System hang in blk-mq layer with mq_hctx->state equal to zero

Issue

System hang in blk-mq layer with mq_hctx->state equal to zero:

PID: 3029837  TASK: ffff8ed105809ec0  CPU: 13  COMMAND: "kworker/13:1"
 #0 [ffffa5efc4fe7b68] __schedule at ffffffff9b090d5f
 #1 [ffffa5efc4fe7c00] schedule at ffffffff9b09118f
 #2 [ffffa5efc4fe7c10] schedule_timeout at ffffffff9b094cf3
 #3 [ffffa5efc4fe7ca8] io_schedule_timeout at ffffffff9b091569
 #4 [ffffa5efc4fe7cc0] wait_for_completion_io_timeout at ffffffff9b091d91
 #5 [ffffa5efc4fe7d28] blk_execute_rq at ffffffff9abea319
 #6 [ffffa5efc4fe7d60] __scsi_execute at ffffffff9adb440b
 #7 [ffffa5efc4fe7da0] scsi_test_unit_ready at ffffffff9adb48ee
 #8 [ffffa5efc4fe7e08] sd_check_events at ffffffffc0232800 [sd_mod]
 #9 [ffffa5efc4fe7e38] disk_check_events at ffffffff9abf83d9
#10 [ffffa5efc4fe7e98] process_one_work at ffffffff9a8ce7d7
#11 [ffffa5efc4fe7ed8] worker_thread at ffffffff9a8ceef0
#12 [ffffa5efc4fe7f10] kthread at ffffffff9a8d4802
#13 [ffffa5efc4fe7f50] ret_from_fork at ffffffff9b20023f

crash> struct blk_mq_hw_ctx.state 0xffff8ecf6cb29400 -x
    state = 0x0