On 7/3/2018 5:11 PM, Jason Gunthorpe wrote:
On Tue, Jul 03, 2018 at 02:17:58PM -0400, Dennis Dalessandro wrote:
On 7/3/2018 1:32 PM, Jason Gunthorpe wrote:
On Mon, Jul 02, 2018 at 08:21:34AM -0700, Dennis Dalessandro wrote:
Hi Doug and Jason,
Here is the TID RDMA series I had mentioned back at OFA. This represents a lot
of hardwork by a number of people. I asked Kaike to provide a cover letter with
some background on what TID RDMA is, I'll paste it here. We'd like this to go
for 4.19 but it is a lot of code to review so won't be surprised if it has to
push off another cycle.
Omni-Path TID RDMA Feature
Intel Omni-Path (OPA) TID RDMA support is a feature that accelerates data
movement between two OPA nodes through the IB Verbs interface. It improves
RDMA READ/WRITE performance by delivering the data payload to a user
buffer directly without any software copying.
If this is verbs based then the rdma-core patches being posted are a
pre-requisite for reviewing this.
Depending on what is in there I may tell you, again, to fix this
drivers uapi before making any changes to the uapi.
It is for verbs, but it is contained within our driver. There are no
rdma-core patches required, no uapi changes. User doesn't have to know about
this at all.
Okay..
I looked over this thing and I don't know what to say.
So you see why Ira and I brought this up at OFA when we were all sitting
around at the networking event. :)
It badly violates all accepted limits for patch size and patch series
length. I don't see what you think anyone else is going to do with
this.
Oh I completely agree. Unfortunately I didn't see a better way to do
this. Kaike did a lot of work to try and break things up into logical
chunks. Any suggestions on how to make it more digestable, we are open
to it?
On the plus side, it seems all contained in the driver and I didn't
notice anything egregiously wrong inside it - but I have no intention
to check any of the details in something this big..
So does that mean you won't take the series?
Perhaps the thing to focus on is that we don't touch any uAPI, we don't
modify the ib_core. Everything is contained in our driver and can't
cause problems for any other drivers or core of the rdma sub system.
-Denny
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
