Re: [bug report] RDMA/rxe: Failure of ibv_query_device() and ibv_query_device_ex() tests in rdma-core

[Zhu Yanjun <yanjun.zhu@linux.dev>]

在 2025/2/26 11:32, Daisuke Matsuda (Fujitsu) 写道:
> Currently, two testcases in rdma-core fail with the latest kernel, leaving the console log below.
> =====
> $ ./build/bin/run_tests.py -k device
> ssssssss....FF........s
> ======================================================================
> FAIL: test_query_device (tests.test_device.DeviceTest.test_query_device)
> Test ibv_query_device()
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>    File "/home/ubuntu/rdma-core/tests/test_device.py", line 63, in test_query_device
>      self.verify_device_attr(attr, dev)
>    File "/home/ubuntu/rdma-core/tests/test_device.py", line 200, in verify_device_attr
>      assert attr.sys_image_guid != 0
>             ^^^^^^^^^^^^^^^^^^^^^^^^
> AssertionError
> 
> ======================================================================
> FAIL: test_query_device_ex (tests.test_device.DeviceTest.test_query_device_ex)
> Test ibv_query_device_ex()
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>    File "/home/ubuntu/rdma-core/tests/test_device.py", line 222, in test_query_device_ex
>      self.verify_device_attr(attr_ex.orig_attr, dev)
>    File "/home/ubuntu/rdma-core/tests/test_device.py", line 200, in verify_device_attr
>      assert attr.sys_image_guid != 0
>             ^^^^^^^^^^^^^^^^^^^^^^^^
> AssertionError
> 
> ----------------------------------------------------------------------
> Ran 23 tests in 0.007s
> 
> FAILED (failures=2, skipped=9)
> =====
> 
> It seems sys_image_guid is set here:
> https://github.com/torvalds/linux/blob/2ac5415022d16d63d912a39a06f32f1f51140261/drivers/infiniband/sw/rxe/rxe.c#L82
> 
> I tried rolling back to commit 57a7138d0627, just before this patch was applied, and found the error resolved.
> [PATCH 1/1] RDMA/rxe: Remove the direct link to net_device
> https://lore.kernel.org/all/20241220222325.2487767-1-yanjun.zhu@linux.dev/

Thanks. The following commits are to fix this problem in upstream and 
for-next.

Because the patchset 
https://patchwork.kernel.org/project/linux-rdma/cover/20250119172831.3123110-1-yanjun.zhu@linux.dev/ 
exists in for-next, but this patchset does not exist in upstream.

Thus, 
https://patchwork.kernel.org/project/linux-rdma/patch/20250301193530.904720-1-yanjun.zhu@linux.dev/ 
is for for-next.

https://patchwork.kernel.org/project/linux-rdma/patch/20250301193351.901749-1-yanjun.zhu@linux.dev/ 
is for upstream.

Thanks,
Zhu Yanjun

> 
> I think the root cause lies in ndev patches applied in the past two months,
> but I am not very sure if it is good idea to revert them. I would like opinions
> from Zhu and other developers.
> 
> Thanks,
> Daisuke Matsuda
>