On Thu, 2019-02-28 at 15:22 +-0200, Leon Romanovsky wrote: +AD4 On Thu, Feb 28, 2019 at 07:39:25AM -0500, Dennis Dalessandro wrote: +AD4 +AD4 On 2/27/2019 2:49 PM, Parav Pandit wrote: +AD4 +AD4 +AD4 Giving low priority to most of them, I think desire to have loopback rdma device are below. +AD4 +AD4 +AD4 1. rxe is not ready for adding IB link types and large code restructure to avoid skb processing in it. Pretty large rewrite to skip skbs. +AD4 +AD4 +AD4 2. stability and reasonable performance +AD4 +AD4 +AD4 3. maintainability +AD4 +AD4 +AD4 +AD4 I don't see how this is more maintainable. We are adding a new driver, a new +AD4 +AD4 user space provider. So I don't see that as being a reason for adding this. +AD4 +AD4 Agree too, it is so tempting to write something new instead of fixing. Hi Leon, Early 2018 there was a discussion at LSF/MM about how to implement a high- performance block driver in user space. A highly optimized RDMA loopback driver in combination with e.g. the NVMeOF kernel initiator driver and a NVMeOF target driver in user space could be used for this purpose. Do you think it is possible to make loopback in the rdma+AF8-rxe driver as fast as in Parav Pandit's driver? See also Matthew Wilcox, +AFs-LSF/MM TOPIC+AF0 A high- performance userspace block driver, January 2018 (https://www.spinics.net/lists/linux-fsdevel/msg120674.html). Thanks, Bart.
