[bug report] rdma_rxe doesn't work with blktests on 5.14.0-rc6

[Zhu Yanjun]

On Wed, Aug 18, 2021 at 12:57 PM Yi Zhang <yi.zhang at redhat.com> wrote:
>
> Hello
>
> I found rdma_rxe doesn't work with blktests on the latest upstream, is
> it a known issue, feel free to let me know if you need any info/test
> for it, thanks.
>
> # nvme_trtype=rdma ./check nvme/008
> nvme/008 (create an NVMeOF host with a block device-backed ns) [failed]
>     runtime  0.323s  ...  0.329s
>     --- tests/nvme/008.out 2021-08-18 00:18:35.380345954 -0400
>     +++ /root/blktests/results/nodev/nvme/008.out.bad 2021-08-18
> 00:35:11.126723074 -0400
>     @@ -1,5 +1,7 @@
>      Running nvme/008
>     -91fdba0d-f87b-4c25-b80f-db7be1418b9e
>     -uuid.91fdba0d-f87b-4c25-b80f-db7be1418b9e
>     -NQN:blktests-subsystem-1 disconnected 1 controller(s)
>     +Failed to write to /dev/nvme-fabrics: Cannot allocate memory
>     +cat: '/sys/class/nvme/nvme*/subsysnqn': No such file or directory
>     +cat: /sys/block/n1/uuid: No such file or directory
>     ...
>     (Run 'diff -u tests/nvme/008.out
> /root/blktests/results/nodev/nvme/008.out.bad' to see the entire diff)
>
>
> [  981.382774] run blktests nvme/008 at 2021-08-18 00:33:21
> [  981.470796] rdma_rxe: loaded
> [  981.474338] infiniband eno1_rxe: set active
> [  981.474340] infiniband eno1_rxe: added eno1
> [  981.476737] eno2 speed is unknown, defaulting to 1000
> [  981.481803] eno2 speed is unknown, defaulting to 1000
> [  981.486865] eno2 speed is unknown, defaulting to 1000
> [  981.492862] infiniband eno2_rxe: set down
> [  981.492864] infiniband eno2_rxe: added eno2
> [  981.492904] eno2 speed is unknown, defaulting to 1000
> [  981.497957] eno2 speed is unknown, defaulting to 1000
> [  981.504338] eno2 speed is unknown, defaulting to 1000
> [  981.510442] eno3 speed is unknown, defaulting to 1000
> [  981.515509] eno3 speed is unknown, defaulting to 1000
> [  981.520580] eno3 speed is unknown, defaulting to 1000
> [  981.526600] infiniband eno3_rxe: set down
> [  981.526601] infiniband eno3_rxe: added eno3
> [  981.526640] eno3 speed is unknown, defaulting to 1000
> [  981.531693] eno3 speed is unknown, defaulting to 1000
> [  981.538052] eno2 speed is unknown, defaulting to 1000
> [  981.543115] eno3 speed is unknown, defaulting to 1000
> [  981.549088] eno4 speed is unknown, defaulting to 1000
> [  981.554149] eno4 speed is unknown, defaulting to 1000
> [  981.559211] eno4 speed is unknown, defaulting to 1000
> [  981.565201] infiniband eno4_rxe: set down
> [  981.565203] infiniband eno4_rxe: added eno4
> [  981.565242] eno4 speed is unknown, defaulting to 1000
> [  981.570306] eno4 speed is unknown, defaulting to 1000
> [  981.576724] eno2 speed is unknown, defaulting to 1000
> [  981.581785] eno3 speed is unknown, defaulting to 1000
> [  981.586848] eno4 speed is unknown, defaulting to 1000
> [  981.599312] loop0: detected capacity change from 0 to 2097152
> [  981.612261] nvmet: adding nsid 1 to subsystem blktests-subsystem-1
> [  981.614215] nvmet_rdma: enabling port 0 (10.16.221.116:4420)
> [  981.622586] nvmet: creating controller 1 for subsystem
> blktests-subsystem-1 for NQN
> nqn.2014-08.org.nvmexpress:uuid:4c4c4544-0035-4b10-8044-b9c04f463333.
> [  981.622830] nvme nvme0: creating 32 I/O queues.
> [  981.634860] nvme nvme0: failed to initialize MR pool sized 128 for
> QID 32         ----------------------- failed here

Recently a lot of commits are merged into linux upstream, can you let
us know the kernel version
on which this problem occurs?

Thanks

Zhu Yanjun


> [  981.772647] eno2 speed is unknown, defaulting to 1000
> [  981.641676] nvme nvme0: rdma connection establishment failed (-12)
> [  981.777715] eno3 speed is unknown, defaulting to 1000
> [  981.782780] eno4 speed is unknown, defaulting to 1000
> [  981.799443] rdma_rxe: unloaded
>
> --
> Best Regards,
>   Yi Zhang
>