hung task timeout issue if flushrq is 1.

[v3nk1]

Hi folks,

I am seeing hung task timeout warning consistently and have a look at below dump.

22734.139015] INFO: task dd:3570 blocked for more than 120 seconds.
[22734.145246] Tainted: G W O 4.9.0 #10
[22734.150879] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[22734.159088] dd D 0 3570 3350 0x00000000
[22734.159141] Call trace:
[22734.159176] [] __switch_to+0x94/0xa8
[22734.159208] [] __schedule+0x180/0x6a0
[22734.159237] [] schedule+0x54/0xc0
[22734.159266] [] schedule_timeout+0x1c4/0x368
[22734.159295] [] io_schedule_timeout+0xa8/0x118
[22734.159324] [] bit_wait_io+0x20/0x68
[22734.159353] [] __wait_on_bit+0xac/0xe8
[22734.159383] [] out_of_line_wait_on_bit+0x74/0x88
[22734.159413] [] __wait_on_buffer+0x38/0x48
[22734.159443] [] __block_write_begin_int+0x198/0x590
[22734.159471] [] block_write_begin+0x60/0xd8
[22734.159501] [] blkdev_write_begin+0x50/0x68
[22734.159530] [] generic_perform_write+0xb8/0x198
[22734.159559] [] __generic_file_write_iter+0x170/0x1d0
[22734.159588] [] blkdev_write_iter+0x78/0xe0
[22734.159618] [] __vfs_write+0xd0/0x128
[22734.159647] [] vfs_write+0xa8/0x1c0
[22734.159676] [] SyS_write+0x54/0xb0
[22734.159705] [] el0_svc_naked+0x20/0x24

Any ideas why it spits this warning, any idea how to resolve it without echoing 0 to /proc/sys/kernel/hung_task_timeout.

Thanks in advance,
Venktesh.

[sectorsize512]

@ezk @krishven @vinothkumarraja can comment more, but I was wondering if you could try the version for the later kernel version: https://github.com/dmdedup/dmdedup4.8
It would be very helpful.

[venktesh-bolla]

I tried it with the latest 4.8. Saw the same issue. Saw 100% iowait.