Hi, I just got an IBM ThinkPad T42 (model 2378FVU) with
Centrino/Pentium M 735 and a Radeon 9600. I'm running 2.6.9-rc3 (with
ipw2200 0.9 wireless drivers) and I've experienced several hard
lockups over the past few days.
The system seems to be completely unresponsive to keyboard and
network, and even "nmi_watchdog=2" didn't produce anything. I'm not
sure what triggers the lockup -- I've had them happen while the system
was idle running an X screensaver, and also while I've been on the
console (non-X) doing nothing but typing through an ssh connection.
Generally it takes a couple of hours for the lockup to happen.
Has anyone seen anything similar? Any ideas on how to go about debugging this?
I'm including my modules and /proc/interrupts, and my .config is attached.
Thanks,
Roland
CPU0
0: 390269 XT-PIC timer
1: 2320 XT-PIC i8042
2: 0 XT-PIC cascade
8: 4 XT-PIC rtc
9: 212 XT-PIC acpi
11: 828 XT-PIC ehci_hcd, uhci_hcd, uhci_hcd,
uhci_hcd, Intel 82801DB-ICH4, yenta, yenta, ipw2200
12: 18184 XT-PIC i8042
14: 7395 XT-PIC ide0
15: 3511 XT-PIC ide1
NMI: 40
LOC: 106169
ERR: 0
MIS: 0
Module Size Used by
ds 17988 4
parport_pc 34112 1
lp 10500 0
parport 39624 2 parport_pc,lp
thermal 12816 0
fan 3852 0
button 6552 0
ac 4748 0
battery 9228 0
ipv6 250496 14
ipw2200 104172 0
ieee80211 21060 1 ipw2200
ieee80211_crypt 5700 1 ieee80211
yenta_socket 20032 0
pcmcia_core 61388 2 ds,yenta_socket
snd_intel8x0 32296 3
snd_ac97_codec 66960 1 snd_intel8x0
snd_pcm_oss 51368 0
snd_mixer_oss 18880 2 snd_pcm_oss
snd_pcm 94628 3 snd_intel8x0,snd_pcm_oss
snd_timer 24836 1 snd_pcm
snd_page_alloc 9800 2 snd_intel8x0,snd_pcm
snd_mpu401_uart 7424 1 snd_intel8x0
snd_rawmidi 24192 1 snd_mpu401_uart
snd_seq_device 7944 1 snd_rawmidi
snd 53220 12
snd_intel8x0,snd_ac97_codec,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer,snd_mpu401_uart,snd_rawmidi,snd_seq_device
intel_agp 21088 1
agpgart 33128 1 intel_agp
e1000 83204 0
On Thu, Sep 30, 2004 at 11:20:17AM -0700, Roland Dreier wrote:
> Hi, I just got an IBM ThinkPad T42 (model 2378FVU) with
> Centrino/Pentium M 735 and a Radeon 9600. I'm running 2.6.9-rc3 (with
> ipw2200 0.9 wireless drivers) and I've experienced several hard
> lockups over the past few days.
>
> The system seems to be completely unresponsive to keyboard and
> network, and even "nmi_watchdog=2" didn't produce anything. I'm not
> sure what triggers the lockup -- I've had them happen while the system
> was idle running an X screensaver, and also while I've been on the
> console (non-X) doing nothing but typing through an ssh connection.
> Generally it takes a couple of hours for the lockup to happen.
I have lockups in X running xlock with my T41p about once a month,
running 2.6.7-rc3-mm1 with atheros and the XFree4.3 radeon driver.
The only thing I noticed is that the hdd-led is constantly on when this
happens.
LLAP, Martin
On Thu, 30 Sep 2004 22:58:51 +0200, Martin Hermanowski <[email protected]> wrote:
> The only thing I noticed is that the hdd-led is constantly on when this
> happens.
Interesting... I just had another lockup and my HDD light was definitely off.
Thanks,
Roland
On 2004-09-30T14:45:48,
Roland Dreier <[email protected]> said:
> > The only thing I noticed is that the hdd-led is constantly on when
> > this happens.
> Interesting... I just had another lockup and my HDD light was
> definitely off.
I'm seeing similar problems on my T30.
Can you confirm that it happens mostly when X has dropped into the
screensaver and tries to come back?
Sincerely,
Lars Marowsky-Br?e <[email protected]>
--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX AG - A Novell company
> Can you confirm that it happens mostly when X has dropped into the
> screensaver and tries to come back?
I'm not sure. I've definitely seen it frozen in the screensaver a few
times (not necessarily when it tries to come back -- I've seen the
screensaver stuck as well). However, I did get a freeze just using
the console, and I've also gotten a freeze with the screensaver set to
just blank.
I guess right now I'm most suspicious about the following things:
- radeon driver. I've switched from radeonfb to vesafb and still had hangs
X11 ati driver still could be an issue
- ipw2200 driver
- interrupt sharing (everything seems to be wired to IRQ 11)
Unfortunately since I haven't been able to get _any_ post-hang
debugging info it's hard to make progress debugging.
Thanks,
Roland
On Thu, Sep 30, 2004 at 10:58:51PM +0200, Martin Hermanowski wrote:
moin,
> I have lockups in X running xlock with my T41p about once a month,
> running 2.6.7-rc3-mm1 with atheros and the XFree4.3 radeon driver.
>
> The only thing I noticed is that the hdd-led is constantly on when this
> happens.
i had that problem on my T41p (kernel oops when unlocking xlock, usually
if constant network traffic went on in the back).
updating madwifi to CVS version helped. (i had the version that comes
with suse 9.1 before).
tty, axel
One thing in common with the hang I've reported on this list seems to be the network traffic bit. My app does quite a bit of
traffic and uses thread mutex primitives, the SCHED_RR scheduler, and select() extensively.
A.
-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Axel Gordon
Grossklaus
Sent: Friday, October 01, 2004 8:47 AM
To: Martin Hermanowski
Cc: Roland Dreier; [email protected];
[email protected]
Subject: Re: Hard lockup on IBM ThinkPad T42
On Thu, Sep 30, 2004 at 10:58:51PM +0200, Martin Hermanowski wrote:
moin,
> I have lockups in X running xlock with my T41p about once a month,
> running 2.6.7-rc3-mm1 with atheros and the XFree4.3 radeon driver.
>
> The only thing I noticed is that the hdd-led is constantly on when this
> happens.
i had that problem on my T41p (kernel oops when unlocking xlock, usually
if constant network traffic went on in the back).
updating madwifi to CVS version helped. (i had the version that comes
with suse 9.1 before).
tty, axel
Last night I upgraded the ipw2200 driver from 0.9 to 0.11, and since
then I haven't had any lockups (13 hours of uptime -- pretty pathetic
in the Linux scheme of things but much better than what I was getting
before).
- Roland