linux-toradex.git/include/linux/hardirq.h, branch v3.4.73

rcu: Track idleness independent of idle tasks

2011-12-11T18:31:24+00:00

Earlier versions of RCU used the scheduling-clock tick to detect idleness
by checking for the idle task, but handled idleness differently for
CONFIG_NO_HZ=y.  But there are now a number of uses of RCU read-side
critical sections in the idle task, for example, for tracing.  A more
fine-grained detection of idleness is therefore required.

This commit presses the old dyntick-idle code into full-time service,
so that rcu_idle_enter(), previously known as rcu_enter_nohz(), is
always invoked at the beginning of an idle loop iteration.  Similarly,
rcu_idle_exit(), previously known as rcu_exit_nohz(), is always invoked
at the end of an idle-loop iteration.  This allows the idle task to
use RCU everywhere except between consecutive rcu_idle_enter() and
rcu_idle_exit() calls, in turn allowing architecture maintainers to
specify exactly where in the idle loop that RCU may be used.

Because some of the userspace upcall uses can result in what looks
to RCU like half of an interrupt, it is not possible to expect that
the irq_enter() and irq_exit() hooks will give exact counts.  This
patch therefore expands the ->dynticks_nesting counter to 64 bits
and uses two separate bitfields to count process/idle transitions
and interrupt entry/exit transitions.  It is presumed that userspace
upcalls do not happen in the idle loop or from usermode execution
(though usermode might do a system call that results in an upcall).
The counter is hard-reset on each process/idle transition, which
avoids the interrupt entry/exit error from accumulating.  Overflow
is avoided by the 64-bitness of the ->dyntick_nesting counter.

This commit also adds warnings if a non-idle task asks RCU to enter
idle state (and these checks will need some adjustment before applying
Frederic's OS-jitter patches (http://lkml.org/lkml/2011/10/7/246).
In addition, validation of ->dynticks and ->dynticks_nesting is added.

Signed-off-by: Paul E. McKenney 
Signed-off-by: Paul E. McKenney 
Reviewed-by: Josh Triplett

sched: Isolate preempt counting in its own config option

2011-06-10T13:15:40+00:00

Create a new CONFIG_PREEMPT_COUNT that handles the inc/dec
of preempt count offset independently. So that the offset
can be updated by preempt_disable() and preempt_enable()
even without the need for CONFIG_PREEMPT beeing set.

This prepares to make CONFIG_DEBUG_SPINLOCK_SLEEP working
with !CONFIG_PREEMPT where it currently doesn't detect
code that sleeps inside explicit preemption disabled
sections.

Signed-off-by: Frederic Weisbecker 
Acked-by: Paul E. McKenney 
Cc: Ingo Molnar 
Cc: Peter Zijlstra

BKL: That's all, folks

2011-03-05T09:56:00+00:00

This removes the implementation of the big kernel lock,
at last. A lot of people have worked on this in the
past, I so the credit for this patch should be with
everyone who participated in the hunt.

The names on the Cc list are the people that were the
most active in this, according to the recorded git
history, in alphabetical order.

Signed-off-by: Arnd Bergmann 
Acked-by: Alan Cox 
Cc: Alessio Igor Bogani 
Cc: Al Viro 
Cc: Andrew Hendry 
Cc: Andrew Morton 
Cc: Christoph Hellwig 
Cc: Eric W. Biederman 
Cc: Frederic Weisbecker 
Cc: Hans Verkuil 
Acked-by: Ingo Molnar 
Cc: Jan Blunck 
Cc: John Kacur 
Cc: Jonathan Corbet 
Cc: Linus Torvalds 
Cc: Matthew Wilcox 
Cc: Oliver Neukum 
Cc: Paul Menage 
Acked-by: Thomas Gleixner 
Cc: Trond Myklebust

hardirq.h: needs sched.h if using BKL

2010-11-18T18:56:29+00:00

This really isn't the right thing to do, and strictly speaking we should
have the BKL depth count in the thread info right next to the preempt
count.  The two really do go together.

However, since that would involve a patch to all architectures, and the
BKL is finally going away, it's simply not worth the effort to do the
RightThing(tm).  Just re-instate the  include that we
used to get accidentally from the smp_lock.h one.

This is all fallout from the same old "BKL: remove extraneous #include
" commit.

Reported-by: Ingo Molnar 
Tested-by: Randy Dunlap 
Cc: Arnd Bergmann 
Signed-off-by: Linus Torvalds

hardirq.h: remove now-empty #ifdef/#endif pair

2010-11-18T02:36:25+00:00

Commit 451a3c24b013 ("BKL: remove extraneous #include ")
removed the #include line that was the only thing that was surrounded by
the #ifdef/#endif.

So now that #ifdef is guarding nothing at all. Just remove it.

Reported-by: Byeong-ryeol Kim 
Signed-off-by: Linus Torvalds

Fix build failure due to hwirq.h needing smp_lock.h

2010-11-17T22:58:36+00:00

Arnd Bergmann did an automated scripting run to find left-over instances
of , and had made it trigger it on the normal BKL use
of lock_kernel and unlock_lernel (and apparently release_kernel_lock and
reacquire_kernel_lock too, used by the scheduler).

That resulted in commit 451a3c24b013 ("BKL: remove extraneous #include
").

However, hardirq.h was the only remaining user of the old
'kernel_locked()' interface, and Arnd's script hadn't checked for that.
So depending on your configuration and what header files had been
included, you would get errors like "implicit declaration of function
'kernel_locked'" during the build.

The right fix is not to just re-instate the smp_lock.h include - it is
to just remove 'kernel_locked()' entirely, since the only use was this
one special low-level detail.  Just make hardirq.h do it directly.

In fact this simplifies and clarifies the code, because some trivial
analysis makes it clear that hardirq.h only ever used _one_ of the two
definitions of kernel_locked(), so we can remove the other one entirely.

Reported-by: Zimny Lech 
Reported-and-acked-by: Randy Dunlap 
Acked-by: Arnd Bergmann 
Signed-off-by: Linus Torvalds

BKL: remove extraneous #include

2010-11-17T16:59:32+00:00

The big kernel lock has been removed from all these files at some point,
leaving only the #include.

Remove this too as a cleanup.

Signed-off-by: Arnd Bergmann 
Signed-off-by: Linus Torvalds

preempt: fix kernel build with !CONFIG_BKL

2010-11-02T12:39:13+00:00

The preempt count logic tries to take the BKL into account, which breaks
when CONFIG_BKL is not set.

Use the same preempt_count offset that we use without CONFIG_PREEMPT
when CONFIG_BKL is disabled.

Signed-off-by: Arnd Bergmann 
Reported-and-tested-by: Kirill A. Shutemov 
Signed-off-by: Linus Torvalds

Merge git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-2.6-irqflags

2010-10-21T21:37:27+00:00

* git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-2.6-irqflags:
  Fix IRQ flag handling naming
  MIPS: Add missing #inclusions of 
  smc91x: Add missing #inclusion of 
  Drop a couple of unnecessary asm/system.h inclusions
  SH: Add missing consts to sys_execve() declaration
  Blackfin: Rename IRQ flags handling functions
  Blackfin: Add missing dep to asm/irqflags.h
  Blackfin: Rename DES PC2() symbol to avoid collision
  Blackfin: Split the BF532 BFIN_*_FIO_FLAG() functions to their own header
  Blackfin: Split PLL code from mach-specific cdef headers

Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

2010-10-21T19:55:43+00:00

* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (29 commits)
  sched: Export account_system_vtime()
  sched: Call tick_check_idle before __irq_enter
  sched: Remove irq time from available CPU power
  sched: Do not account irq time to current task
  x86: Add IRQ_TIME_ACCOUNTING
  sched: Add IRQ_TIME_ACCOUNTING, finer accounting of irq time
  sched: Add a PF flag for ksoftirqd identification
  sched: Consolidate account_system_vtime extern declaration
  sched: Fix softirq time accounting
  sched: Drop group_capacity to 1 only if local group has extra capacity
  sched: Force balancing on newidle balance if local group has capacity
  sched: Set group_imb only a task can be pulled from the busiest cpu
  sched: Do not consider SCHED_IDLE tasks to be cache hot
  sched: Drop all load weight manipulation for RT tasks
  sched: Create special class for stop/migrate work
  sched: Unindent labels
  sched: Comment updates: fix default latency and granularity numbers
  tracing/sched: Add sched_pi_setprio tracepoint
  sched: Give CPU bound RT tasks preference
  sched: Try not to migrate higher priority RT tasks
  ...