As stated in the documentation and implied by its name, the ppo
(preserved program order) relation is intended to link po-earlier
to po-later instructions under certain conditions. However, a
corner case currently allows instructions to be linked by ppo that
are not executed by the same thread, i.e., instructions are being
linked that have no po relation.
This happens due to the mb/strong-fence/fence relations, which (as
one case) provide order when locks are passed between threads
followed by an smp_mb__after_unlock_lock() fence. This is
illustrated in the following litmus test (as can be seen when using
herd7 with `doshow ppo`):
P0(spinlock_t *x, spinlock_t *y)
{
spin_lock(x);
spin_unlock(x);
}
P1(spinlock_t *x, spinlock_t *y)
{
spin_lock(x);
smp_mb__after_unlock_lock();
*y = 1;
}
The ppo relation will link P0's spin_lock(x) and P1's *y=1, because
P0 passes a lock to P1 which then uses this fence.
The patch makes ppo a subrelation of po by letting fence contribute
to ppo only in case the fence links events of the same thread.
Signed-off-by: Jonas Oberhauser <jonas.oberhauser@huaweicloud.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
During a code-reading exercise of linux-kernel.cat CAT file, I generated
a graph to show the to-r relations. While likely not problematic for the
model, I found it confusing that a read-write address dependency would
show as a to-r edge on the graph.
This patch therefore restricts the to-r links derived from addr to only
read-read address dependencies, so that read-write address dependencies don't
show as to-r in the graphs. This should also prevent future users of to-r from
deriving incorrect relations. Note that a read-write address dep, obviously,
still ends up in the ppo relation via the to-w relation.
I verified that a read-read address dependency still shows up as a to-r
link in the graph, as it did before.
For reference, the problematic graph was generated with the following
command:
herd7 -conf linux-kernel.cfg \
-doshow dep -doshow to-r -doshow to-w ./foo.litmus -show all -o OUT/
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
This commit adds support for smp_mb__after_srcu_read_unlock(), which,
when combined with a prior srcu_read_unlock(), implies a full memory
barrier. No ordering is guaranteed to accesses between the two, and
placing accesses between is bad practice in any case.
Tests may be found at https://github.com/paulmckrcu/litmus in files
matching manual/kernel/C-srcu-mb-*.litmus.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
LKMM uses two relations for talking about UNLOCK+LOCK pairings:
1) po-unlock-lock-po, which handles UNLOCK+LOCK pairings
on the same CPU or immediate lock handovers on the same
lock variable
2) po;[UL];(co|po);[LKW];po, which handles UNLOCK+LOCK pairs
literally as described in rcupdate.h#L1002, i.e., even
after a sequence of handovers on the same lock variable.
The latter relation is used only once, to provide the guarantee
defined in rcupdate.h#L1002 by smp_mb__after_unlock_lock(), which
makes any UNLOCK+LOCK pair followed by the fence behave like a full
barrier.
This patch drops this use in favor of using po-unlock-lock-po
everywhere, which unifies the way the model talks about UNLOCK+LOCK
pairings. At first glance this seems to weaken the guarantee given
by LKMM: When considering a long sequence of lock handovers
such as below, where P0 hands the lock to P1, which hands it to P2,
which finally executes such an after_unlock_lock fence, the mb
relation currently links any stores in the critical section of P0
to instructions P2 executes after its fence, but not so after the
patch.
P0(int *x, int *y, spinlock_t *mylock)
{
spin_lock(mylock);
WRITE_ONCE(*x, 2);
spin_unlock(mylock);
WRITE_ONCE(*y, 1);
}
P1(int *y, int *z, spinlock_t *mylock)
{
int r0 = READ_ONCE(*y); // reads 1
spin_lock(mylock);
spin_unlock(mylock);
WRITE_ONCE(*z,1);
}
P2(int *z, int *d, spinlock_t *mylock)
{
int r1 = READ_ONCE(*z); // reads 1
spin_lock(mylock);
spin_unlock(mylock);
smp_mb__after_unlock_lock();
WRITE_ONCE(*d,1);
}
P3(int *x, int *d)
{
WRITE_ONCE(*d,2);
smp_mb();
WRITE_ONCE(*x,1);
}
exists (1:r0=1 /\ 2:r1=1 /\ x=2 /\ d=2)
Nevertheless, the ordering guarantee given in rcupdate.h is actually
not weakened. This is because the unlock operations along the
sequence of handovers are A-cumulative fences. They ensure that any
stores that propagate to the CPU performing the first unlock
operation in the sequence must also propagate to every CPU that
performs a subsequent lock operation in the sequence. Therefore any
such stores will also be ordered correctly by the fence even if only
the final handover is considered a full barrier.
Indeed this patch does not affect the behaviors allowed by LKMM at
all. The mb relation is used to define ordering through:
1) mb/.../ppo/hb, where the ordering is subsumed by hb+ where the
lock-release, rfe, and unlock-acquire orderings each provide hb
2) mb/strong-fence/cumul-fence/prop, where the rfe and A-cumulative
lock-release orderings simply add more fine-grained cumul-fence
edges to substitute a single strong-fence edge provided by a long
lock handover sequence
3) mb/strong-fence/pb and various similar uses in the definition of
data races, where as discussed above any long handover sequence
can be turned into a sequence of cumul-fence edges that provide
the same ordering.
Signed-off-by: Jonas Oberhauser <jonas.oberhauser@huaweicloud.com>
Reviewed-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Viktor (as relayed by Jonas) has pointed out a weakness in the Linux
Kernel Memory Model. Namely, the memory ordering properties of atomic
operations are not monotonic: An atomic op with full-barrier semantics
does not always provide ordering as strong as one with release-barrier
semantics.
The following litmus test illustrates the problem:
--------------------------------------------------
C atomics-not-monotonic
{}
P0(int *x, atomic_t *y)
{
WRITE_ONCE(*x, 1);
smp_wmb();
atomic_set(y, 1);
}
P1(atomic_t *y)
{
int r1;
r1 = atomic_inc_return(y);
}
P2(int *x, atomic_t *y)
{
int r2;
int r3;
r2 = atomic_read(y);
smp_rmb();
r3 = READ_ONCE(*x);
}
exists (2:r2=2 /\ 2:r3=0)
--------------------------------------------------
The litmus test is allowed as shown with atomic_inc_return(), which
has full-barrier semantics. But if the operation is changed to
atomic_inc_return_release(), which only has release-barrier semantics,
the litmus test is forbidden. Clearly this violates monotonicity.
The reason is because the LKMM treats full-barrier atomic ops as if
they were written:
mb();
load();
store();
mb();
(where the load() and store() are the two parts of an atomic RMW op),
whereas it treats release-barrier atomic ops as if they were written:
load();
release_barrier();
store();
The difference is that here the release barrier orders the load part
of the atomic op before the store part with A-cumulativity, whereas
the mb()'s above do not. This means that release-barrier atomics can
effectively extend the cumul-fence relation but full-barrier atomics
cannot.
To resolve this problem we introduce the rmw-sequence relation,
representing an arbitrarily long sequence of atomic RMW operations in
which each operation reads from the previous one, and explicitly allow
it to extend cumul-fence. This modification of the memory model is
sound; it holds for PPC because of B-cumulativity, it holds for TSO
and ARM64 because of other-multicopy atomicity, and we can assume that
atomic ops on all other architectures will be implemented so as to
make it hold for them.
For similar reasons we also allow rmw-sequence to extend the
w-post-bounded relation, which is analogous to cumul-fence in some
ways.
Reported-by: Viktor Vafeiadis <viktor@mpi-sws.org>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reviewed-by: Jonas Oberhauser <jonas.oberhauser@huawei.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
A recent discussion[1] shows that we are in favor of strengthening the
ordering of unlock + lock on the same CPU: a unlock and a po-after lock
should provide the so-called RCtso ordering, that is a memory access S
po-before the unlock should be ordered against a memory access R
po-after the lock, unless S is a store and R is a load.
The strengthening meets programmers' expection that "sequence of two
locked regions to be ordered wrt each other" (from Linus), and can
reduce the mental burden when using locks. Therefore add it in LKMM.
[1]: https://lore.kernel.org/lkml/20210909185937.GA12379@rowland.harvard.edu/
Co-developed-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com> (RISC-V)
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Currently the Linux Kernel Memory Model gives an incorrect response
for the following litmus test:
C plain-WWC
{}
P0(int *x)
{
WRITE_ONCE(*x, 2);
}
P1(int *x, int *y)
{
int r1;
int r2;
int r3;
r1 = READ_ONCE(*x);
if (r1 == 2) {
smp_rmb();
r2 = *x;
}
smp_rmb();
r3 = READ_ONCE(*x);
WRITE_ONCE(*y, r3 - 1);
}
P2(int *x, int *y)
{
int r4;
r4 = READ_ONCE(*y);
if (r4 > 0)
WRITE_ONCE(*x, 1);
}
exists (x=2 /\ 1:r2=2 /\ 2:r4=1)
The memory model says that the plain read of *x in P1 races with the
WRITE_ONCE(*x) in P2.
The problem is that we have a write W and a read R related by neither
fre or rfe, but rather W ->coe W' ->rfe R, where W' is an intermediate
write (the WRITE_ONCE() in P0). In this situation there is no
particular ordering between W and R, so either a wr-vis link from W to
R or an rw-xbstar link from R to W would prove that the accesses
aren't concurrent.
But the LKMM only looks for a wr-vis link, which is equivalent to
assuming that W must execute before R. This is not necessarily true
on non-multicopy-atomic systems, as the WWC pattern demonstrates.
This patch changes the LKMM to accept either a wr-vis or a reverse
rw-xbstar link as a proof of non-concurrency.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Herbert Xu recently reported a problem concerning RCU and compiler
barriers. In the course of discussing the problem, he put forth a
litmus test which illustrated a serious defect in the Linux Kernel
Memory Model's data-race-detection code [1].
The defect was that the LKMM assumed visibility and executes-before
ordering of plain accesses had to be mediated by marked accesses. In
Herbert's litmus test this wasn't so, and the LKMM claimed the litmus
test was allowed and contained a data race although neither is true.
In fact, plain accesses can be ordered by fences even in the absence
of marked accesses. In most cases this doesn't matter, because most
fences only order accesses within a single thread. But the rcu-fence
relation is different; it can order (and induce visibility between)
accesses in different threads -- events which otherwise might be
concurrent. This makes it relevant to data-race detection.
This patch makes two changes to the memory model to incorporate the
new insight:
If a store is separated by a fence from another access,
the store is necessarily visible to the other access (as
reflected in the ww-vis and wr-vis relations). Similarly,
if a load is separated by a fence from another access then
the load necessarily executes before the other access (as
reflected in the rw-xbstar relation).
If a store is separated by a strong fence from a marked access
then it is necessarily visible to any access that executes
after the marked access (as reflected in the ww-vis and wr-vis
relations).
With these changes, the LKMM gives the desired result for Herbert's
litmus test and other related ones [2].
[1] https://lore.kernel.org/lkml/Pine.LNX.4.44L0.1906041026570.1731-100000@iolanthe.rowland.org/
[2] https://github.com/paulmckrcu/litmus/blob/master/manual/plain/C-S-rcunoderef-1.litmushttps://github.com/paulmckrcu/litmus/blob/master/manual/plain/C-S-rcunoderef-2.litmushttps://github.com/paulmckrcu/litmus/blob/master/manual/plain/C-S-rcunoderef-3.litmushttps://github.com/paulmckrcu/litmus/blob/master/manual/plain/C-S-rcunoderef-4.litmushttps://github.com/paulmckrcu/litmus/blob/master/manual/plain/strong-vis.litmus
Reported-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Tested-by: Akira Yokosawa <akiyks@gmail.com>
The rcu-fence relation in the Linux Kernel Memory Model is not well
named. It doesn't act like any other fence relation, in that it does
not relate events before a fence to events after that fence. All it
does is relate certain RCU events to one another (those that are
ordered by the RCU Guarantee); this induces an actual
strong-fence-like relation linking events preceding the first RCU
event to those following the second.
This patch renames rcu-fence, now called rcu-order. It adds a new
definition of rcu-fence, something which should have been present all
along because it is used in the rb relation. And it modifies the
fence and strong-fence relations by making them incorporate the new
rcu-fence.
As a result of this change, there is no longer any need to define
full-fence in the section for detecting data races. It can simply be
replaced by the updated strong-fence relation.
This change should have no effect on the operation of the memory model.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Commit 66be4e66a7 ("rcu: locking and unlocking need to always be at
least barriers") added compiler barriers back into rcu_read_lock() and
rcu_read_unlock(). Furthermore, srcu_read_lock() and
srcu_read_unlock() have always contained compiler barriers.
The Linux Kernel Memory Model ought to know about these barriers.
This patch adds them into the memory model.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This patch adds data-race detection to the Linux-Kernel Memory Model.
As part of this effort, support is added for:
compiler barriers (the barrier() function), and
a new Preserved Program Order term: (addr ; [Plain] ; wmb)
Data races are marked with a special Flag warning in herd. It is
not guaranteed that the model will provide accurate predictions when a
data race is present.
The patch does not include documentation for the data-race detection
facility. The basic design has been explained in various emails, and
a separate documentation patch will be submitted later.
This work is based on an earlier formulation of data races for the
LKMM by Andrea Parri.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reviewed-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This patch adds definitions for marked and plain accesses to the
Linux-Kernel Memory Model. It also modifies the definitions of the
existing parts of the model (including the cumul-fence, prop, hb, pb,
and rb relations) so as to make them apply only to marked accesses.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reviewed-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This patch makes some slight alterations to linux-kernel.cat in
preparation for adding support for data-race detection to the
Linux-Kernel Memory Model.
The definitions of relations involved in Acquire, Release, and
unlock-lock ordering are moved up earlier in the source file.
The rmb relation is factored through the new R4rmb class: the
class of reads to which rmb will apply.
The definition of the fence relation is moved earlier, and it
is split up into read- and write-fences (rmb and wmb) and all
the others.
This should not make any functional changes.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reviewed-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Currently, herdtools version information appears no fewer than three
times in the LKMM source, which is difficult to maintain. This commit
therefore places the required version in one place, namely the
tools/memory-model/README file.
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
This commit checks that the return value of srcu_read_lock() is passed
to the matching srcu_read_unlock(), where "matching" is determined by
nesting. This check operates as follows:
1. srcu_read_lock() creates an integer token, which is stored into
the generated events.
2. srcu_read_unlock() records its second (token) argument into the
generated event.
3. A new herd primitive 'different-values' filters out pairs of events
with identical values from the relation passed as its argument.
4. The bell file applies the above primitive to the (srcu)
read-side-critical-section relation 'srcu-rscs' and flags non-empty
results.
BEWARE: Works only with herd version 7.51+6 and onwards.
Signed-off-by: Luc Maranget <Luc.Maranget@inria.fr>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
[ paulmck: Apply Andrea Parri's off-list feedback. ]
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Add support for SRCU. Herd creates srcu events and linux-kernel.def
associates them with three possible annotations (srcu-lock,
srcu-unlock, and sync-srcu) corresponding to the API routines
srcu_read_lock(), srcu_read_unlock(), and synchronize_srcu().
The linux-kernel.bell file now declares the annotations
and determines matching lock/unlock pairs delimiting SRCU read-side
critical sections, and it also checks for synchronize_srcu() calls
inside an RCU critical section (which would generate a "sleeping in
atomic context" error in real kernel code). The linux-kernel.cat file
now adds SRCU-induced ordering, analogous to the existing RCU-induced
ordering, to the gp and rcu-fence relations.
Curiously enough, these small changes to the model's .cat code are all
that is needed to describe SRCU.
Portions of this patch (linux-kernel.def and the first hunk in
linux-kernel.bell) were written by Luc Maranget.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
CC: Luc Maranget <luc.maranget@inria.fr>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Tested-by: Andrea Parri <andrea.parri@amarulasolutions.com>
In preparation for adding support for SRCU, refactor the definitions
of rcu-fence, rcu-rscsi, rcu-link, and rb by moving the po and po?
terms from the first two to the second two. An rcu-gp relation is
added; it is equivalent to gp with the po and po? terms removed.
This is necessary because for SRCU, we will have to use the loc
relation to check that the terms at the start and end of each disjunct
in the definition of rcu-fence refer to the same srcu_struct
location. If these terms are hidden behind po and po?, there's no way
to carry out this check.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Tested-by: Andrea Parri <andrea.parri@amarulasolutions.com>
In preparation for adding support for SRCU, rename "crit" to
"rcu-rscs", rename "rscs" to "rcu-rscsi", and remove the restriction
to only the outermost level of nesting.
The name change is needed for disambiguating RCU read-side critical
sections from SRCU read-side critical sections. Adding the "i" at the
end of "rcu-rscsi" emphasizes that the relation is inverted; it links
rcu_read_unlock() events to their corresponding preceding
rcu_read_lock() events.
The restriction to outermost nesting levels was never essential; it
was included mostly to show that it could be done. Rather than add
equivalent unnecessary code for SRCU lock nesting, it seemed better to
remove the existing code.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Tested-by: Andrea Parri <andrea.parri@amarulasolutions.com>
The kernel documents smp_mb__after_unlock_lock() the following way:
"Place this after a lock-acquisition primitive to guarantee that
an UNLOCK+LOCK pair acts as a full barrier. This guarantee applies
if the UNLOCK and LOCK are executed by the same CPU or if the
UNLOCK and LOCK operate on the same lock variable."
Formalize in LKMM the above guarantee by defining (new) mb-links according
to the law:
([M] ; po ; [UL] ; (co | po) ; [LKW] ;
fencerel(After-unlock-lock) ; [M])
where the component ([UL] ; co ; [LKW]) identifies "UNLOCK+LOCK pairs on
the same lock variable" and the component ([UL] ; po ; [LKW]) identifies
"UNLOCK+LOCK pairs executed by the same CPU".
In particular, the LKMM forbids the following two behaviors (the second
litmus test below is based on:
Documentation/RCU/Design/Memory-Ordering/Tree-RCU-Memory-Ordering.html
c.f., Section "Tree RCU Grace Period Memory Ordering Building Blocks"):
C after-unlock-lock-same-cpu
(*
* Result: Never
*)
{}
P0(spinlock_t *s, spinlock_t *t, int *x, int *y)
{
int r0;
spin_lock(s);
WRITE_ONCE(*x, 1);
spin_unlock(s);
spin_lock(t);
smp_mb__after_unlock_lock();
r0 = READ_ONCE(*y);
spin_unlock(t);
}
P1(int *x, int *y)
{
int r0;
WRITE_ONCE(*y, 1);
smp_mb();
r0 = READ_ONCE(*x);
}
exists (0:r0=0 /\ 1:r0=0)
C after-unlock-lock-same-lock-variable
(*
* Result: Never
*)
{}
P0(spinlock_t *s, int *x, int *y)
{
int r0;
spin_lock(s);
WRITE_ONCE(*x, 1);
r0 = READ_ONCE(*y);
spin_unlock(s);
}
P1(spinlock_t *s, int *y, int *z)
{
int r0;
spin_lock(s);
smp_mb__after_unlock_lock();
WRITE_ONCE(*y, 1);
r0 = READ_ONCE(*z);
spin_unlock(s);
}
P2(int *z, int *x)
{
int r0;
WRITE_ONCE(*z, 1);
smp_mb();
r0 = READ_ONCE(*x);
}
exists (0:r0=0 /\ 1:r0=0 /\ 2:r0=0)
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Cc: Akira Yokosawa <akiyks@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Daniel Lustig <dlustig@nvidia.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Jade Alglave <j.alglave@ucl.ac.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Luc Maranget <luc.maranget@inria.fr>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linux-arch@vger.kernel.org
Cc: parri.andrea@gmail.com
Link: http://lkml.kernel.org/r/20181203230451.28921-1-paulmck@linux.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
More than one kernel developer has expressed the opinion that the LKMM
should enforce ordering of writes by locking. In other words, given
the following code:
WRITE_ONCE(x, 1);
spin_unlock(&s):
spin_lock(&s);
WRITE_ONCE(y, 1);
the stores to x and y should be propagated in order to all other CPUs,
even though those other CPUs might not access the lock s. In terms of
the memory model, this means expanding the cumul-fence relation.
Locks should also provide read-read (and read-write) ordering in a
similar way. Given:
READ_ONCE(x);
spin_unlock(&s);
spin_lock(&s);
READ_ONCE(y); // or WRITE_ONCE(y, 1);
the load of x should be executed before the load of (or store to) y.
The LKMM already provides this ordering, but it provides it even in
the case where the two accesses are separated by a release/acquire
pair of fences rather than unlock/lock. This would prevent
architectures from using weakly ordered implementations of release and
acquire, which seems like an unnecessary restriction. The patch
therefore removes the ordering requirement from the LKMM for that
case.
There are several arguments both for and against this change. Let us
refer to these enhanced ordering properties by saying that the LKMM
would require locks to be RCtso (a bit of a misnomer, but analogous to
RCpc and RCsc) and it would require ordinary acquire/release only to
be RCpc. (Note: In the following, the phrase "all supported
architectures" is meant not to include RISC-V. Although RISC-V is
indeed supported by the kernel, the implementation is still somewhat
in a state of flux and therefore statements about it would be
premature.)
Pros:
The kernel already provides RCtso ordering for locks on all
supported architectures, even though this is not stated
explicitly anywhere. Therefore the LKMM should formalize it.
In theory, guaranteeing RCtso ordering would reduce the need
for additional barrier-like constructs meant to increase the
ordering strength of locks.
Will Deacon and Peter Zijlstra are strongly in favor of
formalizing the RCtso requirement. Linus Torvalds and Will
would like to go even further, requiring locks to have RCsc
behavior (ordering preceding writes against later reads), but
they recognize that this would incur a noticeable performance
degradation on the POWER architecture. Linus also points out
that people have made the mistake, in the past, of assuming
that locking has stronger ordering properties than is
currently guaranteed, and this change would reduce the
likelihood of such mistakes.
Not requiring ordinary acquire/release to be any stronger than
RCpc may prove advantageous for future architectures, allowing
them to implement smp_load_acquire() and smp_store_release()
with more efficient machine instructions than would be
possible if the operations had to be RCtso. Will and Linus
approve this rationale, hypothetical though it is at the
moment (it may end up affecting the RISC-V implementation).
The same argument may or may not apply to RMW-acquire/release;
see also the second Con entry below.
Linus feels that locks should be easy for people to use
without worrying about memory consistency issues, since they
are so pervasive in the kernel, whereas acquire/release is
much more of an "experts only" tool. Requiring locks to be
RCtso is a step in this direction.
Cons:
Andrea Parri and Luc Maranget think that locks should have the
same ordering properties as ordinary acquire/release (indeed,
Luc points out that the names "acquire" and "release" derive
from the usage of locks). Andrea points out that having
different ordering properties for different forms of acquires
and releases is not only unnecessary, it would also be
confusing and unmaintainable.
Locks are constructed from lower-level primitives, typically
RMW-acquire (for locking) and ordinary release (for unlock).
It is illogical to require stronger ordering properties from
the high-level operations than from the low-level operations
they comprise. Thus, this change would make
while (cmpxchg_acquire(&s, 0, 1) != 0)
cpu_relax();
an incorrect implementation of spin_lock(&s) as far as the
LKMM is concerned. In theory this weakness can be ameliorated
by changing the LKMM even further, requiring
RMW-acquire/release also to be RCtso (which it already is on
all supported architectures).
As far as I know, nobody has singled out any examples of code
in the kernel that actually relies on locks being RCtso.
(People mumble about RCU and the scheduler, but nobody has
pointed to any actual code. If there are any real cases,
their number is likely quite small.) If RCtso ordering is not
needed, why require it?
A handful of locking constructs (qspinlocks, qrwlocks, and
mcs_spinlocks) are built on top of smp_cond_load_acquire()
instead of an RMW-acquire instruction. It currently provides
only the ordinary acquire semantics, not the stronger ordering
this patch would require of locks. In theory this could be
ameliorated by requiring smp_cond_load_acquire() in
combination with ordinary release also to be RCtso (which is
currently true on all supported architectures).
On future weakly ordered architectures, people may be able to
implement locks in a non-RCtso fashion with significant
performance improvement. Meeting the RCtso requirement would
necessarily add run-time overhead.
Overall, the technical aspects of these arguments seem relatively
minor, and it appears mostly to boil down to a matter of opinion.
Since the opinions of senior kernel maintainers such as Linus,
Peter, and Will carry more weight than those of Luc and Andrea, this
patch changes the model in accordance with the maintainers' wishes.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Will Deacon <will.deacon@arm.com>
Reviewed-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: akiyks@gmail.com
Cc: boqun.feng@gmail.com
Cc: dhowells@redhat.com
Cc: j.alglave@ucl.ac.uk
Cc: linux-arch@vger.kernel.org
Cc: luc.maranget@inria.fr
Cc: npiggin@gmail.com
Cc: parri.andrea@gmail.com
Link: http://lkml.kernel.org/r/20180926182920.27644-2-paulmck@linux.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
ASPLOS 2018 was held in March: make sure this is reflected in
header comments and references.
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Akira Yokosawa <akiyks@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Jade Alglave <j.alglave@ucl.ac.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Luc Maranget <luc.maranget@inria.fr>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linux-arch@vger.kernel.org
Cc: parri.andrea@gmail.com
Link: http://lkml.kernel.org/r/1526340837-12222-18-git-send-email-paulmck@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch reorganizes the definition of rb in the Linux Kernel Memory
Consistency Model. The relation is now expressed in terms of
rcu-fence, which consists of a sequence of gp and rscs links separated
by rcu-link links, in which the number of occurrences of gp is >= the
number of occurrences of rscs.
Arguments similar to those published in
http://diy.inria.fr/linux/long.pdf show that rcu-fence behaves like an
inter-CPU strong fence. Furthermore, the definition of rb in terms of
rcu-fence is highly analogous to the definition of pb in terms of
strong-fence, which can help explain why rcu-path expresses a form of
temporal ordering.
This change should not affect the semantics of the memory model, just
its internal organization.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Andrea Parri <parri.andrea@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: akiyks@gmail.com
Cc: dhowells@redhat.com
Cc: j.alglave@ucl.ac.uk
Cc: linux-arch@vger.kernel.org
Cc: luc.maranget@inria.fr
Cc: npiggin@gmail.com
Link: http://lkml.kernel.org/r/1526340837-12222-2-git-send-email-paulmck@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch makes a simple non-functional change to the RCU portion of
the Linux Kernel Memory Consistency Model by renaming the "link" and
"rcu-path" relations to "rcu-link" and "rb", respectively.
The name "link" was an unfortunate choice, because it was too generic
and subject to confusion with other meanings of the same word, which
occur quite often in LKMM documentation. The name "rcu-path" is not
very appropriate, because the relation is analogous to the
happens-before (hb) and propagates-before (pb) relations -- although
that fact won't become apparent until the second patch in this series.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Andrea Parri <parri.andrea@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: akiyks@gmail.com
Cc: boqun.feng@gmail.com
Cc: dhowells@redhat.com
Cc: j.alglave@ucl.ac.uk
Cc: linux-arch@vger.kernel.org
Cc: luc.maranget@inria.fr
Cc: npiggin@gmail.com
Link: http://lkml.kernel.org/r/1526340837-12222-1-git-send-email-paulmck@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Ingo pointed out that:
"The "memory model" name is overly generic, ambiguous and somewhat
misleading, as we usually mean the virtual memory layout/model
when we say "memory model". GCC too uses it in that sense [...]"
Make it clear that tools/memory-model/ uses the term "memory model" as
shorthand for "memory consistency model" by calling out this convention
in tools/memory-model/README.
Stick to the original "memory model" term in sources' headers and for
the subsystem name.
Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: akiyks@gmail.com
Cc: boqun.feng@gmail.com
Cc: dhowells@redhat.com
Cc: j.alglave@ucl.ac.uk
Cc: linux-arch@vger.kernel.org
Cc: luc.maranget@inria.fr
Cc: nborisov@suse.com
Cc: npiggin@gmail.com
Link: http://lkml.kernel.org/r/1519169112-20593-1-git-send-email-paulmck@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There is some reason to believe that Documentation/memory-barriers.txt
could use some help, and a major purpose of this patch is to provide
that help in the form of a design-time tool that can produce all valid
executions of a small fragment of concurrent Linux-kernel code, which is
called a "litmus test". This tool's functionality is roughly similar to
a full state-space search. Please note that this is a design-time tool,
not useful for regression testing. However, we hope that the underlying
Linux-kernel memory model will be incorporated into other tools capable
of analyzing large bodies of code for regression-testing purposes.
The main tool is herd7, together with the linux-kernel.bell,
linux-kernel.cat, linux-kernel.cfg, linux-kernel.def, and lock.cat files
added by this patch. The herd7 executable takes the other files as input,
and all of these files collectively define the Linux-kernel memory memory
model. A brief description of each of these other files is provided
in the README file. Although this tool does have its limitations,
which are documented in the README file, it does improve on the version
reported on in the LWN series (https://lwn.net/Articles/718628/ and
https://lwn.net/Articles/720550/) by supporting locking and arithmetic,
including a much wider variety of read-modify-write atomic operations.
Please note that herd7 is not part of this submission, but is freely
available from http://diy.inria.fr/sources/index.html (and via "git"
at https://github.com/herd/herdtools7).
A second tool is klitmus7, which converts litmus tests to loadable
kernel modules for direct testing. As with herd7, the klitmus7
code is freely available from http://diy.inria.fr/sources/index.html
(and via "git" at https://github.com/herd/herdtools7).
Of course, litmus tests are not always the best way to fully understand a
memory model, so this patch also includes Documentation/explanation.txt,
which describes the memory model in detail. In addition,
Documentation/recipes.txt provides example known-good and known-bad use
cases for those who prefer working by example.
This patch also includes a few sample litmus tests, and a great many
more litmus tests are available at https://github.com/paulmckrcu/litmus.
This patch was the result of a most excellent collaboration founded
by Jade Alglave and also including Alan Stern, Andrea Parri, and Luc
Maranget. For more details on the history of this collaboration, please
refer to the Linux-kernel memory model presentations at 2016 LinuxCon EU,
2016 Kernel Summit, 2016 Linux Plumbers Conference, 2017 linux.conf.au,
or 2017 Linux Plumbers Conference microconference. However, one aspect
of the history does bear repeating due to weak copyright tracking earlier
in this project, which extends back to early 2015. This weakness came
to light in late 2017 after an LKMM presentation by Paul in which an
audience member noted the similarity of some LKMM code to code in early
published papers. This prompted a copyright review.
From Alan Stern:
To say that the model was mine is not entirely accurate.
Pieces of it (especially the Scpv and Atomic axioms) were taken
directly from Jade's models. And of course the Happens-before
and Propagation relations and axioms were heavily based on
Jade and Luc's work, even though they weren't identical to the
earlier versions. Only the RCU portion was completely original.
. . .
One can make a much better case that I wrote the bulk of lock.cat.
However, it was inspired by Luc's earlier version (and still
shares some elements in common), and of course it benefited from
feedback and testing from all members of our group.
The model prior to Alan's was Luc Maranget's. From Luc:
I totally agree on Alan Stern's account of the linux kernel model
genesis. I thank him for his acknowledgments of my participation
to previous model drafts. I'd like to complete Alan Stern's
statement: any bell cat code I have written has its roots in
discussions with Jade Alglave and Paul McKenney. Moreover I
have borrowed cat and bell code written by Jade Alglave freely.
This copyright review therefore resulted in late adds to the copyright
statements of several files.
Discussion of v1 has raised several issues, which we do not believe should
block acceptance given that this level of change will be ongoing, just
as it has been with memory-barriers.txt:
o Under what conditions should ordering provided by pure locking
be seen by CPUs not holding the relevant lock(s)? In particular,
should the message-passing pattern be forbidden?
o Should examples involving C11 release sequences be forbidden?
Note that this C11 is still a moving target for this issue:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0735r0.html
o Some details of the handling of internal dependencies for atomic
read-modify-write atomic operations are still subject to debate.
o Changes recently accepted into mainline greatly reduce the need
to handle DEC Alpha as a special case. These changes add an
smp_read_barrier_depends() to READ_ONCE(), thus causing Alpha
to respect ordering of dependent reads. If these changes stick,
the memory model can be simplified accordingly.
o Will changes be required to accommodate RISC-V?
Differences from v1:
(http://lkml.kernel.org/r/20171113184031.GA26302@linux.vnet.ibm.com)
o Add SPDX notations to .bell and .cat files, replacing
textual license statements.
o Add reference to upcoming ASPLOS paper to .bell and .cat files.
o Updated identifier names in .bell and .cat files to match those
used in the ASPLOS paper.
o Updates to READMEs and other documentation based on review
feedback.
o Added a memory-ordering cheatsheet.
o Update sigs to new Co-Developed-by and add acks and
reviewed-bys.
o Simplify rules detecting nested RCU read-side critical sections.
o Update copyright statements as noted above.
Co-Developed-by: Alan Stern <stern@rowland.harvard.edu>
Co-Developed-by: Andrea Parri <parri.andrea@gmail.com>
Co-Developed-by: Jade Alglave <j.alglave@ucl.ac.uk>
Co-Developed-by: Luc Maranget <luc.maranget@inria.fr>
Co-Developed-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Jade Alglave <j.alglave@ucl.ac.uk>
Signed-off-by: Luc Maranget <luc.maranget@inria.fr>
Signed-off-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Reviewed-by: Boqun Feng <boqun.feng@gmail.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Nicholas Piggin <npiggin@gmail.com>
Acked-by: David Howells <dhowells@redhat.com>
Acked-by: "Reshetova, Elena" <elena.reshetova@intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Akira Yokosawa <akiyks@gmail.com>
Cc: <linux-arch@vger.kernel.org>
Jonas Oberhauser