[PATCH 1/2] be2net: replenish when posting to rx-queue is starved in out of mem conditions

[PATCH 1/2] be2net: replenish when posting to rx-queue is starved in out of mem conditions

[Sathya Perla]

Hi, this is a patch to replenish the rx-queue when it is in a starved
state (due to out-of-mem conditions)

thanks,
-Sathya
P.S: Pls ignore the "company confidential" warning at the end of the
patch emails.


Signed-off-by: Sathya Perla <sathyap@serverengines.com>
---
 drivers/net/benet/be.h      |    1 +
 drivers/net/benet/be_main.c |   50 +++++++++++++++++++++++++------------------
 2 files changed, 30 insertions(+), 21 deletions(-)

diff --git a/drivers/net/benet/be.h b/drivers/net/benet/be.h
index 63d593d..f327be5 100644
--- a/drivers/net/benet/be.h
+++ b/drivers/net/benet/be.h
@@ -194,6 +194,7 @@ struct be_adapter {
 	struct be_eq_obj rx_eq;
 	struct be_rx_obj rx_obj;
 	u32 big_page_size;	/* Compounded page size shared by rx wrbs */
+	bool rx_post_starved;	/* Zero rx frags have been posted to BE */
 
 	struct vlan_group *vlan_grp;
 	u16 num_vlans;
diff --git a/drivers/net/benet/be_main.c b/drivers/net/benet/be_main.c
index 897a63d..80fe1e0 100644
--- a/drivers/net/benet/be_main.c
+++ b/drivers/net/benet/be_main.c
@@ -273,26 +273,6 @@ static void be_rx_eqd_update(struct be_adapter *adapter)
 	rx_eq->cur_eqd = eqd;
 }
 
-static void be_worker(struct work_struct *work)
-{
-	struct be_adapter *adapter =
-		container_of(work, struct be_adapter, work.work);
-	int status;
-
-	/* Check link */
-	be_link_status_update(adapter);
-
-	/* Get Stats */
-	status = be_cmd_get_stats(&adapter->ctrl, &adapter->stats.cmd);
-	if (!status)
-		netdev_stats_update(adapter);
-
-	/* Set EQ delay */
-	be_rx_eqd_update(adapter);
-
-	schedule_delayed_work(&adapter->work, msecs_to_jiffies(1000));
-}
-
 static struct net_device_stats *be_get_stats(struct net_device *dev)
 {
 	struct be_adapter *adapter = netdev_priv(dev);
@@ -900,8 +880,11 @@ static void be_post_rx_frags(struct be_adapter *adapter)
 		page_info->last_page_user = true;
 
 	if (posted) {
-		be_rxq_notify(&adapter->ctrl, rxq->id, posted);
 		atomic_add(posted, &rxq->used);
+		be_rxq_notify(&adapter->ctrl, rxq->id, posted);
+	} else if (atomic_read(&rxq->used) == 0) {
+		/* Let be_worker replenish when memory is available */
+		adapter->rx_post_starved = true;
 	}
 
 	return;
@@ -1305,6 +1288,31 @@ int be_poll_tx(struct napi_struct *napi, int budget)
 	return 1;
 }
 
+static void be_worker(struct work_struct *work)
+{
+	struct be_adapter *adapter =
+		container_of(work, struct be_adapter, work.work);
+	int status;
+
+	/* Check link */
+	be_link_status_update(adapter);
+
+	/* Get Stats */
+	status = be_cmd_get_stats(&adapter->ctrl, &adapter->stats.cmd);
+	if (!status)
+		netdev_stats_update(adapter);
+
+	/* Set EQ delay */
+	be_rx_eqd_update(adapter);
+
+	if (adapter->rx_post_starved) {
+		adapter->rx_post_starved = false;
+		be_post_rx_frags(adapter);
+	}
+
+	schedule_delayed_work(&adapter->work, msecs_to_jiffies(1000));
+}
+
 static void be_msix_enable(struct be_adapter *adapter)
 {
 	int i, status;
-- 
1.5.6.3





___________________________________________________________________________________
This message, together with any attachment(s), contains confidential and proprietary information of
ServerEngines Corporation and is intended only for the designated recipient(s) named above. Any unauthorized
review, printing, retention, copying, disclosure or distribution is strictly prohibited.  If you are not the
intended recipient of this message, please immediately advise the sender by reply email message and
delete all copies of this message and any attachment(s). Thank you.

Re: [PATCH 1/2] be2net: replenish when posting to rx-queue is starved in out of mem conditions

[David Miller]

From: Sathya Perla <sathyap@serverengines.com>
Date: Thu, 19 Mar 2009 14:43:01 +0530

> Hi, this is a patch to replenish the rx-queue when it is in a starved
> state (due to out-of-mem conditions)
> 
> thanks,
> -Sathya
> P.S: Pls ignore the "company confidential" warning at the end of the
> patch emails.

Next you'll be sending us "P.P.S: Please ignore the P.S".

Saying this "ignore the confidential warning" just makes it even more
painful, it's yet another thing to "ignore".  Are there some legal
ramifications for you saying this?  I bet not, so what's the point?
The legalise still applies and it still consumes space in our inboxes.

It's extremely impolite, at the very best.

> Signed-off-by: Sathya Perla <sathyap@serverengines.com>

I'll apply your patches, but use gmail or something if you can't
make this annoying company confidential crap just go away, thanks.

Re: [PATCH 1/2] be2net: replenish when posting to rx-queue is starved in out of mem conditions

[Sathya Perla]

On Fri, Mar 20, 2009 at 12:25 PM, David Miller <davem@davemloft.net> wrote:
>
>
> I'll apply your patches, but use gmail or something if you can't
> make this annoying company confidential crap just go away, thanks.

Sorry about this; I'm re-posting the patch below:

Signed-off-by: Sathya Perla <sathyap@serverengines.com>
---
 drivers/net/benet/be.h      |    1 +
 drivers/net/benet/be_main.c |   50 +++++++++++++++++++++++++------------------
 2 files changed, 30 insertions(+), 21 deletions(-)

diff --git a/drivers/net/benet/be.h b/drivers/net/benet/be.h
index 63d593d..f327be5 100644
--- a/drivers/net/benet/be.h
+++ b/drivers/net/benet/be.h
@@ -194,6 +194,7 @@ struct be_adapter {
 	struct be_eq_obj rx_eq;
 	struct be_rx_obj rx_obj;
 	u32 big_page_size;	/* Compounded page size shared by rx wrbs */
+	bool rx_post_starved;	/* Zero rx frags have been posted to BE */

 	struct vlan_group *vlan_grp;
 	u16 num_vlans;
diff --git a/drivers/net/benet/be_main.c b/drivers/net/benet/be_main.c
index 897a63d..80fe1e0 100644
--- a/drivers/net/benet/be_main.c
+++ b/drivers/net/benet/be_main.c
@@ -273,26 +273,6 @@ static void be_rx_eqd_update(struct be_adapter *adapter)
 	rx_eq->cur_eqd = eqd;
 }

-static void be_worker(struct work_struct *work)
-{
-	struct be_adapter *adapter =
-		container_of(work, struct be_adapter, work.work);
-	int status;
-
-	/* Check link */
-	be_link_status_update(adapter);
-
-	/* Get Stats */
-	status = be_cmd_get_stats(&adapter->ctrl, &adapter->stats.cmd);
-	if (!status)
-		netdev_stats_update(adapter);
-
-	/* Set EQ delay */
-	be_rx_eqd_update(adapter);
-
-	schedule_delayed_work(&adapter->work, msecs_to_jiffies(1000));
-}
-
 static struct net_device_stats *be_get_stats(struct net_device *dev)
 {
 	struct be_adapter *adapter = netdev_priv(dev);
@@ -900,8 +880,11 @@ static void be_post_rx_frags(struct be_adapter *adapter)
 		page_info->last_page_user = true;

 	if (posted) {
-		be_rxq_notify(&adapter->ctrl, rxq->id, posted);
 		atomic_add(posted, &rxq->used);
+		be_rxq_notify(&adapter->ctrl, rxq->id, posted);
+	} else if (atomic_read(&rxq->used) == 0) {
+		/* Let be_worker replenish when memory is available */
+		adapter->rx_post_starved = true;
 	}

 	return;
@@ -1305,6 +1288,31 @@ int be_poll_tx(struct napi_struct *napi, int budget)
 	return 1;
 }

+static void be_worker(struct work_struct *work)
+{
+	struct be_adapter *adapter =
+		container_of(work, struct be_adapter, work.work);
+	int status;
+
+	/* Check link */
+	be_link_status_update(adapter);
+
+	/* Get Stats */
+	status = be_cmd_get_stats(&adapter->ctrl, &adapter->stats.cmd);
+	if (!status)
+		netdev_stats_update(adapter);
+
+	/* Set EQ delay */
+	be_rx_eqd_update(adapter);
+
+	if (adapter->rx_post_starved) {
+		adapter->rx_post_starved = false;
+		be_post_rx_frags(adapter);
+	}
+
+	schedule_delayed_work(&adapter->work, msecs_to_jiffies(1000));
+}
+
 static void be_msix_enable(struct be_adapter *adapter)
 {
 	int i, status;
-- 
1.5.6.3

Re: [PATCH 1/2] be2net: replenish when posting to rx-queue is starved in out of mem conditions

[David Miller]

From: Sathya Perla <sathya.perla@gmail.com>
Date: Fri, 20 Mar 2009 15:49:25 +0530

> On Fri, Mar 20, 2009 at 12:25 PM, David Miller <davem@davemloft.net> wrote:
> >
> >
> > I'll apply your patches, but use gmail or something if you can't
> > make this annoying company confidential crap just go away, thanks.
> 
> Sorry about this; I'm re-posting the patch below:

I said "I'll apply your patches" as in, I've applied these already
but "in the future could you please..."

You didn't need to resubmit these.