Comment 0 for bug 1838939

Revision history for this message
Christian Ehrhardt  (paelzer) wrote :

[Impact]

 * Azure RDMA devices are driven by MLX4 driver (which I assume those devices are derived from).
   But to be enabled properly the system needs to detect them as such.
   That was added in rdma-core v20 hence being fix released in >=Disco.

 * Fix [1] is just a new alias to detect the card as what it is.

[Test Case]

 * TBD: Try to use rdma on "Microsoft Azure RDMA device".
   It would be great if someone could add some real content here.

[Regression Potential]

 * The regression potential should be minimal. Existing systems either already match the HCA
   table or they don't (neither before nor after the fix. An issue I could think of is if
   there are devices that announce as vmbus:3daf2e8ca732094bab99bd1f1c86b501 but are NOT in fact
   this device type - they would then run into issues.
 * Of course the fix will "enable rdma support on some more devices" if someone had those devices
   attached but didn't use/configure them they might now initialize a bit further. But that isn't
   an issue and especially in cloud environments where HW config is just a click away no one
   usually pays for extra devices without using them.
   Never the less we might want Microsoft to ack that this is ok for Azure in general.

[Other Info]

 * This falls under the SRU exception of "other safe cases" for "For Long Term Support releases
   we regularly want to enable new hardware". New modaliases are explicitly listed there as
   cases for this.

[1]: https://github.com/linux-rdma/rdma-core/commit/b98397f3068a42e7d8e67d4ef2a90393b4c4f08a