linux-toradex.git/kernel/rcutorture.c, branch v2.6.30-rc7

cpumask: convert rcutorture.c

2009-03-30T11:35:16+00:00

We're getting rid of cpumasks on the stack.

Simply change tmp_mask to a global, and allocate it in
rcu_torture_init().

Signed-off-by: Rusty Russell 
Acked-by: "Paul E. McKenney" 
Cc: Josh Triplett

rcu: fix bug in rcutorture system-shutdown code

2009-01-07T22:36:25+00:00

This patch fixes an rcutorture bug found by Eric Sesterhenn that
resulted in oopses in response to "rmmod rcutorture".  The problem
was in some new code that attempted to handle the case where a system
is shut down while rcutorture is still running, for example, when
rcutorture is built into the kernel so that it cannot be removed.
The fix causes the rcutorture threads to "park" in an
schedule_timeout_uninterruptible(MAX_SCHEDULE_TIMEOUT) rather than
trying to get them to terminate cleanly.  Concurrent shutdown and
rmmod is illegal.

I believe that this is 2.6.29 material, as it is used in some testing
setups.

For reference, here are the rcutorture operating modes:

CONFIG_RCU_TORTURE_TEST=m

	This is the normal rcutorture build.  Use "modprobe rcutorture"
	(with optional arguments) to start, and "rmmod rcutorture" to
	stop.  If you shut the system down without doing the rmmod, you
	should see console output like:

	rcutorture thread rcu_torture_writer parking due to system shutdown

	One for each rcutorture kthread.

CONFIG_RCU_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST_RUNNABLE=n

	Use this if you want rcutorture built in, but don't want the
	test to start running during early boot.  To start the
	torturing:

		echo 1 > /proc/sys/kernel/rcutorture_runnable

	To stop the torturing, s/1/0/

	You will get "parking" console messages as noted above when
	you shut the system down.

CONFIG_RCU_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST_RUNNABLE=y

	Same as above, except that the torturing starts during early
	boot.  Only for the stout of heart and strong of stomach.
	The same /proc entry noted above may be used to control the
	test.

Located-by: Eric Sesterhenn 
Tested-by: Eric Sesterhenn 
Signed-off-by: Paul E. McKenney 
Signed-off-by: Ingo Molnar

rcu: fix rcutorture bug

2009-01-05T12:09:49+00:00

Fix an rcutorture bug that prevents the shutdown notifier from ever
actually having any effect, due to the fact that kthreads ignore all
signals.

Signed-off-by: Paul E. McKenney 
Signed-off-by: Ingo Molnar

cpumask: convert RCU implementations

2008-12-31T23:42:26+00:00

Impact: use new cpumask API.

rcu_ctrlblk contains a cpumask, and it's highly optimized so I don't want
a cpumask_var_t (ie. a pointer) for the CONFIG_CPUMASK_OFFSTACK case.  It
could use a dangling bitmap, and be allocated in __rcu_init to save memory,
but for the moment we use a bitmap.

(Eventually 'struct cpumask' will be undefined for CONFIG_CPUMASK_OFFSTACK,
so we use a bitmap here to show we really mean it).

We remove on-stack cpumasks, using cpumask_var_t for
rcu_torture_shuffle_tasks() and for_each_cpu_and in force_quiescent_state().

Signed-off-by: Rusty Russell

rcu: fix rcutorture behavior during reboot

2008-12-16T23:04:40+00:00

Impact: fix very rare reboot hang

Because rcutorture ignored all signals, it does not terminate in
response to the signals sent at shutdown time.  This can cause strange
failures due to its continuing to make use of kernel function too late
in the shutdown sequence.  This patch therefore adds a shutdown notifier
to rcutorture, causing it to shut down in response to a reboot or an
orderly shutdown.

Signed-off-by: Paul E. McKenney 
Signed-off-by: Ingo Molnar

byteorder: remove direct includes of linux/byteorder/swab[b].h

2008-10-20T19:51:53+00:00

A consolidated implementation will provide this generically through
asm/byteorder, remove direct includes to avoid breakage when the
changeover to the new implementation occurs.

This hunk was lost from commit 1d8cca44b6a244b7e378546d719041819049a0f9
("byteorder: provide swabb.h generically in asm/byteorder.h")

Signed-off-by: Harvey Harrison 
Signed-off-by: Linus Torvalds

rcu: make rcutorture even more vicious: invoke RCU readers from irq handlers (timers)

2008-06-26T07:24:33+00:00

This patch allows torturing RCU from irq handlers (timers, in this case).
A new module parameter irqreader enables such additional torturing,
and is enabled by default.  Variants of RCU that do not tolerate readers
being called from irq handlers (e.g., SRCU) ignore irqreader.

Signed-off-by: Paul E. McKenney 
Cc: josh@freedesktop.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: dino@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: torvalds@linux-foundation.org
Cc: vegard.nossum@gmail.com
Cc: adobriyan@gmail.com
Cc: oleg@tv-sign.ru
Cc: bunk@kernel.org
Cc: rjw@sisk.pl
Signed-off-by: Ingo Molnar

rcu: make quiescent rcutorture less power-hungry

2008-06-24T11:36:39+00:00

This patch aligns the rcutorture wakeup times to align with all other
multiple-of-a-second wakeups to further decrease power consumption.

Suggested-by: Arjan van de Ven 
Signed-off-by: Paul E. McKenney 
Cc: josh@freedesktop.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: dino@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: torvalds@linux-foundation.org
Cc: vegard.nossum@gmail.com
Cc: adobriyan@gmail.com
Cc: oleg@tv-sign.ru
Cc: bunk@kernel.org
Cc: rjw@sisk.pl
Signed-off-by: Ingo Molnar

rcu, rcutorture: make quiescent rcutorture less power-hungry

2008-06-24T11:36:38+00:00

Signed-off-by: Paul E. McKenney 
Cc: josh@freedesktop.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: dino@in.ibm.com
Cc: akpm@linux-foundation.org
Cc: torvalds@linux-foundation.org
Cc: vegard.nossum@gmail.com
Cc: adobriyan@gmail.com
Cc: oleg@tv-sign.ru
Cc: bunk@kernel.org
Cc: rjw@sisk.pl
Signed-off-by: Ingo Molnar

rcu: make rcutorture more vicious: reinstate boot-time testing

2008-06-19T09:22:15+00:00

This patch re-institutes the ability to build rcutorture directly into
the Linux kernel.  The reason that this capability was removed was that
this could result in your kernel being pretty much useless, as rcutorture
would be running starting from early boot.  This problem has been avoided
by (1) making rcutorture run only three seconds of every six by default,
(2) adding a CONFIG_RCU_TORTURE_TEST_RUNNABLE that permits rcutorture
to be quiesced at boot time, and (3) adding a sysctl in /proc named
/proc/sys/kernel/rcutorture_runnable that permits rcutorture to be
quiesced and unquiesced when built into the kernel.

Please note that this /proc file is -not- available when rcutorture
is built as a module.  Please also note that to get the earlier
take-no-prisoners behavior, you must use the boot command line to set
rcutorture's "stutter" parameter to zero.

The rcutorture quiescing mechanism is currently quite crude: loops
in each rcutorture process that poll a global variable once per tick.
Suggestions for improvement are welcome.  The default action will
be to reduce the polling rate to a few times per second.

Signed-off-by: Paul E. McKenney 
Suggested-by: Ingo Molnar 
Signed-off-by: Ingo Molnar