On Tue, 2018-02-13 at 06:46 +0100, Paolo Valente wrote:
>
> Any chance your gremlin grants you some OOPS?
I crash dumped it, though I'm unlikely to have time to even figure out
what the hell _I'm_ doing in IO land, much less what it's doing :)
If you want it, and have a spot, I could try an over night upload
(vmcore=14GB). Hung tasks are not in short supply.
crash> ps | grep UN
4 2 0 ffff880187f12940 UN 0.0 0 0 [kworker/0:0H]
6 2 0 ffff880187f144c0 UN 0.0 0 0 [mm_percpu_wq]
8 2 1 ffff880187f16040 UN 0.0 0 0 [rcu_sched]
9 2 4 ffff880187f16e00 UN 0.0 0 0 [rcu_bh]
18 2 1 ffff88018740d280 UN 0.0 0 0 [kworker/1:0H]
24 2 2 ffff880187582940 UN 0.0 0 0 [kworker/2:0H]
30 2 3 ffff8801875e8000 UN 0.0 0 0 [kworker/3:0H]
36 2 4 ffff8801875ed280 UN 0.0 0 0 [kworker/4:0H]
42 2 5 ffff88018769a940 UN 0.0 0 0 [kworker/5:0H]
48 2 6 ffff880187700000 UN 0.0 0 0 [kworker/6:0H]
54 2 7 ffff880187705280 UN 0.0 0 0 [kworker/7:0H]
57 2 1 ffff8801877b1b80 UN 0.0 0 0 [netns]
58 2 0 ffff8801877b2940 UN 0.0 0 0 [kworker/0:1]
62 2 2 ffff8801877b6040 UN 0.0 0 0 [writeback]
66 2 3 ffff88041e881b80 UN 0.0 0 0 [crypto]
67 2 5 ffff88041e882940 UN 0.0 0 0 [kintegrityd]
68 2 6 ffff88041e883700 UN 0.0 0 0 [kblockd]
69 2 2 ffff88041e8844c0 UN 0.0 0 0 [kworker/2:1]
73 2 4 ffff88041e910dc0 UN 0.0 0 0 [edac-poller]
74 2 7 ffff88041e911b80 UN 0.0 0 0 [devfreq_wq]
75 2 5 ffff8801877b0000 UN 0.0 0 0 [watchdogd]
91 2 3 ffff8803fa08d280 UN 0.0 0 0 [nvme-wq]
92 2 0 ffff8803fa08e040 UN 0.0 0 0 [ipv6_addrconf]
160 2 1 ffff88041e886e00 UN 0.0 0 0 [kaluad]
163 2 5 ffff8803fa346e00 UN 0.0 0 0 [kmpath_rdacd]
167 2 7 ffff8803fa346040 UN 0.0 0 0 [kmpathd]
168 2 3 ffff8803fa345280 UN 0.0 0 0 [kmpath_handlerd]
290 2 7 ffff8803f784a940 UN 0.0 0 0 [kworker/7:2]
356 2 6 ffff8803f677ee00 UN 0.0 0 0 [ata_sff]
526 2 3 ffff8803fc678000 UN 0.0 0 0 [scsi_tmf_0]
528 2 0 ffff8803fc679b80 UN 0.0 0 0 [scsi_tmf_1]
530 2 3 ffff8803fc67b700 UN 0.0 0 0 [scsi_tmf_2]
532 2 7 ffff8803fc67d280 UN 0.0 0 0 [scsi_tmf_3]
534 2 7 ffff8803fc67ee00 UN 0.0 0 0 [scsi_tmf_4]
536 2 3 ffff8803f6848dc0 UN 0.0 0 0 [scsi_tmf_5]
540 2 2 ffff8803f684c4c0 UN 0.0 0 0 [kworker/u16:6]
541 2 5 ffff8803f684d280 UN 0.0 0 0 [nvkm-disp]
543 2 4 ffff8803f684ee00 UN 0.0 0 0 [ttm_swap]
546 2 3 ffff8803f3ae8000 UN 0.0 0 0 [kworker/3:1H]
547 2 2 ffff8803f3ae8dc0 UN 0.0 0 0 [kworker/2:1H]
550 2 6 ffff8803f3ae9b80 UN 0.0 0 0 [kworker/6:1H]
552 2 5 ffff8803f3aeb700 UN 0.0 0 0 [scsi_tmf_6]
591 2 1 ffff8803f3aeee00 UN 0.0 0 0 [ext4-rsv-conver]
601 2 5 ffff8803f880ee00 UN 0.0 0 0 [kworker/5:1H]
632 2 1 ffff8803fa023700 UN 0.0 0 0 [kworker/1:1H]
643 2 0 ffff8803fa025280 UN 0.0 0 0 [kworker/0:1H]
659 2 4 ffff8803f677d280 UN 0.0 0 0 [kworker/4:1H]
663 2 7 ffff8803f6779b80 UN 0.0 0 0 [kworker/7:1H]
701 2 0 ffff8803f9eea940 UN 0.0 0 0 [kworker/0:2]
702 2 1 ffff8803f9eeb700 UN 0.0 0 0 [rpciod]
703 2 7 ffff8803f9eee040 UN 0.0 0 0 [xprtiod]
1020 2 4 ffff8803fc1c44c0 UN 0.0 0 0 [acpi_thermal_pm]
1170 2 3 ffff8803f1708dc0 UN 0.0 0 0 [jbd2/sdd1-8]
1171 2 6 ffff8803f170c4c0 UN 0.0 0 0 [ext4-rsv-conver]
1176 2 3 ffff8803fac58000 UN 0.0 0 0 [ext4-rsv-conver]
4233 2 5 ffff8803a8be8dc0 UN 0.0 0 0 [kworker/5:1]
8808 10134 6 ffff88037e789b80 UN 0.0 8104 3248 make
15603 2 5 ffff88041e886040 UN 0.0 0 0 [kworker/5:0]
15815 2 1 ffff88037e788dc0 UN 0.0 0 0 [kworker/1:0]
19344 8808 6 ffff8803c6658dc0 UN 0.0 7720 2960 make
19633 19632 3 ffff88041e916e00 UN 0.0 9924 2296 gcc
19651 19649 3 ffff8803a8bee040 UN 0.0 9896 2276 gcc
19683 19682 3 ffff8803b6e29b80 UN 0.0 9904 2300 gcc
19709 19687 2 ffff8803a8a41b80 UN 0.0 5852 1868 rm
19728 19727 2 ffff8803a8be9b80 UN 0.0 9932 2256 gcc
19732 19731 5 ffff880396dc0dc0 UN 0.0 9920 2296 gcc
19739 19606 4 ffff8803f394d280 UN 0.0 4560 1608 rm
19740 1 5 ffff8803b6ec2940 UN 0.1 49792 19732 cc1
19744 19656 2 ffff880396ceb700 UN 0.0 4524 1828 recordmcount
19922 1 3 ffff8803a8a40dc0 UN 0.1 388140 13996 pool
20278 2 2 ffff880396dc2940 UN 0.0 0 0 [kworker/u16:1]
20505 2 2 ffff8803b6eb1b80 UN 0.0 0 0 [kworker/2:2]
20648 2 3 ffff88037e78a940 UN 0.0 0 0 [kworker/3:1]
20766 2 6 ffff880187f11b80 UN 0.0 0 0 [kworker/u16:0]
21007 2 6 ffff8803c97844c0 UN 0.0 0 0 [kworker/6:0]
21385 2 4 ffff8803a8a46040 UN 0.0 0 0 [kworker/4:1]
21562 2 1 ffff8803f7b61b80 UN 0.0 0 0 [kworker/1:2]
21730 2 7 ffff8803d6525280 UN 0.0 0 0 [kworker/7:3]
22775 2 4 ffff880396dbe040 UN 0.0 0 0 [kworker/4:2]
22879 2 6 ffff880396db8000 UN 0.0 0 0 [kworker/6:1]
22937 2 7 ffff880396dbc4c0 UN 0.0 0 0 [kworker/7:0]
22952 2 4 ffff880396dbee00 UN 0.0 0 0 [kworker/u16:2]
23004 2 5 ffff880396dbb700 UN 0.0 0 0 [kworker/5:2]
23070 2 1 ffff8803d6520dc0 UN 0.0 0 0 [kworker/1:1]
23152 4478 2 ffff8803a8a444c0 UN 0.0 5748 1656 sync
31874 2 3 ffff88018740c4c0 UN 0.0 0 0 [kworker/3:0]
> BTW, the only other commit in that series that affects bfq interaction
> with the rest of the system is:
> a52a69ea89dc block, bfq: limit tags for writes and async I/O
I'll see if that changes things on the side. Work beckons.
> And the other commits that do something beyond changing some calculation are:
> 0d52af590552 block, bfq: release oom-queue ref to root group on exit
> 52257ffbfcaf block, bfq: put async queues for root bfq groups too
>
> Thanks,
> Paolo
>
> > -Mike
>