This commit implements an SRCU state machine in support of call_srcu().
The state machine is preemptible, light-weight, and single-threaded,
minimizing synchronization overhead. In particular, there is no longer
any need for synchronize_srcu() to be guarded by a mutex.
Expedited processing is handled, at least in the absence of concurrent
grace-period operations on that same srcu_struct structure, by having
the synchronize_srcu_expedited() thread take on the role of the
workqueue thread for one iteration.
There is a reasonable probability that a given SRCU callback will
be invoked on the same CPU that registered it, however, there is no
guarantee. Concurrent SRCU grace-period primitives can cause callbacks
to be executed elsewhere, even in absence of CPU-hotplug operations.
Callbacks execute in process context, but under the influence of
local_bh_disable(), so it is illegal to sleep in an SRCU callback
function.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The earlier algorithm used an "expedited" flag combined with a "trycount"
counter to differentiate between normal and expedited SRCU grace periods.
However, the difference can be encoded into a single counter with a cutoff
value and different initial values for expedited and normal SRCU grace
periods. This commit makes that change.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Conflicts:
kernel/srcu.c
Expand the calls to srcu_readers_active_idx() from srcu_readers_active()
inline. This change improves cache locality by interating over the CPUs
once rather than twice.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit implements a variant of Peter's algorithm, which may be found
at https://lkml.org/lkml/2012/2/1/119.
o Make the checking lock-free to enable parallel checking.
Parallel checking is required when (1) the original checking
task is preempted for a long time, (2) sychronize_srcu_expedited()
starts during an ongoing SRCU grace period, or (3) we wish to
avoid acquiring a lock.
o Since the checking is lock-free, we avoid a mutex in state machine
for call_srcu().
o Remove the SRCU_REF_MASK and remove the coupling with the flipping.
This might allow us to remove the preempt_disable() in future
versions, though such removal will need great care because it
rescinds the one-old-reader-per-CPU guarantee.
o Remove a smp_mb(), simplify the comments and make the smp_mb() pairs
more intuitive.
Inspired-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The safety of SRCU is provided byy wait_idx() rather than flipping.
The flipping actually prevents starvation.
This commit therefore updates the comments to more accurately and
precisely describe what is going on.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This is an optimization of the SRCU grace period. To guard against
preempted readers with old values of the counter, it suffices to scan the
old counters once more, then flip ->completed only one time. The reason
this works is that the old readers must have incremented the old set of
counters (if they have not yet incremented, then their critical section
starts after this grace period, so they may be safely ignored).
This commit therefore optimizes the second flip out in favor of a simple
rescan.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The purpose of the upper bit of SRCU's per-CPU counters is to guarantee
that no reasonable series of srcu_read_lock() and srcu_read_unlock()
operations can return the value of the counter to its original value.
This guarantee is require only after the index has been switched to
the other set of counters, so at most one srcu_read_lock() can affect
a given CPU's counter. The number of srcu_read_unlock() operations
on a given counter is limited to the number of tasks in the system,
which given the Linux kernel's current structure is limited to far less
than 2^30 on 32-bit systems and far less than 2^62 on 64-bit systems.
(Something about a limited number of bytes in the kernel's address space.)
Therefore, if srcu_read_lock() increments the upper bits, then
srcu_read_unlock() need not do so. In this case, an srcu_read_lock() and
an srcu_read_unlock() will flip the lower bit of the upper field of the
counter. An unreasonably large additional number of srcu_read_unlock()
operations would be required to return the counter to its initial value,
thus preserving the guarantee.
This commit takes this approach, which further allows it to shrink
the size of the upper field to one bit, making the number of
srcu_read_unlock() operations required to return the counter to its
initial value even more unreasonable than before.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The fastpath in __synchronize_srcu() is designed to handle cases where
there are a large number of concurrent calls for the same srcu_struct
structure. However, the Linux kernel currently does not use SRCU in
this manner, so remove the fastpath checks for simplicity.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The current implementation of synchronize_srcu_expedited() can cause
severe OS jitter due to its use of synchronize_sched(), which in turn
invokes try_stop_cpus(), which causes each CPU to be sent an IPI.
This can result in severe performance degradation for real-time workloads
and especially for short-interation-length HPC workloads. Furthermore,
because only one instance of try_stop_cpus() can be making forward progress
at a given time, only one instance of synchronize_srcu_expedited() can
make forward progress at a time, even if they are all operating on
distinct srcu_struct structures.
This commit, inspired by an earlier implementation by Peter Zijlstra
(https://lkml.org/lkml/2012/1/31/211) and by further offline discussions,
takes a strictly algorithmic bits-in-memory approach. This has the
disadvantage of requiring one explicit memory-barrier instruction in
each of srcu_read_lock() and srcu_read_unlock(), but on the other hand
completely dispenses with OS jitter and furthermore allows SRCU to be
used freely by CPUs that RCU believes to be idle or offline.
The update-side implementation handles the single read-side memory
barrier by rechecking the per-CPU counters after summing them and
by running through the update-side state machine twice.
This implementation has passed moderate rcutorture testing on both
x86 and Power. Also updated to use this_cpu_ptr() instead of per_cpu_ptr(),
as suggested by Peter Zijlstra.
Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Reviewed-by: Lai Jiangshan <laijs@cn.fujitsu.com>
The expedited RCU primitives can be quite useful, but they have some
high costs as well. This commit updates and creates docbook comments
calling out the costs, and updates the RCU documentation as well.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
It is illegal to have a grace period within a same-flavor RCU read-side
critical section, so this commit adds lockdep-RCU checks to splat when
such abuse is encountered. This commit does not detect more elaborate
RCU deadlock situations. These situations might be a job for lockdep
enhancements.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The changed files were only including linux/module.h for the
EXPORT_SYMBOL infrastructure, and nothing else. Revector them
onto the isolated export header for faster compile times.
Nothing to see here but a whole lot of instances of:
-#include <linux/module.h>
+#include <linux/export.h>
This commit is only changing the kernel dir; next targets
will probably be mm, fs, the arch dirs, etc.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Because the adaptive synchronize_srcu_expedited() approach has
worked very well in testing, remove the kernel parameter and
replace it by a C-preprocessor macro. If someone finds problems
with this approach, a more complex and aggressively adaptive
approach might be required.
Longer term, SRCU will be merged with the other RCU implementations,
at which point synchronize_srcu_expedited() will be event driven,
just as synchronize_sched_expedited() currently is. At that point,
there will be no need for this adaptive approach.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The synchronize_srcu_expedited() function is currently quick if there
are no active readers, but will delay a full jiffy if there are any.
If these readers leave their SRCU read-side critical sections quickly,
this is way too long to wait. So this commit first waits ten microseconds,
and only then falls back to jiffy-at-a-time waiting.
Reported-by: Avi Kivity <avi@redhat.com>
Reported-by: Marcelo Tosatti <mtosatti@redhat.com>
Tested-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
The CONFIG_DEBUG_LOCK_ALLOC ifdef isn't necessary at this point, because it is
checked in an outer ifdef level already and has no effect here.
Signed-off-by: Christian Dietrich <qy03fugy@stud.informatik.uni-erlangen.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Inspection is proving insufficient to catch all RCU misuses,
which is understandable given that rcu_dereference() might be
protected by any of four different flavors of RCU (RCU, RCU-bh,
RCU-sched, and SRCU), and might also/instead be protected by any
of a number of locking primitives. It is therefore time to
enlist the aid of lockdep.
This set of patches is inspired by earlier work by Peter
Zijlstra and Thomas Gleixner, and takes the following approach:
o Set up separate lockdep classes for RCU, RCU-bh, and RCU-sched.
o Set up separate lockdep classes for each instance of SRCU.
o Create primitives that check for being in an RCU read-side
critical section. These return exact answers if lockdep is
fully enabled, but if unsure, report being in an RCU read-side
critical section. (We want to avoid false positives!)
The primitives are:
For RCU: rcu_read_lock_held(void)
For RCU-bh: rcu_read_lock_bh_held(void)
For RCU-sched: rcu_read_lock_sched_held(void)
For SRCU: srcu_read_lock_held(struct srcu_struct *sp)
o Add rcu_dereference_check(), which takes a second argument
in which one places a boolean expression based on the above
primitives and/or lockdep_is_held().
o A new kernel configuration parameter, CONFIG_PROVE_RCU, enables
rcu_dereference_check(). This depends on CONFIG_PROVE_LOCKING,
and should be quite helpful during the transition period while
CONFIG_PROVE_RCU-unaware patches are in flight.
The existing rcu_dereference() primitive does no checking, but
upcoming patches will change that.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: mathieu.desnoyers@polymtl.ca
Cc: josh@joshtriplett.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
Cc: Valdis.Kletnieks@vt.edu
Cc: dhowells@redhat.com
LKML-Reference: <1266887105-1528-1-git-send-email-paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This patch creates a synchronize_srcu_expedited() that uses
synchronize_sched_expedited() where synchronize_srcu()
uses synchronize_sched(). The synchronize_srcu() and
synchronize_srcu_expedited() functions become one-liners that
pass synchronize_sched() or synchronize_sched_expedited(),
repectively, to a new __synchronize_srcu() function.
While in the file, move the EXPORT_SYMBOL_GPL()s to immediately
follow the corresponding functions.
Requested-by: Avi Kivity <avi@redhat.com>
Tested-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: dipankar@in.ibm.com
Cc: mathieu.desnoyers@polymtl.ca
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
Cc: Valdis.Kletnieks@vt.edu
Cc: dhowells@redhat.com
Cc: avi@redhat.com
LKML-Reference: <12565226354038-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Make the needlessly global srcu_readers_active() static.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently the init_srcu_struct() routine has no way to report out-of-memory
errors. This patch (as761) makes it return -ENOMEM when the per-cpu data
allocation fails.
The patch also makes srcu_init_notifier_head() report a BUG if a notifier
head can't be initialized. Perhaps it should return -ENOMEM instead, but
in the most likely cases where this might occur I don't think any recovery
is possible. Notifier chains generally are not created dynamically.
[akpm@osdl.org: avoid statement-with-side-effect in macro]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Paul E. McKenney <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Updated patch adding a variant of RCU that permits sleeping in read-side
critical sections. SRCU is as follows:
o Each use of SRCU creates its own srcu_struct, and each
srcu_struct has its own set of grace periods. This is
critical, as it prevents one subsystem with a blocking
reader from holding up SRCU grace periods for other
subsystems.
o The SRCU primitives (srcu_read_lock(), srcu_read_unlock(),
and synchronize_srcu()) all take a pointer to a srcu_struct.
o The SRCU primitives must be called from process context.
o srcu_read_lock() returns an int that must be passed to
the matching srcu_read_unlock(). Realtime RCU avoids the
need for this by storing the state in the task struct,
but SRCU needs to allow a given code path to pass through
multiple SRCU domains -- storing state in the task struct
would therefore require either arbitrary space in the
task struct or arbitrary limits on SRCU nesting. So I
kicked the state-storage problem up to the caller.
Of course, it is not permitted to call synchronize_srcu()
while in an SRCU read-side critical section.
o There is no call_srcu(). It would not be hard to implement
one, but it seems like too easy a way to OOM the system.
(Hey, we have enough trouble with call_rcu(), which does
-not- permit readers to sleep!!!) So, if you want it,
please tell me why...
[josht@us.ibm.com: sparse notation]
Signed-off-by: Paul E. McKenney <paulmck@us.ibm.com>
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
