Skip to contentSkip to conversation

aoe: ignore vendor extension AoE responses

PATCH In review Ed Cashin 3 participants 5 messages Feb 12, 2009

Subsystems (2)
lore
Share & tools
lore

Ed Cashin

author
#0 lore raw

PATCH 1/1 aoe: ignore vendor extension AoE responses

The Welland ME-747K-SI AoE target generates unsolicited AoE responses that are marked as vendor extensions. Instead of ignoring these packets, the aoe driver was generating kernel messages for each unrecognized response received. This patch corrects the behavior.

---

I will place this patch in the new aoe quilt tree for linux-next when this patch has passed LKML review.

 drivers/block/aoe/aoe.h    |    1 +
 drivers/block/aoe/aoenet.c |    2 ++
 2 files changed, 3 insertions(+), 0 deletions(-)
drivers/block/aoe/aoe.h +1 -0
index c237527..5e41e6d 100644
@@ -18,6 +18,7 @@
 enum {
 	AOECMD_ATA,
 	AOECMD_CFG,
+	AOECMD_VEND_MIN = 0xf0,
 
 	AOEFL_RSP = (1<<3),
 	AOEFL_ERR = (1<<2),
drivers/block/aoe/aoenet.c +2 -0
index 30de5b1..c6099ba 100644
@@ -142,6 +142,8 @@ aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt,
 		aoecmd_cfg_rsp(skb);
 		break;
 	default:
+		if (h->cmd >= AOECMD_VEND_MIN)
+			break;	/* don't complain about vendor commands */
 		printk(KERN_INFO "aoe: unknown cmd %d\n", h->cmd);
 	}
 exit:
Signature · 1 line
1.5.6.5
#1 lore raw
Quoted 35 lines · On Wed, 11 Feb 2009 19:16:44 -0500 Ed Cashin <[email protected]> wrote:

The Welland ME-747K-SI AoE target generates unsolicited AoE responses that are marked as vendor extensions. Instead of ignoring these packets, the aoe driver was generating kernel messages for each unrecognized response received. This patch corrects the behavior.

---

I will place this patch in the new aoe quilt tree for linux-next when this patch has passed LKML review.

 drivers/block/aoe/aoe.h    |    1 +
 drivers/block/aoe/aoenet.c |    2 ++
 2 files changed, 3 insertions(+), 0 deletions(-)
drivers/block/aoe/aoe.h +1 -0
index c237527..5e41e6d 100644
@@ -18,6 +18,7 @@
 enum {
 	AOECMD_ATA,
 	AOECMD_CFG,
+	AOECMD_VEND_MIN = 0xf0,
 
 	AOEFL_RSP = (1<<3),
 	AOEFL_ERR = (1<<2),
drivers/block/aoe/aoenet.c +0 -0
index 30de5b1..c6099ba 100644

drivers/block/aoe/aoenet.c aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt, · line 149

Quoted 8 lines
drivers/block/aoe/aoenet.c +2 -0
@@ -142,6 +142,8 @@ aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt,
 		aoecmd_cfg_rsp(skb);
 		break;
 	default:
+		if (h->cmd >= AOECMD_VEND_MIN)
+			break;	/* don't complain about vendor commands */
 		printk(KERN_INFO "aoe: unknown cmd %d\n", h->cmd);
 	}
 exit:

Do you believe that this fix should be backported into earlier kernel versions? (2.6.28.x, 2.6.27.x, etc)

#2 lore raw

On Wed, 11 Feb 2009 16:29:44 -0800, I waved a wand and this message magically appears in front of Andrew Morton:

Quoted 2 lines

Do you believe that this fix should be backported into earlier kernel versions? (2.6.28.x, 2.6.27.x, etc)

Yes, please - I'll be using AoE on these kernels. It might also be a good opportunity to update the AoE module, as I believe the latest out of the kernel tree module is currently at 70, whilst the version in 2.6.27 is at 47.

Signature · 4 lines
http://www.munted.org.uk

Fearsome grindings.
.
author
#3 lore raw
Quoted 10 lines · Alex Buell

On Wed, 11 Feb 2009 16:29:44 -0800, I waved a wand and this message magically appears in front of Andrew Morton:

Quoted 2 lines

Do you believe that this fix should be backported into earlier kernel versions? (2.6.28.x, 2.6.27.x, etc)

Yes, please - I'll be using AoE on these kernels. It might also be a good opportunity to update the AoE module, as I believe the latest out of the kernel tree module is currently at 70, whilst the version in 2.6.27 is at 47.

Yes, my intention is to regularly push patches now that I understand linux-next and have a quilt tree in place to push patches to it.

Signature · 1 line
  Ed Cashin <[email protected]>
author
#4 lore raw
In reply to Andrew Morton
Quoted 8 lines · Andrew Morton
Quoted 5 lines · On Wed, 11 Feb 2009 19:16:44 -0500 Ed Cashin <[email protected]> wrote:

The Welland ME-747K-SI AoE target generates unsolicited AoE responses that are marked as vendor extensions. Instead of ignoring these packets, the aoe driver was generating kernel messages for each unrecognized response received. This patch corrects the behavior.

...

Quoted 2 lines

Do you believe that this fix should be backported into earlier kernel versions? (2.6.28.x, 2.6.27.x, etc)

It shouldn't hurt, it can help, and it's well tested.

I am not sure how many environments have AoE packets that have vendor extension command types, though, so I can't gauge how pressing the need is for backporting.

If the cost of backporting is low, it probably should be backported.

Signature · 1 line
  Ed Cashin <[email protected]>