[PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

[PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

[Mike Marciniszyn]

From: Mike Marciniszyn <mike.marciniszyn-h88ZbnxC6KDQT0dZR+AlfA@public.gmane.org>

0b088e00 ("RDS: Use page_remainder_alloc() for recv bufs")
added uses of sg_dma_len() and sg_dma_address().

IB ulps should use ib_sg_dma_len() and ib_sg_dma_address
respectively since some HCAs overload ib_sg_dma* operations.

Signed-off-by: Mike Marciniszyn <mike.marciniszyn-h88ZbnxC6KDQT0dZR+AlfA@public.gmane.org>
---
 net/rds/ib_recv.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c
index 8d19491..cd44ed3 100644
--- a/net/rds/ib_recv.c
+++ b/net/rds/ib_recv.c
@@ -339,8 +339,8 @@ static int rds_ib_recv_refill_one(struct rds_connection *conn,
 	sge->length = sizeof(struct rds_header);
 
 	sge = &recv->r_sge[1];
-	sge->addr = sg_dma_address(&recv->r_frag->f_sg);
-	sge->length = sg_dma_len(&recv->r_frag->f_sg);
+	sge->addr = ib_sg_dma_address(ic->i_cm_id->device, &recv->r_frag->f_sg);
+	sge->length = ib_sg_dma_len(ic->i_cm_id->device, &recv->r_frag->f_sg);
 
 	ret = 0;
 out:
@@ -381,7 +381,10 @@ void rds_ib_recv_refill(struct rds_connection *conn, int prefill)
 		ret = ib_post_recv(ic->i_cm_id->qp, &recv->r_wr, &failed_wr);
 		rdsdebug("recv %p ibinc %p page %p addr %lu ret %d\n", recv,
 			 recv->r_ibinc, sg_page(&recv->r_frag->f_sg),
-			 (long) sg_dma_address(&recv->r_frag->f_sg), ret);
+			 (long) ib_sg_dma_address(
+				ic->i_cm_id->device,
+				&recv->r_frag->f_sg),
+			ret);
 		if (ret) {
 			rds_ib_conn_error(conn, "recv post on "
 			       "%pI4 returned %d, disconnecting and "


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

[Venkat Venkatsubra]

On 4/19/2012 11:15 AM, Mike Marciniszyn wrote:
> From: Mike Marciniszyn<mike.marciniszyn-h88ZbnxC6KDQT0dZR+AlfA@public.gmane.org>
>
> 0b088e00 ("RDS: Use page_remainder_alloc() for recv bufs")
> added uses of sg_dma_len() and sg_dma_address().
>
> IB ulps should use ib_sg_dma_len() and ib_sg_dma_address
> respectively since some HCAs overload ib_sg_dma* operations.
>
> Signed-off-by: Mike Marciniszyn<mike.marciniszyn-h88ZbnxC6KDQT0dZR+AlfA@public.gmane.org>
>
Acked-by: Venkat Venkatsubra <venkat.x.venkatsubra-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>

Venkat
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

FW: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

[Marciniszyn, Mike]

I haven’t seen this patch get upstream.

Is there something I should do to help?

Mike

-----Original Message-----
From: linux-rdma-owner@vger.kernel.org [mailto:linux-rdma-owner@vger.kernel.org] On Behalf Of Mike Marciniszyn
Sent: Thursday, April 19, 2012 12:15 PM
To: venkat.x.venkatsubra@oracle.com
Cc: linux-rdma@vger.kernel.org; rds-devel@oss.oracle.com
Subject: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

From: Mike Marciniszyn <mike.marciniszyn@qlogic.com>

0b088e00 ("RDS: Use page_remainder_alloc() for recv bufs") added uses of sg_dma_len() and sg_dma_address().

IB ulps should use ib_sg_dma_len() and ib_sg_dma_address respectively since some HCAs overload ib_sg_dma* operations.

Signed-off-by: Mike Marciniszyn <mike.marciniszyn@qlogic.com>
---
 net/rds/ib_recv.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c index 8d19491..cd44ed3 100644
--- a/net/rds/ib_recv.c
+++ b/net/rds/ib_recv.c
@@ -339,8 +339,8 @@ static int rds_ib_recv_refill_one(struct rds_connection *conn,
 	sge->length = sizeof(struct rds_header);
 
 	sge = &recv->r_sge[1];
-	sge->addr = sg_dma_address(&recv->r_frag->f_sg);
-	sge->length = sg_dma_len(&recv->r_frag->f_sg);
+	sge->addr = ib_sg_dma_address(ic->i_cm_id->device, &recv->r_frag->f_sg);
+	sge->length = ib_sg_dma_len(ic->i_cm_id->device, &recv->r_frag->f_sg);
 
 	ret = 0;
 out:
@@ -381,7 +381,10 @@ void rds_ib_recv_refill(struct rds_connection *conn, int prefill)
 		ret = ib_post_recv(ic->i_cm_id->qp, &recv->r_wr, &failed_wr);
 		rdsdebug("recv %p ibinc %p page %p addr %lu ret %d\n", recv,
 			 recv->r_ibinc, sg_page(&recv->r_frag->f_sg),
-			 (long) sg_dma_address(&recv->r_frag->f_sg), ret);
+			 (long) ib_sg_dma_address(
+				ic->i_cm_id->device,
+				&recv->r_frag->f_sg),
+			ret);
 		if (ret) {
 			rds_ib_conn_error(conn, "recv post on "
 			       "%pI4 returned %d, disconnecting and "


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: FW: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

[Venkat Venkatsubra]

On 6/22/2012 1:24 PM, Marciniszyn, Mike wrote:
> I haven’t seen this patch get upstream.
>
> Is there something I should do to help?
>
> Mike
>
> -----Original Message-----
> From: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org [mailto:linux-rdma-owner@vger.kernel.org] On Behalf Of Mike Marciniszyn
> Sent: Thursday, April 19, 2012 12:15 PM
> To: venkat.x.venkatsubra-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org
> Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; rds-devel-N0ozoZBvEnrZJqsBc5GL+g@public.gmane.org
> Subject: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len
>
> From: Mike Marciniszyn<mike.marciniszyn-h88ZbnxC6KDQT0dZR+AlfA@public.gmane.org>
>
> 0b088e00 ("RDS: Use page_remainder_alloc() for recv bufs") added uses of sg_dma_len() and sg_dma_address().
>
> IB ulps should use ib_sg_dma_len() and ib_sg_dma_address respectively since some HCAs overload ib_sg_dma* operations.
>
> Signed-off-by: Mike Marciniszyn<mike.marciniszyn-h88ZbnxC6KDQT0dZR+AlfA@public.gmane.org>
> ---
>   net/rds/ib_recv.c |    9 ++++++---
>   1 files changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c index 8d19491..cd44ed3 100644
> --- a/net/rds/ib_recv.c
> +++ b/net/rds/ib_recv.c
> @@ -339,8 +339,8 @@ static int rds_ib_recv_refill_one(struct rds_connection *conn,
>   	sge->length = sizeof(struct rds_header);
>
>   	sge =&recv->r_sge[1];
> -	sge->addr = sg_dma_address(&recv->r_frag->f_sg);
> -	sge->length = sg_dma_len(&recv->r_frag->f_sg);
> +	sge->addr = ib_sg_dma_address(ic->i_cm_id->device,&recv->r_frag->f_sg);
> +	sge->length = ib_sg_dma_len(ic->i_cm_id->device,&recv->r_frag->f_sg);
>
>   	ret = 0;
>   out:
> @@ -381,7 +381,10 @@ void rds_ib_recv_refill(struct rds_connection *conn, int prefill)
>   		ret = ib_post_recv(ic->i_cm_id->qp,&recv->r_wr,&failed_wr);
>   		rdsdebug("recv %p ibinc %p page %p addr %lu ret %d\n", recv,
>   			 recv->r_ibinc, sg_page(&recv->r_frag->f_sg),
> -			 (long) sg_dma_address(&recv->r_frag->f_sg), ret);
> +			 (long) ib_sg_dma_address(
> +				ic->i_cm_id->device,
> +				&recv->r_frag->f_sg),
> +			ret);
>   		if (ret) {
>   			rds_ib_conn_error(conn, "recv post on "
>   			       "%pI4 returned %d, disconnecting and "
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>
I had Acked it according to my April 19th email. ;-)

Venkat
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: FW: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len

[Marciniszyn, Mike]

What is the process for getting patches into RDS.

The original messages and ack's are shown in http://oss.oracle.com/pipermail/rds-devel/2012-April/thread.html.

I can repost the patches if necessary.

Mike 

> -----Original Message-----
> From: Venkat Venkatsubra [mailto:venkat.x.venkatsubra@oracle.com]
> Sent: Friday, June 22, 2012 3:35 PM
> To: Marciniszyn, Mike
> Cc: linux-rdma@vger.kernel.org; Mascarenhas, Edward
> Subject: Re: FW: [PATCH] RDS: correct ib api use with
> sg_dma_address/sg_dma_len
> 
> On 6/22/2012 1:24 PM, Marciniszyn, Mike wrote:
> > I haven’t seen this patch get upstream.
> >
> > Is there something I should do to help?
> >
> > Mike
> >
> > -----Original Message-----
> > From: linux-rdma-owner@vger.kernel.org [mailto:linux-rdma-
> owner@vger.kernel.org] On Behalf Of Mike Marciniszyn
> > Sent: Thursday, April 19, 2012 12:15 PM
> > To: venkat.x.venkatsubra@oracle.com
> > Cc: linux-rdma@vger.kernel.org; rds-devel@oss.oracle.com
> > Subject: [PATCH] RDS: correct ib api use with sg_dma_address/sg_dma_len
> >
> > From: Mike Marciniszyn<mike.marciniszyn@qlogic.com>
> >
> > 0b088e00 ("RDS: Use page_remainder_alloc() for recv bufs") added uses of
> sg_dma_len() and sg_dma_address().
> >
> > IB ulps should use ib_sg_dma_len() and ib_sg_dma_address respectively
> since some HCAs overload ib_sg_dma* operations.
> >
> > Signed-off-by: Mike Marciniszyn<mike.marciniszyn@qlogic.com>
> > ---
> >   net/rds/ib_recv.c |    9 ++++++---
> >   1 files changed, 6 insertions(+), 3 deletions(-)
> >
> > diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c index 8d19491..cd44ed3
> 100644
> > --- a/net/rds/ib_recv.c
> > +++ b/net/rds/ib_recv.c
> > @@ -339,8 +339,8 @@ static int rds_ib_recv_refill_one(struct
> rds_connection *conn,
> >   	sge->length = sizeof(struct rds_header);
> >
> >   	sge =&recv->r_sge[1];
> > -	sge->addr = sg_dma_address(&recv->r_frag->f_sg);
> > -	sge->length = sg_dma_len(&recv->r_frag->f_sg);
> > +	sge->addr = ib_sg_dma_address(ic->i_cm_id->device,&recv->r_frag-
> >f_sg);
> > +	sge->length = ib_sg_dma_len(ic->i_cm_id->device,&recv->r_frag-
> >f_sg);
> >
> >   	ret = 0;
> >   out:
> > @@ -381,7 +381,10 @@ void rds_ib_recv_refill(struct rds_connection
> *conn, int prefill)
> >   		ret = ib_post_recv(ic->i_cm_id->qp,&recv->r_wr,&failed_wr);
> >   		rdsdebug("recv %p ibinc %p page %p addr %lu ret %d\n",
> recv,
> >   			 recv->r_ibinc, sg_page(&recv->r_frag->f_sg),
> > -			 (long) sg_dma_address(&recv->r_frag->f_sg), ret);
> > +			 (long) ib_sg_dma_address(
> > +				ic->i_cm_id->device,
> > +				&recv->r_frag->f_sg),
> > +			ret);
> >   		if (ret) {
> >   			rds_ib_conn_error(conn, "recv post on "
> >   			       "%pI4 returned %d, disconnecting and "
> >
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the
> body of a message to majordomo@vger.kernel.org More majordomo info at
> http://vger.kernel.org/majordomo-info.html
> >
> >
> I had Acked it according to my April 19th email. ;-)
> 
> Venkat