[BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Bradley Chapman]

Mr. Morton,

I am experiencing defunct event/0 kernel daemons under 2.6.0-test6-mm1
with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20 with synaptics
support, and XFree86 4.3.0-10. Moving the touchpad in either X or with
gpm causes defunct event/0 processes to be created. This didn't happen
with 2.6.0-test5-mm4, using either synaptics_drv 0.11.6 or 0.11.7, and
I'm pretty sure that it didn't happen with 2.6.0-test6, but I will check
that.

I see in your e-mail here: 

http://www.ussg.iu.edu/hypermail/linux/kernel/0309.3/0926.html

that you merged a synaptics update in the -mm1 patchset. Do you have a
link to this patch so that I can revert it and see if it still occurs?

TIA

Brad Chapman

=====
Brad Chapman

Permanent e-mail: kakadu_croc@yahoo.com

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Andrew Morton]

Bradley Chapman <kakadu_croc@yahoo.com> wrote:
>
> I am experiencing defunct event/0 kernel daemons under 2.6.0-test6-mm1
>  with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20 with synaptics
>  support, and XFree86 4.3.0-10. Moving the touchpad in either X or with
>  gpm causes defunct event/0 processes to be created. 

Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?

You could try reverting synaptics-reconnect.patch, and then serio-reconnect.patch from

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test6/2.6.0-test6-mm1/broken-out

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Bradley Chapman]

Mr. Morton,

--- Andrew Morton <akpm@osdl.org> wrote:
> Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> >
> > I am experiencing defunct event/0 kernel daemons under 2.6.0-test6-mm1
> >  with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20 with synaptics
> >  support, and XFree86 4.3.0-10. Moving the touchpad in either X or with
> >  gpm causes defunct event/0 processes to be created. 
> 
> Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?

That was one of the first things I checked. There were no Oopses and nothing
else seemed to suffer, i.e. no high CPU usage or other nasty dmesg messages.

> 
> You could try reverting synaptics-reconnect.patch, and then serio-reconnect.patch
> from
> 
>
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test6/2.6.0-test6-mm1/broken-out
> 

OK. I will report back once I have rebooted.

Brad


=====
Brad Chapman

Permanent e-mail: kakadu_croc@yahoo.com

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Bradley Chapman]

Mr. Morton,

--- Andrew Morton <akpm@osdl.org> wrote:
> Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> >
> > I am experiencing defunct event/0 kernel daemons under 2.6.0-test6-mm1
> >  with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20 with synaptics
> >  support, and XFree86 4.3.0-10. Moving the touchpad in either X or with
> >  gpm causes defunct event/0 processes to be created. 
> 
> Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
> 
> You could try reverting synaptics-reconnect.patch, and then serio-reconnect.patch
> from
> 
>
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test6/2.6.0-test6-mm1/broken-out

I reverted these two patches and rebooted. No such luck; I noticed that immediately
after starting gpm and X and moving my Logitech MX700 mouse to start the Konsole,
I had many events/0 processes that were defunct. No Oopses at all. Here is the
relevant part of dmesg:

drivers/usb/core/usb.c: registered new driver hiddev
drivers/usb/core/usb.c: registered new driver hid
drivers/usb/input/hid-core.c: v2.0:USB HID core driver
mice: PS/2 mouse device common for all mice
i8042.c: Detected active multiplexing controller, rev 1.1.
serio: i8042 AUX0 port at 0x60,0x64 irq 12
serio: i8042 AUX1 port at 0x60,0x64 irq 12
serio: i8042 AUX2 port at 0x60,0x64 irq 12
Synaptics Touchpad, model: 1
 Firmware: 5.9
 Sensor: 28
 new absolute packet format
 Touchpad has extended capability bits
 -> four buttons
 -> multifinger detection
 -> palm detection
mousedev: attached device SynPS/2 Synaptics TouchPad at input/mouse0 <----- *
input: SynPS/2 Synaptics TouchPad on isa0060/serio4
serio: i8042 AUX3 port at 0x60,0x64 irq 12
input: AT Translated Set 2 keyboard on isa0060/serio0
serio: i8042 KBD port at 0x60,0x64 irq 1
...
hub 1-0:1.0: new USB device on port 1, assigned address 2
mousedev: attached device Logitech USB Receiver at input/mouse1 <---------- *
input: USB HID v1.10 Mouse [Logitech USB Receiver] on usb-0000:00:1d.0-1

If you want the rest of the USB parts, I can send that too.

Brad

[*] - This is a patch I added to mousedev.c at the end of mousedev_connect() -
when the input core connects a new device to mousedev, this message prints the
devnode it made it available on. It's just a debugging patch.


=====
Brad Chapman

Permanent e-mail: kakadu_croc@yahoo.com

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Dmitry Torokhov]

On Monday 29 September 2003 11:41 am, Andrew Morton wrote:
> Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> > I am experiencing defunct event/0 kernel daemons under
> > 2.6.0-test6-mm1 with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20
> > with synaptics support, and XFree86 4.3.0-10. Moving the touchpad in
> > either X or with gpm causes defunct event/0 processes to be created.
>
> Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
>
> You could try reverting synaptics-reconnect.patch, and then
> serio-reconnect.patch from
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-tes
>t6/2.6.0-test6-mm1/broken-out
>

Input subsystem uses only one kernel thread called kseriod, not eventsX.
I think it's not synaptics/serio reconnect but other patch you mentioned
(call_usermodehelper-retval-fix-2.patch)

Dmitry

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Bradley Chapman]

Mr. Torokhov,

--- Dmitry Torokhov <dtor_core@ameritech.net> wrote:
> On Monday 29 September 2003 11:41 am, Andrew Morton wrote:
> > Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> > > I am experiencing defunct event/0 kernel daemons under
> > > 2.6.0-test6-mm1 with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20
> > > with synaptics support, and XFree86 4.3.0-10. Moving the touchpad in
> > > either X or with gpm causes defunct event/0 processes to be created.
> >
> > Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
> >
> > You could try reverting synaptics-reconnect.patch, and then
> > serio-reconnect.patch from
> >
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-tes
> >t6/2.6.0-test6-mm1/broken-out
> >
> 
> Input subsystem uses only one kernel thread called kseriod, not eventsX.
> I think it's not synaptics/serio reconnect but other patch you mentioned
> (call_usermodehelper-retval-fix-2.patch)

OK. I'll try reverting that one too. I'm inclined to agree that it could be
the problem, because I noticed one more thing before I rebooted to 2.6.0-test6.

I recently installed the latest RH9 hotplug packages from the hotplug Sourceforge
website. Under 2.6.0-test6-mm1 the hotplug initscript leaves a defunct hotplug
process along with the other events/0 defunct processes. Does this mean anything?

The hotplug version is hotplug-2003_08_05-1.

> 
> Dmitry

Brad

=====
Brad Chapman

Permanent e-mail: kakadu_croc@yahoo.com

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Dmitry Torokhov]

On Monday 29 September 2003 12:49 pm, Bradley Chapman wrote:
> Mr. Torokhov,
>
> --- Dmitry Torokhov <dtor_core@ameritech.net> wrote:
> > On Monday 29 September 2003 11:41 am, Andrew Morton wrote:
> > > Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> > > > I am experiencing defunct event/0 kernel daemons under
> > > > 2.6.0-test6-mm1 with synaptics_drv 0.11.7, Dmitry Torokhov's
> > > > gpm-1.20 with synaptics support, and XFree86 4.3.0-10. Moving the
> > > > touchpad in either X or with gpm causes defunct event/0 processes
> > > > to be created.
> > >
> > > Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
> > >
> > > You could try reverting synaptics-reconnect.patch, and then
> > > serio-reconnect.patch from
> > >
> > > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0
> > >-tes t6/2.6.0-test6-mm1/broken-out
> >
> > Input subsystem uses only one kernel thread called kseriod, not
> > eventsX. I think it's not synaptics/serio reconnect but other patch
> > you mentioned (call_usermodehelper-retval-fix-2.patch)
>
> OK. I'll try reverting that one too. I'm inclined to agree that it
> could be the problem, because I noticed one more thing before I
> rebooted to 2.6.0-test6.
>
> I recently installed the latest RH9 hotplug packages from the hotplug
> Sourceforge website. Under 2.6.0-test6-mm1 the hotplug initscript
> leaves a defunct hotplug process along with the other events/0 defunct
> processes. Does this mean anything?
>
> The hotplug version is hotplug-2003_08_05-1.

Othe people reported that reverting call_suserhelper patch resolved their
eventsX/hotplug zombies problem.

Btw, GPM does not work with test6 or test6-mm yet, I will hopefully have
the upated version later tonight.

Dmitry

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Bradley Chapman]

Mr. Torokhov,

--- Dmitry Torokhov <dtor_core@ameritech.net> wrote:
> On Monday 29 September 2003 12:49 pm, Bradley Chapman wrote:
> > Mr. Torokhov,
> >
> > --- Dmitry Torokhov <dtor_core@ameritech.net> wrote:
> > > On Monday 29 September 2003 11:41 am, Andrew Morton wrote:
> > > > Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> > > > > I am experiencing defunct event/0 kernel daemons under
> > > > > 2.6.0-test6-mm1 with synaptics_drv 0.11.7, Dmitry Torokhov's
> > > > > gpm-1.20 with synaptics support, and XFree86 4.3.0-10. Moving the
> > > > > touchpad in either X or with gpm causes defunct event/0 processes
> > > > > to be created.
> > > >
> > > > Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
> > > >
> > > > You could try reverting synaptics-reconnect.patch, and then
> > > > serio-reconnect.patch from
> > > >
> > > > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0
> > > >-tes t6/2.6.0-test6-mm1/broken-out
> > >
> > > Input subsystem uses only one kernel thread called kseriod, not
> > > eventsX. I think it's not synaptics/serio reconnect but other patch
> > > you mentioned (call_usermodehelper-retval-fix-2.patch)
> >
> > OK. I'll try reverting that one too. I'm inclined to agree that it
> > could be the problem, because I noticed one more thing before I
> > rebooted to 2.6.0-test6.
> >
> > I recently installed the latest RH9 hotplug packages from the hotplug
> > Sourceforge website. Under 2.6.0-test6-mm1 the hotplug initscript
> > leaves a defunct hotplug process along with the other events/0 defunct
> > processes. Does this mean anything?
> >
> > The hotplug version is hotplug-2003_08_05-1.
> 
> Othe people reported that reverting call_suserhelper patch resolved their
> eventsX/hotplug zombies problem.

Yup! I'm running 2.6.0-test6-mm1 with synaptics-reconnect, serio-reconnect
and call_usermodehlper-retval-fix reverted, and no more zombies. It was the
call_usermodehelper reversion that fixed it. I looked at the patch and it
appears to mess with SIGCHLD signal handling and the kernel stack. How would
that mess up hotplug and the events/0 kernel thread?

> 
> Btw, GPM does not work with test6 or test6-mm yet, I will hopefully have
> the upated version later tonight.

Yes it does ;-)

On both 2.6.0-test6, 2.6.0-test6-mm1 full and 2.6.0-test6-mm1 minus the three
patches I mentioned above, the touchpad does work (at least it moves the cursor
and clicks). 

> 
> Dmitry

Brad

=====
Brad Chapman

Permanent e-mail: kakadu_croc@yahoo.com

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Chris Wright]

* Andrew Morton (akpm@osdl.org) wrote:
> Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> >
> > I am experiencing defunct event/0 kernel daemons under 2.6.0-test6-mm1
> >  with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20 with synaptics
> >  support, and XFree86 4.3.0-10. Moving the touchpad in either X or with
> >  gpm causes defunct event/0 processes to be created. 
> 
> Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
> 
> You could try reverting synaptics-reconnect.patch, and then serio-reconnect.patch from

Andrew, I wonder if this isn't caused by the call_usermodehelper patch.
Looks like you were right ;-)

thanks,
-chris
-- 
Linux Security Modules     http://lsm.immunix.org     http://lsm.bkbits.net

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Bradley Chapman]

Mr. Wright,

--- Chris Wright <chrisw@osdl.org> wrote:
> * Andrew Morton (akpm@osdl.org) wrote:
> > Bradley Chapman <kakadu_croc@yahoo.com> wrote:
> > >
> > > I am experiencing defunct event/0 kernel daemons under 2.6.0-test6-mm1
> > >  with synaptics_drv 0.11.7, Dmitry Torokhov's gpm-1.20 with synaptics
> > >  support, and XFree86 4.3.0-10. Moving the touchpad in either X or with
> > >  gpm causes defunct event/0 processes to be created. 
> > 
> > Defunct is odd.  Have you run `dmesg' to see if the kernel oopsed?
> > 
> > You could try reverting synaptics-reconnect.patch, and then
> serio-reconnect.patch from
> 
> Andrew, I wonder if this isn't caused by the call_usermodehelper patch.
> Looks like you were right ;-)

He is right. I reverted call_usermodehelper-retval-fix-2.patch and everything
works again. Why would that break the source of events/0 and hotplug?

I'd post a link to my reponse but the USSG archive is slow ;-)

> 
> thanks,
> -chris

Brad


=====
Brad Chapman

Permanent e-mail: kakadu_croc@yahoo.com

__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com

Re: [BUG] Defunct event/0 processes under 2.6.0-test6-mm1

[Chris Wright]

* Bradley Chapman (kakadu_croc@yahoo.com) wrote:
> He is right. I reverted call_usermodehelper-retval-fix-2.patch and everything
> works again. Why would that break the source of events/0 and hotplug?

That patch changes the SIGCHLD handler for kernel threads that are
started via wait_for_helper().  These threads are cloned from keventd
(hence events/0 as the process name), and they call hotplug.  The child
is not being reaped properly.

thanks,
-chris
-- 
Linux Security Modules     http://lsm.immunix.org     http://lsm.bkbits.net