Potential iwlwifi locking problem in 2.6.26-git9 - Kernel

This is a discussion on Potential iwlwifi locking problem in 2.6.26-git9 - Kernel ; Is this a new one? I don't think I've seen this on LKML.... - Ted [ 728.456792] [ 728.456795] ============================================= [ 728.456810] [ INFO: possible recursive locking detected ] [ 728.456818] 2.6.26-05781-g44ea66a #25 [ 728.456823] --------------------------------------------- [ 728.456830] NetworkManager/8287 is ...

+ Reply to Thread
Results 1 to 8 of 8

Thread: Potential iwlwifi locking problem in 2.6.26-git9

  1. Potential iwlwifi locking problem in 2.6.26-git9


    Is this a new one? I don't think I've seen this on LKML....

    - Ted

    [ 728.456792]
    [ 728.456795] =============================================
    [ 728.456810] [ INFO: possible recursive locking detected ]
    [ 728.456818] 2.6.26-05781-g44ea66a #25
    [ 728.456823] ---------------------------------------------
    [ 728.456830] NetworkManager/8287 is trying to acquire lock:
    [ 728.456836] (&dev->addr_list_lock){-...}, at: [] dev_mc_sync+0x1c/0x59
    [ 728.456867]
    [ 728.456869] but task is already holding lock:
    [ 728.456875] (&dev->addr_list_lock){-...}, at: [] dev_set_rx_mode+0x19/0x2b
    [ 728.456894]
    [ 728.456895] other info that might help us debug this:
    [ 728.456902] 2 locks held by NetworkManager/8287:
    [ 728.456908] #0: (rtnl_mutex){--..}, at: [] rtnl_lock+0x14/0x16
    [ 728.456929] #1: (&dev->addr_list_lock){-...}, at: [] dev_set_rx_mode+0x19/0x2b
    [ 728.456949]
    [ 728.456951] stack backtrace:
    [ 728.456961] Pid: 8287, comm: NetworkManager Not tainted 2.6.26-05781-g44ea66a #25
    [ 728.456999] [] __lock_acquire+0x99f/0xcc5
    [ 728.457037] [] ? trace_hardirqs_on+0xb/0xd
    [ 728.457054] [] ? trace_hardirqs_on_caller+0xd2/0x11d
    [ 728.457076] [] lock_acquire+0x6a/0x90
    [ 728.457091] [] ? dev_mc_sync+0x1c/0x59
    [ 728.457109] [] _spin_lock_bh+0x26/0x53
    [ 728.457126] [] ? dev_mc_sync+0x1c/0x59
    [ 728.457143] [] dev_mc_sync+0x1c/0x59
    [ 728.457159] [] ieee80211_set_multicast_list+0xca/0xd2 [mac80211]
    [ 728.457223] [] __dev_set_rx_mode+0x7e/0x81
    [ 728.457239] [] dev_set_rx_mode+0x20/0x2b
    [ 728.457255] [] dev_open+0x84/0x9e
    [ 728.457269] [] dev_change_flags+0xa1/0x14f
    [ 728.457286] [] do_setlink+0x250/0x300
    [ 728.457308] [] ? _read_unlock+0x22/0x25
    [ 728.457327] [] rtnl_setlink+0xa9/0xc1
    [ 728.457364] [] ? trace_hardirqs_on_caller+0xe8/0x11d
    [ 728.457384] [] ? rtnl_setlink+0x0/0xc1
    [ 728.457401] [] rtnetlink_rcv_msg+0x197/0x1b1
    [ 728.457424] [] ? rtnetlink_rcv_msg+0x0/0x1b1
    [ 728.457441] [] netlink_rcv_skb+0x35/0x7a
    [ 728.457460] [] rtnetlink_rcv+0x1c/0x24
    [ 728.457476] [] netlink_unicast+0x1ad/0x20e
    [ 728.457499] [] netlink_sendmsg+0x248/0x255
    [ 728.457530] [] sock_sendmsg+0xe3/0xfe
    [ 728.457562] [] ? autoremove_wake_function+0x0/0x35
    [ 728.457583] [] ? native_sched_clock+0xd5/0xf4
    [ 728.457609] [] ? sched_clock+0xb/0x1c
    [ 728.457628] [] ? lock_release_holdtime+0x3f/0x44
    [ 728.457652] [] ? copy_from_user+0x45/0x68
    [ 728.457672] [] ? verify_iovec+0x45/0x73
    [ 728.457693] [] sys_sendmsg+0x152/0x1ad
    [ 728.457713] [] ? sys_recvmsg+0x177/0x184
    [ 728.457734] [] ? fget_light+0x95/0xb5
    [ 728.457757] [] ? copy_to_user+0x47/0x50
    [ 728.457776] [] ? move_addr_to_user+0x5b/0x73
    [ 728.457796] [] ? sys_getsockname+0x70/0x7b
    [ 728.457814] [] ? _spin_unlock_bh+0x2a/0x2d
    [ 728.457831] [] ? release_sock+0xb3/0xbb
    [ 728.457851] [] ? sock_setsockopt+0x49d/0x4a8
    [ 728.457874] [] ? fget_light+0xab/0xb5
    [ 728.457909] [] sys_socketcall+0x168/0x183
    [ 728.457935] [] sysenter_past_esp+0x78/0xc5
    [ 728.457969] =======================
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  2. Re: Potential iwlwifi locking problem in 2.6.26-git9

    On Wed, Jul 23, 2008 at 05:36:30PM -0400, Theodore Ts'o wrote:
    >
    > Is this a new one? I don't think I've seen this on LKML....


    The context of this was I had just booted to 2.6.26-git9, noted with
    annoyance that that the iwl4965 driver needed a new microcode file, so
    I popped in my EVDO card, downloaded the firmware, moved the firmware
    into place, and then got the above error message.

    I've since noted that it seems to be utterly refusing to associate
    with my wireless access point. This was working just fine with
    2.6.26-git6 kernel, so it looks like a recent regression? I just
    noticed someone else had recently complained about a similar "refusal
    to associate" with the iwl3945 driver.

    - Ted

    >
    > [ 728.456792]
    > [ 728.456795] =============================================
    > [ 728.456810] [ INFO: possible recursive locking detected ]
    > [ 728.456818] 2.6.26-05781-g44ea66a #25
    > [ 728.456823] ---------------------------------------------
    > [ 728.456830] NetworkManager/8287 is trying to acquire lock:
    > [ 728.456836] (&dev->addr_list_lock){-...}, at: [] dev_mc_sync+0x1c/0x59
    > [ 728.456867]
    > [ 728.456869] but task is already holding lock:
    > [ 728.456875] (&dev->addr_list_lock){-...}, at: [] dev_set_rx_mode+0x19/0x2b
    > [ 728.456894]
    > [ 728.456895] other info that might help us debug this:
    > [ 728.456902] 2 locks held by NetworkManager/8287:
    > [ 728.456908] #0: (rtnl_mutex){--..}, at: [] rtnl_lock+0x14/0x16
    > [ 728.456929] #1: (&dev->addr_list_lock){-...}, at: [] dev_set_rx_mode+0x19/0x2b
    > [ 728.456949]
    > [ 728.456951] stack backtrace:
    > [ 728.456961] Pid: 8287, comm: NetworkManager Not tainted 2.6.26-05781-g44ea66a #25
    > [ 728.456999] [] __lock_acquire+0x99f/0xcc5
    > [ 728.457037] [] ? trace_hardirqs_on+0xb/0xd
    > [ 728.457054] [] ? trace_hardirqs_on_caller+0xd2/0x11d
    > [ 728.457076] [] lock_acquire+0x6a/0x90
    > [ 728.457091] [] ? dev_mc_sync+0x1c/0x59
    > [ 728.457109] [] _spin_lock_bh+0x26/0x53
    > [ 728.457126] [] ? dev_mc_sync+0x1c/0x59
    > [ 728.457143] [] dev_mc_sync+0x1c/0x59
    > [ 728.457159] [] ieee80211_set_multicast_list+0xca/0xd2 [mac80211]
    > [ 728.457223] [] __dev_set_rx_mode+0x7e/0x81
    > [ 728.457239] [] dev_set_rx_mode+0x20/0x2b
    > [ 728.457255] [] dev_open+0x84/0x9e
    > [ 728.457269] [] dev_change_flags+0xa1/0x14f
    > [ 728.457286] [] do_setlink+0x250/0x300
    > [ 728.457308] [] ? _read_unlock+0x22/0x25
    > [ 728.457327] [] rtnl_setlink+0xa9/0xc1
    > [ 728.457364] [] ? trace_hardirqs_on_caller+0xe8/0x11d
    > [ 728.457384] [] ? rtnl_setlink+0x0/0xc1
    > [ 728.457401] [] rtnetlink_rcv_msg+0x197/0x1b1
    > [ 728.457424] [] ? rtnetlink_rcv_msg+0x0/0x1b1
    > [ 728.457441] [] netlink_rcv_skb+0x35/0x7a
    > [ 728.457460] [] rtnetlink_rcv+0x1c/0x24
    > [ 728.457476] [] netlink_unicast+0x1ad/0x20e
    > [ 728.457499] [] netlink_sendmsg+0x248/0x255
    > [ 728.457530] [] sock_sendmsg+0xe3/0xfe
    > [ 728.457562] [] ? autoremove_wake_function+0x0/0x35
    > [ 728.457583] [] ? native_sched_clock+0xd5/0xf4
    > [ 728.457609] [] ? sched_clock+0xb/0x1c
    > [ 728.457628] [] ? lock_release_holdtime+0x3f/0x44
    > [ 728.457652] [] ? copy_from_user+0x45/0x68
    > [ 728.457672] [] ? verify_iovec+0x45/0x73
    > [ 728.457693] [] sys_sendmsg+0x152/0x1ad
    > [ 728.457713] [] ? sys_recvmsg+0x177/0x184
    > [ 728.457734] [] ? fget_light+0x95/0xb5
    > [ 728.457757] [] ? copy_to_user+0x47/0x50
    > [ 728.457776] [] ? move_addr_to_user+0x5b/0x73
    > [ 728.457796] [] ? sys_getsockname+0x70/0x7b
    > [ 728.457814] [] ? _spin_unlock_bh+0x2a/0x2d
    > [ 728.457831] [] ? release_sock+0xb3/0xbb
    > [ 728.457851] [] ? sock_setsockopt+0x49d/0x4a8
    > [ 728.457874] [] ? fget_light+0xab/0xb5
    > [ 728.457909] [] sys_socketcall+0x168/0x183
    > [ 728.457935] [] sysenter_past_esp+0x78/0xc5
    > [ 728.457969] =======================

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

  3. Re: Potential iwlwifi locking problem in 2.6.26-git9

    On Thu, Jul 24, 2008 at 12:49 AM, Theodore Tso wrote:
    > On Wed, Jul 23, 2008 at 05:36:30PM -0400, Theodore Ts'o wrote:
    >>
    >> Is this a new one? I don't think I've seen this on LKML....

    >
    > The context of this was I had just booted to 2.6.26-git9, noted with
    > annoyance that that the iwl4965 driver needed a new microcode file, so
    > I popped in my EVDO card, downloaded the firmware, moved the firmware
    > into place, and then got the above error message.
    >
    > I've since noted that it seems to be utterly refusing to associate
    > with my wireless access point. This was working just fine with
    > 2.6.26-git6 kernel, so it looks like a recent regression? I just
    > noticed someone else had recently complained about a similar "refusal
    > to associate" with the iwl3945 driver.


    I'm not sure but this can be connected to what is described in thread:
    'New: dhcp doesn't work with iwl4965'
    Tomas
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  4. Re: Potential iwlwifi locking problem in 2.6.26-git9

    From: "Theodore Ts'o"
    Date: Wed, 23 Jul 2008 17:36:30 -0400

    >
    > Is this a new one? I don't think I've seen this on LKML....


    Yes, we have, we're trying to work it out on netdev.
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  5. Re: Potential iwlwifi locking problem in 2.6.26-git9

    Theodore Tso wrote:
    > On Wed, Jul 23, 2008 at 05:36:30PM -0400, Theodore Ts'o wrote:
    >
    >> Is this a new one? I don't think I've seen this on LKML....
    >>

    >
    > The context of this was I had just booted to 2.6.26-git9, noted with
    > annoyance that that the iwl4965 driver needed a new microcode file, so
    > I popped in my EVDO card, downloaded the firmware, moved the firmware
    > into place, and then got the above error message.
    >
    > I've since noted that it seems to be utterly refusing to associate
    > with my wireless access point. This was working just fine with
    > 2.6.26-git6 kernel, so it looks like a recent regression? I just
    > noticed someone else had recently complained about a similar "refusal
    > to associate" with the iwl3945 driver.
    >


    I noticed the same thing last night; a current git build goes through
    all the motions, but then fails to associate.

    J
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  6. Re: Potential iwlwifi locking problem in 2.6.26-git9

    On Thu 24.Jul'08 at 1:02:15 +0300, Tomas Winkler wrote:
    > On Thu, Jul 24, 2008 at 12:49 AM, Theodore Tso wrote:
    > > On Wed, Jul 23, 2008 at 05:36:30PM -0400, Theodore Ts'o wrote:
    > >>
    > >> Is this a new one? I don't think I've seen this on LKML....

    > >
    > > The context of this was I had just booted to 2.6.26-git9, noted with
    > > annoyance that that the iwl4965 driver needed a new microcode file, so
    > > I popped in my EVDO card, downloaded the firmware, moved the firmware
    > > into place, and then got the above error message.
    > >
    > > I've since noted that it seems to be utterly refusing to associate
    > > with my wireless access point. This was working just fine with
    > > 2.6.26-git6 kernel, so it looks like a recent regression? I just
    > > noticed someone else had recently complained about a similar "refusal
    > > to associate" with the iwl3945 driver.

    >
    > I'm not sure but this can be connected to what is described in thread:
    > 'New: dhcp doesn't work with iwl4965'
    > Tomas


    I can't associate with the access point anymore, and it is a very
    recent regression for me too (two or three days ago).

    I have the new firmware iwlwifi-4965-2.ucode inside /lib/firmware
    and when I try to connect (using the connect button in drakroam from
    Mandriva) I get this (digged from the syslog):

    Jul 23 13:46:01 localhost drakroam[3801]: running: /sbin/ifdown wlan0 daemon
    Jul 23 13:46:01 localhost kernel: wlan0: associate with AP 00:1b:11:3c:f6:db
    Jul 23 13:46:01 localhost ifplugd(wlan0)[3170]: Exiting.
    Jul 23 13:46:04 localhost kernel: iwl4965 0000:06:00.0: PCI INT A disabled
    Jul 23 13:46:04 localhost drakroam[3801]: running: /sbin/ifup wlan0 daemon
    Jul 23 13:46:04 localhost kernel: iwl4965 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
    Jul 23 13:46:04 localhost kernel: iwl4965 0000:06:00.0: restoring config space at offset 0x1 (was 0x100102, writing 0x100106)
    Jul 23 13:46:04 localhost kernel: ------------[ cut here ]------------
    Jul 23 13:46:04 localhost kernel: WARNING: at net/core/dev.c:1344 __netif_schedule+0x6f/0x80()
    Jul 23 13:46:04 localhost kernel: Modules linked in: snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss loop nvram snd_hda_intel snd_pcm snd_timer ohci1394 sg snd_page_alloc ieee1394 sky2 iwl4965 snd_hwdep sr_mod i2c_i801 evdev ide_disk ide_core ata_piix ahci libata dock sd_mod scsi_mod uhci_hcd ohci_hcd ehci_hcd usbcore [last unloaded: scsi_wait_scan]
    Jul 23 13:46:04 localhost kernel: Pid: 1613, comm: iwl4965/0 Not tainted 2.6.26-06077-gc010b2f #147
    Jul 23 13:46:04 localhost kernel:
    Jul 23 13:46:04 localhost kernel: Call Trace:
    Jul 23 13:46:04 localhost kernel: [] warn_on_slowpath+0x5f/0x80
    Jul 23 13:46:04 localhost kernel: [] __netif_schedule+0x6f/0x80
    Jul 23 13:46:04 localhost kernel: [] ieee80211_wake_queue+0x5a/0x70
    Jul 23 13:46:04 localhost kernel: [] ieee80211_wake_queues+0x2f/0x50
    Jul 23 13:46:04 localhost kernel: [] iwl_bg_alive_start+0x216/0x4b0 [iwl4965]
    Jul 23 13:46:04 localhost kernel: [] ? iwl_bg_alive_start+0x0/0x4b0 [iwl4965]
    Jul 23 13:46:04 localhost kernel: [] run_workqueue+0x98/0x140
    Jul 23 13:46:04 localhost kernel: [] worker_thread+0xa3/0x110
    Jul 23 13:46:04 localhost kernel: [] ? autoremove_wake_function+0x0/0x40
    Jul 23 13:46:04 localhost kernel: [] ? worker_thread+0x0/0x110
    Jul 23 13:46:04 localhost kernel: [] kthread+0x4d/0x80
    Jul 23 13:46:04 localhost kernel: [] child_rip+0xa/0x11
    Jul 23 13:46:04 localhost kernel: [] ? kthread+0x0/0x80
    Jul 23 13:46:04 localhost kernel: [] ? child_rip+0x0/0x11
    Jul 23 13:46:04 localhost kernel:
    Jul 23 13:46:04 localhost kernel: ---[ end trace 9c04ce6cf2b007a2 ]---
    Jul 23 13:46:04 localhost kernel: Registered led device: iwl-phy0:radio
    Jul 23 13:46:04 localhost kernel: Registered led device: iwl-phy0:assoc
    Jul 23 13:46:04 localhost kernel: Registered led device: iwl-phy0:RX
    Jul 23 13:46:04 localhost kernel: Registered led device: iwl-phy0:TX
    Jul 23 13:46:04 localhost kernel: iwl4965: TX Power requested while scanning!

    I am trying to bisect it, but I am facing some difficulties with compilation
    problems and non-booting kernels.

    In the thread Thomas Winkler pointed out the reporter says that using dhclient instead
    of dhcp makes it work. However Mandriva uses dhclient and it does not work here.


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

  7. Re: Potential iwlwifi locking problem in 2.6.26-git9

    From: "Carlos R. Mafra"
    Date: Wed, 23 Jul 2008 17:59:06 -0500

    > I am trying to bisect it, but I am facing some difficulties with compilation
    > problems and non-booting kernels.


    We know what the problem is and are discussing it on netdev
    and linux-wireless already.
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  8. Re: Potential iwlwifi locking problem in 2.6.26-git9

    On Wed 23.Jul'08 at 16:11:00 -0700, David Miller wrote:
    > From: "Carlos R. Mafra"
    > Date: Wed, 23 Jul 2008 17:59:06 -0500
    >
    > > I am trying to bisect it, but I am facing some difficulties with compilation
    > > problems and non-booting kernels.

    >
    > We know what the problem is and are discussing it on netdev
    > and linux-wireless already.


    Thank you for letting me know, David!
    I appreciate that deeply.
    --
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

+ Reply to Thread