Impact: fix truncated recursion bug message printout

When recursion_bug is true, kernel discards original message because printk_buf
contains recursion_bug_msg with NULL terminator. The sizeof(recursion_bug_msg)
makes this, use strlen() to get correct length without NULL terminator.
Reported-by: NToshikazu Nakayama <nakayama.ts@ncos.nec.co.jp>
Signed-off-by: NHiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

both log_buf_copy() and log_buf_len are unused.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Add documentation in kerneldoc for new printk format extensions

This patch documents the new %pS/%pF options in printk in kernel doc.

Hope I didn't miss any other extension.
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

People can use the real name an an index into MAINTAINERS to find the
current email address.
Signed-off-by: NFrancois Cami <francois.cami@free.fr>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add documentation in kerneldoc for new printk format extensions

This patch documents the new %pS/%pF options in printk in kernel doc.

Hope I didn't miss any other extension.
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NJonathan Corbet <corbet@lwn.net>

This is pure tty code so put it in the tty layer where it can be with the
locking relevant material it uses
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Remove the comment describing the possibility of printk() deadlocking on
runqueue lock.
Signed-off-by: NJiri Kosina <jkosina@suse.cz>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Dmitry Adamushko reported:

> [*] btw., with DEBUG being enabled, pr_debug() generates [1] when
> debug_smp_processor_id() is used (CONFIG_DEBUG_PREEMPT).
>
> the problem seems to be caused by the following commit:
> commit b845b517
> Author: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Date:   Fri Aug 8 21:47:09 2008 +0200
>
>     printk: robustify printk
>
>
> wake_up_klogd() -> __get_cpu_var() -> smp_processor_id()
>
> and that's being called from release_console_sem() which is, in turn,
> said to be "may be called from any context" [2]
>
> and in this case, it seems to be called from some non-preemptible
> context (although, it can't be printk()...
> although, I haven't looked carefully yet).
>
> Provided [2], __get_cpu_var() is perhaps not the right solution there.
>
>
> [1]
>
> [ 7697.942005] BUG: using smp_processor_id() in preemptible [00000000] code: syslogd/3542
> [ 7697.942005] caller is wake_up_klogd+0x1b/0x50
> [ 7697.942005] Pid: 3542, comm: syslogd Not tainted 2.6.27-rc3-tip-git #2
> [ 7697.942005] Call Trace:
> [ 7697.942005]  [<ffffffff8036b398>] debug_smp_processor_id+0xe8/0xf0
> [ 7697.942005]  [<ffffffff80239d3b>] wake_up_klogd+0x1b/0x50
> [ 7697.942005]  [<ffffffff8023a047>] release_console_sem+0x1e7/0x200
> [ 7697.942005]  [<ffffffff803c0f17>] do_con_write+0xb7/0x1f30
> [ 7697.942005]  [<ffffffff8020d920>] ? show_trace+0x10/0x20
> [ 7697.942005]  [<ffffffff8020dc42>] ? dump_stack+0x72/0x80
> [ 7697.942005]  [<ffffffff8036392d>] ? __ratelimit+0xbd/0xe0
> [ 7697.942005]  [<ffffffff8036b398>] ? debug_smp_processor_id+0xe8/0xf0
> [ 7697.942005]  [<ffffffff80239d3b>] ? wake_up_klogd+0x1b/0x50
> [ 7697.942005]  [<ffffffff8023a047>] ? release_console_sem+0x1e7/0x200
> [ 7697.942005]  [<ffffffff803c2de9>] con_write+0x19/0x30
> [ 7697.942005]  [<ffffffff803b37b6>] write_chan+0x276/0x3c0
> [ 7697.942005]  [<ffffffff80232b20>] ? default_wake_function+0x0/0x10
> [ 7697.942005]  [<ffffffff804cb872>] ? _spin_lock_irqsave+0x22/0x50
> [ 7697.942005]  [<ffffffff803b1334>] tty_write+0x194/0x260
> [ 7697.942005]  [<ffffffff803b3540>] ? write_chan+0x0/0x3c0
> [ 7697.942005]  [<ffffffff803b14a4>] redirected_tty_write+0xa4/0xb0
> [ 7697.942005]  [<ffffffff803b1400>] ? redirected_tty_write+0x0/0xb0
> [ 7697.942005]  [<ffffffff802a88c2>] do_loop_readv_writev+0x52/0x80
> [ 7697.942005]  [<ffffffff802a939d>] do_readv_writev+0x1bd/0x1d0
> [ 7697.942005]  [<ffffffff802a93e9>] vfs_writev+0x39/0x60
> [ 7697.942005]  [<ffffffff802a9870>] sys_writev+0x50/0x90
> [ 7697.942005]  [<ffffffff8020bb3b>] system_call_fastpath+0x16/0x1b
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Reported-by: NDmitry Adamushko <dmitry.adamushko@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Avoid deadlocks against rq->lock and xtime_lock by deferring the klogd
wakeup by polling from the timer tick.
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The comment assumed the burst to be one and the ratelimit used to be named
printk_ratelimit_jiffies.
Signed-off-by: NUwe Kleine-König <Uwe.Kleine-Koenig@digi.com>
Cc: Dave Young <hidave.darkstar@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

All ratelimit user use same jiffies and burst params, so some messages
(callbacks) will be lost.

For example:
a call printk_ratelimit(5 * HZ, 1)
b call printk_ratelimit(5 * HZ, 1) before the 5*HZ timeout of a, then b will
will be supressed.

- rewrite __ratelimit, and use a ratelimit_state as parameter.  Thanks for
  hints from andrew.

- Add WARN_ON_RATELIMIT, update rcupreempt.h

- remove __printk_ratelimit

- use __ratelimit in net_ratelimit
Signed-off-by: NDave Young <hidave.darkstar@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Dave Young <hidave.darkstar@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Tell the user about the no_console_suspend option, so that we don't have to
tell each bug reporter personally.

[akpm@linux-foundation.org: clarify the text a little]
Signed-off-by: NPavel Machek <pavel@suse.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NDaniel Guilak <daniel@danielguilak.com>
Acked-by: NJosh Triplett <josh@kernel.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

this symbol is needed by drivers/video/xen-fbfront.ko.

[ cherry-picked from tip/core/printk ]
Signed-off-by: NIngo Molnar <mingo@elte.hu>

It's not explicitly marked as asmlinkage, but invoked from x86_32
startup code with parameters on stack.

No other architectures define early_printk and none of them are affected
by this change, since defines asmlinkage as empty token.
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

this symbol is needed by drivers/video/xen-fbfront.ko.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Without console= arguments on the kernel command line, the first
console to register becomes enabled and the preferred console (the one
behind /dev/console).  This is normally tty (assuming
CONFIG_VT_CONSOLE is enabled, which it commonly is).

This is okay as long tty is a useful console.  But unless we have the
PV framebuffer, and it is enabled for this domain, tty0 in domU is
merely a dummy.  In that case, we want the preferred console to be the
Xen console hvc0, and we want it without having to fiddle with the
kernel command line.  Commit b8c2d3df
did that for us.

Since we now have the PV framebuffer, we want to enable and prefer tty
again, but only when PVFB is enabled.  But even then we still want to
enable the Xen console as well.

Problem: when tty registers, we can't yet know whether the PVFB is
enabled.  By the time we can know (xenstore is up), the console setup
game is over.

Solution: enable console tty by default, but keep hvc as the preferred
console.  Change the preferred console to tty when PVFB probes
successfully, unless we've been given console kernel parameters.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

printk(KERN_ALERT "Danger Will Robinson!\nAlien Approaching!\n");

At present this will result in one message at ALERT level and one
at the current default message loglevel (e.g. WARNING). This is
non-intuitive.

Modify vprintk() to remember the message loglevel each time it
is specified and use it for subsequent lines of output which do
not specify one, within the same call to printk.
Signed-off-by: NNick Andrew <nick@nick-andrew.net>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Restructure the logic of vprintk() so the processing of the leading
3 characters of each input line is in one place, regardless whether
printk_time is enabled. This makes the code smaller and easier to
understand.

size reduction in kernel/printk.o:

   text	   data	    bss	    dec	    hex	filename
   6157	    397	1049804	1056358	 101e66	printk.o.before
   6117	    397	1049804	1056318	 101e3e	printk.o.after

and some style uncleanlinesses removed as well as a side-effect:

 Before:
    total: 19 errors, 22 warnings, 1340 lines checked
 After:
    total: 17 errors, 22 warnings, 1333 lines checked
Signed-off-by: NNick Andrew <nick@nick-andrew.net>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

console election: If some console happens to be registered first which does
not provide a tty binding (!console->device), it prevents that more suited
consoles which are registered later on can enter the candidate pool for
console_device().  This is observable with KGDB's console which may already
be registered (and exploited!) during early debugger connections, that is
before any regular console registration.

This patch fixes the issue by postponing the final, automated
preferred_console selection until someone with a non-NULL device handler
comes around.
Signed-off-by: NJan Kiszka <jan.kiszka@web.de>
Cc: Jason Wessel <jason.wessel@windriver.com>
Cc: Gerd Hoffmann <kraxel@suse.de>
Cc: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Make printk_recursion_bug_msg static and drop printk prefix from recursion
variables.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

[ Stephen Rothwell <sfr@canb.auug.org.au>: build fix ]
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

This patch adds latency tracing for critical timings
(how long interrupts are disabled for).

 "irqsoff" is added to /debugfs/tracing/available_tracers

Note:
  tracing_max_latency
    also holds the max latency for irqsoff (in usecs).
   (default to large number so one must start latency tracing)

  tracing_thresh
    threshold (in usecs) to always print out if irqs off
    is detected to be longer than stated here.
    If irq_thresh is non-zero, then max_irq_latency
    is ignored.

Here's an example of a trace with ftrace_enabled = 0

=======
preemption latency trace v1.1.5 on 2.6.24-rc7
Signed-off-by: NIngo Molnar <mingo@elte.hu>
--------------------------------------------------------------------
 latency: 100 us, #3/3, CPU#1 | (M:rt VP:0, KP:0, SP:0 HP:0 #P:2)
    -----------------
    | task: swapper-0 (uid:0 nice:0 policy:0 rt_prio:0)
    -----------------
 => started at: _spin_lock_irqsave+0x2a/0xb7
 => ended at:   _spin_unlock_irqrestore+0x32/0x5f

                 _------=> CPU#
                / _-----=> irqs-off
               | / _----=> need-resched
               || / _---=> hardirq/softirq
               ||| / _--=> preempt-depth
               |||| /
               |||||     delay
   cmd     pid ||||| time  |   caller
      \   /    |||||   \   |   /
 swapper-0     1d.s3    0us+: _spin_lock_irqsave+0x2a/0xb7 (e1000_update_stats+0x47/0x64c [e1000])
 swapper-0     1d.s3  100us : _spin_unlock_irqrestore+0x32/0x5f (e1000_update_stats+0x641/0x64c [e1000])
 swapper-0     1d.s3  100us : trace_hardirqs_on_caller+0x75/0x89 (_spin_unlock_irqrestore+0x32/0x5f)

vim:ft=help
=======

And this is a trace with ftrace_enabled == 1

=======
preemption latency trace v1.1.5 on 2.6.24-rc7
--------------------------------------------------------------------
 latency: 102 us, #12/12, CPU#1 | (M:rt VP:0, KP:0, SP:0 HP:0 #P:2)
    -----------------
    | task: swapper-0 (uid:0 nice:0 policy:0 rt_prio:0)
    -----------------
 => started at: _spin_lock_irqsave+0x2a/0xb7
 => ended at:   _spin_unlock_irqrestore+0x32/0x5f

                 _------=> CPU#
                / _-----=> irqs-off
               | / _----=> need-resched
               || / _---=> hardirq/softirq
               ||| / _--=> preempt-depth
               |||| /
               |||||     delay
   cmd     pid ||||| time  |   caller
      \   /    |||||   \   |   /
 swapper-0     1dNs3    0us+: _spin_lock_irqsave+0x2a/0xb7 (e1000_update_stats+0x47/0x64c [e1000])
 swapper-0     1dNs3   46us : e1000_read_phy_reg+0x16/0x225 [e1000] (e1000_update_stats+0x5e2/0x64c [e1000])
 swapper-0     1dNs3   46us : e1000_swfw_sync_acquire+0x10/0x99 [e1000] (e1000_read_phy_reg+0x49/0x225 [e1000])
 swapper-0     1dNs3   46us : e1000_get_hw_eeprom_semaphore+0x12/0xa6 [e1000] (e1000_swfw_sync_acquire+0x36/0x99 [e1000])
 swapper-0     1dNs3   47us : __const_udelay+0x9/0x47 (e1000_read_phy_reg+0x116/0x225 [e1000])
 swapper-0     1dNs3   47us+: __delay+0x9/0x50 (__const_udelay+0x45/0x47)
 swapper-0     1dNs3   97us : preempt_schedule+0xc/0x84 (__delay+0x4e/0x50)
 swapper-0     1dNs3   98us : e1000_swfw_sync_release+0xc/0x55 [e1000] (e1000_read_phy_reg+0x211/0x225 [e1000])
 swapper-0     1dNs3   99us+: e1000_put_hw_eeprom_semaphore+0x9/0x35 [e1000] (e1000_swfw_sync_release+0x50/0x55 [e1000])
 swapper-0     1dNs3  101us : _spin_unlock_irqrestore+0xe/0x5f (e1000_update_stats+0x641/0x64c [e1000])
 swapper-0     1dNs3  102us : _spin_unlock_irqrestore+0x32/0x5f (e1000_update_stats+0x641/0x64c [e1000])
 swapper-0     1dNs3  102us : trace_hardirqs_on_caller+0x75/0x89 (_spin_unlock_irqrestore+0x32/0x5f)

vim:ft=help
=======
Signed-off-by: NSteven Rostedt <srostedt@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Fix update_console_cmdline() not to to read beyond the terminating zero of its
name argument.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This adds a minimalistic braille screen reader support.  This is meant to
be used by blind people e.g.  on boot failures or when / cannot be mounted
etc and thus the userland screen readers can not work.

[akpm@linux-foundation.org: fix exports]
Signed-off-by: NSamuel Thibault <samuel.thibault@ens-lyon.org>
Cc: Jiri Kosina <jikos@jikos.cz>
Cc: Dmitry Torokhov <dtor@mail.ru>
Acked-by: NAlan Cox <alan@redhat.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

- Operations are now a shared const function block as with most other Linux
  objects

- Introduce wrappers for some optional functions to get consistent behaviour

- Wrap put_char which used to be patched by the tty layer

- Document which functions are needed/optional

- Make put_char report success/fail

- Cache the driver->ops pointer in the tty as tty->ops

- Remove various surplus lock calls we no longer need

- Remove proc_write method as noted by Alexey Dobriyan

- Introduce some missing sanity checks where certain driver/ldisc
  combinations would oops as they didn't check needed methods were present

[akpm@linux-foundation.org: fix fs/compat_ioctl.c build]
[akpm@linux-foundation.org: fix isicom]
[akpm@linux-foundation.org: fix arch/ia64/hp/sim/simserial.c build]
[akpm@linux-foundation.org: fix kgdb]
Signed-off-by: NAlan Cox <alan@redhat.com>
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Cc: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Due to the rcupreempt.h WARN_ON trigged, I got 2G syslog file.  For some
serious complaining of kernel, we need repeat the warnings, so here I isolate
the ratelimit part of printk.c to a standalone file.
Signed-off-by: NDave Young <hidave.darkstar@gmail.com>
Acked-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When I cleaned up printk() and split up the printk locking logic in
commit 266c2e0a ("Make printk() console
semaphore accesses sensible") I had incorrectly moved the call to
have_callable_console() outside of the console semaphore.

That was buggy.  The console semaphore protects the console_drivers list
that is used by have_callable_console().

Thanks go to Bongani Hlope who saw this as a hang on shutdown and reboot
and bisected the bug to the right commit, and tested this patch. See

	http://lkml.org/lkml/2008/4/11/315Bisected-and-tested-by: NBongani Hlope <bonganilinux@mweb.co.za>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The printk() logic on when/how to get the console semaphore was
unreadable, this splits the code up into a few helper functions and
makes it easier to follow what is going on.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

printk recursion detection prepends message to printk_buf and offsets
printk_buf when actual message is printed but it forgets to trim buffer
length accordingly. This can result in overrun in extreme cases. Fix it.

[ mingo@elte.hu:

  bug was introduced by me via:

   commit 32a76006
   Author: Ingo Molnar <mingo@elte.hu>
   Date:   Fri Jan 25 21:07:58 2008 +0100

       printk: make printk more robust by not allowing recursion
]
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Makes an embedded image a bit smaller.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The scheduled removal of the 'time' option.
Signed-off-by: NAdrian Bunk <bunk@kernel.org>
Acked-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Remove the duplicate inclusion of linux/jiffies.h from kernel/printk.c
Signed-off-by: NJesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Stop using unsigned _longs_ for printk buffer indexes.  Log buffer is way
smaller than 2 gigabytes and unsigned ints will work too .  Indeed, they do
work nicely on all 32-bit platforms where longs and ints are the same.

With this patch, we have following size savings on amd64:

   text    data     bss     dec     hex filename
   5997     313   17736   24046    5dee 2.6.23.1.t64/kernel/printk.o
   5858     313   17700   23871    5d3f 2.6.23.1.printk.t64/kernel/printk.o
Signed-off-by: NDenys Vlasenko <vda.linux@googlemail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

i was debugging early crashes and wondered where all the printks
went. The reason: ignore_loglevel_setup() was not called yet ...
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Enable CONFIG_EMBEDDED to select CONFIG_EARLY_PRINTK on 64-bit as well.

saves ~2K:

   text    data     bss     dec     hex filename
   7290283 3672091 1907848 12870222         c4624e vmlinux.before
   7288373 3671795 1907848 12868016         c459b0 vmlinux.after
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

revert 19ef9309.

Kevin Winchester reported a lockup during X startup an bisected
it to this commit.
Reported-by: NKevin Winchester <kjwinchester@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

printk timestamps: use ktime_get().

Some platforms have a functioning clocksource function only after
they are done with early bootup, so delay this until out of
SYSTEM_BOOTING state.

it's also inherently safe now, as any bugs in this area will be
caught by the printk recursion checks.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

printk_clock() is obsolete - it has been replaced with cpu_clock().
Signed-off-by: NIngo Molnar <mingo@elte.hu>