During failover nvme-tcp tests in a HA setup a crash orccured. iostat was running in the background while the test was executed. I am not sure if the fix is correct. I am unsure if using blk_queue_enter() is the right thing. But when starring at the code snipped I saw in older version blk_queue_enter() was used but got evenutally replaced by the rcu_read_lock() as far I understand at the moment... So any tip here would be highly appreciated. Daniel Daniel Wagner (2): block: remove commented out code from diskstats_show block: hold queue lock while iterating in diskstats_show block/genhd.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) -- 2.29.2