Re: ACPI causes page fault/panic upon reboot

From: John Baldwin (jhb_at_FreeBSD.org)
Date: 12/31/03

  • Next message: Andre Guibert de Bruet: "Re: 4.9-stable to 5.x" 
    Date: Wed, 31 Dec 2003 12:05:55 -0500 (EST)
To: Dan Hulme <d@diefree.com>

    On 31-Dec-2003 Dan Hulme wrote:
    > John Baldwin wrote:
    >
    >>Can you build a debug kernel, reproduce the panic, and then do a list
    >>from gdb on your kernel.debug of the instruction pointer? For example,
    >>from the above you do:
    >>
    >>gdb -k kernel.debug
    >>l *0xc05fd8c5
    >>
    >>
    > First I should say that turning off acpi seemed to fix the problem.
    > However, the problem appears to be in dc0.
    >
    > Ok, rebuilt my kernel with
    >
    > makeoptions DEBUG=-g
    > options DDB
    >
    > Rebooted, did a trace when it went to the db prompt.:
    >
    > ***********************
    >
    > Shutting down ACPI
    >
    > Fatal trap 12: page fault while in kernel mode
    > fault virtual address = 0x18
    > fault code = supervisor write, page not present
    > instruction pointer = 0x8:0xc0603bc5
    > stack pointer = 0x10:0xd032cc90
    > frame pointer = 0x10:0xd032ccc0
    > code segment = base 0x0, limit 0xfffff, type 0x1b
    > = DPL 0, pres 1, def32 1, gran 1
    > processor eflags = interrupt enabled, resume, IOPL = 0
    > current process = 19 (irq9: rl0 dc0 acpi0)
    > kernel: type 12 trap, code=0
    > Stopped at dc_rxeof+0x1b5: movl %edx,0x18(%eax)
    >
    > db>trace
    >
    > dc_rxeof(c33bb000,0,c0700bbe,c43,c33bb000) at dc_rxeof+0x1b5
    > dc_intr(c33bb000,0,c06f4d21,21f,c333d388) at dc_intr+0x100
    > ithread_loop(c1843980,d032cd48,c06f4be6,311,0) at ithread_loop+0x162
    > fork_trampoline() at fork_trampoline+0x8
    > --- trap 0x1, eip = 0, esp = 0xd032cd7c, ebp = 0 ---
    >
    > *****************
    >
    > Did a gdb -k kernel.debug, and listed the code:
    >
    > *****************
    >
    > (kgdb) l *0xc0603bc5
    > 0xc0603bc5 is in dc_rxeof (../../../pci/if_dc.c:2804).
    > 2799 * If we are on an architecture with alignment
    > problems, or
    > 2800 * if the allocation fails, then use m_devget
    > and leave the
    > 2801 * existing buffer in the receive ring.
    > 2802 */
    > 2803 if (dc_quick && dc_newbuf(sc, i, 1) == 0) {
    > 2804 m->m_pkthdr.rcvif = ifp;
    > 2805 m->m_pkthdr.len = m->m_len = total_len;
    > 2806 DC_INC(i, DC_RX_LIST_CNT);
    > 2807 } else
    > 2808 #endif
    >
    > *****************
    >
    > I hope this helps.

    Looks like m is NULL. One thing you might try is setting the
    hw.acpi.disable_on_poweroff sysctl to 0. If that works, you
    can set it in /boot/loader.conf to make it permament. 

    -- 
John Baldwin <jhb@FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve!"  -  http://www.FreeBSD.org/
_______________________________________________
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"
  • Next message: Andre Guibert de Bruet: "Re: 4.9-stable to 5.x"

    Relevant Pages