Commit dbd1e66e authored by Robert Richter's avatar Robert Richter
Browse files

Merge commit 'linux-2.6/master' (early part) into oprofile/core 



This branch depends on these apic patches:

      apic, x86: Use BIOS settings for IBS and MCE threshold interrupt LVT offsets
      apic, x86: Check if EILVT APIC registers are available (AMD only)

Signed-off-by: default avatarRobert Richter <robert.richter@amd.com>
parents 328b8f1b 4a60cfa9

Documentation/DocBook/genericirq.tmpl

+52 −32
Original line number Diff line number Diff line
@@ -28,7 +28,7 @@ 
  </authorgroup>



  <copyright>

   <year>2005-2006</year>

   <year>2005-2010</year>

   <holder>Thomas Gleixner</holder>

  </copyright>

  <copyright>
@@ -100,6 +100,10 @@ 
	  <listitem><para>Edge type</para></listitem>

	  <listitem><para>Simple type</para></listitem>

	</itemizedlist>

	During the implementation we identified another type:

	<itemizedlist>

	  <listitem><para>Fast EOI type</para></listitem>

	</itemizedlist>

	In the SMP world of the __do_IRQ() super-handler another type

	was identified:

	<itemizedlist>
@@ -153,6 +157,7 @@ 
	is still available. This leads to a kind of duality for the time

	being. Over time the new model should be used in more and more

	architectures, as it enables smaller and cleaner IRQ subsystems.

	It's deprecated for three years now and about to be removed.

	</para>

  </chapter>

  <chapter id="bugs">
@@ -217,6 +222,7 @@ 
	  <itemizedlist>

	  <listitem><para>handle_level_irq</para></listitem>

	  <listitem><para>handle_edge_irq</para></listitem>

	  <listitem><para>handle_fasteoi_irq</para></listitem>

	  <listitem><para>handle_simple_irq</para></listitem>

	  <listitem><para>handle_percpu_irq</para></listitem>

	  </itemizedlist>
@@ -233,33 +239,33 @@ 
		are used by the default flow implementations.

		The following helper functions are implemented (simplified excerpt):

		<programlisting>

default_enable(irq)

default_enable(struct irq_data *data)

{

	desc->chip->unmask(irq);

	desc->chip->irq_unmask(data);

}



default_disable(irq)

default_disable(struct irq_data *data)

{

	if (!delay_disable(irq))

		desc->chip->mask(irq);

	if (!delay_disable(data))

		desc->chip->irq_mask(data);

}



default_ack(irq)

default_ack(struct irq_data *data)

{

	chip->ack(irq);

	chip->irq_ack(data);

}



default_mask_ack(irq)

default_mask_ack(struct irq_data *data)

{

	if (chip->mask_ack) {

		chip->mask_ack(irq);

	if (chip->irq_mask_ack) {

		chip->irq_mask_ack(data);

	} else {

		chip->mask(irq);

		chip->ack(irq);

		chip->irq_mask(data);

		chip->irq_ack(data);

	}

}



noop(irq)

noop(struct irq_data *data))

{

}
@@ -278,9 +284,24 @@ noop(irq) 
		<para>

		The following control flow is implemented (simplified excerpt):

		<programlisting>

desc->chip->start();

desc->chip->irq_mask();

handle_IRQ_event(desc->action);

desc->chip->irq_unmask();

		</programlisting>

		</para>

	    </sect3>

	    <sect3 id="Default_FASTEOI_IRQ_flow_handler">

		<title>Default Fast EOI IRQ flow handler</title>

		<para>

		handle_fasteoi_irq provides a generic implementation

		for interrupts, which only need an EOI at the end of

		the handler

		</para>

		<para>

		The following control flow is implemented (simplified excerpt):

		<programlisting>

handle_IRQ_event(desc->action);

desc->chip->end();

desc->chip->irq_eoi();

		</programlisting>

		</para>

	    </sect3>
@@ -294,20 +315,19 @@ desc->chip->end(); 
		The following control flow is implemented (simplified excerpt):

		<programlisting>

if (desc->status &amp; running) {

	desc->chip->hold();

	desc->chip->irq_mask();

	desc->status |= pending | masked;

	return;

}

desc->chip->start();

desc->chip->irq_ack();

desc->status |= running;

do {

	if (desc->status &amp; masked)

		desc->chip->enable();

		desc->chip->irq_unmask();

	desc->status &amp;= ~pending;

	handle_IRQ_event(desc->action);

} while (status &amp; pending);

desc->status &amp;= ~running;

desc->chip->end();

		</programlisting>

		</para>

   	    </sect3>
@@ -342,9 +362,9 @@ handle_IRQ_event(desc->action); 
		<para>

		The following control flow is implemented (simplified excerpt):

		<programlisting>

desc->chip->start();

handle_IRQ_event(desc->action);

desc->chip->end();

if (desc->chip->irq_eoi)

        desc->chip->irq_eoi();

		</programlisting>

		</para>

   	    </sect3>
@@ -375,8 +395,7 @@ desc->chip->end(); 
	mechanism. (It's necessary to enable CONFIG_HARDIRQS_SW_RESEND when

	you want to use the delayed interrupt disable feature and your

	hardware is not capable of retriggering	an interrupt.)

	The delayed interrupt disable can be runtime enabled, per interrupt,

	by setting the IRQ_DELAYED_DISABLE flag in the irq_desc status field.

	The delayed interrupt disable is not configurable.

	</para>

	</sect2>

    </sect1>
@@ -387,13 +406,13 @@ desc->chip->end(); 
	contains all the direct chip relevant functions, which

	can be utilized by the irq flow implementations.

	  <itemizedlist>

	  <listitem><para>ack()</para></listitem>

	  <listitem><para>mask_ack() - Optional, recommended for performance</para></listitem>

	  <listitem><para>mask()</para></listitem>

	  <listitem><para>unmask()</para></listitem>

	  <listitem><para>retrigger() - Optional</para></listitem>

	  <listitem><para>set_type() - Optional</para></listitem>

	  <listitem><para>set_wake() - Optional</para></listitem>

	  <listitem><para>irq_ack()</para></listitem>

	  <listitem><para>irq_mask_ack() - Optional, recommended for performance</para></listitem>

	  <listitem><para>irq_mask()</para></listitem>

	  <listitem><para>irq_unmask()</para></listitem>

	  <listitem><para>irq_retrigger() - Optional</para></listitem>

	  <listitem><para>irq_set_type() - Optional</para></listitem>

	  <listitem><para>irq_set_wake() - Optional</para></listitem>

	  </itemizedlist>

	These primitives are strictly intended to mean what they say: ack means

	ACK, masking means masking of an IRQ line, etc. It is up to the flow
@@ -458,6 +477,7 @@ desc->chip->end(); 
     <para>

     This chapter contains the autogenerated documentation of the internal functions.

     </para>

!Ikernel/irq/irqdesc.c

!Ikernel/irq/handle.c

!Ikernel/irq/chip.c

  </chapter>

Documentation/DocBook/kernel-locking.tmpl

+4 −10
Original line number Diff line number Diff line
@@ -1645,7 +1645,9 @@ the amount of locking which needs to be done. 
      all the readers who were traversing the list when we deleted the

      element are finished.  We use <function>call_rcu()</function> to

      register a callback which will actually destroy the object once

      the readers are finished.

      all pre-existing readers are finished.  Alternatively,

      <function>synchronize_rcu()</function> may be used to block until

      all pre-existing are finished.

    </para>

    <para>

      But how does Read Copy Update know when the readers are
@@ -1714,7 +1716,7 @@ the amount of locking which needs to be done. 
-        object_put(obj);

+        list_del_rcu(&amp;obj-&gt;list);

         cache_num--;

+        call_rcu(&amp;obj-&gt;rcu, cache_delete_rcu, obj);

+        call_rcu(&amp;obj-&gt;rcu, cache_delete_rcu);

 }



 /* Must be holding cache_lock */
@@ -1725,14 +1727,6 @@ the amount of locking which needs to be done. 
         if (++cache_num > MAX_CACHE_SIZE) {

                 struct object *i, *outcast = NULL;

                 list_for_each_entry(i, &amp;cache, list) {

@@ -85,6 +94,7 @@

         obj-&gt;popularity = 0;

         atomic_set(&amp;obj-&gt;refcnt, 1); /* The cache holds a reference */

         spin_lock_init(&amp;obj-&gt;lock);

+        INIT_RCU_HEAD(&amp;obj-&gt;rcu);



         spin_lock_irqsave(&amp;cache_lock, flags);

         __cache_add(obj);

@@ -104,12 +114,11 @@

 struct object *cache_find(int id)

 {

Documentation/RCU/checklist.txt

+39 −7
Original line number Diff line number Diff line
@@ -218,13 +218,22 @@ over a rather long period of time, but improvements are always welcome! 
	include:



	a.	Keeping a count of the number of data-structure elements

		used by the RCU-protected data structure, including those

		waiting for a grace period to elapse.  Enforce a limit

		on this number, stalling updates as needed to allow

		previously deferred frees to complete.



		Alternatively, limit only the number awaiting deferred

		free rather than the total number of elements.

		used by the RCU-protected data structure, including

		those waiting for a grace period to elapse.  Enforce a

		limit on this number, stalling updates as needed to allow

		previously deferred frees to complete.	Alternatively,

		limit only the number awaiting deferred free rather than

		the total number of elements.



		One way to stall the updates is to acquire the update-side

		mutex.	(Don't try this with a spinlock -- other CPUs

		spinning on the lock could prevent the grace period

		from ever ending.)  Another way to stall the updates

		is for the updates to use a wrapper function around

		the memory allocator, so that this wrapper function

		simulates OOM when there is too much memory awaiting an

		RCU grace period.  There are of course many other

		variations on this theme.



	b.	Limiting update rate.  For example, if updates occur only

		once per hour, then no explicit rate limiting is required,
@@ -365,3 +374,26 @@ over a rather long period of time, but improvements are always welcome! 
	and the compiler to freely reorder code into and out of RCU

	read-side critical sections.  It is the responsibility of the

	RCU update-side primitives to deal with this.



17.	Use CONFIG_PROVE_RCU, CONFIG_DEBUG_OBJECTS_RCU_HEAD, and

	the __rcu sparse checks to validate your RCU code.  These

	can help find problems as follows:



	CONFIG_PROVE_RCU: check that accesses to RCU-protected data

		structures are carried out under the proper RCU

		read-side critical section, while holding the right

		combination of locks, or whatever other conditions

		are appropriate.



	CONFIG_DEBUG_OBJECTS_RCU_HEAD: check that you don't pass the

		same object to call_rcu() (or friends) before an RCU

		grace period has elapsed since the last time that you

		passed that same object to call_rcu() (or friends).



	__rcu sparse checks: tag the pointer to the RCU-protected data

		structure with __rcu, and sparse will warn you if you

		access that pointer without the services of one of the

		variants of rcu_dereference().



	These debugging aids can help you find problems that are

	otherwise extremely difficult to spot.

Documentation/RCU/stallwarn.txt

+18 −0
Original line number Diff line number Diff line
@@ -80,6 +80,24 @@ o A CPU looping with bottom halves disabled. This condition can 
o	For !CONFIG_PREEMPT kernels, a CPU looping anywhere in the kernel

	without invoking schedule().



o	A CPU-bound real-time task in a CONFIG_PREEMPT kernel, which might

	happen to preempt a low-priority task in the middle of an RCU

	read-side critical section.   This is especially damaging if

	that low-priority task is not permitted to run on any other CPU,

	in which case the next RCU grace period can never complete, which

	will eventually cause the system to run out of memory and hang.

	While the system is in the process of running itself out of

	memory, you might see stall-warning messages.



o	A CPU-bound real-time task in a CONFIG_PREEMPT_RT kernel that

	is running at a higher priority than the RCU softirq threads.

	This will prevent RCU callbacks from ever being invoked,

	and in a CONFIG_TREE_PREEMPT_RCU kernel will further prevent

	RCU grace periods from ever completing.  Either way, the

	system will eventually run out of memory and hang.  In the

	CONFIG_TREE_PREEMPT_RCU case, you might see stall-warning

	messages.



o	A bug in the RCU implementation.



o	A hardware failure.  This is quite unlikely, but has occurred

Documentation/RCU/trace.txt

+12 −1
Original line number Diff line number Diff line
@@ -125,6 +125,17 @@ o "b" is the batch limit for this CPU. If more than this number 
	of RCU callbacks is ready to invoke, then the remainder will

	be deferred.



o	"ci" is the number of RCU callbacks that have been invoked for

	this CPU.  Note that ci+ql is the number of callbacks that have

	been registered in absence of CPU-hotplug activity.



o	"co" is the number of RCU callbacks that have been orphaned due to

	this CPU going offline.



o	"ca" is the number of RCU callbacks that have been adopted due to

	other CPUs going offline.  Note that ci+co-ca+ql is the number of

	RCU callbacks registered on this CPU.



There is also an rcu/rcudata.csv file with the same information in

comma-separated-variable spreadsheet format.
@@ -180,7 +191,7 @@ o "s" is the "signaled" state that drives force_quiescent_state()'s 


o	"jfq" is the number of jiffies remaining for this grace period

	before force_quiescent_state() is invoked to help push things

	along.  Note that CPUs in dyntick-idle mode thoughout the grace

	along.  Note that CPUs in dyntick-idle mode throughout the grace

	period will not report on their own, but rather must be check by

	some other CPU via force_quiescent_state().