nbd/002: wait for the device node to show up

[yizhanglinux]

Fix the nbd/002 failure recently found by CKI:

nbd/002 (tests on partition handling for an nbd device)
nbd/002 (tests on partition handling for an nbd device) [failed]
runtime ... 3.679s
--- tests/nbd/002.out 2024-03-07 15:21:19.796849151 -0500
+++ /mnt/tests/gitlab.com/redhat/centos-stream/tests/kernel/kernel-tests/-/archive/production/kernel-tests-production.zip/storage/blktests/blk/blktests/results/nodev/nbd/002.out.bad 2024-03-07 16:01:23.838920252 -0500
@@ -1,4 +1,4 @@
Running nbd/002
Testing IOCTL path
Testing the netlink path
-Test complete
+Didn't have partition on the netlink path

[yizhanglinux]

So this change is not enough, the failure still can be observed during the stress test.

=======================333
nbd/002 (tests on partition handling for an nbd device)      [failed]ime  1.542s  ...
    runtime  1.542s  ...  0.813s
    --- tests/nbd/002.out	2024-03-08 07:02:03.709843571 -0500
    +++ /mnt/tests/gitlab.com/redhat/centos-stream/tests/kernel/kernel-tests/-/archive/production/kernel-tests-production.zip/storage/blktests/blk/blktests/results/nodev/nbd/002.out.bad	2024-03-11 01:48:44.582135222 -0400
    @@ -1,4 +1,4 @@
     Running nbd/002
     Testing IOCTL path
     Testing the netlink path
    -Test complete
    +Didn't have partition on the netlink path

[yizhanglinux]

Seems the issue cannot be reproduced now, close it now.
https://datawarehouse.cki-project.org/issue/2568

[kawasaki]

@yizhanglinux Sorry for the slow response.

Thanks for the report, and effort to fix this issue. I took a look into this, and was able to recreate the failure. To recreate it, I needed to disable KASAN and lock related debug options from my kernel config. When I repeat nbd/002 10 times, the failure is recreated always. So I think still we have the possibility to see the failure in the future CKI blktests runs.

I think the failure is not kernel side issue. Will post a fix candidate patch in blktests side.

[yizhanglinux]

@yizhanglinux Sorry for the slow response.

Thanks for the report, and effort to fix this issue. I took a look into this, and was able to recreate the failure. To recreate it, I needed to disable KASAN and lock related debug options from my kernel config. When I repeat nbd/002 10 times, the failure is recreated always. So I think still we have the possibility to see the failure in the future CKI blktests runs.

I think the failure is not kernel side issue. Will post a fix candidate patch in blktests side.

Yes, you are right, I changed to use one non-debug kernel and reproduced it again:

=======================49
nbd/002 (tests on partition handling for an nbd device)      [failed]
    runtime  1.687s  ...  1.029s
    --- tests/nbd/002.out	2024-03-19 04:51:34.051614893 +0100
    +++ /root/blktests/results/nodev/nbd/002.out.bad	2024-03-19 05:05:57.913874474 +0100
    @@ -1,4 +1,4 @@
     Running nbd/002
     Testing IOCTL path
     Testing the netlink path
    -Test complete
    +Didn't have partition on the netlink path
[root@storageqe-75 blktests]# uname -r
6.9.0-0.rc0.20240315gite5eb28f6d1af.7.fc41.x86_64