| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <1239479565.8206.25.camel@localhost.localdomain> Date: Sat, 11 Apr 2009 21:52:44 +0200 From: Marcel Holtmann <marcel@...tmann.org> To: Dave Young <hidave.darkstar@...il.com> Cc: Thomas Gleixner <tglx@...utronix.de>, Maciej Rutecki <maciej.rutecki@...il.com>, Andrew Morton <akpm@...ux-foundation.org>, Linux Kernel Mailing List <linux-kernel@...r.kernel.org>, linux-bluetooth@...r.kernel.org Subject: Re: [2.6.29-git12] BUG: scheduling while atomic: swapper/0/0x10000100 Hi Dave, > > On Fri, 10 Apr 2009, Maciej Rutecki wrote: > >> 2009/4/10 Thomas Gleixner <tglx@...utronix.de>: > >> > >> [...] > >> > Please upload trace.bz2 file to some place. > >> > >> http://unixy.pl/maciek/download/kernel/2.6.30-rc1-git3/laptop/trace-2.6.30-rc1-git3.bz2 > > > > Ok. That gives us the necessary information: > > > > <idle>-0 1d... 76597263us : do_softirq <-irq_exit > > <idle>-0 1d... 76597263us : __do_softirq <-do_softirq > > <idle>-0 1..s. 76597264us : tasklet_action <-__do_softirq > > <idle>-0 1..s. 76597265us : hci_rx_task <-tasklet_action > > <idle>-0 1..s. 76597266us : _read_lock <-hci_rx_task > > <idle>-0 1..s. 76597267us : skb_dequeue <-hci_rx_task > > <idle>-0 1..s. 76597268us : _spin_lock_irqsave <-skb_dequeue > > <idle>-0 1d.s. 76597269us : _spin_unlock_irqrestore <-skb_dequeue > > <idle>-0 1..s. 76597270us : hci_send_to_sock <-hci_rx_task > > <idle>-0 1..s. 76597270us+: _read_lock <-hci_send_to_sock > > <idle>-0 1..s. 76597272us : skb_clone <-hci_send_to_sock > > <idle>-0 1..s. 76597273us : kmem_cache_alloc <-skb_clone > > <idle>-0 1..s. 76597274us : __skb_clone <-skb_clone > > <idle>-0 1..s. 76597275us : __copy_skb_header <-__skb_clone > > <idle>-0 1..s. 76597276us : skb_push <-hci_send_to_sock > > <idle>-0 1..s. 76597277us : sock_queue_rcv_skb <-hci_send_to_sock > > <idle>-0 1..s. 76597278us : sk_filter <-sock_queue_rcv_skb > > <idle>-0 1..s. 76597279us : security_sock_rcv_skb <-sk_filter > > <idle>-0 1..s. 76597280us : cap_socket_sock_rcv_skb <-security_sock_rcv_skb > > <idle>-0 1..s. 76597280us : local_bh_disable <-sk_filter > > <idle>-0 1..s. 76597281us : local_bh_enable <-sk_filter > > <idle>-0 1..s. 76597282us : skb_queue_tail <-sock_queue_rcv_skb > > <idle>-0 1..s. 76597283us : _spin_lock_irqsave <-skb_queue_tail > > <idle>-0 1d.s. 76597284us : _spin_unlock_irqrestore <-skb_queue_tail > > <idle>-0 1..s. 76597285us : sock_def_readable <-sock_queue_rcv_skb > > <idle>-0 1..s. 76597286us : _read_lock <-sock_def_readable > > <idle>-0 1..s. 76597286us : __wake_up_sync_key <-sock_def_readable > > <idle>-0 1..s. 76597287us : _spin_lock_irqsave <-__wake_up_sync_key > > <idle>-0 1d.s. 76597288us : __wake_up_common <-__wake_up_sync_key > > <idle>-0 1d.s. 76597289us : pollwake <-__wake_up_common > > <idle>-0 1d.s. 76597290us : default_wake_function <-pollwake > > <idle>-0 1d.s. 76597290us : try_to_wake_up <-default_wake_function > > <idle>-0 1d.s. 76597292us : update_shares <-try_to_wake_up > > <idle>-0 1d.s. 76597293us : walk_tg_tree <-update_shares > > <idle>-0 1d.s. 76597294us : tg_nop <-walk_tg_tree > > <idle>-0 1d.s. 76597294us : tg_nop <-walk_tg_tree > > <idle>-0 1d.s. 76597295us : tg_shares_up <-walk_tg_tree > > <idle>-0 1d.s. 76597297us : tg_nop <-walk_tg_tree > > <idle>-0 1d.s. 76597297us : tg_shares_up <-walk_tg_tree > > <idle>-0 1d.s. 76597298us : tg_nop <-walk_tg_tree > > <idle>-0 1d.s. 76597299us : tg_shares_up <-walk_tg_tree > > <idle>-0 1d.s. 76597300us : tg_nop <-walk_tg_tree > > <idle>-0 1d.s. 76597301us : tg_shares_up <-walk_tg_tree > > <idle>-0 1d.s. 76597302us : tg_nop <-walk_tg_tree > > <idle>-0 1d.s. 76597303us : tg_shares_up <-walk_tg_tree > > <idle>-0 1d.s. 76597304us : tg_shares_up <-walk_tg_tree > > <idle>-0 1d.s. 76597305us+: task_rq_lock <-try_to_wake_up > > <idle>-0 1d.s. 76597314us : _spin_lock <-task_rq_lock > > <idle>-0 1d.s. 76597315us : select_task_rq_fair <-try_to_wake_up > > <idle>-0 1d.s. 76597316us : activate_task <-try_to_wake_up > > <idle>-0 1d.s. 76597317us : enqueue_task_fair <-activate_task > > <idle>-0 1d.s. 76597317us : enqueue_entity <-enqueue_task_fair > > <idle>-0 1d.s. 76597318us : update_curr <-enqueue_entity > > <idle>-0 1d.s. 76597319us : place_entity <-enqueue_entity > > <idle>-0 1d.s. 76597320us : enqueue_entity <-enqueue_task_fair > > <idle>-0 1d.s. 76597321us : update_curr <-enqueue_entity > > <idle>-0 1d.s. 76597321us : place_entity <-enqueue_entity > > <idle>-0 1d.s. 76597322us : check_preempt_curr_idle <-try_to_wake_up > > <idle>-0 1d.s. 76597323us : resched_task <-check_preempt_curr_idle > > <idle>-0 1dNs. 76597324us : _spin_unlock_irqrestore <-try_to_wake_up > > <idle>-0 1dNs. 76597325us+: _spin_unlock_irqrestore <-__wake_up_sync_key > > <idle>-0 1.Ns. 76597326us : hci_event_packet <-hci_rx_task > > <idle>-0 1.Ns. 76597327us : skb_pull <-hci_event_packet > > <idle>-0 1.Ns. 76597328us : _spin_lock <-hci_event_packet > > <idle>-0 1.Ns. 76597330us : del_timer <-hci_event_packet > > <idle>-0 1.Ns. 76597331us : hci_conn_add_sysfs <-hci_event_packet > > <idle>-0 1.Ns. 76597332us+: dev_set_name <-hci_conn_add_sysfs > > Marcel, this is caused by hci conn id changes, originally the conn id > is like "acl:xx:xx:xx:xx:xx:xx", then it was changed to > "hdev_name:connection_handle", yes use the handle looks better, but it > need to be done in the event handling function. > > What do you think about change back to original way plus hdev_name? > like: "hdev_name:xx:xx:xx:xx:xx:xx" that is actually a bad idea since in theory we can have more than one SCO connection in the end. So we should use the handle. Also with the upcoming High Speed or Low Energy stuff we need to differentiate by the handle since more and more connections are possible. We should fix the real problem here and to proper reference counting for the kobject with kref here. This code might need bigger fixing since when it was creating for 2.4, the driver model was not present. Feel free to propose something. I am open for ideas. Regards Marcel -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@...r.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists