addresses in xmon trace output?

addresses in xmon trace output?

[Steven Hanley]

All

what are the addresses you see in an xmon trace output? I had a crash
yesterday with benh 2.4.6-pre3 (rsynced two days ago) where typing x in xmon
didnt recover normal operation of the machine.

So the T command seemed to give some interesting numbers I assumed were
addresses in the system map I could look up once I rebooted.

[11:35:34] 22 shiva sjh ~>cat mon_crash
c67b2098 c00e1bd8 c00e1080 c011f9df c00de590 c00de7ec c003b6a4 c0003b3c
exception:c00 [c67a1e90] fe0f640 feeff40 feeff44 fecef4f4 fecfd4c feca62c feaaab8 10001808 fd9675c 0
[11:35:47] 23 shiva sjh ~>for i in c67b2098 c00e1bd8 c00e1080 c011f9df c00de590 c00de7ec c003b6a4 c0003b3c ; do grep -i $i /boot/System.map-2.4.6-pre3 ; done
c0003b3c T ret_from_syscall_1
[11:36:25] 24 shiva sjh ~>for i in fe0f640 feeff40 feeff44 fecef4f4 fecfd4c feca62c feaaab8 10001808 fd9675c ; do grep -i $i /boot/System.map-2.4.6-pre3 ; done
[11:37:18] 25 shiva sjh ~>grep c67a1e90 /boot/System.map-2.4.6-pre3
[11:37:34] 26 shiva sjh ~>

so they are not addresses in there (well except possibly for the
ret_from_syscall_1

This is on a pismo, the crash happened when I opened the lid (woke up form
sleep) in X (4.1.0), the mosue wasnt appearing and I couldnt change desktops
with the control arrow combo (sawfish) so I switched to console, then back to
X, as I came back to X the xmon console appeared overwriting my X display.

I will see if this is reporducable tonight (currently I am running 2.4.4 benh
as it doesnt exhibit this crash) as I expect to see paulus tonight so can
maybe show him the xmon console after the crash and get hi mto look at it.

Anyway what are the addresses in the xmon trace?

        See You
            Steve

--
sjh@wibble.net http://wibble.net/~sjh/
Look Up In The Sky
   Is it a bird?  No
      Is it a plane?  No
         Is it a small blue banana?
YES

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: addresses in xmon trace output?

[Daniel Jacobowitz]

On Thu, Jun 28, 2001 at 12:03:46PM +1000, Steven Hanley wrote:
>
> All
>
> what are the addresses you see in an xmon trace output? I had a crash
> yesterday with benh 2.4.6-pre3 (rsynced two days ago) where typing x in xmon
> didnt recover normal operation of the machine.
>
> So the T command seemed to give some interesting numbers I assumed were
> addresses in the system map I could look up once I rebooted.
>
> [11:35:34] 22 shiva sjh ~>cat mon_crash
> c67b2098 c00e1bd8 c00e1080 c011f9df c00de590 c00de7ec c003b6a4 c0003b3c
> exception:c00 [c67a1e90] fe0f640 feeff40 feeff44 fecef4f4 fecfd4c feca62c feaaab8 10001808 fd9675c 0
> [11:35:47] 23 shiva sjh ~>for i in c67b2098 c00e1bd8 c00e1080 c011f9df c00de590 c00de7ec c003b6a4 c0003b3c ; do grep -i $i /boot/System.map-2.4.6-pre3 ; done
> c0003b3c T ret_from_syscall_1
> [11:36:25] 24 shiva sjh ~>for i in fe0f640 feeff40 feeff44 fecef4f4 fecfd4c feca62c feaaab8 10001808 fd9675c ; do grep -i $i /boot/System.map-2.4.6-pre3 ; done
> [11:37:18] 25 shiva sjh ~>grep c67a1e90 /boot/System.map-2.4.6-pre3
> [11:37:34] 26 shiva sjh ~>
>
> so they are not addresses in there (well except possibly for the
> ret_from_syscall_1

They're addresses all right.  System.map lists the beginning of each
function; those addresses will be somewhere in the middle.  It's
sorted; find the closest match before each.

--
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: addresses in xmon trace output?

[Steven Hanley]

On Wed, Jun 27, 2001 at 07:11:03PM -0700, Daniel Jacobowitz wrote:
> They're addresses all right.  System.map lists the beginning of each
> function; those addresses will be somewhere in the middle.  It's
> sorted; find the closest match before each.

okay, that was easy, although the first address doesnt seem to be in the map
anywhere, maybe I copied it down wrong I suppose.

[13:04:08] 30 shiva sjh ~>./findinsmap.pl /boot/System.map-2.4.6-pre3
c67b2098 c00e1bd8 c00e1080 c011f9df c00de590 c00de7ec c003b6a4 c0003b3c
Addr: c00e1bd8 A: c00e1aa8 B: c00e1cb0 Function: c00e1aa8 T alloc_skb
Addr: c00e1080 A: c00e1004 B: c00e1170 Function: c00e1004 T sock_alloc_send_skb
Addr: c011f9df A: c011f8b4 B: c011fc74 Function: c011f8b4 t unix_stream_sendmsg
Addr: c00de590 A: c00de4f4 B: c00de5c0 Function: c00de4f4 T sock_sendmsg
Addr: c00de7ec A: c00de758 B: c00de7fc Function: c00de758 t sock_write
Addr: c003b6a4 A: c003b5dc B: c003b6f0 Function: c003b5dc T sys_write
Addr: c0003b3c A: c0003b3c B: c0003bf0 Function: c0003b3c T ret_from_syscall_1

        See You
            Steve

--
sjh@wibble.net http://wibble.net/~sjh/
Look Up In The Sky
   Is it a bird?  No
      Is it a plane?  No
         Is it a small blue banana?
YES

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: addresses in xmon trace output?

[Benjamin Herrenschmidt]

>what are the addresses you see in an xmon trace output? I had a crash
>yesterday with benh 2.4.6-pre3 (rsynced two days ago) where typing x in xmon
>didnt recover normal operation of the machine.
>
>So the T command seemed to give some interesting numbers I assumed were
>addresses in the system map I could look up once I rebooted.

Paulus helped me fix a bug in the wakup code that could cause
the kernel MMU mappings to be trashed (VSIDs initialized
incorrectly) leading to all sort of crashes on wakeup and
possibly bogus xmon output.

If you resync the tree, you should get the fix.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: addresses in xmon trace output?

[Steven Hanley]

On Thu, Jun 28, 2001 at 11:38:49AM +0200, Benjamin Herrenschmidt wrote:
> Paulus helped me fix a bug in the wakup code that could cause
> the kernel MMU mappings to be trashed (VSIDs initialized
> incorrectly) leading to all sort of crashes on wakeup and
> possibly bogus xmon output.
>
> If you resync the tree, you should get the fix.

thanks, paulus told me what to change in sleep.S a few hours ago. I will
rsync again sometime tomorrow probably.

	See You
	    Steve

--
sjh@wibble.net http://wibble.net/~sjh
Look Up In The Sky
    Is it a bird?   No
        Is it a plane    No
            Is it a small blue banana?
Yes

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/