doc: update ICE doc

Message ID 1555040027-66786-1-git-send-email-wenzhuo.lu@intel.com
State Accepted
Delegated to: Qi Zhang
Headers show
Series
  • doc: update ICE doc
Related show

Checks

Context Check Description
ci/Intel-compilation success Compilation OK
ci/checkpatch success coding style OK

Commit Message

Wenzhuo Lu April 12, 2019, 3:33 a.m.
Update ICE document to describe a MDD event.

Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
---
 doc/guides/nics/ice.rst | 9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Rami Rosen April 15, 2019, 9:54 p.m. | #1
Hi,
In I40E DPDK nic guide, https://doc.dpdk.org/guides/nics/i40e.html,
MDD does not appear.

In IXGBE DPDK nic guide,  https://doc.dpdk.org/guides/nics/ixgbe.html,
MDD appears, but in the known issues section.

I think MDD is supported on both.

just wonder, for the sake of consistency, is it worth to add such a section
also
for these nics ?

Other than that,

Acked-by: Rami Rosen <ramirose@gmail.com>
Wenzhuo Lu April 16, 2019, 4:39 a.m. | #2
Hi Rami,

From: Rami Rosen [mailto:ramirose@gmail.com]
Sent: Tuesday, April 16, 2019 5:55 AM
To: Lu, Wenzhuo <wenzhuo.lu@intel.com>
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] doc: update ICE doc

Hi,
In I40E DPDK nic guide, https://doc.dpdk.org/guides/nics/i40e.html,
MDD does not appear.
[Wenzhuo] As I know, this feature is not supported by i40e. I40e can send such packets. That’s why we want to mentions this different behavior for ice.

In IXGBE DPDK nic guide,  https://doc.dpdk.org/guides/nics/ixgbe.html,
MDD appears, but in the known issues section.
[Wenzhuo] Actually, for ixgbe, we want to mention MDD as a limitation (the same section as  known issue). Because it’s more like our driver’s robust is not so strong. We have to ask the APP to provide the right L2/L3 length. (The reason is it’s not effective to let the driver inspect the packet to get the right length.)
But this ice behavior is more like a feature. To my opinion, it’s reasonable to drop such packets. That’s why I put it in the feature section.

I think MDD is supported on both.

just wonder, for the sake of consistency, is it worth to add such a section also
for these nics ?

Other than that,

Acked-by: Rami Rosen <ramirose@gmail.com<mailto:ramirose@gmail.com>>
Rami Rosen April 16, 2019, 7:59 a.m. | #3
Hi, Wenzhuo,

Regarding MDD for I40E, I am basing this on what I saw on the web regarding
MDD on I40E:
https://lists.osuosl.org/pipermail/intel-wired-lan/Week-of-Mon-20170925/010131.html
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1772675
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1723127
and more.

Regards,
Rami Rosen
Wenzhuo Lu April 17, 2019, 5:16 a.m. | #4
Hi Rami,

From: Rami Rosen [mailto:ramirose@gmail.com]
Sent: Tuesday, April 16, 2019 3:59 PM
To: Lu, Wenzhuo <wenzhuo.lu@intel.com>
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] doc: update ICE doc

Hi, Wenzhuo,

Regarding MDD for I40E, I am basing this on what I saw on the web regarding MDD on I40E:
https://lists.osuosl.org/pipermail/intel-wired-lan/Week-of-Mon-20170925/010131.html
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1772675
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1723127
and more.
[Wenzhuo] Have to say this patch is only for ICE. I’d like to leave this question to our i40e maintainers.

Regards,
Rami Rosen
Zhang, Qi Z April 23, 2019, 3:31 a.m. | #5
> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Lu, Wenzhuo
> Sent: Tuesday, April 16, 2019 12:40 PM
> To: Rami Rosen <ramirose@gmail.com>
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] doc: update ICE doc
> 
> Hi Rami,
> 
> From: Rami Rosen [mailto:ramirose@gmail.com]
> Sent: Tuesday, April 16, 2019 5:55 AM
> To: Lu, Wenzhuo <wenzhuo.lu@intel.com>
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] doc: update ICE doc
> 
> Hi,
> In I40E DPDK nic guide, https://doc.dpdk.org/guides/nics/i40e.html,
> MDD does not appear.
> [Wenzhuo] As I know, this feature is not supported by i40e. I40e can send
> such packets. That’s why we want to mentions this different behavior for
> ice.
> 
> In IXGBE DPDK nic guide,  https://doc.dpdk.org/guides/nics/ixgbe.html,
> MDD appears, but in the known issues section.
> [Wenzhuo] Actually, for ixgbe, we want to mention MDD as a limitation
> (the same section as  known issue). Because it’s more like our driver’s
> robust is not so strong. We have to ask the APP to provide the right L2/L3
> length. (The reason is it’s not effective to let the driver inspect the packet
> to get the right length.) But this ice behavior is more like a feature. To my
> opinion, it’s reasonable to drop such packets. That’s why I put it in the
> feature section.
> 
> I think MDD is supported on both.
> 
> just wonder, for the sake of consistency, is it worth to add such a section
> also for these nics ?
> 
> Other than that,
> 
> Acked-by: Rami Rosen
> <ramirose@gmail.com<mailto:ramirose@gmail.com>>

Applied to dpdk-next-net-intel.

Thanks
Qi

Patch

diff --git a/doc/guides/nics/ice.rst b/doc/guides/nics/ice.rst
index 15c2766..666b1b2 100644
--- a/doc/guides/nics/ice.rst
+++ b/doc/guides/nics/ice.rst
@@ -82,6 +82,15 @@  are chosen based on 2 conditions.
   If any not supported features are used, ICE vector PMD is disabled and the
   normal paths are chosen.
 
+Malicious driver detection (MDD)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+It's not appropriate to send a packet, if this packet's destination MAC address
+is just this port's MAC address. If SW tries to send such packets, HW will
+report a MDD event and drop the packets.
+
+The APPs based on DPDK should avoid providing such packets.
+
 Sample Application Notes
 ------------------------