Please enable JavaScript.

Coggle requires JavaScript to display documents.

nvme pt uring hang - Coggle Diagram

- - - - on which cpu or hctx?
        
        228 ? D 1:15 [cpuhp/35]
        
        hctx13: default 35
        
        SCHED_RESTART|INACTIVE
    - - last offline cpu is 4
    - - cpu24 is already down
      - INACTIVE
  - - - --numjobs=1 --ioengine=io_uring_cmd
      - --filename=/dev/ng4n1
        
        irq 202, cpu list 4,24, effective list 0
        
        0 is online
        
        poll_queues = 0
      - --iodepth=16 --bs=4k
  - - - RQF_DONTPREP
        
        __nvme_submit_sync_cmd
      - RQF_SCHED_TAGS
      - RQF_QUIET
      - RQF_IO_STAT
      - RQF_STATS
    - - REQ_ALLOC_CACHE
      - REQ_FAILFAST_TRANSPORT
    - - must be done alreayd, but sched tag isn't released
  - - - plug is empty