Comment 9 for bug 1854842

Hi Mohammad,

Thanks for testing those kernels.

The second one did not include 0aa1d18615c163f92935b806dcaff9157645233a, which is why you still see the splat. Re-reading it, db849faa9bef993a1379dc510623f750a72fa7ce depends on 0aa1d18615c163f92935b806dcaff9157645233a to function, and I probably should have included it in the kernel for you to test.

I fixed the backport of db849faa9bef993a1379dc510623f750a72fa7ce slightly to integrate with 0aa1d18615c163f92935b806dcaff9157645233a better, and fixed the offset calculation in include/linux/mlx5/mlx5_ifc.h which I had previously missed on my initial backport. 0x1b - 0x1 = 0x1a instead of the 0xc from the upstream commit, since the 4.15 kernel does not have the max_vxlan_udp_ports[0x8] or reserved_at_38[0x6] offsets.

I have submitted both commits for SRU into the 4.15 bionic kernel, and they are now on the Ubuntu kernel mailing list:

https://lists.ubuntu.com/archives/kernel-team/2019-December/106293.html
https://lists.ubuntu.com/archives/kernel-team/2019-December/106294.html
https://lists.ubuntu.com/archives/kernel-team/2019-December/106295.html

The next step is for the kernel team to review the patches, and ack them. To be accepted into the next SRU cycle, they need a minimum of 2 acks. When they do get 2 acks, they will be applied to master-next and be a part of the next SRU cycle.

Once the patches have been acked and applied, the kernel will be built and available in -proposed. I will ask you to verify that the kernel in -proposed fixes the problem. I will be sure to message you when the time comes for this to happen.

Looking at https://kernel.ubuntu.com/, we see the last day for commits is the 1st of Jan, with eventual release somewhere around the 27th of Jan, give or take a few days.

I'll keep you informed about the progress of these patches.

Thanks,
Matthew