2.6.5/iBook G3: Oops on resume from sleep

2.6.5/iBook G3: Oops on resume from sleep

[Kiko Piris]

Hi,

Yesterday I compiled vanilla 2.6.5 (plus laptop-mode patch).

Today I noticed an oops when resuming from sleep.

Here's what I get on the screen console:

---8<---
adb: starting probe task...
adb devices: [2]: 2 c4 [3]: 3 1 [7] 1f
ADB keyboard at 2, handler 1
ADB mouse at 3, handler set to 4 (trackpad)
adb: finished probe task...
tumbler: i2c is not initialized
Oops: kernel access of bad area, sig: 11 [#1]
NIP: DA378C3C LR: DA468E0C SP: D4F13D70 REGS: d4f13cc0 TRAP: 0301    Not tainted
MSR: 0000b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
DAR: 00000006, DSISR: 40000000
TASK = d74920c0[1233] 'pmud' Last syscall: 54
GPR00: DA468E0C D4F13D70 D74920C0 00000000 00000040 00000000 00000000 00000040
GPR08: 00000002 FFFFFFFF D74920C0 00000000 2DE985CF
Call trace:
 [da468e0c] snapper_set_capture_source+0x28/0x38 [snd_powermac]
 [da46936c] tumbler_resume+0xa0/0x138 [snd_powermac]
 [da466318] snd_pmac_resume+0xf4/0xf8 [snd_powermac]
 [da466364] snd_pmac_sleep_notify+0x48/0x54 [snd_powermac]
 [c030788c] 0xc030788c
 [c0307fe0] 0xc0307fe0
 [c03084c4] 0xc03084c4
 [c0308d2c] 0xc0308d2c
 [c0067c04] sys_ioctl+0xdc/0x2f4
 [c0007600] ret_from_syscall+0x0/0x44
eth0: Link is up at 100 Mbps, full-duplex.
eth0: Pause is disabled
--->8---

And the system is completely hung (I've copied it by hand, so I could
have made a mistake typing).

I've noticed that if I put the iBook to sleep having unloaded alsa, it
resumes ok. Having alsa loaded, I can reproduce it 100% of the times.


Some info about my system:

* Debian Sid up2date.

* Kernel config, if needed, is uploaded at
http://www.pirispons.net/bulma/llista/iBook/

$ cat /proc/version
Linux version 2.6.5 (root@sacarino) (gcc version 3.3.3 (Debian 20040401)) #1 Sun Apr 4 14:23:10 CEST 2004

$ cat /proc/cpuinfo
processor       : 0
cpu             : 750FX
temperature     : 47 C (uncalibrated)
clock           : 400MHz
revision        : 2.3 (pvr 7000 0203)
bogomips        : 793.71
machine         : PowerBook4,3
motherboard     : PowerBook4,3 MacRISC2 MacRISC Power Macintosh
detected as     : 257 (iBook 2 rev. 2)
pmac flags      : 0000000b
L2 cache        : 512K unified
memory          : 384MB
pmac-generation : NewWorld


Please, let me know if I can provide any additional information.

Many thanks in advance

--
Kiko

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Martin Habets]

What sound chip do you actually have? The kernel seems to think
you have a snapper.
You can find out with
cat /proc/asound/cards
after ALSA is loaded.

This capture_source thing was added pretty recently.

Martin

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Kiko Piris]

On 06/04/2004 at 00:13, Martin Habets wrote:

> What sound chip do you actually have? The kernel seems to think
> you have a snapper.
> You can find out with
> cat /proc/asound/cards
> after ALSA is loaded.

It seems I do:

---8<---
0 [Snapper        ]: PMac Snapper - PowerMac Snapper
                     PowerMac Snapper (Dev 24) Sub-frame 0
--->8---

> This capture_source thing was added pretty recently.

I don't quite understand you, sorry.

--
Kiko

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Martin Habets]

On Tue, Apr 06, 2004 at 09:37:41AM +0200, Kiko Piris wrote:
> It seems I do:
>
> ---8<---
> 0 [Snapper        ]: PMac Snapper - PowerMac Snapper
>                      PowerMac Snapper (Dev 24) Sub-frame 0
> --->8---

Good, same chip I have.

> > This capture_source thing was added pretty recently.
>
> I don't quite understand you, sorry.

Well, the code is so new that the bug fixes have not made it into the
kernel yet. But they are in the alsa cvs node.
What I did is pick up revision 1.29 (the latest) from
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/ppc/tumbler.c

Make a backup of your existing tumbler.c file, just in case... Put the
new one in an recompile.

Martin

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Kiko Piris]

On 06/04/2004 at 13:01, Martin Habets wrote:

> Well, the code is so new that the bug fixes have not made it into the
> kernel yet. But they are in the alsa cvs node.
> What I did is pick up revision 1.29 (the latest) from
> http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/ppc/tumbler.c
>
> Make a backup of your existing tumbler.c file, just in case... Put the
> new one in an recompile.

Revision 1.29 of that file, AFAICS is the latest one.

This is exactly the same file that comes with 2.6.5.

If I understood you correctly...

Thanks for replying, btw :-)

--
Kiko

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Martin Habets]

Well, I'm out of clues then. Must be something else affecting
this. I'm still on 2.6.3-benh and by the looks of it I'll stay on
that for a while and get some other work done.

Probably not relevant, but does the input source in alsa mixer work
okay for you?

Martin

On Tue, Apr 06, 2004 at 02:43:25PM +0200, Kiko Piris wrote:
>
> On 06/04/2004 at 13:01, Martin Habets wrote:
>
> > Well, the code is so new that the bug fixes have not made it into the
> > kernel yet. But they are in the alsa cvs node.
> > What I did is pick up revision 1.29 (the latest) from
> > http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/ppc/tumbler.c
> >
> > Make a backup of your existing tumbler.c file, just in case... Put the
> > new one in an recompile.
>
> Revision 1.29 of that file, AFAICS is the latest one.
>
> This is exactly the same file that comes with 2.6.5.
>
> If I understood you correctly...
>
> Thanks for replying, btw :-)
>
> --
> Kiko

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Kiko Piris]

On 06/04/2004 at 23:18, Martin Habets wrote:

> Well, I'm out of clues then. Must be something else affecting

No worries. As BenH pointed, I filed a bug at alsa-project bugtracking
page (no response yet).

> Probably not relevant, but does the input source in alsa mixer work
> okay for you?

What do you mean? If I can record sound with the mic? Yes I can.

In fact I had never used the mic that comes with the iBook before, I
just apt-get installed audacity and recorded something. OK, no problem.

The funny thing is that with alsa, playback is horrible (too low and
*very* poor quality). I just have both dmasound_pmac and alsa compiled
as modules[*] and I load those modules when I want to listen to some
audio file (sound is not a crucial issue for me).

[*] My system is set up so that oss's dmasound_pmac tries to get loaded
on boot. Then I've "hacked" a bit the modprobe.d configuration file for
alsa modules as you can see in [1]. When I want to disable alsa I just
put that #alsa.off# comment and, with that configuration, I can use oss
and disable alsa on-the-fly.

It's quite handy when, like now, alsa causes system hangs on waking up
from sleep; just disable it and be happy (as I said before, sound is not
a vital issue for me on my laptop).

[1] http://www.pirispons.net/bulma/llista/iBook/modprobe.d/kiko

Kind regards.

--
Kiko

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Benjamin Herrenschmidt]

On Tue, 2004-04-06 at 07:53, Kiko Piris wrote:
> Hi,
>
> Yesterday I compiled vanilla 2.6.5 (plus laptop-mode patch).
>
> Today I noticed an oops when resuming from sleep.
>
> Here's what I get on the screen console:

Please, send this report to the driver author/maintainer, I'm
not sure the Alsa folks are following this list.

Ben.

> ---8<---
> adb: starting probe task...
> adb devices: [2]: 2 c4 [3]: 3 1 [7] 1f
> ADB keyboard at 2, handler 1
> ADB mouse at 3, handler set to 4 (trackpad)
> adb: finished probe task...
> tumbler: i2c is not initialized
> Oops: kernel access of bad area, sig: 11 [#1]
> NIP: DA378C3C LR: DA468E0C SP: D4F13D70 REGS: d4f13cc0 TRAP: 0301    Not tainted
> MSR: 0000b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
> DAR: 00000006, DSISR: 40000000
> TASK = d74920c0[1233] 'pmud' Last syscall: 54
> GPR00: DA468E0C D4F13D70 D74920C0 00000000 00000040 00000000 00000000 00000040
> GPR08: 00000002 FFFFFFFF D74920C0 00000000 2DE985CF
> Call trace:
>  [da468e0c] snapper_set_capture_source+0x28/0x38 [snd_powermac]
>  [da46936c] tumbler_resume+0xa0/0x138 [snd_powermac]
>  [da466318] snd_pmac_resume+0xf4/0xf8 [snd_powermac]
>  [da466364] snd_pmac_sleep_notify+0x48/0x54 [snd_powermac]
>  [c030788c] 0xc030788c
>  [c0307fe0] 0xc0307fe0
>  [c03084c4] 0xc03084c4
>  [c0308d2c] 0xc0308d2c
>  [c0067c04] sys_ioctl+0xdc/0x2f4
>  [c0007600] ret_from_syscall+0x0/0x44
> eth0: Link is up at 100 Mbps, full-duplex.
> eth0: Pause is disabled
> --->8---
>
> And the system is completely hung (I've copied it by hand, so I could
> have made a mistake typing).
>
> I've noticed that if I put the iBook to sleep having unloaded alsa, it
> resumes ok. Having alsa loaded, I can reproduce it 100% of the times.
>
>
> Some info about my system:
>
> * Debian Sid up2date.
>
> * Kernel config, if needed, is uploaded at
> http://www.pirispons.net/bulma/llista/iBook/
>
> $ cat /proc/version
> Linux version 2.6.5 (root@sacarino) (gcc version 3.3.3 (Debian 20040401)) #1 Sun Apr 4 14:23:10 CEST 2004
>
> $ cat /proc/cpuinfo
> processor       : 0
> cpu             : 750FX
> temperature     : 47 C (uncalibrated)
> clock           : 400MHz
> revision        : 2.3 (pvr 7000 0203)
> bogomips        : 793.71
> machine         : PowerBook4,3
> motherboard     : PowerBook4,3 MacRISC2 MacRISC Power Macintosh
> detected as     : 257 (iBook 2 rev. 2)
> pmac flags      : 0000000b
> L2 cache        : 512K unified
> memory          : 384MB
> pmac-generation : NewWorld
>
>
> Please, let me know if I can provide any additional information.
>
> Many thanks in advance
>
> --
> Kiko
>
--
Benjamin Herrenschmidt <benh@kernel.crashing.org>


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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Aaron Lunansky]

I'm having similar issues with an ibook G3 (600MHz) - only my system
seems to fail to sleep, and this is regardless of modules (I have
sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
get a blank screen and it does not respond to any input. I'm not sure
where it is when it dies though - what's the best way to figure this
out? (how can I make SysRq work on an ibook (which has no SysRq key)).

I get this with every 2.6.* kernel I've tried, including 2.6.5. I
previously had sleep working on my ibook with a 2.6 kernel, but I've
since replaced the (dead) hard drive. I can get my ibook to sleep with
a 2.4.24 kernel.

Thanks,
Aaron

On Mon, Apr 05, 2004 at 11:53:18PM +0200, Kiko Piris wrote:
>
> Hi,
>
> Yesterday I compiled vanilla 2.6.5 (plus laptop-mode patch).
>
> Today I noticed an oops when resuming from sleep.
>
> Here's what I get on the screen console:
>
> ---8<---
> adb: starting probe task...
> adb devices: [2]: 2 c4 [3]: 3 1 [7] 1f
> ADB keyboard at 2, handler 1
> ADB mouse at 3, handler set to 4 (trackpad)
> adb: finished probe task...
> tumbler: i2c is not initialized
> Oops: kernel access of bad area, sig: 11 [#1]
> NIP: DA378C3C LR: DA468E0C SP: D4F13D70 REGS: d4f13cc0 TRAP: 0301    Not tainted
> MSR: 0000b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
> DAR: 00000006, DSISR: 40000000
> TASK = d74920c0[1233] 'pmud' Last syscall: 54
> GPR00: DA468E0C D4F13D70 D74920C0 00000000 00000040 00000000 00000000 00000040
> GPR08: 00000002 FFFFFFFF D74920C0 00000000 2DE985CF
> Call trace:
>  [da468e0c] snapper_set_capture_source+0x28/0x38 [snd_powermac]
>  [da46936c] tumbler_resume+0xa0/0x138 [snd_powermac]
>  [da466318] snd_pmac_resume+0xf4/0xf8 [snd_powermac]
>  [da466364] snd_pmac_sleep_notify+0x48/0x54 [snd_powermac]
>  [c030788c] 0xc030788c
>  [c0307fe0] 0xc0307fe0
>  [c03084c4] 0xc03084c4
>  [c0308d2c] 0xc0308d2c
>  [c0067c04] sys_ioctl+0xdc/0x2f4
>  [c0007600] ret_from_syscall+0x0/0x44
> eth0: Link is up at 100 Mbps, full-duplex.
> eth0: Pause is disabled
> --->8---
>
> And the system is completely hung (I've copied it by hand, so I could
> have made a mistake typing).
>
> I've noticed that if I put the iBook to sleep having unloaded alsa, it
> resumes ok. Having alsa loaded, I can reproduce it 100% of the times.
>
>
> Some info about my system:
>
> * Debian Sid up2date.
>
> * Kernel config, if needed, is uploaded at
> http://www.pirispons.net/bulma/llista/iBook/
>
> $ cat /proc/version
> Linux version 2.6.5 (root@sacarino) (gcc version 3.3.3 (Debian 20040401)) #1 Sun Apr 4 14:23:10 CEST 2004
>
> $ cat /proc/cpuinfo
> processor       : 0
> cpu             : 750FX
> temperature     : 47 C (uncalibrated)
> clock           : 400MHz
> revision        : 2.3 (pvr 7000 0203)
> bogomips        : 793.71
> machine         : PowerBook4,3
> motherboard     : PowerBook4,3 MacRISC2 MacRISC Power Macintosh
> detected as     : 257 (iBook 2 rev. 2)
> pmac flags      : 0000000b
> L2 cache        : 512K unified
> memory          : 384MB
> pmac-generation : NewWorld
>
>
> Please, let me know if I can provide any additional information.
>
> Many thanks in advance
>
> --
> Kiko
>

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Benjamin Herrenschmidt]

On Thu, 2004-04-08 at 05:17, Aaron Lunansky wrote:
> I'm having similar issues with an ibook G3 (600MHz) - only my system
> seems to fail to sleep, and this is regardless of modules (I have
> sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
> get a blank screen and it does not respond to any input. I'm not sure
> where it is when it dies though - what's the best way to figure this
> out? (how can I make SysRq work on an ibook (which has no SysRq key)).

No, once the screen is off, there isn't much you can do. I suggest
you do some experiments in console mode, use "snooze -f" (that will
bypass pmud or pbbuttonsd, which tend to cut the backlight early).

Also, did it die on sleep or on wakeup ? That is, did it reach the
sleep state where the led starts "snoring" ?

> I get this with every 2.6.* kernel I've tried, including 2.6.5. I
> previously had sleep working on my ibook with a 2.6 kernel, but I've
> since replaced the (dead) hard drive. I can get my ibook to sleep with
> a 2.4.24 kernel.
>
> Thanks,
> Aaron
>
> On Mon, Apr 05, 2004 at 11:53:18PM +0200, Kiko Piris wrote:
> >
> > Hi,
> >
> > Yesterday I compiled vanilla 2.6.5 (plus laptop-mode patch).
> >
> > Today I noticed an oops when resuming from sleep.
> >
> > Here's what I get on the screen console:
> >
> > ---8<---
> > adb: starting probe task...
> > adb devices: [2]: 2 c4 [3]: 3 1 [7] 1f
> > ADB keyboard at 2, handler 1
> > ADB mouse at 3, handler set to 4 (trackpad)
> > adb: finished probe task...
> > tumbler: i2c is not initialized
> > Oops: kernel access of bad area, sig: 11 [#1]
> > NIP: DA378C3C LR: DA468E0C SP: D4F13D70 REGS: d4f13cc0 TRAP: 0301    Not tainted
> > MSR: 0000b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
> > DAR: 00000006, DSISR: 40000000
> > TASK = d74920c0[1233] 'pmud' Last syscall: 54
> > GPR00: DA468E0C D4F13D70 D74920C0 00000000 00000040 00000000 00000000 00000040
> > GPR08: 00000002 FFFFFFFF D74920C0 00000000 2DE985CF
> > Call trace:
> >  [da468e0c] snapper_set_capture_source+0x28/0x38 [snd_powermac]
> >  [da46936c] tumbler_resume+0xa0/0x138 [snd_powermac]
> >  [da466318] snd_pmac_resume+0xf4/0xf8 [snd_powermac]
> >  [da466364] snd_pmac_sleep_notify+0x48/0x54 [snd_powermac]
> >  [c030788c] 0xc030788c
> >  [c0307fe0] 0xc0307fe0
> >  [c03084c4] 0xc03084c4
> >  [c0308d2c] 0xc0308d2c
> >  [c0067c04] sys_ioctl+0xdc/0x2f4
> >  [c0007600] ret_from_syscall+0x0/0x44
> > eth0: Link is up at 100 Mbps, full-duplex.
> > eth0: Pause is disabled
> > --->8---
> >
> > And the system is completely hung (I've copied it by hand, so I could
> > have made a mistake typing).
> >
> > I've noticed that if I put the iBook to sleep having unloaded alsa, it
> > resumes ok. Having alsa loaded, I can reproduce it 100% of the times.
> >
> >
> > Some info about my system:
> >
> > * Debian Sid up2date.
> >
> > * Kernel config, if needed, is uploaded at
> > http://www.pirispons.net/bulma/llista/iBook/
> >
> > $ cat /proc/version
> > Linux version 2.6.5 (root@sacarino) (gcc version 3.3.3 (Debian 20040401)) #1 Sun Apr 4 14:23:10 CEST 2004
> >
> > $ cat /proc/cpuinfo
> > processor       : 0
> > cpu             : 750FX
> > temperature     : 47 C (uncalibrated)
> > clock           : 400MHz
> > revision        : 2.3 (pvr 7000 0203)
> > bogomips        : 793.71
> > machine         : PowerBook4,3
> > motherboard     : PowerBook4,3 MacRISC2 MacRISC Power Macintosh
> > detected as     : 257 (iBook 2 rev. 2)
> > pmac flags      : 0000000b
> > L2 cache        : 512K unified
> > memory          : 384MB
> > pmac-generation : NewWorld
> >
> >
> > Please, let me know if I can provide any additional information.
> >
> > Many thanks in advance
> >
> > --
> > Kiko
> >
>
--
Benjamin Herrenschmidt <benh@kernel.crashing.org>


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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Aaron Lunansky]

So what I've figured out so far is that the system is not crashed- the
screen is still on in fact, blinking a cursor in the top left. I can ssh
in to the machine but a cursory review of dmesg shows nothing of
significance.

Unfortunately I can't do much with an ssh session before it seems to lock up,
but I can launch another ssh session if I want. Also, if I disconnect my
USB keyboard while the screen is in this blinking state, I get a "USB
device soandso" disconnected message.

Since my machines not crashing I can poke at it some more, but I could
use some pointers as to where I should poke.

Thanks,
Aaron

On Thu, Apr 08, 2004 at 08:49:31AM +1000, Benjamin Herrenschmidt wrote:
> On Thu, 2004-04-08 at 05:17, Aaron Lunansky wrote:
> > I'm having similar issues with an ibook G3 (600MHz) - only my system
> > seems to fail to sleep, and this is regardless of modules (I have
> > sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
> > get a blank screen and it does not respond to any input. I'm not sure
> > where it is when it dies though - what's the best way to figure this
> > out? (how can I make SysRq work on an ibook (which has no SysRq key)).
>
> No, once the screen is off, there isn't much you can do. I suggest
> you do some experiments in console mode, use "snooze -f" (that will
> bypass pmud or pbbuttonsd, which tend to cut the backlight early).
>
> Also, did it die on sleep or on wakeup ? That is, did it reach the
> sleep state where the led starts "snoring" ?
>
> > I get this with every 2.6.* kernel I've tried, including 2.6.5. I
> > previously had sleep working on my ibook with a 2.6 kernel, but I've
> > since replaced the (dead) hard drive. I can get my ibook to sleep with
> > a 2.4.24 kernel.
> >
> > Thanks,
> > Aaron

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Benjamin Herrenschmidt]

On Thu, 2004-04-08 at 23:14, Aaron Lunansky wrote:
> So what I've figured out so far is that the system is not crashed- the
> screen is still on in fact, blinking a cursor in the top left. I can ssh
> in to the machine but a cursory review of dmesg shows nothing of
> significance.
>
> Unfortunately I can't do much with an ssh session before it seems to lock up,
> but I can launch another ssh session if I want. Also, if I disconnect my
> USB keyboard while the screen is in this blinking state, I get a "USB
> device soandso" disconnected message.
>
> Since my machines not crashing I can poke at it some more, but I could
> use some pointers as to where I should poke.

You didn't reply to my questions.

> Thanks,
> Aaron
>
> On Thu, Apr 08, 2004 at 08:49:31AM +1000, Benjamin Herrenschmidt wrote:
> > On Thu, 2004-04-08 at 05:17, Aaron Lunansky wrote:
> > > I'm having similar issues with an ibook G3 (600MHz) - only my system
> > > seems to fail to sleep, and this is regardless of modules (I have
> > > sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
> > > get a blank screen and it does not respond to any input. I'm not sure
> > > where it is when it dies though - what's the best way to figure this
> > > out? (how can I make SysRq work on an ibook (which has no SysRq key)).
> >
> > No, once the screen is off, there isn't much you can do. I suggest
> > you do some experiments in console mode, use "snooze -f" (that will
> > bypass pmud or pbbuttonsd, which tend to cut the backlight early).
> >
> > Also, did it die on sleep or on wakeup ? That is, did it reach the
> > sleep state where the led starts "snoring" ?
> >
> > > I get this with every 2.6.* kernel I've tried, including 2.6.5. I
> > > previously had sleep working on my ibook with a 2.6 kernel, but I've
> > > since replaced the (dead) hard drive. I can get my ibook to sleep with
> > > a 2.4.24 kernel.
> > >
> > > Thanks,
> > > Aaron
--
Benjamin Herrenschmidt <benh@kernel.crashing.org>


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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Aaron Lunansky]

What I was saying is the system doesn't seem to crash, but ends up in a bad state.
The system doesn't make it to sleep (the led does not 'snore').

Did I miss anything?

Thanks,
Aaron

On Fri, Apr 09, 2004 at 04:03:32PM +1000, Benjamin Herrenschmidt wrote:
>
> On Thu, 2004-04-08 at 23:14, Aaron Lunansky wrote:
> > So what I've figured out so far is that the system is not crashed- the
> > screen is still on in fact, blinking a cursor in the top left. I can ssh
> > in to the machine but a cursory review of dmesg shows nothing of
> > significance.
> >
> > Unfortunately I can't do much with an ssh session before it seems to lock up,
> > but I can launch another ssh session if I want. Also, if I disconnect my
> > USB keyboard while the screen is in this blinking state, I get a "USB
> > device soandso" disconnected message.
> >
> > Since my machines not crashing I can poke at it some more, but I could
> > use some pointers as to where I should poke.
>
> You didn't reply to my questions.
>
> > Thanks,
> > Aaron
> >
> > On Thu, Apr 08, 2004 at 08:49:31AM +1000, Benjamin Herrenschmidt wrote:
> > > On Thu, 2004-04-08 at 05:17, Aaron Lunansky wrote:
> > > > I'm having similar issues with an ibook G3 (600MHz) - only my system
> > > > seems to fail to sleep, and this is regardless of modules (I have
> > > > sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
> > > > get a blank screen and it does not respond to any input. I'm not sure
> > > > where it is when it dies though - what's the best way to figure this
> > > > out? (how can I make SysRq work on an ibook (which has no SysRq key)).
> > >
> > > No, once the screen is off, there isn't much you can do. I suggest
> > > you do some experiments in console mode, use "snooze -f" (that will
> > > bypass pmud or pbbuttonsd, which tend to cut the backlight early).
> > >
> > > Also, did it die on sleep or on wakeup ? That is, did it reach the
> > > sleep state where the led starts "snoring" ?

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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Benjamin Herrenschmidt]

On Fri, 2004-04-09 at 23:47, Aaron Lunansky wrote:
> What I was saying is the system doesn't seem to crash, but ends up in a bad state.
> The system doesn't make it to sleep (the led does not 'snore').
>
> Did I miss anything?

Ah ok, that's what I wanted to know. I tdid _not_ snore.

At this point, I don't know what's up. Make sure you don't have a
broken driver (like Alsa). Eventually remove pmac_zilog too (it
should work but then, I did hear about problems with earlier versions)


> Thanks,
> Aaron
>
> On Fri, Apr 09, 2004 at 04:03:32PM +1000, Benjamin Herrenschmidt wrote:
> >
> > On Thu, 2004-04-08 at 23:14, Aaron Lunansky wrote:
> > > So what I've figured out so far is that the system is not crashed- the
> > > screen is still on in fact, blinking a cursor in the top left. I can ssh
> > > in to the machine but a cursory review of dmesg shows nothing of
> > > significance.
> > >
> > > Unfortunately I can't do much with an ssh session before it seems to lock up,
> > > but I can launch another ssh session if I want. Also, if I disconnect my
> > > USB keyboard while the screen is in this blinking state, I get a "USB
> > > device soandso" disconnected message.
> > >
> > > Since my machines not crashing I can poke at it some more, but I could
> > > use some pointers as to where I should poke.
> >
> > You didn't reply to my questions.
> >
> > > Thanks,
> > > Aaron
> > >
> > > On Thu, Apr 08, 2004 at 08:49:31AM +1000, Benjamin Herrenschmidt wrote:
> > > > On Thu, 2004-04-08 at 05:17, Aaron Lunansky wrote:
> > > > > I'm having similar issues with an ibook G3 (600MHz) - only my system
> > > > > seems to fail to sleep, and this is regardless of modules (I have
> > > > > sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
> > > > > get a blank screen and it does not respond to any input. I'm not sure
> > > > > where it is when it dies though - what's the best way to figure this
> > > > > out? (how can I make SysRq work on an ibook (which has no SysRq key)).
> > > >
> > > > No, once the screen is off, there isn't much you can do. I suggest
> > > > you do some experiments in console mode, use "snooze -f" (that will
> > > > bypass pmud or pbbuttonsd, which tend to cut the backlight early).
> > > >
> > > > Also, did it die on sleep or on wakeup ? That is, did it reach the
> > > > sleep state where the led starts "snoring" ?
--
Benjamin Herrenschmidt <benh@kernel.crashing.org>


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

[resolved] Re: 2.6.5/iBook G3: Oops on resume from sleep

[Aaron Lunansky]

heh, good call, it was pmac_zilog that was causing the weird behaviour.
Removed it from my kernel (it was compiled in) and my system sleeps fine now.

Thanks,
Aaron

On Sat, Apr 10, 2004 at 09:52:40AM +1000, Benjamin Herrenschmidt wrote:
>
> On Fri, 2004-04-09 at 23:47, Aaron Lunansky wrote:
> > What I was saying is the system doesn't seem to crash, but ends up in a bad state.
> > The system doesn't make it to sleep (the led does not 'snore').
> >
> > Did I miss anything?
>
> Ah ok, that's what I wanted to know. I tdid _not_ snore.
>
> At this point, I don't know what's up. Make sure you don't have a
> broken driver (like Alsa). Eventually remove pmac_zilog too (it
> should work but then, I did hear about problems with earlier versions)

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

Re: [resolved] Re: 2.6.5/iBook G3: Oops on resume from sleep

[Benjamin Herrenschmidt]

On Sat, 2004-04-10 at 11:20, Aaron Lunansky wrote:
> heh, good call, it was pmac_zilog that was causing the weird behaviour.
> Removed it from my kernel (it was compiled in) and my system sleeps fine now.

Can you enable debugging in the driver (turn #undef DEBUG to #define
DEBUG at the beginning of the driver, you may also need to boot the
kernel with "debug" on the command line to get the messages).

See what messages you get from the driver both during boot and when
trying to suspend. Eventually, add some more output to pmz_suspend()
function to find out where it dies

Ben.


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

Re: 2.6.5/iBook G3: Oops on resume from sleep

[Michael Schmitz]

> > seems to fail to sleep, and this is regardless of modules (I have
> > sound/airport/ethernet compiled as modules). By fails to sleep, I mean I
> > get a blank screen and it does not respond to any input. I'm not sure
> > where it is when it dies though - what's the best way to figure this
> > out? (how can I make SysRq work on an ibook (which has no SysRq key)).
>
> No, once the screen is off, there isn't much you can do. I suggest
> you do some experiments in console mode, use "snooze -f" (that will
> bypass pmud or pbbuttonsd, which tend to cut the backlight early).

The latest Debian pmud doesn't play with backlight anymore (as per your
bug report), rather leaves backlight control to the kernel.

	Michael


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