very much.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/