lkml.org 
[lkml]   [2006]   [Oct]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectUserspace process may be able to DoS kernel
    [I'm not subscribed on this list - please CC answers to me.]

    Hello,

    It seems that the latest version of Skype may exhibit a problem in the
    kernel where a non-root userspace process is able to block the whole
    system for durations of up to several minutes. If someone is able to
    reproduce the steps which cause the problem he may be able to DoS a
    system by consecutively causing soft lockups.

    There were some reports of this problem on other lists before, but
    mostly on tainted systems. I was able to reproduce this problem on a
    non-tainted mostly vanilla 2.6.17.6 kernel (it includes the suspend2
    patches). As most users who reported this problem are using Ubuntu,
    the problem may be related to one of the settings in Ubuntu's kernel
    config. The configuration of my kernel is also based on the Ubuntu
    kernel config. As I am not using the patched kernel by Ubuntu I hope
    that the LKML is the right place to report this issue.

    The lockup usually occurs when Skype 1.3.x for Linux (I'm using
    1.3.0.53) sits around idle for some time and then (presumably) uses
    the sound device (i.e. for me it happens when I call a contact -
    others reported this problem occurs for incoming messages [there may
    be an audio notification of the messages enabled]). The lockup can
    take from several seconds (where it is not detected by the kernel) up
    to some minutes. The whole system seems to be blocked - i.e. there is
    not even disk IO.

    dmesg reports the following:
    BUG: soft lockup detected on CPU#0!
    <c01562cd> softlockup_tick+0xad/0xf0 <c012e609> update_process_times+0x39/0x90
    <c011600b> smp_apic_timer_interrupt+0x5b/0x70 <c0110037>
    get_offset_pmtmr+0x97/0x1060
    <c0103d20> apic_timer_interrupt+0x1c/0x24 <c013d390> hrtimer_get_res+0x0/0x60
    <c0110037> get_offset_pmtmr+0x97/0x1060 <c0106b9f> do_gettimeofday+0x1f/0xd0
    <c0129654> getnstimeofday+0x14/0x40 <c01398d1> sys_clock_gettime+0x31/0xb0
    <c01031e7> sysenter_past_esp+0x54/0x75

    A copy of my kernel config is available at:
    http://virtual.sysfrog.org/~gst/config.txt

    The hardware where this problem occurs here is a X41 Thinkpad (Pentium
    M Dothan). There are also reports on other non-Intel CPUs e.g. AMD
    Turion 64.

    Please see the more extensive documentation of this problem on
    https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15/+bug/53216
    (although some of the people there use tainted kernels).

    I will upgrade my 2.6.17.6 kernel to 2.6.18 and try to reproduce the
    problem there in the following days (but I fear that I won't have much
    time before the weekend).

    bye,
    /gst
    -
    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/

    \
     
     \ /
      Last update: 2006-10-11 18:57    [W:0.022 / U:0.520 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site