Re: 4.9-RC panic on 24 hours

From: Kris Kennaway (kris_at_obsecurity.org)
Date: 10/27/03

  • Next message: Dan Strick: "Re: Final 4.9-RC (i386) available now, please help test."
    Date: Mon, 27 Oct 2003 10:11:01 -0800
    To: "Michael W. Oliver" <michael@gargantuan.com>
    
    
    

    On Sun, Oct 26, 2003 at 03:02:36PM -0500, Michael W. Oliver wrote:
    > Hi again stable folk.
    >
    > CVSUP performed at: 2003.10.23.at.22.21.18 (EDT)
    >
    > Now, I want to save some folks some time here and dispell some of the usual
    > reasons that are tossed my way when I bring up these 24 hour panics:
    >
    > 1) It is not happening at 3:00 AM, and I can run periodic anytime of the day
    > without incident.
    >
    > 2) The panics happened in the past, as this time, almost exactly on 24
    > hours, regardless of the time of day of the last panic or reboot.
    >
    > 3) The panics do not happen to coincide with any specific task in cron,
    > network activity, etc.
    >
    > I am seriously contemplating moving back to RELENG_4_8 since I have had so
    > many problems with STABLE since the PAE import. This isn't to say that my
    > problems are related to PAE, just that that's about the time that this
    > craziness started.
    >
    > I hope that someone finds this information useful, and I am open to any
    > suggestions that anyone would like to present.
    >
    > Thank you much!
    >
    > ---> backtrace, uname, dmesg
    >
    > http://michael.gargantuan.com/phoenix_panic.log

    #16 0xc01f45b5 in arptimer (ignored_arg=0x0) at /usr/src/sys/netinet/if_ether.c:152
            rt = (struct rtentry *) 0x0
            s = 4194304
            la = (struct llinfo_arp *) 0x620000
            ola = (struct llinfo_arp *) 0x0
    #17 0xc01a8259 in softclock () at /usr/src/sys/kern/kern_timeout.c:131
            c_func = (void (*)()) 0xc01f4580 <arptimer>
            c_arg = (void *) 0x0
            c = (struct callout *) 0x0
            bucket = (struct callout_tailq *) 0xcf0e3bd0
            s = 4194304
            curticks = 8640950
            steps = 0

    I wonder if this is related to the (security-related) ARP changes from a few weeks ago.

    Kris

    
    



  • Next message: Dan Strick: "Re: Final 4.9-RC (i386) available now, please help test."

    Relevant Pages

    • 4.9-RC panic on 24 hours
      ... I want to save some folks some time here and dispell some of the usual ... reasons that are tossed my way when I bring up these 24 hour panics: ... The panics happened in the past, as this time, almost exactly on 24 ... many problems with STABLE since the PAE import. ...
      (freebsd-stable)
    • Re: aac & PAE not happy in -current
      ... It's fine with 2GB of ram, but when I tried bumping it to 6GB and turning on PAE, the aac driver panics on boot. ... Booting the same kernel without PAE I get the same thing: ...
      (freebsd-current)
    • Re: FreeBSD 5.4-RELEASE-p5 panic
      ... > would die while trying to unlock a vnode lock that the thread didn't own. ... These couple of panics are relatively easy to reproduce on demand. ... When you boot these servers without PAE ...
      (freebsd-questions)