Re: [Bug 60555] New: Some amount of ifconfig cause load average increase and system freeze

Re: [Bug 60555] New: Some amount of ifconfig cause load average increase and system freeze

[Bjorn Helgaas]

[+cc Thomas, e1000e driver folks, linux-pci, lkml]

On Mon, Jul 15, 2013 at 2:31 AM,  <bugzilla-daemon@bugzilla.kernel.org> wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=60555
>
>             Bug ID: 60555
>            Summary: Some amount of ifconfig cause load average increase
>                     and system freeze
>            Product: Drivers
>            Version: 2.5
>     Kernel Version: 3.10.0
>           Hardware: All
>                 OS: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: normal
>           Priority: P1
>          Component: PCI
>           Assignee: drivers_pci@kernel-bugs.osdl.org
>           Reporter: onorua@gmail.com
>         Regression: No
>
> I've faced with the situation when system stopped to process any request.
> Further analysis showed that the load average was around 20, and was constantly
> increasing (the normal LA for this system is 0.3-1).

Thanks for the report; sorry it fell through the cracks.

Is this problem reproducible?  If so, can you try to reproduce it on
v3.11?  Can you please attach a complete dmesg log and "lspci -vv" (as
root) output to the bugzilla?

The BUG_ON you're hitting is here, but I don't know what it means:

    #ifdef CONFIG_GENERIC_HARDIRQS
                for (i = 0; i < nvec; i++)
                        BUG_ON(irq_has_action(entry->irq + i));
    #endif

> dmesg showed following:
> ================================
> kernel BUG at drivers/pci/msi.c:342!
> invalid opcode: 0000 [#1] PREEMPT SMP
> Modules linked in: btusb binfmt_misc uinput ipv6 uvcvideo videobuf2_vmalloc
> videobuf2_memops videobuf2_core snd_hda_codec_hdmi arc4 snd_hda_codec_conexant
> snd_hda_intel iwldvm snd_hda_codec mac80211 snd_hwdep thinkpad_acpi iwlwifi
> cfg80211 mperf coretemp kvm_intel kvm e1000e ptp pps_core
> CPU: 0 PID: 23655 Comm: ifconfig Not tainted 3.10.0 #1
> Hardware name: LENOVO 4290PW4/4290PW4, BIOS 8DET66WW (1.36 ) 10/22/2012
> task: ffff880362cadb00 ti: ffff8803935b0000 task.ti: ffff8803935b0000
> RIP: 0010:[<ffffffff8127308a>]  [<ffffffff8127308a>] free_msi_irqs+0x56/0x10a
> RSP: 0018:ffff8803935b1c80  EFLAGS: 00010282
> RAX: ffff8803a8dcb900 RBX: ffff880393653e00 RCX: 00000000fffffffa
> RDX: 0000000000000000 RSI: 000000000000002a RDI: 0000000000000000
> RBP: ffff8803935b1ca8 R08: ffff88040d400000 R09: 000000000000002c
> R10: 00000000000000c8 R11: ffff8803935b1c5c R12: ffff880407823000
> R13: ffff8804078236a8 R14: 0000000000000000 R15: 0000000000000001
> FS:  00007f537f15d700(0000) GS:ffff88041e200000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00000000004026d4 CR3: 00000003a8faf000 CR4: 00000000000407f0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> Stack:
>  ffff880407823000 ffff88040629c000 00000000fffffff0 ffff880407823098
>  0000000000000001 ffff8803935b1cc0 ffffffff81273a5f ffff88040629c800
>  ffff8803935b1cd8 ffffffffa00283dc ffff88040629c800 ffff8803935b1d00
> Call Trace:
>  [<ffffffff81273a5f>] pci_disable_msi+0x2c/0x46
>  [<ffffffffa00283dc>] e1000e_reset_interrupt_capability+0x4c/0x59 [e1000e]
>  [<ffffffffa00286cf>] e1000_request_irq+0x1f1/0x249 [e1000e]
>  [<ffffffffa002c45a>] e1000_open+0xde/0x440 [e1000e]
>  [<ffffffff814b7a76>] __dev_open+0x8b/0xd0
>  [<ffffffff814b7c9a>] __dev_change_flags+0xa4/0x126
>  [<ffffffff814b7d8a>] dev_change_flags+0x18/0x4c
>  [<ffffffff81504e85>] devinet_ioctl+0x25b/0x548
>  [<ffffffff8150632d>] inet_ioctl+0x81/0x9e
>  [<ffffffff814a24b0>] sock_do_ioctl+0x20/0x3d
>  [<ffffffff814a28d6>] sock_ioctl+0x1fd/0x207
>  [<ffffffff810f058d>] vfs_ioctl+0x18/0x34
>  [<ffffffff810f0d49>] do_vfs_ioctl+0x32f/0x3e4
>  [<ffffffff810eb690>] ? final_putname+0x2f/0x32
>  [<ffffffff810f0e50>] SyS_ioctl+0x52/0x7d
>  [<ffffffff81023c20>] ? do_page_fault+0x9/0xb
>  [<ffffffff8157a652>] system_call_fastpath+0x16/0x1b
> Code: 83 7b 0c 00 74 2f 8a 0b 41 bf 01 00 00 00 45 31 f6 d0 e9 83 e1 07 41 d3
> e7 8b 7b 0c 44 01 f7 e8 f1 93 e1 ff 48 83 78 58 00 74 02 <0f> 0b 41 ff c6 45 39
> fe 7c e4 48 8b 5b 10 48 83 eb 10 eb b8 4c
> RIP  [<ffffffff8127308a>] free_msi_irqs+0x56/0x10a
>  RSP <ffff8803935b1c80>
> ==================
>
> --
> You are receiving this mail because:
> You are watching the assignee of the bug.

Re: [Bug 60555] New: Some amount of ifconfig cause load average increase and system freeze

[Bjorn Helgaas]

On Tue, Sep 10, 2013 at 3:04 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> [+cc Thomas, e1000e driver folks, linux-pci, lkml]
>
> On Mon, Jul 15, 2013 at 2:31 AM,  <bugzilla-daemon@bugzilla.kernel.org> wrote:
>> https://bugzilla.kernel.org/show_bug.cgi?id=60555
>>
>>             Bug ID: 60555
>>            Summary: Some amount of ifconfig cause load average increase
>>                     and system freeze
>>            Product: Drivers
>>            Version: 2.5
>>     Kernel Version: 3.10.0
>>           Hardware: All
>>                 OS: Linux
>>               Tree: Mainline
>>             Status: NEW
>>           Severity: normal
>>           Priority: P1
>>          Component: PCI
>>           Assignee: drivers_pci@kernel-bugs.osdl.org
>>           Reporter: onorua@gmail.com
>>         Regression: No
>>
>> I've faced with the situation when system stopped to process any request.
>> Further analysis showed that the load average was around 20, and was constantly
>> increasing (the normal LA for this system is 0.3-1).
>
> Thanks for the report; sorry it fell through the cracks.
>
> Is this problem reproducible?  If so, can you try to reproduce it on
> v3.11?  Can you please attach a complete dmesg log and "lspci -vv" (as
> root) output to the bugzilla?
>
> The BUG_ON you're hitting is here, but I don't know what it means:
>
>     #ifdef CONFIG_GENERIC_HARDIRQS
>                 for (i = 0; i < nvec; i++)
>                         BUG_ON(irq_has_action(entry->irq + i));
>     #endif

What's the status of this?  Is it still a problem?  Should we just
close the bug as unreproducible?

Bjorn

Re: [Bug 60555] New: Some amount of ifconfig cause load average increase and system freeze

[Bjorn Helgaas]

On Fri, Nov 8, 2013 at 5:24 AM, Yaroslav Molochko <onorua@gmail.com> wrote:
> Sorry. I was not using this machine for quite a bit, I've updated to the new
> kernel and updated bug with the information you asked

Thanks.

Note that this appears to be vanilla v3.12, which has e1000e version
2.3.2-k in it.  Intel maintains newer drivers out-of-tree at
http://sourceforge.net/projects/e1000/, and it's possible this is some
bug that has already been fixed.  The current version there looks like
e1000e-2.5.4, released 2013-09-05.

I think only the Intel folks can help you out with this.

Bjorn

>> > On Mon, Jul 15, 2013 at 2:31 AM,  <bugzilla-daemon@bugzilla.kernel.org>
>> > wrote:
>> >> https://bugzilla.kernel.org/show_bug.cgi?id=60555
>> >>
>> >>             Bug ID: 60555
>> >>            Summary: Some amount of ifconfig cause load average increase
>> >>                     and system freeze
>> >>            Product: Drivers
>> >>            Version: 2.5
>> >>     Kernel Version: 3.10.0
>> >>           Hardware: All
>> >>                 OS: Linux
>> >>               Tree: Mainline
>> >>             Status: NEW
>> >>           Severity: normal
>> >>           Priority: P1
>> >>          Component: PCI
>> >>           Assignee: drivers_pci@kernel-bugs.osdl.org
>> >>           Reporter: onorua@gmail.com
>> >>         Regression: No
>> >>
>> >> I've faced with the situation when system stopped to process any
>> >> request.
>> >> Further analysis showed that the load average was around 20, and was
>> >> constantly
>> >> increasing (the normal LA for this system is 0.3-1).
>> >
>> > Thanks for the report; sorry it fell through the cracks.
>> >
>> > Is this problem reproducible?  If so, can you try to reproduce it on
>> > v3.11?  Can you please attach a complete dmesg log and "lspci -vv" (as
>> > root) output to the bugzilla?
>> >
>> > The BUG_ON you're hitting is here, but I don't know what it means:
>> >
>> >     #ifdef CONFIG_GENERIC_HARDIRQS
>> >                 for (i = 0; i < nvec; i++)
>> >                         BUG_ON(irq_has_action(entry->irq + i));
>> >     #endif
>>
>> What's the status of this?  Is it still a problem?  Should we just
>> close the bug as unreproducible?
>>
>> Bjorn
>
>