Re: lockup with voluntary preempt R0 and VP, KP, etc, disabled

From: Florian Schmidt (mista.tapas_at_gmx.net)
Date: 09/03/04

  • Next message: Ingo Molnar: "Re: [patch] voluntary-preempt-2.6.9-rc1-bk4-R1"
    Date:	Fri, 3 Sep 2004 13:59:19 +0200
    To: Ingo Molnar <mingo@elte.hu>
    
    

    On Fri, 3 Sep 2004 12:32:44 +0200
    Ingo Molnar <mingo@elte.hu> wrote:

    >
    > * Florian Schmidt <mista.tapas@gmx.net> wrote:
    >
    > > > best would be to enable the NMI watchdog (nmi_watchdog=1 (or 2)
    > > > boot option) - check that it's working via the attached
    > > > lockupcli.c userspace code (run it as root). To have an NMI
    > > > watchdog you need APIC/IOAPIC support in the .config.
    > >
    > > Hmm, my local APIC kinda sux and i had all kinds of trouble using it
    > > [ERR count went up, stuff didn't work; XT-PIC always worked very
    > > well here]. So i'm kinda reluctant to compile with APIC. I suppose i
    > > can compile with it but turn it off via boot option? Or does the
    > > watchdog need a APIC that actually does stuff?
    >
    > in theory you should be able to boot with just the local APIC enabled
    > and "nmi_watchdog=2" and get a working NMI watchdog. You'll still have
    > the normal PIC IRQ handling.
    >
    > Ingo
    >

    ok, here you go [mental note to myself: get a second computer or a
    digital camera :), maybe my ZX81 will do]. This is what is saw on my
    console [sometimes i got lazy in between, i hope not at the wrong
    spots]:

    [<c0105e29>] do_signal+0x79/0x110
    [<c01162a0>] default_wake_function+0x0/0x20
      c012fgd7 do_futex+0x47/0xa0
      c012fb20 sys_futex
      c0103f07 do_notify
      c01060e6 work_notifysig
    Code: 96 54 01 00 00 8e e0 8e e8 85 d2 74 11 c7 86 54 01 00 00 00 00 00
    00 89 d0 e8 bb e4 ff 8b 9e 5c 01 00 00 85 db 74 09 8b 45 0c <8b> 40 20
    48 78 08 8b 5d f8 8b 75 fc c9 c3 c7 86 56 01 00 00 00
    <6> note: scsynth exited with preempt count 1
    Unable to handle kernel NULL pointer dereference at virtual adress
    00000020 printing eip:
    c0117ff2
    *pde = 00000000
    Oops = 0000[#21]
    PREEMPT
    Modules linked in: [you wanna know em?]
    CPU: 0
    EIP: 0060:[<c0117ff2>] Not tainted VLI
    EFLAGS: 00010282 (2.6.9-rc1-VP-R0)
    EIP is at mm_release+0x42/0xb0

    That's all except for the modules list..
    -
    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/


  • Next message: Ingo Molnar: "Re: [patch] voluntary-preempt-2.6.9-rc1-bk4-R1"

    Relevant Pages

    • Re: Linux 2.4.25-rc4
      ... So I suspect that sending PM events is not exactly SMP safe, ... I booted with 'nosmp', APIC was disabled? ... it might indicate that we need another boot option to adapt to different ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • 2.6.12-rc4 OOPS EIP is at free_block
      ... saa7115: decoder disable output ... netconsole lirc_i2c lirc_dev msp3400 saa7115 ivtv ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: [patch 0/3] j_state_lock, j_list_lock, remove-bitlocks
      ... BUG: Unable to handle kernel NULL pointer dereference at virtual address ... TAS_BUFFER_FNS(RevokeValid, revokevalid) ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • [2.6.0-test11] Ext3 journalling Oops on file system error
      ... issues (grumble), and I had skipped the fsck accidentally. ... Aborting journal on device md0. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: 2.6.9rc2-mm4 oops
      ... > I found a struct acpi_driver in hpet.c that was erroneously marked ... Sorry, I did, and still get the oops. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)