linux-stable.git/kernel/rcu, branch v4.13.4

srcu: Provide ordering for CPU not involved in grace period

2017-09-27T12:43:11+00:00

commit 35732cf9dd38b1efb0f2f22c91c61b51337d1ac3 upstream.

Tree RCU guarantees that every online CPU has a memory barrier between
any given grace period and any of that CPU's RCU read-side sections that
must be ordered against that grace period.  Since RCU doesn't always
know where read-side critical sections are, the actual implementation
guarantees order against prior and subsequent non-idle non-offline code,
whether in an RCU read-side critical section or not.  As a result, there
does not need to be a memory barrier at the end of synchronize_rcu()
and friends because the ordering internal to the grace period has
ordered every CPU's post-grace-period execution against each CPU's
pre-grace-period execution, again for all non-idle online CPUs.

In contrast, SRCU can have non-idle online CPUs that are completely
uninvolved in a given SRCU grace period, for example, a CPU that
never runs any SRCU read-side critical sections and took no part in
the grace-period processing.  It is in theory possible for a given
synchronize_srcu()'s wakeup to be delivered to a CPU that was completely
uninvolved in the prior SRCU grace period, which could mean that the
code following that synchronize_srcu() would end up being unordered with
respect to both the grace period and any pre-existing SRCU read-side
critical sections.

This commit therefore adds an smp_mb() to the end of __synchronize_srcu(),
which prevents this scenario from occurring.

Reported-by: Lance Roy 
Signed-off-by: Paul E. McKenney 
Acked-by: Lance Roy 
Signed-off-by: Greg Kroah-Hartman

rcu: Remove RCU CPU stall warnings from Tiny RCU

2017-06-09T01:52:45+00:00

Tiny RCU's job is to be tiny, so this commit removes its RCU CPU
stall warning code.  After this, there is no longer any need for
rcu_sched_ctrlblk and rcu_bh_ctrlblk to be in tiny_plugin.h, so this
commit also moves them to tiny.c.

Signed-off-by: Paul E. McKenney

rcu: Remove event tracing from Tiny RCU

2017-06-09T01:52:45+00:00

This commit saves a few lines by getting rid of Tiny RCU's event tracing.

Signed-off-by: Paul E. McKenney

rcu: Move RCU debug Kconfig options to kernel/rcu

2017-06-09T01:52:44+00:00

RCU's debugging Kconfig options are in the unintuitive location
lib/Kconfig.debug, and there are enough of them that it would be good for
them to be more centralized.  This commit therefore extracts RCU's Kconfig
options from init/Kconfig into a new kernel/rcu/Kconfig.debug file.

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

rcu: Move RCU non-debug Kconfig options to kernel/rcu

2017-06-09T01:52:44+00:00

RCU's Kconfig options are scattered, and there are enough of them
that it would be good for them to be more centralized.  This commit
therefore extracts RCU's Kconfig options from init/Kconfig into a new
kernel/rcu/Kconfig file.

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

rcu: Eliminate NOCBs CPU-state Kconfig options

2017-06-09T01:52:43+00:00

The CONFIG_RCU_NOCB_CPU_ALL, CONFIG_RCU_NOCB_CPU_NONE, and
CONFIG_RCU_NOCB_CPU_ZERO Kconfig options are used only in testing and
are redundant with the rcu_nocbs= boot parameter.  This commit therefore
removes these three Kconfig options and adjusts the rcutorture scripts
to use the boot parameter instead.

Signed-off-by: Paul E. McKenney

rcu: Remove debugfs tracing

2017-06-09T01:52:43+00:00

RCU's debugfs tracing used to be the only reasonable low-level debug
information available, but ftrace and event tracing has since surpassed
the RCU debugfs level of usefulness.  This commit therefore removes
RCU's debugfs tracing.

Signed-off-by: Paul E. McKenney

srcu: Remove Classic SRCU

2017-06-09T01:52:42+00:00

Classic SRCU was only ever intended to be a fallback in case of issues
with Tree/Tiny SRCU, and the latter two are doing quite well in testing.
This commit therefore removes Classic SRCU.

Signed-off-by: Paul E. McKenney

srcu: Fix rcutorture-statistics typo

2017-06-09T01:52:42+00:00

The function srcutorture_get_gp_data() duplicated the check for
sp->batch_check0.head instead of also checking sp->batch_check1.head.
The only effect of this typo would be for rcutorture statistics to
understate the fraction of time that an SRCU grace period was in flight,
and only for Classic SRCU.  This commit fixes this typo.

Reported-by: David Binderman 
Signed-off-by: Paul E. McKenney

rcu: Remove the now-obsolete PROVE_RCU_REPEATEDLY Kconfig option

2017-06-09T01:52:41+00:00

The PROVE_RCU_REPEATEDLY Kconfig option was initially added due to
the volume of messages from PROVE_RCU: Doing just one per boot would
have required excessive numbers of boots to locate them all.  However,
PROVE_RCU messages are now relatively rare, so there is no longer any
reason to need more than one such message per boot.  This commit therefore
removes the PROVE_RCU_REPEATEDLY Kconfig option.

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