[PATCH 1/5] drm/msm: Fix improper uses of smp_mb__{before,after}_atomic()

[PATCH 1/5] drm/msm: Fix improper uses of smp_mb__{before,after}_atomic()

[Andrea Parri]

These barriers only apply to the read-modify-write operations; in
particular, they do not apply to the atomic_set() primitive.

Replace the barriers with smp_mb()s.

Fixes: b1fc2839d2f92 ("drm/msm: Implement preemption for A5XX targets")
Cc: stable@vger.kernel.org
Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Cc: Rob Clark <robdclark@gmail.com>
Cc: Sean Paul <sean@poorly.run>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Jordan Crouse <jcrouse@codeaurora.org>
Cc: linux-arm-msm@vger.kernel.org
Cc: dri-devel@lists.freedesktop.org
Cc: freedreno@lists.freedesktop.org
---
 drivers/gpu/drm/msm/adreno/a5xx_preempt.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c
index 3d62310a535fb..ee0820ee0c664 100644
--- a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c
@@ -39,10 +39,10 @@ static inline void set_preempt_state(struct a5xx_gpu *gpu,
 	 * preemption or in the interrupt handler so barriers are needed
 	 * before...
 	 */
-	smp_mb__before_atomic();
+	smp_mb();
 	atomic_set(&gpu->preempt_state, new);
 	/* ... and after*/
-	smp_mb__after_atomic();
+	smp_mb();
 }
 
 /* Write the most recent wptr for the given ring into the hardware */
-- 
2.7.4

[PATCH 2/5] bio: fix improper use of smp_mb__before_atomic()

[Andrea Parri]

This barrier only applies to the read-modify-write operations; in
particular, it does not apply to the atomic_set() primitive.

Replace the barrier with an smp_mb().

Fixes: dac56212e8127 ("bio: skip atomic inc/dec of ->bi_cnt for most use cases")
Cc: stable@vger.kernel.org
Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: linux-block@vger.kernel.org
---
 include/linux/bio.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/bio.h b/include/linux/bio.h
index e584673c18814..5becbafb84e8a 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -224,7 +224,7 @@ static inline void bio_cnt_set(struct bio *bio, unsigned int count)
 {
 	if (count != 1) {
 		bio->bi_flags |= (1 << BIO_REFFED);
-		smp_mb__before_atomic();
+		smp_mb();
 	}
 	atomic_set(&bio->__bi_cnt, count);
 }
-- 
2.7.4

Re: [PATCH 2/5] bio: fix improper use of smp_mb__before_atomic()

[Peter Zijlstra]

On Mon, Apr 29, 2019 at 10:14:58PM +0200, Andrea Parri wrote:
> This barrier only applies to the read-modify-write operations; in
> particular, it does not apply to the atomic_set() primitive.
> 
> Replace the barrier with an smp_mb().

> @@ -224,7 +224,7 @@ static inline void bio_cnt_set(struct bio *bio, unsigned int count)
>  {
>  	if (count != 1) {
>  		bio->bi_flags |= (1 << BIO_REFFED);
> -		smp_mb__before_atomic();

Maybe also add:

		/*
		 * XXX the comment that explain this barrier goes here
		 */
> +		smp_mb();
>  	}
>  	atomic_set(&bio->__bi_cnt, count);
>  }

Re: [PATCH 2/5] bio: fix improper use of smp_mb__before_atomic()

[Ming Lei]

On Mon, Apr 29, 2019 at 10:14:58PM +0200, Andrea Parri wrote:
> This barrier only applies to the read-modify-write operations; in
> particular, it does not apply to the atomic_set() primitive.
> 
> Replace the barrier with an smp_mb().
> 
> Fixes: dac56212e8127 ("bio: skip atomic inc/dec of ->bi_cnt for most use cases")
> Cc: stable@vger.kernel.org
> Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
> Reported-by: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
> Cc: Jens Axboe <axboe@kernel.dk>
> Cc: linux-block@vger.kernel.org
> ---
>  include/linux/bio.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/bio.h b/include/linux/bio.h
> index e584673c18814..5becbafb84e8a 100644
> --- a/include/linux/bio.h
> +++ b/include/linux/bio.h
> @@ -224,7 +224,7 @@ static inline void bio_cnt_set(struct bio *bio, unsigned int count)
>  {
>  	if (count != 1) {
>  		bio->bi_flags |= (1 << BIO_REFFED);
> -		smp_mb__before_atomic();
> +		smp_mb();
>  	}
>  	atomic_set(&bio->__bi_cnt, count);
>  }
> -- 
> 2.7.4
> 

Looks fine,

Reviewed-by: Ming Lei <ming.lei@redhat.com>

Thanks,
Ming

[PATCH 3/5] sbitmap: fix improper use of smp_mb__before_atomic()

[Andrea Parri]

This barrier only applies to the read-modify-write operations; in
particular, it does not apply to the atomic_set() primitive.

Replace the barrier with an smp_mb().

Fixes: 6c0ca7ae292ad ("sbitmap: fix wakeup hang after sbq resize")
Cc: stable@vger.kernel.org
Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Omar Sandoval <osandov@fb.com>
Cc: linux-block@vger.kernel.org
---
 lib/sbitmap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/sbitmap.c b/lib/sbitmap.c
index 155fe38756ecf..4a7fc4915dfc6 100644
--- a/lib/sbitmap.c
+++ b/lib/sbitmap.c
@@ -435,7 +435,7 @@ static void sbitmap_queue_update_wake_batch(struct sbitmap_queue *sbq,
 		 * to ensure that the batch size is updated before the wait
 		 * counts.
 		 */
-		smp_mb__before_atomic();
+		smp_mb();
 		for (i = 0; i < SBQ_WAIT_QUEUES; i++)
 			atomic_set(&sbq->ws[i].wait_cnt, 1);
 	}
-- 
2.7.4

Re: [PATCH 3/5] sbitmap: fix improper use of smp_mb__before_atomic()

[Ming Lei]

On Mon, Apr 29, 2019 at 10:14:59PM +0200, Andrea Parri wrote:
> This barrier only applies to the read-modify-write operations; in
> particular, it does not apply to the atomic_set() primitive.
> 
> Replace the barrier with an smp_mb().
> 
> Fixes: 6c0ca7ae292ad ("sbitmap: fix wakeup hang after sbq resize")
> Cc: stable@vger.kernel.org
> Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
> Reported-by: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
> Cc: Jens Axboe <axboe@kernel.dk>
> Cc: Omar Sandoval <osandov@fb.com>
> Cc: linux-block@vger.kernel.org
> ---
>  lib/sbitmap.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/sbitmap.c b/lib/sbitmap.c
> index 155fe38756ecf..4a7fc4915dfc6 100644
> --- a/lib/sbitmap.c
> +++ b/lib/sbitmap.c
> @@ -435,7 +435,7 @@ static void sbitmap_queue_update_wake_batch(struct sbitmap_queue *sbq,
>  		 * to ensure that the batch size is updated before the wait
>  		 * counts.
>  		 */
> -		smp_mb__before_atomic();
> +		smp_mb();
>  		for (i = 0; i < SBQ_WAIT_QUEUES; i++)
>  			atomic_set(&sbq->ws[i].wait_cnt, 1);
>  	}
> -- 
> 2.7.4
> 

sbitmap_queue_update_wake_batch() won't be called in fast path, and
the fix is correct too, so:

Reviewed-by: Ming Lei <ming.lei@redhat.com>

thanks,
Ming

[PATCH 4/5] ceph: fix improper use of smp_mb__before_atomic()

[Andrea Parri]

This barrier only applies to the read-modify-write operations; in
particular, it does not apply to the atomic64_set() primitive.

Replace the barrier with an smp_mb().

Fixes: fdd4e15838e59 ("ceph: rework dcache readdir")
Cc: stable@vger.kernel.org
Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Cc: "Yan, Zheng" <zyan@redhat.com>
Cc: Sage Weil <sage@redhat.com>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: ceph-devel@vger.kernel.org
---
 fs/ceph/super.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 16c03188578ea..b5c782e6d62f1 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -541,7 +541,7 @@ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci,
 					   long long release_count,
 					   long long ordered_count)
 {
-	smp_mb__before_atomic();
+	smp_mb();
 	atomic64_set(&ci->i_complete_seq[0], release_count);
 	atomic64_set(&ci->i_complete_seq[1], ordered_count);
 }
-- 
2.7.4

Re: [PATCH 4/5] ceph: fix improper use of smp_mb__before_atomic()

[Peter Zijlstra]

On Mon, Apr 29, 2019 at 10:15:00PM +0200, Andrea Parri wrote:
> This barrier only applies to the read-modify-write operations; in
> particular, it does not apply to the atomic64_set() primitive.
> 
> Replace the barrier with an smp_mb().
> 

> @@ -541,7 +541,7 @@ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci,
>  					   long long release_count,
>  					   long long ordered_count)
>  {
> -	smp_mb__before_atomic();

same
	/*
	 * XXX: the comment that explain this barrier goes here.
	 */

> +	smp_mb();

>  	atomic64_set(&ci->i_complete_seq[0], release_count);
>  	atomic64_set(&ci->i_complete_seq[1], ordered_count);
>  }
> -- 
> 2.7.4
> 

Re: [PATCH 4/5] ceph: fix improper use of smp_mb__before_atomic()

[Yan, Zheng]

On Tue, Apr 30, 2019 at 4:26 PM Peter Zijlstra <peterz@infradead.org> wrote:
>
> On Mon, Apr 29, 2019 at 10:15:00PM +0200, Andrea Parri wrote:
> > This barrier only applies to the read-modify-write operations; in
> > particular, it does not apply to the atomic64_set() primitive.
> >
> > Replace the barrier with an smp_mb().
> >
>
> > @@ -541,7 +541,7 @@ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci,
> >                                          long long release_count,
> >                                          long long ordered_count)
> >  {
> > -     smp_mb__before_atomic();
>
> same
>         /*
>          * XXX: the comment that explain this barrier goes here.
>          */
>

makes sure operations that setup readdir cache (update page cache and
i_size) are strongly ordered with following atomic64_set.

> > +     smp_mb();
>
> >       atomic64_set(&ci->i_complete_seq[0], release_count);
> >       atomic64_set(&ci->i_complete_seq[1], ordered_count);
> >  }
> > --
> > 2.7.4
> >

[PATCH 5/5] IB/hfi1: Fix improper uses of smp_mb__before_atomic()

[Andrea Parri]

This barrier only applies to the read-modify-write operations; in
particular, it does not apply to the atomic_read() primitive.

Replace the barrier with an smp_mb().

Fixes: 856cc4c237add ("IB/hfi1: Add the capability for reserved operations")
Cc: stable@vger.kernel.org
Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Cc: Dennis Dalessandro <dennis.dalessandro@intel.com>
Cc: Mike Marciniszyn <mike.marciniszyn@intel.com>
Cc: Doug Ledford <dledford@redhat.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: linux-rdma@vger.kernel.org
---
 drivers/infiniband/sw/rdmavt/qp.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/sw/rdmavt/qp.c b/drivers/infiniband/sw/rdmavt/qp.c
index a34b9a2a32b60..b64fd151d31fb 100644
--- a/drivers/infiniband/sw/rdmavt/qp.c
+++ b/drivers/infiniband/sw/rdmavt/qp.c
@@ -1863,11 +1863,11 @@ static inline int rvt_qp_is_avail(
 	u32 reserved_used;
 
 	/* see rvt_qp_wqe_unreserve() */
-	smp_mb__before_atomic();
+	smp_mb();
 	reserved_used = atomic_read(&qp->s_reserved_used);
 	if (unlikely(reserved_op)) {
 		/* see rvt_qp_wqe_unreserve() */
-		smp_mb__before_atomic();
+		smp_mb();
 		if (reserved_used >= rdi->dparms.reserved_operations)
 			return -ENOMEM;
 		return 0;
@@ -1882,7 +1882,7 @@ static inline int rvt_qp_is_avail(
 		avail = slast - qp->s_head;
 
 	/* see rvt_qp_wqe_unreserve() */
-	smp_mb__before_atomic();
+	smp_mb();
 	reserved_used = atomic_read(&qp->s_reserved_used);
 	avail =  avail - 1 -
 		(rdi->dparms.reserved_operations - reserved_used);
-- 
2.7.4

RE: [PATCH 5/5] IB/hfi1: Fix improper uses of smp_mb__before_atomic()

[Ruhl, Michael J]

>-----Original Message-----
>From: linux-rdma-owner@vger.kernel.org [mailto:linux-rdma-
>owner@vger.kernel.org] On Behalf Of Andrea Parri
>Sent: Monday, April 29, 2019 4:15 PM
>To: linux-kernel@vger.kernel.org
>Cc: Andrea Parri <andrea.parri@amarulasolutions.com>;
>stable@vger.kernel.org; Dalessandro, Dennis
><dennis.dalessandro@intel.com>; Marciniszyn, Mike
><mike.marciniszyn@intel.com>; Doug Ledford <dledford@redhat.com>;
>Jason Gunthorpe <jgg@ziepe.ca>; linux-rdma@vger.kernel.org
>Subject: [PATCH 5/5] IB/hfi1: Fix improper uses of smp_mb__before_atomic()
>
>This barrier only applies to the read-modify-write operations; in
>particular, it does not apply to the atomic_read() primitive.
>
>Replace the barrier with an smp_mb().

This is one of a couple of barrier issues that we are currently looking into.

See:

[PATCH for-next 6/9] IB/rdmavt: Add new completion inline

We will take a look at this one as well.

Thanks,

Mike

>Fixes: 856cc4c237add ("IB/hfi1: Add the capability for reserved operations")
>Cc: stable@vger.kernel.org
>Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
>Reported-by: Peter Zijlstra <peterz@infradead.org>
>Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
>Cc: Dennis Dalessandro <dennis.dalessandro@intel.com>
>Cc: Mike Marciniszyn <mike.marciniszyn@intel.com>
>Cc: Doug Ledford <dledford@redhat.com>
>Cc: Jason Gunthorpe <jgg@ziepe.ca>
>Cc: linux-rdma@vger.kernel.org
>---
> drivers/infiniband/sw/rdmavt/qp.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
>diff --git a/drivers/infiniband/sw/rdmavt/qp.c
>b/drivers/infiniband/sw/rdmavt/qp.c
>index a34b9a2a32b60..b64fd151d31fb 100644
>--- a/drivers/infiniband/sw/rdmavt/qp.c
>+++ b/drivers/infiniband/sw/rdmavt/qp.c
>@@ -1863,11 +1863,11 @@ static inline int rvt_qp_is_avail(
> 	u32 reserved_used;
>
> 	/* see rvt_qp_wqe_unreserve() */
>-	smp_mb__before_atomic();
>+	smp_mb();
> 	reserved_used = atomic_read(&qp->s_reserved_used);
> 	if (unlikely(reserved_op)) {
> 		/* see rvt_qp_wqe_unreserve() */
>-		smp_mb__before_atomic();
>+		smp_mb();
> 		if (reserved_used >= rdi->dparms.reserved_operations)
> 			return -ENOMEM;
> 		return 0;
>@@ -1882,7 +1882,7 @@ static inline int rvt_qp_is_avail(
> 		avail = slast - qp->s_head;
>
> 	/* see rvt_qp_wqe_unreserve() */
>-	smp_mb__before_atomic();
>+	smp_mb();
> 	reserved_used = atomic_read(&qp->s_reserved_used);
> 	avail =  avail - 1 -
> 		(rdi->dparms.reserved_operations - reserved_used);
>--
>2.7.4

Re: [PATCH 5/5] IB/hfi1: Fix improper uses of smp_mb__before_atomic()

[Peter Zijlstra]

On Mon, Apr 29, 2019 at 10:15:01PM +0200, Andrea Parri wrote:
> This barrier only applies to the read-modify-write operations; in
> particular, it does not apply to the atomic_read() primitive.
> 
> Replace the barrier with an smp_mb().
> 
> Fixes: 856cc4c237add ("IB/hfi1: Add the capability for reserved operations")
> Cc: stable@vger.kernel.org
> Reported-by: "Paul E. McKenney" <paulmck@linux.ibm.com>
> Reported-by: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
> Cc: Dennis Dalessandro <dennis.dalessandro@intel.com>
> Cc: Mike Marciniszyn <mike.marciniszyn@intel.com>
> Cc: Doug Ledford <dledford@redhat.com>
> Cc: Jason Gunthorpe <jgg@ziepe.ca>
> Cc: linux-rdma@vger.kernel.org
> ---
>  drivers/infiniband/sw/rdmavt/qp.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/infiniband/sw/rdmavt/qp.c b/drivers/infiniband/sw/rdmavt/qp.c
> index a34b9a2a32b60..b64fd151d31fb 100644
> --- a/drivers/infiniband/sw/rdmavt/qp.c
> +++ b/drivers/infiniband/sw/rdmavt/qp.c
> @@ -1863,11 +1863,11 @@ static inline int rvt_qp_is_avail(
>  	u32 reserved_used;
>  
>  	/* see rvt_qp_wqe_unreserve() */

I see a completely bogus comment in rvf_op_wqe_unreserve(), referring to
bogus comments makes this barrier bogus too.

> -	smp_mb__before_atomic();
> +	smp_mb();
>  	reserved_used = atomic_read(&qp->s_reserved_used);
>  	if (unlikely(reserved_op)) {
>  		/* see rvt_qp_wqe_unreserve() */
> -		smp_mb__before_atomic();

This was before, but there is nothing _after_ this. Which means this
barrier was complete garbage anyway.

> +		smp_mb();
>  		if (reserved_used >= rdi->dparms.reserved_operations)
>  			return -ENOMEM;
>  		return 0;
> @@ -1882,7 +1882,7 @@ static inline int rvt_qp_is_avail(
>  		avail = slast - qp->s_head;
>  
>  	/* see rvt_qp_wqe_unreserve() */
> -	smp_mb__before_atomic();
> +	smp_mb();

Same as the first.

>  	reserved_used = atomic_read(&qp->s_reserved_used);
>  	avail =  avail - 1 -
>  		(rdi->dparms.reserved_operations - reserved_used);