[PATCH] IB/srpt: Support HCAs with more than two ports

[PATCH] IB/srpt: Support HCAs with more than two ports

[Bart Van Assche]

Since there are adapters that have four ports, increase the size of
the srpt_device.port[] array. This patch avoids that the following
warning is hit with quad port Chelsio adapters:

    WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));

Reported-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Steve Wise <swise@opengridcomputing.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: <stable@vger.kernel.org>
---
 drivers/infiniband/ulp/srpt/ib_srpt.c | 6 +++---
 drivers/infiniband/ulp/srpt/ib_srpt.h | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 3081c629a7f7..50f4f9806ac6 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -2969,10 +2969,12 @@ static void srpt_add_one(struct ib_device *device)
 
 	pr_debug("device = %p\n", device);
 
-	sdev = kzalloc(sizeof(*sdev), GFP_KERNEL);
+	sdev = kzalloc(sizeof(*sdev) + device->phys_port_cnt *
+		       sizeof(*sdev->port), GFP_KERNEL);
 	if (!sdev)
 		goto err;
 
+	sdev->port = (void *)(sdev + 1);
 	sdev->device = device;
 	mutex_init(&sdev->sdev_mutex);
 
@@ -3023,8 +3025,6 @@ static void srpt_add_one(struct ib_device *device)
 			      srpt_event_handler);
 	ib_register_event_handler(&sdev->event_handler);
 
-	WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));
-
 	for (i = 1; i <= sdev->device->phys_port_cnt; i++) {
 		sport = &sdev->port[i - 1];
 		INIT_LIST_HEAD(&sport->nexus_list);
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.h b/drivers/infiniband/ulp/srpt/ib_srpt.h
index 2361483476a0..b20ad6dbf966 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.h
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.h
@@ -410,7 +410,7 @@ struct srpt_device {
 	struct mutex		sdev_mutex;
 	bool			use_srq;
 	struct srpt_recv_ioctx	**ioctx_ring;
-	struct srpt_port	port[2];
+	struct srpt_port	*port;
 	struct ib_event_handler	event_handler;
 	struct list_head	list;
 };
-- 
2.17.1

RE: [PATCH] IB/srpt: Support HCAs with more than two ports

[Steve Wise]

> Subject: [PATCH] IB/srpt: Support HCAs with more than two ports
> 
> Since there are adapters that have four ports, increase the size of
> the srpt_device.port[] array. This patch avoids that the following
> warning is hit with quad port Chelsio adapters:
> 
>     WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));
> 
> Reported-by: Steve Wise <swise@opengridcomputing.com>
> Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
> Cc: Steve Wise <swise@opengridcomputing.com>
> Cc: Christoph Hellwig <hch@infradead.org>
> Cc: <stable@vger.kernel.org>

Looks good.

Reviewed-by: Steve Wise <swise@opengridcomputing.com>

Re: [PATCH] IB/srpt: Support HCAs with more than two ports

[Jason Gunthorpe]

On Tue, Jun 26, 2018 at 03:24:48PM -0700, Bart Van Assche wrote:
> Since there are adapters that have four ports, increase the size of
> the srpt_device.port[] array. This patch avoids that the following
> warning is hit with quad port Chelsio adapters:
> 
>     WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));
> 
> Reported-by: Steve Wise <swise@opengridcomputing.com>
> Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
> Cc: Steve Wise <swise@opengridcomputing.com>
> Cc: Christoph Hellwig <hch@infradead.org>
> Cc: <stable@vger.kernel.org>
> ---
>  drivers/infiniband/ulp/srpt/ib_srpt.c | 6 +++---
>  drivers/infiniband/ulp/srpt/ib_srpt.h | 2 +-
>  2 files changed, 4 insertions(+), 4 deletions(-)

Can we write it this way instead? More typesafe.

diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 1b0b285a0ae021..36d9fab7c99800 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -2970,7 +2970,8 @@ static void srpt_add_one(struct ib_device *device)
 
 	pr_debug("device = %p\n", device);
 
-	sdev = kzalloc(sizeof(*sdev), GFP_KERNEL);
+	sdev = kzalloc(struct_size(sdev, port, device->phys_port_cnt),
+		       GFP_KERNEL);
 	if (!sdev)
 		goto err;
 
@@ -3024,8 +3025,6 @@ static void srpt_add_one(struct ib_device *device)
 			      srpt_event_handler);
 	ib_register_event_handler(&sdev->event_handler);
 
-	WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));
-
 	for (i = 1; i <= sdev->device->phys_port_cnt; i++) {
 		sport = &sdev->port[i - 1];
 		INIT_LIST_HEAD(&sport->nexus_list);
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.h b/drivers/infiniband/ulp/srpt/ib_srpt.h
index 2361483476a025..444dfd7281b5fe 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.h
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.h
@@ -396,9 +396,9 @@ struct srpt_port {
  * @sdev_mutex:	   Serializes use_srq changes.
  * @use_srq:       Whether or not to use SRQ.
  * @ioctx_ring:    Per-HCA SRQ.
- * @port:          Information about the ports owned by this HCA.
  * @event_handler: Per-HCA asynchronous IB event handler.
  * @list:          Node in srpt_dev_list.
+ * @port:          Information about the ports owned by this HCA.
  */
 struct srpt_device {
 	struct ib_device	*device;
@@ -410,9 +410,9 @@ struct srpt_device {
 	struct mutex		sdev_mutex;
 	bool			use_srq;
 	struct srpt_recv_ioctx	**ioctx_ring;
-	struct srpt_port	port[2];
 	struct ib_event_handler	event_handler;
 	struct list_head	list;
+	struct srpt_port        port[];
 };
 
 #endif				/* IB_SRPT_H */

Re: [PATCH] IB/srpt: Support HCAs with more than two ports

[Bart Van Assche]

On 06/29/18 09:21, Jason Gunthorpe wrote:
> On Tue, Jun 26, 2018 at 03:24:48PM -0700, Bart Van Assche wrote:
>> Since there are adapters that have four ports, increase the size of
>> the srpt_device.port[] array. This patch avoids that the following
>> warning is hit with quad port Chelsio adapters:
>>
>>      WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));
>>
>> Reported-by: Steve Wise <swise@opengridcomputing.com>
>> Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
>> Cc: Steve Wise <swise@opengridcomputing.com>
>> Cc: Christoph Hellwig <hch@infradead.org>
>> Cc: <stable@vger.kernel.org>
>> ---
>>   drivers/infiniband/ulp/srpt/ib_srpt.c | 6 +++---
>>   drivers/infiniband/ulp/srpt/ib_srpt.h | 2 +-
>>   2 files changed, 4 insertions(+), 4 deletions(-)
> 
> Can we write it this way instead? More typesafe.

Hello Jason,

The patch you posted looks fine to me. Do you want me to add your 
Signed-off-by after I have retested this patch when I repost it?

Thanks,

Bart.

Re: [PATCH] IB/srpt: Support HCAs with more than two ports

[Jason Gunthorpe]

On Fri, Jun 29, 2018 at 09:32:42AM -0700, Bart Van Assche wrote:
> On 06/29/18 09:21, Jason Gunthorpe wrote:
> >On Tue, Jun 26, 2018 at 03:24:48PM -0700, Bart Van Assche wrote:
> >>Since there are adapters that have four ports, increase the size of
> >>the srpt_device.port[] array. This patch avoids that the following
> >>warning is hit with quad port Chelsio adapters:
> >>
> >>     WARN_ON(sdev->device->phys_port_cnt > ARRAY_SIZE(sdev->port));
> >>
> >>Reported-by: Steve Wise <swise@opengridcomputing.com>
> >>Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
> >>Cc: Steve Wise <swise@opengridcomputing.com>
> >>Cc: Christoph Hellwig <hch@infradead.org>
> >>Cc: <stable@vger.kernel.org>
> >>  drivers/infiniband/ulp/srpt/ib_srpt.c | 6 +++---
> >>  drivers/infiniband/ulp/srpt/ib_srpt.h | 2 +-
> >>  2 files changed, 4 insertions(+), 4 deletions(-)
> >
> >Can we write it this way instead? More typesafe.
> 
> Hello Jason,
> 
> The patch you posted looks fine to me. Do you want me to add your
> Signed-off-by after I have retested this patch when I repost it?

If you can retest it today just send me a note and I'll use it as
posted, otherwise yes please.

Jason

Re: [PATCH] IB/srpt: Support HCAs with more than two ports

[Bart Van Assche]

On 06/29/18 10:20, Jason Gunthorpe wrote:
> If you can retest it today just send me a note and I'll use it as
> posted, otherwise yes please.

Hello Jason,

The patch you posted works fine on my test setup.

Thanks,

Bart.