Re: panic in ithread_loop()
On 10-Dec-2002 Dag-Erling Smorgrav wrote: > John Baldwin <[EMAIL PROTECTED]> writes: >> On 10-Dec-2002 Dag-Erling Smorgrav wrote: >> > John Baldwin <[EMAIL PROTECTED]> writes: >> > > This is where it faulted for some reason or another. It was running >> > > a registered interrupt handler. Do you have any kernel modules in this >> > > system? >> > Not that were loaded at the time of the panic. >> Had you unloaded any modules prior to the panic? > > No, the panic occurs during boot. Sorry for not making that clear. Oh, hrm. :( -- John Baldwin <[EMAIL PROTECTED]> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: panic in ithread_loop()
John Baldwin <[EMAIL PROTECTED]> writes: > On 10-Dec-2002 Dag-Erling Smorgrav wrote: > > John Baldwin <[EMAIL PROTECTED]> writes: > > > This is where it faulted for some reason or another. It was running > > > a registered interrupt handler. Do you have any kernel modules in this > > > system? > > Not that were loaded at the time of the panic. > Had you unloaded any modules prior to the panic? No, the panic occurs during boot. Sorry for not making that clear. DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: panic in ithread_loop()
On Wed, 11 Dec 2002, Mitsuru IWASAKI wrote: > > Fatal trap 12: page fault while in kernel mode > > fault virtual address = 0xc01e8d > > fault code = supervisor write, page not present > > instruction pointer = 0x8:0xc045dc80 > > stack pointer = 0x10:0xd536dce4 > > frame pointer = 0x10:0xd536dd00 > > code segment= base 0x0, limit 0xf, type 0x1b > > = DPL 0, pres 1, def32 1, gran 1 > > processor eflags= interrupt enabled, resume, IOPL = 0 > > current process = 12 (swi6: tty:sio clock) > > kernel: type 12 trap, code=0 > > Stopped at 0xc045dc80: movb%al,0xc01e8d > > I saw this panics many times recently. After totally fresh > buildworld and buildkernel (i.e. rm -rf /usr/obj/* and w/o -DNOCLEAN), > problem was gone. > > Please note that recently we had gcc changes. > > I've noticed -freg-struct-return brakages, some important ports which > using ndbm library (such as qpopper) is broken right now... I am suspicious about the fault address. If you add a lsb, you get a valid kernel address, which seems to indicate a packing/alignment problem. -Nate To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: panic in ithread_loop()
> This is 100% reproducible with a top-of-tree kernel, but didn't happen > with Wednesday's sources: > > Fatal trap 12: page fault while in kernel mode > fault virtual address = 0xc01e8d > fault code = supervisor write, page not present > instruction pointer = 0x8:0xc045dc80 > stack pointer = 0x10:0xd536dce4 > frame pointer = 0x10:0xd536dd00 > code segment= base 0x0, limit 0xf, type 0x1b > = DPL 0, pres 1, def32 1, gran 1 > processor eflags= interrupt enabled, resume, IOPL = 0 > current process = 12 (swi6: tty:sio clock) > kernel: type 12 trap, code=0 > Stopped at 0xc045dc80: movb%al,0xc01e8d I saw this panics many times recently. After totally fresh buildworld and buildkernel (i.e. rm -rf /usr/obj/* and w/o -DNOCLEAN), problem was gone. Please note that recently we had gcc changes. I've noticed -freg-struct-return brakages, some important ports which using ndbm library (such as qpopper) is broken right now... Thanks To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: panic in ithread_loop()
On 10-Dec-2002 Dag-Erling Smorgrav wrote: > John Baldwin <[EMAIL PROTECTED]> writes: >> This is where it faulted for some reason or another. It was running >> a registered interrupt handler. Do you have any kernel modules in this >> system? > > Not that were loaded at the time of the panic. Had you unloaded any modules prior to the panic? -- John Baldwin <[EMAIL PROTECTED]> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: panic in ithread_loop()
John Baldwin <[EMAIL PROTECTED]> writes: > This is where it faulted for some reason or another. It was running > a registered interrupt handler. Do you have any kernel modules in this > system? Not that were loaded at the time of the panic. DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
RE: panic in ithread_loop()
On 08-Dec-2002 Dag-Erling Smorgrav wrote:
> This is 100% reproducible with a top-of-tree kernel, but didn't happen
> with Wednesday's sources:
>
> Fatal trap 12: page fault while in kernel mode
> fault virtual address = 0xc01e8d
> fault code = supervisor write, page not present
> instruction pointer = 0x8:0xc045dc80
> stack pointer = 0x10:0xd536dce4
> frame pointer = 0x10:0xd536dd00
> code segment= base 0x0, limit 0xf, type 0x1b
> = DPL 0, pres 1, def32 1, gran 1
> processor eflags= interrupt enabled, resume, IOPL = 0
> current process = 12 (swi6: tty:sio clock)
> kernel: type 12 trap, code=0
> Stopped at 0xc045dc80: movb%al,0xc01e8d
> db> trace
> _end(0) at 0xc045dc80
> ithread_loop(c152ab00,d536dd48,c1537b60,c01d3d80,0) at ithread_loop+0x11c
> fork_exit(c01d3d80,c152ab00,d536dd48) at fork_exit+0x8f
> fork_trampoline() at fork_trampoline+0x1a
> --- trap 0x1, eip = 0, esp = 0xd536dd7c, ebp = 0 ---
>
> and according to gdb:
>
> (gdb) list _end
> No line number known for _end.
> (gdb) list *0xc045dc80
> No source file for address 0xc045dc80.
This is where it faulted for some reason or another. It was running
a registered interrupt handler. Do you have any kernel modules in this
system?
> (gdb) l *(ithread_loop + 0x11c)
> 0xc01d3e9c is in ithread_loop (../../../kern/kern_intr.c:536).
> 531 goto restart;
> 532 }
> 533 if ((ih->ih_flags & IH_MPSAFE) == 0)
> 534 mtx_lock(&Giant);
> 535 ih->ih_handler(ih->ih_argument);
> 536 if ((ih->ih_flags & IH_MPSAFE) == 0)
> 537 mtx_unlock(&Giant);
> 538 }
> 539 }
> 540
> (gdb) l *(fork_exit+0x8f)
> 0xc01d326f is in fork_exit (../../../kern/kern_fork.c:872).
> 867
> 868 /*
> 869 * Check if a kernel thread misbehaved and returned from its main
> 870 * function.
> 871 */
> 872 PROC_LOCK(p);
> 873 if (p->p_flag & P_KTHREAD) {
> 874 PROC_UNLOCK(p);
> 875 mtx_lock(&Giant);
> 876 printf("Kernel thread \"%s\" (pid %d) exited prematurely.\n",
> (gdb) l *(fork_trampoline+0x1a)
> 0xc02db94e is at {standard input}:152.
> 147 {standard input}: No such file or directory.
> in {standard input}
>
> DES
> --
> Dag-Erling Smorgrav - [EMAIL PROTECTED]
>
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-current" in the body of the message
--
John Baldwin <[EMAIL PROTECTED]> <>< http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve!" - http://www.FreeBSD.org/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
panic in ithread_loop()
This is 100% reproducible with a top-of-tree kernel, but didn't happen
with Wednesday's sources:
Fatal trap 12: page fault while in kernel mode
fault virtual address = 0xc01e8d
fault code = supervisor write, page not present
instruction pointer = 0x8:0xc045dc80
stack pointer = 0x10:0xd536dce4
frame pointer = 0x10:0xd536dd00
code segment= base 0x0, limit 0xf, type 0x1b
= DPL 0, pres 1, def32 1, gran 1
processor eflags= interrupt enabled, resume, IOPL = 0
current process = 12 (swi6: tty:sio clock)
kernel: type 12 trap, code=0
Stopped at 0xc045dc80: movb%al,0xc01e8d
db> trace
_end(0) at 0xc045dc80
ithread_loop(c152ab00,d536dd48,c1537b60,c01d3d80,0) at ithread_loop+0x11c
fork_exit(c01d3d80,c152ab00,d536dd48) at fork_exit+0x8f
fork_trampoline() at fork_trampoline+0x1a
--- trap 0x1, eip = 0, esp = 0xd536dd7c, ebp = 0 ---
and according to gdb:
(gdb) list _end
No line number known for _end.
(gdb) list *0xc045dc80
No source file for address 0xc045dc80.
(gdb) l *(ithread_loop + 0x11c)
0xc01d3e9c is in ithread_loop (../../../kern/kern_intr.c:536).
531 goto restart;
532 }
533 if ((ih->ih_flags & IH_MPSAFE) == 0)
534 mtx_lock(&Giant);
535 ih->ih_handler(ih->ih_argument);
536 if ((ih->ih_flags & IH_MPSAFE) == 0)
537 mtx_unlock(&Giant);
538 }
539 }
540
(gdb) l *(fork_exit+0x8f)
0xc01d326f is in fork_exit (../../../kern/kern_fork.c:872).
867
868 /*
869 * Check if a kernel thread misbehaved and returned from its main
870 * function.
871 */
872 PROC_LOCK(p);
873 if (p->p_flag & P_KTHREAD) {
874 PROC_UNLOCK(p);
875 mtx_lock(&Giant);
876 printf("Kernel thread \"%s\" (pid %d) exited prematurely.\n",
(gdb) l *(fork_trampoline+0x1a)
0xc02db94e is at {standard input}:152.
147 {standard input}: No such file or directory.
in {standard input}
DES
--
Dag-Erling Smorgrav - [EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
