Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1680343 > unrolled thread

Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with lock/unlock pair

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair Alan Stern <stern@rowland.harvard.edu> - 2017-07-03 16:50 +0200
    Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-07-03 19:20 +0200
      Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair Manfred Spraul <manfred@colorfullife.com> - 2017-07-03 21:10 +0200
        Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair Alan Stern <stern@rowland.harvard.edu> - 2017-07-03 22:00 +0200
          Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair Manfred Spraul <manfred@colorfullife.com> - 2017-07-06 20:50 +0200
      Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair Alan Stern <stern@rowland.harvard.edu> - 2017-07-03 22:10 +0200
        Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with  lock/unlock pair "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-07-03 23:00 +0200

#1680343 — Re: [PATCH RFC 01/26] netfilter: Replace spin_unlock_wait() with lock/unlock pair

On Sat, 1 Jul 2017, Manfred Spraul wrote:

> As we want to remove spin_unlock_wait() and replace it with explicit
> spin_lock()/spin_unlock() calls, we can use this to simplify the
> locking.
> 
> In addition:
> - Reading nf_conntrack_locks_all needs ACQUIRE memory ordering.
> - The new code avoids the backwards loop.
> 
> Only slightly tested, I did not manage to trigger calls to
> nf_conntrack_all_lock().
> 
> Fixes: b16c29191dc8
> Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
> Cc: <stable@vger.kernel.org>
> Cc: Sasha Levin <sasha.levin@oracle.com>
> Cc: Pablo Neira Ayuso <pablo@netfilter.org>
> Cc: netfilter-devel@vger.kernel.org
> ---
>  net/netfilter/nf_conntrack_core.c | 44 +++++++++++++++++++++------------------
>  1 file changed, 24 insertions(+), 20 deletions(-)
> 
> diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
> index e847dba..1193565 100644
> --- a/net/netfilter/nf_conntrack_core.c
> +++ b/net/netfilter/nf_conntrack_core.c
> @@ -96,19 +96,24 @@ static struct conntrack_gc_work conntrack_gc_work;
>  
>  void nf_conntrack_lock(spinlock_t *lock) __acquires(lock)
>  {
> +	/* 1) Acquire the lock */
>  	spin_lock(lock);
> -	while (unlikely(nf_conntrack_locks_all)) {
> -		spin_unlock(lock);
>  
> -		/*
> -		 * Order the 'nf_conntrack_locks_all' load vs. the
> -		 * spin_unlock_wait() loads below, to ensure
> -		 * that 'nf_conntrack_locks_all_lock' is indeed held:
> -		 */
> -		smp_rmb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
> -		spin_unlock_wait(&nf_conntrack_locks_all_lock);
> -		spin_lock(lock);
> -	}
> +	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics */
> +	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
> +		return;

As far as I can tell, this read does not need to have ACQUIRE
semantics.

You need to guarantee that two things can never happen:

    (1) We read nf_conntrack_locks_all == false, and this routine's
	critical section for nf_conntrack_locks[i] runs after the
	(empty) critical section for that lock in 
	nf_conntrack_all_lock().

    (2) We read nf_conntrack_locks_all == true, and this routine's 
	critical section for nf_conntrack_locks_all_lock runs before 
	the critical section in nf_conntrack_all_lock().

In fact, neither one can happen even if smp_load_acquire() is replaced
with READ_ONCE().  The reason is simple enough, using this property of
spinlocks:

	If critical section CS1 runs before critical section CS2 (for 
	the same lock) then: (a) every write coming before CS1's
	spin_unlock() will be visible to any read coming after CS2's
	spin_lock(), and (b) no write coming after CS2's spin_lock()
	will be visible to any read coming before CS1's spin_unlock().

Thus for (1), assuming the critical sections run in the order mentioned
above, since nf_conntrack_all_lock() writes to nf_conntrack_locks_all
before releasing nf_conntrack_locks[i], and since nf_conntrack_lock()
acquires nf_conntrack_locks[i] before reading nf_conntrack_locks_all,
by (a) the read will always see the write.

Similarly for (2), since nf_conntrack_all_lock() acquires 
nf_conntrack_locks_all_lock before writing to nf_conntrack_locks_all, 
and since nf_conntrack_lock() reads nf_conntrack_locks_all before 
releasing nf_conntrack_locks_all_lock, by (b) the read cannot see the 
write.

Alan Stern

> +
> +	/* fast path failed, unlock */
> +	spin_unlock(lock);
> +
> +	/* Slow path 1) get global lock */
> +	spin_lock(&nf_conntrack_locks_all_lock);
> +
> +	/* Slow path 2) get the lock we want */
> +	spin_lock(lock);
> +
> +	/* Slow path 3) release the global lock */
> +	spin_unlock(&nf_conntrack_locks_all_lock);
>  }
>  EXPORT_SYMBOL_GPL(nf_conntrack_lock);
>  
> @@ -149,18 +154,17 @@ static void nf_conntrack_all_lock(void)
>  	int i;
>  
>  	spin_lock(&nf_conntrack_locks_all_lock);
> -	nf_conntrack_locks_all = true;
>  
> -	/*
> -	 * Order the above store of 'nf_conntrack_locks_all' against
> -	 * the spin_unlock_wait() loads below, such that if
> -	 * nf_conntrack_lock() observes 'nf_conntrack_locks_all'
> -	 * we must observe nf_conntrack_locks[] held:
> -	 */
> -	smp_mb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
> +	nf_conntrack_locks_all = true;
>  
>  	for (i = 0; i < CONNTRACK_LOCKS; i++) {
> -		spin_unlock_wait(&nf_conntrack_locks[i]);
> +		spin_lock(&nf_conntrack_locks[i]);
> +
> +		/* This spin_unlock provides the "release" to ensure that
> +		 * nf_conntrack_locks_all==true is visible to everyone that
> +		 * acquired spin_lock(&nf_conntrack_locks[]).
> +		 */
> +		spin_unlock(&nf_conntrack_locks[i]);
>  	}
>  }

[toc] | [next] | [standalone]

#1680509

On Mon, Jul 03, 2017 at 10:39:49AM -0400, Alan Stern wrote:
> On Sat, 1 Jul 2017, Manfred Spraul wrote:
> 
> > As we want to remove spin_unlock_wait() and replace it with explicit
> > spin_lock()/spin_unlock() calls, we can use this to simplify the
> > locking.
> > 
> > In addition:
> > - Reading nf_conntrack_locks_all needs ACQUIRE memory ordering.
> > - The new code avoids the backwards loop.
> > 
> > Only slightly tested, I did not manage to trigger calls to
> > nf_conntrack_all_lock().
> > 
> > Fixes: b16c29191dc8
> > Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
> > Cc: <stable@vger.kernel.org>
> > Cc: Sasha Levin <sasha.levin@oracle.com>
> > Cc: Pablo Neira Ayuso <pablo@netfilter.org>
> > Cc: netfilter-devel@vger.kernel.org
> > ---
> >  net/netfilter/nf_conntrack_core.c | 44 +++++++++++++++++++++------------------
> >  1 file changed, 24 insertions(+), 20 deletions(-)
> > 
> > diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
> > index e847dba..1193565 100644
> > --- a/net/netfilter/nf_conntrack_core.c
> > +++ b/net/netfilter/nf_conntrack_core.c
> > @@ -96,19 +96,24 @@ static struct conntrack_gc_work conntrack_gc_work;
> >  
> >  void nf_conntrack_lock(spinlock_t *lock) __acquires(lock)
> >  {
> > +	/* 1) Acquire the lock */
> >  	spin_lock(lock);
> > -	while (unlikely(nf_conntrack_locks_all)) {
> > -		spin_unlock(lock);
> >  
> > -		/*
> > -		 * Order the 'nf_conntrack_locks_all' load vs. the
> > -		 * spin_unlock_wait() loads below, to ensure
> > -		 * that 'nf_conntrack_locks_all_lock' is indeed held:
> > -		 */
> > -		smp_rmb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
> > -		spin_unlock_wait(&nf_conntrack_locks_all_lock);
> > -		spin_lock(lock);
> > -	}
> > +	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics */
> > +	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
> > +		return;
> 
> As far as I can tell, this read does not need to have ACQUIRE
> semantics.
> 
> You need to guarantee that two things can never happen:
> 
>     (1) We read nf_conntrack_locks_all == false, and this routine's
> 	critical section for nf_conntrack_locks[i] runs after the
> 	(empty) critical section for that lock in 
> 	nf_conntrack_all_lock().
> 
>     (2) We read nf_conntrack_locks_all == true, and this routine's 
> 	critical section for nf_conntrack_locks_all_lock runs before 
> 	the critical section in nf_conntrack_all_lock().
> 
> In fact, neither one can happen even if smp_load_acquire() is replaced
> with READ_ONCE().  The reason is simple enough, using this property of
> spinlocks:
> 
> 	If critical section CS1 runs before critical section CS2 (for 
> 	the same lock) then: (a) every write coming before CS1's
> 	spin_unlock() will be visible to any read coming after CS2's
> 	spin_lock(), and (b) no write coming after CS2's spin_lock()
> 	will be visible to any read coming before CS1's spin_unlock().
> 
> Thus for (1), assuming the critical sections run in the order mentioned
> above, since nf_conntrack_all_lock() writes to nf_conntrack_locks_all
> before releasing nf_conntrack_locks[i], and since nf_conntrack_lock()
> acquires nf_conntrack_locks[i] before reading nf_conntrack_locks_all,
> by (a) the read will always see the write.
> 
> Similarly for (2), since nf_conntrack_all_lock() acquires 
> nf_conntrack_locks_all_lock before writing to nf_conntrack_locks_all, 
> and since nf_conntrack_lock() reads nf_conntrack_locks_all before 
> releasing nf_conntrack_locks_all_lock, by (b) the read cannot see the 
> write.

And the Linux kernel memory model (https://lwn.net/Articles/718628/
and https://lwn.net/Articles/720550/) agrees with Alan.  Here is
a litmus test, which emulates spin_lock() with xchg_acquire() and
spin_unlock() with smp_store_release():

------------------------------------------------------------------------

C C-ManfredSpraul-L1G1xchgnr.litmus

(* Expected result: Never.  *)

{
}

P0(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
{
	/* Acquire local lock. */
	r10 = xchg_acquire(lcl, 1);
	r1 = READ_ONCE(*nfcla);
	if (r1) {
		smp_store_release(lcl, 0);
		r11 = xchg_acquire(gbl, 1);
		r12 = xchg_acquire(lcl, 1);
		smp_store_release(gbl, 0);
	}
	r2 = READ_ONCE(*gbl_held);
	WRITE_ONCE(*lcl_held, 1);
	WRITE_ONCE(*lcl_held, 0);
	smp_store_release(lcl, 0);
}

P1(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
{
	/* Acquire global lock. */
	r10 = xchg_acquire(gbl, 1);
	WRITE_ONCE(*nfcla, 1);
	r11 = xchg_acquire(lcl, 1);
	smp_store_release(lcl, 0);
	r2 = READ_ONCE(*lcl_held);
	WRITE_ONCE(*gbl_held, 1);
	WRITE_ONCE(*gbl_held, 0);
	smp_store_release(gbl, 0);
}

exists
((0:r2=1 \/ 1:r2=1) /\ 0:r10=0 /\ 0:r11=0 /\ 0:r12=0 /\ 1:r10=0 /\ 1:r11=0)

------------------------------------------------------------------------

The memory model says that the forbidden state does not happen:

------------------------------------------------------------------------

States 25
0:r10=0; 0:r11=0; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=0; 0:r11=0; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=1; 1:r2=0;
0:r10=0; 0:r11=0; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=1; 1:r2=1;
0:r10=0; 0:r11=0; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=1; 1:r2=0;
0:r10=0; 0:r11=0; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=1; 1:r2=1;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=1; 1:r2=0;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=1; 1:r2=1;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=1; 1:r2=0;
0:r10=0; 0:r11=1; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=1; 1:r2=1;
0:r10=0; 0:r11=1; 0:r12=1; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=0; 0:r11=1; 0:r12=1; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=0; 0:r11=1; 0:r12=1; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=0; 0:r11=1; 0:r12=1; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=1; 0:r11=0; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=1; 0:r11=0; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=1; 0:r11=0; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=1; 0:r11=0; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=1; 0:r11=1; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=1; 0:r11=1; 0:r12=0; 0:r2=0; 1:r10=0; 1:r11=0; 1:r2=1;
0:r10=1; 0:r11=1; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=0;
0:r10=1; 0:r11=1; 0:r12=0; 0:r2=1; 1:r10=0; 1:r11=0; 1:r2=1;
No
Witnesses
Positive: 0 Negative: 260
Condition exists ((0:r2=1 \/ 1:r2=1) /\ 0:r10=0 /\ 0:r11=0 /\ 0:r12=0 /\ 1:r10=0 /\ 1:r11=0)
Observation C-ManfredSpraul-L1G1xchgnr Never 0 260

------------------------------------------------------------------------

(Note the line "Positive: 0 Negative: 260", in other words, there
were no scenarios that matched the "exists" clause and 260 that did
not match.)

Of course, testing is also required.  ;-)

Manfred, any objections to my changing your patch as Alan suggests?

							Thanx, Paul

[toc] | [prev] | [next] | [standalone]

#1680541

On 07/03/2017 07:14 PM, Paul E. McKenney wrote:
> On Mon, Jul 03, 2017 at 10:39:49AM -0400, Alan Stern wrote:
>> On Sat, 1 Jul 2017, Manfred Spraul wrote:
>>
>>> As we want to remove spin_unlock_wait() and replace it with explicit
>>> spin_lock()/spin_unlock() calls, we can use this to simplify the
>>> locking.
>>>
>>> In addition:
>>> - Reading nf_conntrack_locks_all needs ACQUIRE memory ordering.
>>> - The new code avoids the backwards loop.
>>>
>>> Only slightly tested, I did not manage to trigger calls to
>>> nf_conntrack_all_lock().
>>>
>>> Fixes: b16c29191dc8
>>> Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
>>> Cc: <stable@vger.kernel.org>
>>> Cc: Sasha Levin <sasha.levin@oracle.com>
>>> Cc: Pablo Neira Ayuso <pablo@netfilter.org>
>>> Cc: netfilter-devel@vger.kernel.org
>>> ---
>>>   net/netfilter/nf_conntrack_core.c | 44 +++++++++++++++++++++------------------
>>>   1 file changed, 24 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
>>> index e847dba..1193565 100644
>>> --- a/net/netfilter/nf_conntrack_core.c
>>> +++ b/net/netfilter/nf_conntrack_core.c
>>> @@ -96,19 +96,24 @@ static struct conntrack_gc_work conntrack_gc_work;
>>>   
>>>   void nf_conntrack_lock(spinlock_t *lock) __acquires(lock)
>>>   {
>>> +	/* 1) Acquire the lock */
>>>   	spin_lock(lock);
>>> -	while (unlikely(nf_conntrack_locks_all)) {
>>> -		spin_unlock(lock);
>>>   
>>> -		/*
>>> -		 * Order the 'nf_conntrack_locks_all' load vs. the
>>> -		 * spin_unlock_wait() loads below, to ensure
>>> -		 * that 'nf_conntrack_locks_all_lock' is indeed held:
>>> -		 */
>>> -		smp_rmb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
>>> -		spin_unlock_wait(&nf_conntrack_locks_all_lock);
>>> -		spin_lock(lock);
>>> -	}
>>> +	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics */
>>> +	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
>>> +		return;
>> As far as I can tell, this read does not need to have ACQUIRE
>> semantics.
>>
>> You need to guarantee that two things can never happen:
>>
>>      (1) We read nf_conntrack_locks_all == false, and this routine's
>> 	critical section for nf_conntrack_locks[i] runs after the
>> 	(empty) critical section for that lock in
>> 	nf_conntrack_all_lock().
>>
>>      (2) We read nf_conntrack_locks_all == true, and this routine's
>> 	critical section for nf_conntrack_locks_all_lock runs before
>> 	the critical section in nf_conntrack_all_lock().
I was looking at nf_conntrack_all_unlock:
There is a smp_store_release() - which memory barrier does this pair with?

nf_conntrack_all_unlock()
     <arbitrary writes>
     smp_store_release(a, false)
     spin_unlock(b);

nf_conntrack_lock()
     spin_lock(c);
     xx=read_once(a)
     if (xx==false)
         return
     <arbitrary read>

>> In fact, neither one can happen even if smp_load_acquire() is replaced
>> with READ_ONCE().  The reason is simple enough, using this property of
>> spinlocks:
>>
>> 	If critical section CS1 runs before critical section CS2 (for
>> 	the same lock) then: (a) every write coming before CS1's
>> 	spin_unlock() will be visible to any read coming after CS2's
>> 	spin_lock(), and (b) no write coming after CS2's spin_lock()
>> 	will be visible to any read coming before CS1's spin_unlock().
Does this apply? The locks are different.
>> Thus for (1), assuming the critical sections run in the order mentioned
>> above, since nf_conntrack_all_lock() writes to nf_conntrack_locks_all
>> before releasing nf_conntrack_locks[i], and since nf_conntrack_lock()
>> acquires nf_conntrack_locks[i] before reading nf_conntrack_locks_all,
>> by (a) the read will always see the write.
>>
>> Similarly for (2), since nf_conntrack_all_lock() acquires
>> nf_conntrack_locks_all_lock before writing to nf_conntrack_locks_all,
>> and since nf_conntrack_lock() reads nf_conntrack_locks_all before
>> releasing nf_conntrack_locks_all_lock, by (b) the read cannot see the
>> write.
> And the Linux kernel memory model (https://lwn.net/Articles/718628/
> and https://lwn.net/Articles/720550/) agrees with Alan.  Here is
> a litmus test, which emulates spin_lock() with xchg_acquire() and
> spin_unlock() with smp_store_release():
>
> ------------------------------------------------------------------------
>
> C C-ManfredSpraul-L1G1xchgnr.litmus
>
> (* Expected result: Never.  *)
>
> {
> }
>
> P0(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
> {
> 	/* Acquire local lock. */
> 	r10 = xchg_acquire(lcl, 1);
> 	r1 = READ_ONCE(*nfcla);
> 	if (r1) {
> 		smp_store_release(lcl, 0);
> 		r11 = xchg_acquire(gbl, 1);
> 		r12 = xchg_acquire(lcl, 1);
> 		smp_store_release(gbl, 0);
> 	}
> 	r2 = READ_ONCE(*gbl_held);
> 	WRITE_ONCE(*lcl_held, 1);
> 	WRITE_ONCE(*lcl_held, 0);
> 	smp_store_release(lcl, 0);
> }
>
> P1(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
> {
> 	/* Acquire global lock. */
> 	r10 = xchg_acquire(gbl, 1);
> 	WRITE_ONCE(*nfcla, 1);
> 	r11 = xchg_acquire(lcl, 1);
> 	smp_store_release(lcl, 0);
> 	r2 = READ_ONCE(*lcl_held);
> 	WRITE_ONCE(*gbl_held, 1);
> 	WRITE_ONCE(*gbl_held, 0);
Where is the write that resets nfcla=0?
> 	smp_store_release(gbl, 0);
> }
>
> exists
> ((0:r2=1 \/ 1:r2=1) /\ 0:r10=0 /\ 0:r11=0 /\ 0:r12=0 /\ 1:r10=0 /\ 1:r11=0)
>
> ------------------------------------------------------------------------
>
> The memory model says that the forbidden state does not happen:
[...]
> Manfred, any objections to my changing your patch as Alan suggests?
I tried to pair the memory barriers:
nf_conntrack_all_unlock() contains a smp_store_release().
What does that pair with?

--
     Manfred

[toc] | [prev] | [next] | [standalone]

#1680553

On Mon, 3 Jul 2017, Manfred Spraul wrote:

> >>> +	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics */
> >>> +	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
> >>> +		return;
> >> As far as I can tell, this read does not need to have ACQUIRE
> >> semantics.
> >>
> >> You need to guarantee that two things can never happen:
> >>
> >>      (1) We read nf_conntrack_locks_all == false, and this routine's
> >> 	critical section for nf_conntrack_locks[i] runs after the
> >> 	(empty) critical section for that lock in
> >> 	nf_conntrack_all_lock().
> >>
> >>      (2) We read nf_conntrack_locks_all == true, and this routine's
> >> 	critical section for nf_conntrack_locks_all_lock runs before
> >> 	the critical section in nf_conntrack_all_lock().
> I was looking at nf_conntrack_all_unlock:
> There is a smp_store_release() - which memory barrier does this pair with?
> 
> nf_conntrack_all_unlock()
>      <arbitrary writes>
>      smp_store_release(a, false)
>      spin_unlock(b);
> 
> nf_conntrack_lock()
>      spin_lock(c);
>      xx=read_once(a)
>      if (xx==false)
>          return
>      <arbitrary read>

Ah, I see your point.  Yes, I did wonder about what would happen when
nf_conntrack_locks_all was set back to false.  But I didn't think about
it any further, because the relevant code wasn't in your patch.

> I tried to pair the memory barriers:
> nf_conntrack_all_unlock() contains a smp_store_release().
> What does that pair with?

You are right, this does need to be smp_load_acquire() after all.  
Perhaps the preceding comment should mention that it pairs with the 
smp_store_release() from an earlier invocation of 
nf_conntrack_all_unlock().

(Alternatively, you could make nf_conntrack_all_unlock() do a
lock+unlock on all the locks in the array, just like
nf_conntrack_all_lock().  But of course, that would be a lot less
efficient.)

Alan Stern

[toc] | [prev] | [next] | [standalone]

#1682644

[Multipart message — attachments visible in raw view] — view raw

Hi Alan,

On 07/03/2017 09:57 PM, Alan Stern wrote:
>
> (Alternatively, you could make nf_conntrack_all_unlock() do a
> lock+unlock on all the locks in the array, just like
> nf_conntrack_all_lock().  But of course, that would be a lot less
> efficient.)
Hmmmm.

Someone with a weakly ordered system who can test this?
semop() has a very short hotpath.

Either with aim9.shared_memory.ops_per_sec or

#sem-scalebench -t 10 -m 0
https://github.com/manfred-colorfu/ipcscale/blob/master/sem-scalebench.cpp
--
     Manfred

[toc] | [prev] | [next] | [standalone]

#1680557

On Mon, 3 Jul 2017, Paul E. McKenney wrote:

> On Mon, Jul 03, 2017 at 10:39:49AM -0400, Alan Stern wrote:
> > On Sat, 1 Jul 2017, Manfred Spraul wrote:
> > 
> > > As we want to remove spin_unlock_wait() and replace it with explicit
> > > spin_lock()/spin_unlock() calls, we can use this to simplify the
> > > locking.
> > > 
> > > In addition:
> > > - Reading nf_conntrack_locks_all needs ACQUIRE memory ordering.
> > > - The new code avoids the backwards loop.
> > > 
> > > Only slightly tested, I did not manage to trigger calls to
> > > nf_conntrack_all_lock().
> > > 
> > > Fixes: b16c29191dc8
> > > Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
> > > Cc: <stable@vger.kernel.org>
> > > Cc: Sasha Levin <sasha.levin@oracle.com>
> > > Cc: Pablo Neira Ayuso <pablo@netfilter.org>
> > > Cc: netfilter-devel@vger.kernel.org
> > > ---
> > >  net/netfilter/nf_conntrack_core.c | 44 +++++++++++++++++++++------------------
> > >  1 file changed, 24 insertions(+), 20 deletions(-)
> > > 
> > > diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
> > > index e847dba..1193565 100644
> > > --- a/net/netfilter/nf_conntrack_core.c
> > > +++ b/net/netfilter/nf_conntrack_core.c
> > > @@ -96,19 +96,24 @@ static struct conntrack_gc_work conntrack_gc_work;
> > >  
> > >  void nf_conntrack_lock(spinlock_t *lock) __acquires(lock)
> > >  {
> > > +	/* 1) Acquire the lock */
> > >  	spin_lock(lock);
> > > -	while (unlikely(nf_conntrack_locks_all)) {
> > > -		spin_unlock(lock);
> > >  
> > > -		/*
> > > -		 * Order the 'nf_conntrack_locks_all' load vs. the
> > > -		 * spin_unlock_wait() loads below, to ensure
> > > -		 * that 'nf_conntrack_locks_all_lock' is indeed held:
> > > -		 */
> > > -		smp_rmb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
> > > -		spin_unlock_wait(&nf_conntrack_locks_all_lock);
> > > -		spin_lock(lock);
> > > -	}
> > > +	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics */
> > > +	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
> > > +		return;
> > 
> > As far as I can tell, this read does not need to have ACQUIRE
> > semantics.
> > 
> > You need to guarantee that two things can never happen:
> > 
> >     (1) We read nf_conntrack_locks_all == false, and this routine's
> > 	critical section for nf_conntrack_locks[i] runs after the
> > 	(empty) critical section for that lock in 
> > 	nf_conntrack_all_lock().
> > 
> >     (2) We read nf_conntrack_locks_all == true, and this routine's 
> > 	critical section for nf_conntrack_locks_all_lock runs before 
> > 	the critical section in nf_conntrack_all_lock().
> > 
> > In fact, neither one can happen even if smp_load_acquire() is replaced
> > with READ_ONCE().  The reason is simple enough, using this property of
> > spinlocks:
> > 
> > 	If critical section CS1 runs before critical section CS2 (for 
> > 	the same lock) then: (a) every write coming before CS1's
> > 	spin_unlock() will be visible to any read coming after CS2's
> > 	spin_lock(), and (b) no write coming after CS2's spin_lock()
> > 	will be visible to any read coming before CS1's spin_unlock().
> > 
> > Thus for (1), assuming the critical sections run in the order mentioned
> > above, since nf_conntrack_all_lock() writes to nf_conntrack_locks_all
> > before releasing nf_conntrack_locks[i], and since nf_conntrack_lock()
> > acquires nf_conntrack_locks[i] before reading nf_conntrack_locks_all,
> > by (a) the read will always see the write.
> > 
> > Similarly for (2), since nf_conntrack_all_lock() acquires 
> > nf_conntrack_locks_all_lock before writing to nf_conntrack_locks_all, 
> > and since nf_conntrack_lock() reads nf_conntrack_locks_all before 
> > releasing nf_conntrack_locks_all_lock, by (b) the read cannot see the 
> > write.
> 
> And the Linux kernel memory model (https://lwn.net/Articles/718628/
> and https://lwn.net/Articles/720550/) agrees with Alan.  Here is
> a litmus test, which emulates spin_lock() with xchg_acquire() and
> spin_unlock() with smp_store_release():
> 
> ------------------------------------------------------------------------
> 
> C C-ManfredSpraul-L1G1xchgnr.litmus
> 
> (* Expected result: Never.  *)
> 
> {
> }
> 
> P0(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
> {
> 	/* Acquire local lock. */
> 	r10 = xchg_acquire(lcl, 1);
> 	r1 = READ_ONCE(*nfcla);
> 	if (r1) {
> 		smp_store_release(lcl, 0);
> 		r11 = xchg_acquire(gbl, 1);
> 		r12 = xchg_acquire(lcl, 1);
> 		smp_store_release(gbl, 0);
> 	}
> 	r2 = READ_ONCE(*gbl_held);
> 	WRITE_ONCE(*lcl_held, 1);
> 	WRITE_ONCE(*lcl_held, 0);
> 	smp_store_release(lcl, 0);
> }
> 
> P1(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
> {
> 	/* Acquire global lock. */
> 	r10 = xchg_acquire(gbl, 1);
> 	WRITE_ONCE(*nfcla, 1);
> 	r11 = xchg_acquire(lcl, 1);
> 	smp_store_release(lcl, 0);
> 	r2 = READ_ONCE(*lcl_held);
> 	WRITE_ONCE(*gbl_held, 1);

This litmus test is incomplete, because it omits the assignment setting
nf_conntrack_locks_all back to false when the global lock is released.  
You should insert

	smp_store_release(*nfcla, 0);

right here.

> 	WRITE_ONCE(*gbl_held, 0);
> 	smp_store_release(gbl, 0);
> }
> 
> exists
> ((0:r2=1 \/ 1:r2=1) /\ 0:r10=0 /\ 0:r11=0 /\ 0:r12=0 /\ 1:r10=0 /\ 1:r11=0)

With that addition, the litmus test fails unless the read of nfcla in 
P0 is an smp_load_acquire.  So Manfred's patch should not be changed.

Alan Stern

[toc] | [prev] | [next] | [standalone]

#1680564

On Mon, Jul 03, 2017 at 04:04:14PM -0400, Alan Stern wrote:
> On Mon, 3 Jul 2017, Paul E. McKenney wrote:
> 
> > On Mon, Jul 03, 2017 at 10:39:49AM -0400, Alan Stern wrote:
> > > On Sat, 1 Jul 2017, Manfred Spraul wrote:
> > > 
> > > > As we want to remove spin_unlock_wait() and replace it with explicit
> > > > spin_lock()/spin_unlock() calls, we can use this to simplify the
> > > > locking.
> > > > 
> > > > In addition:
> > > > - Reading nf_conntrack_locks_all needs ACQUIRE memory ordering.
> > > > - The new code avoids the backwards loop.
> > > > 
> > > > Only slightly tested, I did not manage to trigger calls to
> > > > nf_conntrack_all_lock().
> > > > 
> > > > Fixes: b16c29191dc8
> > > > Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
> > > > Cc: <stable@vger.kernel.org>
> > > > Cc: Sasha Levin <sasha.levin@oracle.com>
> > > > Cc: Pablo Neira Ayuso <pablo@netfilter.org>
> > > > Cc: netfilter-devel@vger.kernel.org
> > > > ---
> > > >  net/netfilter/nf_conntrack_core.c | 44 +++++++++++++++++++++------------------
> > > >  1 file changed, 24 insertions(+), 20 deletions(-)
> > > > 
> > > > diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
> > > > index e847dba..1193565 100644
> > > > --- a/net/netfilter/nf_conntrack_core.c
> > > > +++ b/net/netfilter/nf_conntrack_core.c
> > > > @@ -96,19 +96,24 @@ static struct conntrack_gc_work conntrack_gc_work;
> > > >  
> > > >  void nf_conntrack_lock(spinlock_t *lock) __acquires(lock)
> > > >  {
> > > > +	/* 1) Acquire the lock */
> > > >  	spin_lock(lock);
> > > > -	while (unlikely(nf_conntrack_locks_all)) {
> > > > -		spin_unlock(lock);
> > > >  
> > > > -		/*
> > > > -		 * Order the 'nf_conntrack_locks_all' load vs. the
> > > > -		 * spin_unlock_wait() loads below, to ensure
> > > > -		 * that 'nf_conntrack_locks_all_lock' is indeed held:
> > > > -		 */
> > > > -		smp_rmb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
> > > > -		spin_unlock_wait(&nf_conntrack_locks_all_lock);
> > > > -		spin_lock(lock);
> > > > -	}
> > > > +	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics */
> > > > +	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
> > > > +		return;
> > > 
> > > As far as I can tell, this read does not need to have ACQUIRE
> > > semantics.
> > > 
> > > You need to guarantee that two things can never happen:
> > > 
> > >     (1) We read nf_conntrack_locks_all == false, and this routine's
> > > 	critical section for nf_conntrack_locks[i] runs after the
> > > 	(empty) critical section for that lock in 
> > > 	nf_conntrack_all_lock().
> > > 
> > >     (2) We read nf_conntrack_locks_all == true, and this routine's 
> > > 	critical section for nf_conntrack_locks_all_lock runs before 
> > > 	the critical section in nf_conntrack_all_lock().
> > > 
> > > In fact, neither one can happen even if smp_load_acquire() is replaced
> > > with READ_ONCE().  The reason is simple enough, using this property of
> > > spinlocks:
> > > 
> > > 	If critical section CS1 runs before critical section CS2 (for 
> > > 	the same lock) then: (a) every write coming before CS1's
> > > 	spin_unlock() will be visible to any read coming after CS2's
> > > 	spin_lock(), and (b) no write coming after CS2's spin_lock()
> > > 	will be visible to any read coming before CS1's spin_unlock().
> > > 
> > > Thus for (1), assuming the critical sections run in the order mentioned
> > > above, since nf_conntrack_all_lock() writes to nf_conntrack_locks_all
> > > before releasing nf_conntrack_locks[i], and since nf_conntrack_lock()
> > > acquires nf_conntrack_locks[i] before reading nf_conntrack_locks_all,
> > > by (a) the read will always see the write.
> > > 
> > > Similarly for (2), since nf_conntrack_all_lock() acquires 
> > > nf_conntrack_locks_all_lock before writing to nf_conntrack_locks_all, 
> > > and since nf_conntrack_lock() reads nf_conntrack_locks_all before 
> > > releasing nf_conntrack_locks_all_lock, by (b) the read cannot see the 
> > > write.
> > 
> > And the Linux kernel memory model (https://lwn.net/Articles/718628/
> > and https://lwn.net/Articles/720550/) agrees with Alan.  Here is
> > a litmus test, which emulates spin_lock() with xchg_acquire() and
> > spin_unlock() with smp_store_release():
> > 
> > ------------------------------------------------------------------------
> > 
> > C C-ManfredSpraul-L1G1xchgnr.litmus
> > 
> > (* Expected result: Never.  *)
> > 
> > {
> > }
> > 
> > P0(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
> > {
> > 	/* Acquire local lock. */
> > 	r10 = xchg_acquire(lcl, 1);
> > 	r1 = READ_ONCE(*nfcla);
> > 	if (r1) {
> > 		smp_store_release(lcl, 0);
> > 		r11 = xchg_acquire(gbl, 1);
> > 		r12 = xchg_acquire(lcl, 1);
> > 		smp_store_release(gbl, 0);
> > 	}
> > 	r2 = READ_ONCE(*gbl_held);
> > 	WRITE_ONCE(*lcl_held, 1);
> > 	WRITE_ONCE(*lcl_held, 0);
> > 	smp_store_release(lcl, 0);
> > }
> > 
> > P1(int *nfcla, spinlock_t *gbl, int *gbl_held, spinlock_t *lcl, int *lcl_held)
> > {
> > 	/* Acquire global lock. */
> > 	r10 = xchg_acquire(gbl, 1);
> > 	WRITE_ONCE(*nfcla, 1);
> > 	r11 = xchg_acquire(lcl, 1);
> > 	smp_store_release(lcl, 0);
> > 	r2 = READ_ONCE(*lcl_held);
> > 	WRITE_ONCE(*gbl_held, 1);
> 
> This litmus test is incomplete, because it omits the assignment setting
> nf_conntrack_locks_all back to false when the global lock is released.  
> You should insert
> 
> 	smp_store_release(*nfcla, 0);
> 
> right here.
> 
> > 	WRITE_ONCE(*gbl_held, 0);
> > 	smp_store_release(gbl, 0);
> > }
> > 
> > exists
> > ((0:r2=1 \/ 1:r2=1) /\ 0:r10=0 /\ 0:r11=0 /\ 0:r12=0 /\ 1:r10=0 /\ 1:r11=0)
> 
> With that addition, the litmus test fails unless the read of nfcla in 
> P0 is an smp_load_acquire.  So Manfred's patch should not be changed.

Very good!  Aside from updating litmus tests, my work is done!  ;-)

							Thanx, Paul

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web