blktests failures with v6.5

Shinichiro Kawasaki shinichiro.kawasaki at wdc.com
Tue Sep 5 03:36:52 PDT 2023


Hi all,

FYI, here's the run results with latest blktests (git hash: 2045e8d3df86) and
kernel v6.5.

Comparing with the failures observed with v6.5-rc1 [1], I observed one new
failure, #5 in the list below. Also I observed two failures got resolved (#1 and
#2 in [1]). Comparing with report with v6.4 [2], I found that the failures at
block/024 and nvme/030,031 are no longer observed (#2 and #4 in [2]).

List of failures
================
#1: block/011
#2: nvme/003 (fabrics transport)
#3: nvme/* (fc transport)
#4: zbd/010
#5: srp/002, 011 (rdma_rxe driver)

[1] https://lore.kernel.org/linux-block/mklc3tlep35diuya2nnthlx2n5g75kikcqvmims4uox5qsszay@crbdp52wenhc/
[2] https://lore.kernel.org/linux-block/lkmloyrqpebispffur5udxdiubmevvodtsvnap3jz7tv5ihstr@jg7ejye3bein/

Failure description
===================

#1: block/011

   The test case fails with NVME devices due to lockdep WARNING "possible
   circular locking dependency detected". Reported in Sep/2022 [3]. In LSF
   2023, it was noted that this failure should be fixed.

   [3] https://lore.kernel.org/linux-block/20220930001943.zdbvolc3gkekfmcv@shindev/

#2: nvme/003 (fabrics transport)

   When nvme test group is run with trtype=rdma or tcp, the test case fails
   due to lockdep WARNING "possible circular locking dependency detected".
   Reported in May/2023. Fix discussions were held [4][5].

   [4] https://lore.kernel.org/linux-nvme/20230511150321.103172-1-bvanassche@acm.org/
   [5] https://lore.kernel.org/linux-nvme/5cb63b10-72ab-4b7c-7574-48583ea8ffd1@grimberg.me/

#3: nvme/* (fc transport)

   With trtype=fc configuration, test run on nvme test group hangs. Daniel is
   driving fix work.

#4: zbd/010

   f2fs mount fails with zoned block devices. A fix patch is already
   up-streamed for v6.6-rc1 [6].

   [6] https://lkml.kernel.org/linux-f2fs-devel/20230711050101.GA19128@lst.de/

#5: srp/002, 011 (rdma_rxe driver)

   Test process hang is observed occasionally. Reported to the relevant mailing
   lists in Aug/2023 [8]. It is suggested to change default driver of the srp
   test from rdma_rxe to siw so that we can take time to fix the failure.

   [7] https://lore.kernel.org/linux-rdma/18a3ae8c-145b-4c7f-a8f5-67840feeb98c@acm.org/T/#mee9882c2cfd0cfff33caa04e75418576f4c7a789


More information about the Linux-nvme mailing list