2011-02-28 21:41:00

by Alexey Dobriyan

[permalink] [raw]
Subject: 2.6.38-rc6: SysRq broken?

It just strikes me that SysRq is broken here:
Alt+SysRq+h doesn't print anything

CONFIG_MAGIC_SYSRQ=y
CONFIG_INPUT_EVDEV=y
kernel.sysrq=1
echo h >/proc/sysrq-trigger works

Keyboard is USB one, "Logitech HID compliant keyboard".


2011-02-28 21:57:30

by Dmitry Torokhov

[permalink] [raw]
Subject: Re: 2.6.38-rc6: SysRq broken?

On Monday, February 28, 2011 01:40:54 pm Alexey Dobriyan wrote:
> It just strikes me that SysRq is broken here:
> Alt+SysRq+h doesn't print anything
>
> CONFIG_MAGIC_SYSRQ=y
> CONFIG_INPUT_EVDEV=y
> kernel.sysrq=1
> echo h >/proc/sysrq-trigger works
>
> Keyboard is USB one, "Logitech HID compliant keyboard".

What does /proc/bus/input/devices show?

--
Dmitry

2011-02-28 22:00:06

by Bruno Prémont

[permalink] [raw]
Subject: Re: 2.6.38-rc6: SysRq broken?

On Mon, 28 February 2011 Alexey Dobriyan <[email protected]> wrote:
> It just strikes me that SysRq is broken here:
> Alt+SysRq+h doesn't print anything
>
> CONFIG_MAGIC_SYSRQ=y
> CONFIG_INPUT_EVDEV=y
> kernel.sysrq=1
> echo h >/proc/sysrq-trigger works
>
> Keyboard is USB one, "Logitech HID compliant keyboard".

I've seen same kind of issue with built-in keybaord (PS2) of my laptop.

Sysrq-Sync, Sysrq-Umount, Sysrq-Help didn't work but Sysrq-reBoot did.

In my case it was during "stuck" shutdown process, caps key was still
working and toggled caps LED anything else didn't trigger anything visible.

Bruno

2011-02-28 22:04:54

by Alexey Dobriyan

[permalink] [raw]
Subject: Re: 2.6.38-rc6: SysRq broken?

On Mon, Feb 28, 2011 at 01:57:21PM -0800, Dmitry Torokhov wrote:
> On Monday, February 28, 2011 01:40:54 pm Alexey Dobriyan wrote:
> > It just strikes me that SysRq is broken here:
> > Alt+SysRq+h doesn't print anything
> >
> > CONFIG_MAGIC_SYSRQ=y
> > CONFIG_INPUT_EVDEV=y
> > kernel.sysrq=1
> > echo h >/proc/sysrq-trigger works
> >
> > Keyboard is USB one, "Logitech HID compliant keyboard".
>
> What does /proc/bus/input/devices show?

I: Bus=0019 Vendor=0000 Product=0001 Version=0000
N: Name="Power Button"
P: Phys=PNP0C0C/button/input0
S: Sysfs=/devices/LNXSYSTM:00/device:00/PNP0C0C:00/input/input0
U: Uniq=
H: Handlers=kbd event0
B: PROP=0
B: EV=3
B: KEY=10000000000000 0

I: Bus=0019 Vendor=0000 Product=0001 Version=0000
N: Name="Power Button"
P: Phys=LNXPWRBN/button/input0
S: Sysfs=/devices/LNXSYSTM:00/LNXPWRBN:00/input/input1
U: Uniq=
H: Handlers=kbd event1
B: PROP=0
B: EV=3
B: KEY=10000000000000 0

I: Bus=0003 Vendor=093a Product=2510 Version=0111
N: Name="PIXART USB OPTICAL MOUSE"
P: Phys=usb-0000:00:1d.0-1.1/input0
S: Sysfs=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.1/2-1.1:1.0/input/input2
U: Uniq=
H: Handlers=mouse0 event2
B: PROP=0
B: EV=17
B: KEY=70000 0 0 0 0
B: REL=103
B: MSC=10

I: Bus=0003 Vendor=046d Product=c30e Version=0110
N: Name="Logitech HID compliant keyboard"
P: Phys=usb-0000:00:1d.0-1.2/input0
S: Sysfs=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/input/input3
U: Uniq=
H: Handlers=sysrq kbd event3
B: PROP=0
B: EV=120013
B: KEY=1000000000007 ff800000000007ff febeffdff3cfffff fffffffffffffffe
B: MSC=10
B: LED=7

I: Bus=0003 Vendor=046d Product=c30e Version=0110
N: Name="Logitech HID compliant keyboard"
P: Phys=usb-0000:00:1d.0-1.2/input1
S: Sysfs=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.1/input/input4
U: Uniq=
H: Handlers=kbd event4
B: PROP=0
B: EV=13
B: KEY=fff ffffffffffffffff 2000000 387ad800d001 1e000000000000 0
B: MSC=10

2011-03-01 18:04:42

by Dmitry Torokhov

[permalink] [raw]
Subject: Re: 2.6.38-rc6: SysRq broken?

On Tue, Mar 01, 2011 at 12:04:47AM +0200, Alexey Dobriyan wrote:
> On Mon, Feb 28, 2011 at 01:57:21PM -0800, Dmitry Torokhov wrote:
> > On Monday, February 28, 2011 01:40:54 pm Alexey Dobriyan wrote:
> > > It just strikes me that SysRq is broken here:
> > > Alt+SysRq+h doesn't print anything
> > >
> > > CONFIG_MAGIC_SYSRQ=y
> > > CONFIG_INPUT_EVDEV=y
> > > kernel.sysrq=1
> > > echo h >/proc/sysrq-trigger works
> > >
> > > Keyboard is USB one, "Logitech HID compliant keyboard".
> >
> > What does /proc/bus/input/devices show?
>
>
> I: Bus=0003 Vendor=046d Product=c30e Version=0110
> N: Name="Logitech HID compliant keyboard"
> P: Phys=usb-0000:00:1d.0-1.2/input0
> S: Sysfs=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/input/input3
> U: Uniq=
> H: Handlers=sysrq kbd event3
> B: PROP=0
> B: EV=120013
> B: KEY=1000000000007 ff800000000007ff febeffdff3cfffff fffffffffffffffe
> B: MSC=10
> B: LED=7
>

Hmm, looks like we bound sysrq handler here and the device has both alt
and SysRq keys. Is it only help that is not working? What abouth other
functions (reboot, etc)?

--
Dmitry

2011-03-01 21:00:57

by Alexey Dobriyan

[permalink] [raw]
Subject: Re: 2.6.38-rc6: SysRq broken?

On Tue, Mar 01, 2011 at 10:04:29AM -0800, Dmitry Torokhov wrote:
> On Tue, Mar 01, 2011 at 12:04:47AM +0200, Alexey Dobriyan wrote:
> > On Mon, Feb 28, 2011 at 01:57:21PM -0800, Dmitry Torokhov wrote:
> > > On Monday, February 28, 2011 01:40:54 pm Alexey Dobriyan wrote:
> > > > It just strikes me that SysRq is broken here:
> > > > Alt+SysRq+h doesn't print anything
> > > >
> > > > CONFIG_MAGIC_SYSRQ=y
> > > > CONFIG_INPUT_EVDEV=y
> > > > kernel.sysrq=1
> > > > echo h >/proc/sysrq-trigger works
> > > >
> > > > Keyboard is USB one, "Logitech HID compliant keyboard".
> > >
> > > What does /proc/bus/input/devices show?
> >
> >
> > I: Bus=0003 Vendor=046d Product=c30e Version=0110
> > N: Name="Logitech HID compliant keyboard"
> > P: Phys=usb-0000:00:1d.0-1.2/input0
> > S: Sysfs=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/input/input3
> > U: Uniq=
> > H: Handlers=sysrq kbd event3
> > B: PROP=0
> > B: EV=120013
> > B: KEY=1000000000007 ff800000000007ff febeffdff3cfffff fffffffffffffffe
> > B: MSC=10
> > B: LED=7
> >
>
> Hmm, looks like we bound sysrq handler here and the device has both alt
> and SysRq keys.

Yes, both.

> Is it only help that is not working? What abouth other functions (reboot, etc)?

Seems so, SysRq+Sync works, SysRq+L works as well.

2011-03-01 21:40:27

by Bruno Prémont

[permalink] [raw]
Subject: Re: 2.6.38-rc6: SysRq broken?

On Mon, 28 February 2011 Bruno Prémont <[email protected]> wrote:
> On Mon, 28 February 2011 Alexey Dobriyan <[email protected]> wrote:
> > It just strikes me that SysRq is broken here:
> > Alt+SysRq+h doesn't print anything
> >
> > CONFIG_MAGIC_SYSRQ=y
> > CONFIG_INPUT_EVDEV=y
> > kernel.sysrq=1
> > echo h >/proc/sysrq-trigger works
> >
> > Keyboard is USB one, "Logitech HID compliant keyboard".
>
> I've seen same kind of issue with built-in keybaord (PS2) of my laptop.
>
> Sysrq-Sync, Sysrq-Umount, Sysrq-Help didn't work but Sysrq-reBoot did.
>
> In my case it was during "stuck" shutdown process, caps key was still
> working and toggled caps LED anything else didn't trigger anything visible.
>
> Bruno

Well it worked (half way?) though fb was probably frozen (vt-deadlock?)
as I have traces in my log:

[14382.063476] pcmanfm[28117]: segfault at 8 ip b772e5b5 sp bfcfba30 error 4 in libfm-gtk.so.0.0.0[b7700000+34000]
[43437.896858] SysRq : Emergency Sync
[43572.221361] SysRq : Keyboard mode set to system default
[43575.574551] SysRq : Emergency Sync
[43590.513441] SysRq : Emergency Remount R/O
[43605.923769] SysRq : Emergency Sync
[43607.320324] SysRq : Emergency Remount R/O
[43627.329605] SysRq : HELP : loglevel(0-9) reBoot Crash terminate-all-tasks(E) memory-full-oom-kill(F) kill-all-tasks(I)
thaw-filesystems(J) saK show-memory-usage(M) nice-all-RT-tasks(N) powerOff show-registers(P) show-all-timers(Q) unRaw Sync show-task-states(T) Unmount force-fb
(V) show-blocked-tasks(W)
[43676.561306] SysRq : Emergency Sync
[43690.319213] SysRq : SAK
[43690.319245] SAK: killed process 1926 (Xorg): task_session(p)==tty->session
[43690.319313] SAK: killed process 1926 (Xorg): task_session(p)==tty->session
[43690.319345] SAK: killed process 2801 (rc): fd#0 opened to the tty
[43690.319358] SAK: killed process 2802 (rc): fd#0 opened to the tty
[43690.319368] SAK: killed process 2890 (ntpd): fd#0 opened to the tty
[43690.319378] SAK: killed process 2891 (runscript.sh): fd#0 opened to the tty
[43690.319387] SAK: killed process 2898 (start-stop-daem): fd#0 opened to the tty
[43694.809374] SysRq : Keyboard mode set to system default
[43797.368728] SysRq : SAK
[43797.368768] SAK: killed process 1926 (Xorg): task_session(p)==tty->session
[43797.368829] SAK: killed process 1926 (Xorg): task_session(p)==tty->session
[43797.368864] SAK: killed process 2898 (start-stop-daem): fd#0 opened to the tty

But then, "Emergency Remount R/O" should have prevented anything further
from hitting the disk. (and after sysrq-reboot filesystems were dirty)

Bruno